Ron Ghosh, October 2018
The D11 instrument (Schmatz, Ibel), a pinhole instrument with a 64x64 1cm pixel multidetector was a world beating instrument from its first use. With up to 40m collimation distance and 40m distance from sample to detector, and neutron wavelengths from 4.5A to over 30A there was enormous demand from a wide range of experimenters.
For the first time using mixtures of deuterated and hydrogenous polymers it was possible to explicitly demonstrate Florey's hypothesis that amorphous polymers exhibit the same conformation in the solid state as in the gas phase.
Polymer groups from Strabourg/Saclay (Benoit,Janninck, de Gennes, Farnoux, Cotton, Ober, Higgins), Mainz and Freiburg, Manchester (Allen, Edwards, Richards) with Americans King and Ullman queued for time to study amorphous polymers and thermodynamic properties. Crystalline polymers were studied in Aachen and Bristol and Sheffield.
Metallurgic problems like superalloys were studied by groups from Stuttgart and the ILL (Kostorz, Goetz et al)
For biology partial deuteration of specific subunits lead to many studies of structures at low resolution in solution - in marked contrast to crystallographic measurements where crystals were available.
A major study of ribosome structure using triangulation on a large number of separately deuterated subunits, alas, was interrupted by the shutdown in 1991.
Such was the demand for D11 a second instrument, D17, was added in 1976; a higher resolution detector (5x5mm) was intended to attract a class of low resolution crystallographic experiments. The major fraction of the program was however small angle scattering, and later reflectometry, before finally being rebuilt as a dedicated reflectometer. A limitation was its shortest wavelength being about 9A, due to the initial bent guides necessary to site the instrument in the guide hall.
Later new instruments, D22 and D33 now offer further small angle capabilities to D11.
Software for treating data is available at ftp://ftp.ill.fr/pub/cs/sans
D11
Two helical slot velocity selectors were available from the start. Adele (A) had a 50% wavelength band and was used when the highest (formidable) fluxes were required; it was little used after 1976. The second, Brunhilde (B), was most used; it had a 10% wavelength pass band. Adele was always scheduled at the beginning of a cycle, getting less active, simplifying exchange. These were housed in a casemate close to IN10, taking the bottom 5cm of the cold beam from the H15 guide.
There were three sections of guide between the end of the selector and the sample with lengths of 20 m, 10 m, and 5 m, mounted on micrometer screw traverses which were linked with chain drives to ensusre parallel alignment, with springs cushioning arrival at the critical in-beam position. Any misalignement had a strong effect on the effective pitch of the helix of the selector, changing the nominal wavelength. The minimum wavelength was 4.5 Å (5000rpm), though there was also provision for two crystals in dogleg configuration to obtain shorter wavelengths. Between each section of guide there was a box which could contain a diaphragm. The whole assembly was within a long steel box which was evacuated. The last guide in-place towards the sample acted as the source, hence it was possible to move this source back from 5 to 40 m.
The sample facilities comprised either an evacuated cloche, into which a finger ("chimney") with quartz windows could be mounted, or alternative fingers which could be lowered into the beam path within the evacuated detector tank. A small gate-valve separated the cloche from the incident beam box. A major cost of this simple instrument was the 64 cm VAT gate valve leading to the main detector tube. While later instruments introduced sample changers from the start, apart from the rarely used D11B holder samples were changed manually until about 1982, when the translation rack from D17 was introduced on D11. By then the 1 cm Helma quartz cells had become standard for many experiments. The rack could be heated or cooled over a limited temperature range.
Initially the detector was lowered into the detector tube by removing a cover plate; the top of the detector mount serving as the cover plate, while the original was moved to cover the hole from which the detector had been taken. There were such locations 2, 5 10 20 and 40 m from the sample. The discriminator and decoding electronics were mounted in air on the top of the cover-plate which had vacuum feed-throughs to the detector. After moving the detector evacuating the whole tube took between 3 and 4 hours. A beam stop was placed in front of the detector on a thin alunminium rod, and this, in theory, could be replaced through an airlock from the top. To minimise moisture and reduce pumping times the detector tube was always let up to atmospheric pressure using dry nitrogen gas (from heating liquid nitrogen).
The parallel x and y signals were taken from the detector through video quality cables to the cabin, with plugs at each position; after passing through a FIFO buffer they were set to the Telefunken TR86 (NICOLE) system with simple handshake tests, and incremented a section of dual port memory. The maximum count rate accceptable exceeded 150 000 events per second. The detector electronics included a mandatory deadtime of one microsecond for coding each x-y event.
The detector weighed about 1 tonne. An electric crane, capacity 1 tonne, was mounted on a gantry which could be pushed the length of the tube. In practice a block and tackle was the most reliable and gentle way of moving the fragile detector by the users, though this was exhausting during summer when temperatures in the guide hall regularly exceeded 33 C.
