A detector employing neutron sensitive image plates is comparatively cheap, is capable of high spatial resolution, has good homogeneity, a large dynamical range, extended linearity and no dead-time, and can be made to subtend very large angles at the specimen. The neutron sensitive plates are based on the same storage phosphor (BaFBr doped with Eu2+ ions) which is used for X-ray image plates, but with Gd2O3 added. This enables the Gd nuclei to act as neutron scintillators by creating a cascade of g-rays and conversion electrons.

The use of neutron image plates to construct a 2p Laue detector using a "white" neutron beam has been demonstrated with the prototype "cold-LADI". The present proposal is to construct a similar machine for thermal neutrons using already proved components.

The objective of the new machine is to extend the technique to problems of physics and chemistry on thermal LADI, in addition to biology on cold LADI. This is a technique in which a large part of reciprocal space can be examined; traditional diffractometers "only see what they are looking for". In this LADI is similar to powder diffraction, except that it should be even more powerful, since a 2D projection is obtained instead of a 1D diagram.

Thermal LADI is also like the single crystal machines on pulsed neutron sources, except that the time averaged flux on the sample can be considerably greater. Like these TOF machines it will probably not compete with traditional single crystal diffractometers where the highest precision is needed (often necessary for work complementary to X-ray measurements), but it should open up whole new fields of application.

For example it will be possible to follow structural transitions as a function of temperature, looking for superstructure spots, to examine the onset of anti-ferromagnetism, especially in complex cases of spirals and incommensurable modulations, to map out diffuse scattering as an indication of defects and short range order etc. We believe that it would become indispensable for the first work on new materials, either elastic scattering or 3-axis measurements, much as is now the case with neutron powder diffraction.
 

Applications to Physical and Chemical (small unit cell) problems.

During the periods when it is not scheduled for visitor use, LADI has been moved to a thermal beam for test experiments on problems in physics and chemistry involving smaller unit cells. An outline of the variety of applications which have so far been explored is given below.

Magnetism.

The LADI instrument is equipped with an Edwards Displex cryostat and a small Joule-Thompson expansion refrigerator, which was used to study temperature driven magnetic transitions in 4 % Ga doped FeGe2 (figure 2). The large detector surface combined with the white beam allowed simultaneously the rapid investigation of extensive parts of reciprocal space, and due to the high spatial resolution of the detector, even very subtle changes in the positions of the reflections could be followed as the temperature changed, a drastic departure in style compared with previous experiments (J. B. Forsyth et al. J. Mag. Magn. Mat. 177-181 (1998) 1395).

Charge Density Waves.

The puzzle of the diffuse layers of reflections in monochromatic diffraction patterns from hexagonal single crystals of La2Co1.7 has been solved using the ability of LADI to survey rapidly large volumes of reciprocal space, and found to be due to a charge (nuclear) density wave with a propagation vector t = 0.113a*,0,0.203c*. The relative strengths of the groups of satellites associated with fundamental reflections show that the displacement of nuclei is close to the c axis direction and that the effect persists from room temperature down to 15K ( P. Schobinger-Papamantellos et al. Collected Abstracts, ECM Meeting, Prague, August 1998).

High-pressure experiments

Large gains in speed can also be obtained for high-pressure experiments, and the first test with a diamond anvil cell on a known sample, a small 0.04 mm3 crystal of KH2PO4 at 2 GPa, gave an estimate of a factor hundred above similar measurements on a conventional diffractometer equipped with a position sensitive detector (Kuhs and Ahsbahs, ILL report, 1997). In addition, the combination of the large detector surface and Laue techniques alleviated the problems that often occur due to changes of crystal orientation during the pressure experiments. From this experiment it was concluded that with a modified cell approximately 10 GPa can be reached using this set-up; pressures presently inaccessible to single crystal neutron diffraction experiments at the ILL.

Rapid structural studies.

Due to the large wavelength range and increased detector area over conventional monochromatic experiments, Laue diffraction data can be collected on LADI on smaller crystals and in times which are typically one tenth of those needed for the equivalent monochromatic experiment. Due to the increased background levels in Laue diffraction , the weak reflection data are normally less precise than those of a monochromatic experiment, but the subsequent refinement of the crystal structure shows that the parameters obtained are sufficiently accurate most purposes.

A typical test of the data quality was the measurement on asparagine.H2O, which was done in 12 hours using a 1.5 mm3 crystal and a thermal, white beam with a useful wavelength range from 1.1 to 1.9 Å (D.A.A. Myles et al. Physica B241-243 (1998) 1122). The final crystallographic agreement factor was 0.029 for 1060 reflections (figure 3) and while the precision of the bond lengths was almost as good as in a previous monochromatic experiment, the data were collected fifteen times faster on a crystal of only one tenth of the volume.

In further tests of the increased data collection rate a non-linear optical material, deuterated zinc (tris)thiourea sulphate, was studied (J Cole, Ph.D. Thesis (1998) University of Durham). Again a white beam was used, but because of the small crystal cross-section of around 1 mm and the good spatial resolution there was no problem with overlap of neighbouring reflections. It was found that even though the estimated standard deviations on bond distances and angles were larger than from a conventional diffractometer experiment, the non-bonded contacts were of sufficient quality to give the same structural conclusions concerning the material as the earlier experiment, although the crystal was smaller and the measurement time was an order of magnitude less than previously.

This increase in measurement rate was also evident in a study of ND3-density distributions in the oriental disorder of Ni(ND3)6Cl2, where again a factor of over hundred in combined speed and crystal size was observed (Schiebel, Büttner and Kearley, ILL report, 1997). In this case it was found that scattering density maps were very similar to earlier studies (figure 4). For these compounds it is of interest to compare the crystallographic observations with molecular dynamics modelling, and with the proposed new detector this could be done as a function of temperature - for example near a phase transition - thus adding a completely new dimension to these types of studies.