The detector (Jacobe) employed 10BF3 gas as neutron detecting medium at a pressure of about 700 mm pressure. The electrodes were etched from a copper coating with a 1 cm spacing on glass plates for x and y, and a grid of anodes in the 1 cm drift space. To maintain the planarity of the gas filled chamber a thin diaphragm separated it from a dome volume containing CO2. The gas detector was about 50% efficient at 6 Å, rising to nearly black at 10 Å and longer. One consequence was the discrimination against thermal neutrons from background, and thermalisation of cold neutrons in the sample. With a characteristc wavelength around 1.6 Å these were only counted with low efficiency.
Evidently the initially conceived design was cumbersome and inefficient in use; it was costly in measuring time to move the detector. As a contribution to the instrument the RAL designed and built a detector tank introduced before the main detector tube to offer diffuse scattering possibility for simultaneous measurements with SANS. This was designated D11B. There was a vertical sample changer included, and 32 detectors were placed about this new sample position. For compatibility the cloche was retained. A blockhouse in front of the sample position was built to house choppers (which were never installed) for the diffuse measurements. Wih great pressure from the SANS community there was little time to really test the combination of detectors; the counting times from the two measurements were quite disparate, The diffuse scattering instrument, D7, with a conventional sample arrangement for cryostats etc. was never really challenged and was easier to enhance with polarisation analysis etc..
New sample diaphragms were added at the blockhouse in front of the sample to simplify following the beam which dropped through gravity over long source to sample distances. Most importantly, and at the cost of losses through an additional window in front of the detector, this was mounted in air, and was itself used to serve as a bung for the detector tube. For distances less than 40 m it was only necessary to evacuate the section between detector and the 64 cm VAT gate valve.
New electronics with a CAMAC memory system was introduced, copying that built for D17. The instrument was transferred from NICOLE to a dedicated PDP11/40 running control programs developed for D17. Without the strain of the high data rates from D11, NICOLE managed to operate the time of flight instruments for another two years, although it no longer benefited from the 24/7 presence of experimenters on D11 to restart the system after failure.
After IN6 was constructed, using the full 15cm heght of the H15 guide, lights were introduced on D11 indicating which of the four wavelengths IN6 was using (4.1-5.6Å) to help avoid using the missing wavelengths..
The major upgrade in 1983 replaced the detector tube by one of sufficient diameter to place the detector and associated electronics inside in a new containment chamber at atmospheric pressure. Water cooling took away the heat generated by the decoding electronics; the data were sent by serial link to a receiver on the CAMAC, over 80m away. Though the overall length of the detector tube was reduced to 36 m, the gain in operational simplicity far outweighed this limited loss.
In 1985 the PDP11/40 was replace by a VAX11/730 32-bit computer. The software was later used without change to update D17 with a second-hand VAX. The low level system software was re-used on all the other ILL instruments which adopted the cheaper and faster microvax II computers when these became available.
D17
Even in 1975, after two years of D11 operations it was clear that the demand for SANS from metallurgists, polymer scientists, virus biologists, and other communities exceeded the measurement times available by several fold.
By curving and extending the H17 guide from the reactor through a new hole in the containment building enough space was gained to introduce D17 (Roth), a new SANS instrument, though with a short wavelength limit of 9.5 Å due to the guide curvature. The initial design had a sample-detector distance of 2.8 m, though this was later extended first to 3 m and then 3.5, though this could only be achieved using a fixed arm extension rail beyond the air-pad floor. The detector developed for this instrument had 128x128 cells, each 5 mm x 5mm , with a correspondingly smaller drift space to match the higher spatial resolution. The LETI technology followed that of D11, again using 10BF3 gas. The long wavelengths compensated for the shorter drift space, but the deector was slightly less efficient than the D11 detector. The higher resolution was intended to extend use to low resoution crystallography experiments for biological systems again using isotope substitution. The sample table (on a pillar copied from the two-armed IN1 spectrometer) could include a Huber four circle goniometer. Two helical velocity selectors were originally built with 10% and 5% wavelength resolution, the latter intended for the crystallographic experiments. In practice the lower intensity made crystal alignment very difficult using the 5% selector.
The detector vacuum tank arm was fitted with air pads to allow controlled movement to high angles, and by removing cylindrical sections the distance to the detector could vary 0.8 m to 2.8 m. The detector was often used off-axis to extend the q-range. The beam stop was motorised in the x direction, and could be exchanged through an airlock into the main tank.
The incident collimation after the selector was between diaphragms with an evacuated flight path, but other optical components could also be introduced in this space. Because a standard sample table in air was used it was easy to introduce a greater variety of sample environments that on D11. The instrument offered great flexibility to try out configurations, converging collimators, polarisation analysis using supermirrors, and finally a large number of reflection measurements were made before it was rebuilt as a dedicated reflectometer.
The initial electronics were based on a CAMAC memory system (Axmann) and were controlled by a PDP11/40 minicomputer running the RSX11/D system (Grevaz), transferring data by serial line (Lesourne, Tillier) to the PDP10 central computer. The same systems were introduced to D11 when removed from the NICOLE system. Both were upgraded to use the RSX11/M system in 1979, copying some software routines from the NICOLE-II replacement project of the "Deuxieme Souffle".