The instrument is now completely automatic in crystal rotation and read-out, and there is no obstacle to use the detector on a monochromatic beam line. The only condition will be, that the individual recording is longer than the read-out time of about two minutes, which is indeed often the case. It is also obvious that such a detector would be ideally suited for many types of diffuse scattering experiments, but these applications have yet to be explored .
 

The proposed diffractometer.

The diffractometer proposed would be for diffraction physics and chemistry, where it obviously can be used for many different purposes. We propose the following unit (see figure 5), which is a vertical detector cylinder covered on the inside with neutron sensitive image plates; the reader head is also placed on the inside. This will give detector quantum efficiencies of at least 58%, a value measured for 'front' exposure and readout on the LADI detector. There will be free access from the top to insert many different sample holders such as cryostats, furnaces, magnets and pressure cells (this is not possible with LADI, which is horizontally mounted). The sample holder is integrated into a high precision sample rotation unit, which sits above the detector and can be lowered into this for the measurement.

Dimensions: 40 cm high, 100 cm circumference.

Pixel dimensions: 200, 400 or 800 m

Read-out time: 2 minutes

The detector will be controlled either by VME electronics connected to a HP workstation (the same as on the two present instruments - a 16-plate, automatic reader of large X-ray plates in operation at the ESRF and LADI at the ILL - which have identical, fully developed sets of electronics), or by Windows NT and a PC, which is rapidly becoming the preferred option for instrument control.
 

Estimated cost.

The detailed design study for thermal LADI is already finished (SICN Veurey, in consultation with F. Cipriani EMBL), and the complete machine can be constructed and delivered by an external company for a fixed price (2.9 MF). Demand on ILL technical services will be minimal.
 

Utilisation

Because this kind of machine is so new, many of its applications remain to be developed, and for this reason we see it not as an additional scheduled instrument, but as a tool for new investigations.

We propose that the machine be set up on a fixed position at the end of thermal guide tube H22 (behind D1A/D1B) in the position now used occasionally by cold LADI. Tests prove that a clean neutron beam can be delivered to this position with a very low gamma background (image plates are sensitive to X-rays and would need expensive protection if not placed on a guide tube). The measured intensity on H22 is sufficient with the size of crystals usually available (less of a problem than with cold LADI).

Technical assistance from ILL would be limited to sharing the technician presently employed on the neighbouring D10. It is proposed that scientific support and utilisation would be provided by external scientists (including students and post-docs) detached from user laboratories in return for beam time to develop new applications of the machine. Already several external laboratories have expressed interest in such an arrangement.

This proposal satisfies current ILL policy of developing new projects with active participation from outside users, and has many of the advantages of the CRG system while keeping a large part of the instrument time available to the general community.

The prototype cylindrical neutron image-plate detector (F. Cipriani, J.-C. Castagna, C. Wilkinson, P. Oleinek and M.S. Lehmann: Cold Neutron Protein Crystallography using a Large Position-Sensitive Detector based on Image-Plate Technology. Journal of Neutron Research. 4 (1996) 79).

1: Image plate on drum. 2: Drum. 3: Sample holder. 4: Crystal. 5: Transmission belt to drive drum. Motor is under table. 6: Carrier for reading head with photomultiplier. 7: He-Ne laser. 8: Mirrors for bringing the laser light to the reader head. 9: Reader head with photomultiplier. 10: Encoder for drum rotation. 11: Cover.

The sample crystal is mounted on a goniometer head on the cylinder axis, and can be rotated around this axis. The neutron beam, which enters and leaves via opposed holes in the cylinder, produces Bragg reflections and other, more general scattering patterns, which pass through the aluminium wall and are recorded on the image plates mounted on the outside cylinder surface. The detector, which has a radius of 159.2 mm and a length of 400 mm is read off in a phonographic mode with the reading head tracking slowly horizontally while the cylinder rotates at high speed. The readout time is four minutes, giving an image of the pattern recorded on the plates comprising 4000 x 2000 square pixels 200 mm on edge.

The sample axis is horizontal and the opening for the sample holder has only a diameter of 10 cm, so this detector would generally be unsuitable for physics/chemistry type work, where bulky sample environment units - often with vertical geometry - are used.

Two pictures of the phase transition of 4% Ga doped FeGe2. The first picture shows part of the room-temperature Laue diffraction pattern. This material undergoes a number of transitions and below 180K it locks into a non-harmonic spiral phase with a propagation vector of exactly 1/10. This gives rise to additional Bragg-peaks, which are shown in the second picture. The recording times of a few minutes are sufficient to give integrated intensities of good statistical quality.

Data from a small (1.5 mm3) single crystal of asparagine was measured in 12hrs. Errors on bond lengths are 0.002 to 0.003 Å, and the picture of the molecule shows that the thermal motion parameters derived are meaningful. This should be compared to the original measurement done at the Brookhaven high-flux reactor (J.J. Verbist, M.S. Lehmann, T.F. Koetzle and W.C. Hamilton. Acta Cryst. B28 (1972) 3006) which took over a week and employed a 15 mm3 crystal.

Figures comparing scattering density maps derived using maximum entropy techniques of, to the left: Ni(ND3)6Br2 (P. Schiebel et al. J. Phys: Condens. Matter 6 (1994) 10989) and to the right: Ni(ND3)6Cl2 (cylinder detector). The left hand picture only shows the central part, which is the distribution of deuterium atoms, but clearly the two distributions are very similar.

Schematic diagram of the rack containing the detector and the sample holder. The two pictures show the whole unit seen from the sides. The sample holder is integrated into a high precision sample rotation unit, which sits above the detector and can be lowered into this for the measurement.