An achievement whose impact on the scientific life at the ILL has been, and still is huge.
As a whole series of early letters and reports attest, cryogenics was a crucial issue for the ILL from the beginning. For example:
- In 1972 the ILL was obliged to borrow a cryostat from the CEA for D2 .
- In March 1973 the directors brought the scientists and cryogenics staff together to discuss a strategy . They favoured commercial cryostats adapted specifically to the requirements of the instruments. The specifications were for an autonomy of only 30 to 50 hours, with temperature fixed or variable across a limited range. The rise from 1.5 K to 300 K required changing the cryogenic fluid, etc. And, to make things worse, sample alignment in the neutron beam meant leaning the cryostat, and neither the cryostats nor the ILL goniometers were designed for that.
- In octobre 1973 vigorous exchange of letters highlights the ILL's discontent with regard to its supplier Stoehr . Brochier strongly criticised Management, who were following Harwell and other centres, disregarding the specificities of the ILL. He was already envisaging variable temperature cryostats operating continuously between 1.5 and 300 K, non-existent at the time.
It's clear that the time had come for a new solution.
Although never patented and never published (except for one almost impossible-to-find technical report ), the much-copied Orange cryostat conquered world neutronics (ILL: 70, HFIR: 13, HZB: 15, LLB: 10, ISIS: 8, NCNR: 5, ANSTO, JAEA, LANSCE, PNPI, SNS, etc., at least 200 cryostats), synchrotrons (ESRF, etc.) and industry.
Dominique explains how he built it with Serge Pujol in 1975:
When I did the calculations for it - pushed on by Serge - I felt I was designing a steam locomotive in the age of electric trains. It was when those very promising cryocoolers were coming onto the market.
The ILL's cryogenics service was struggling with badly designed commercial cryostats; they were difficult to use, not flexible enough for our users, and especially far too slow. It would take half a day to change the sample or move from liquid helium temperature to liquid nitrogen. This was absurd for a high-flux reactor where experiments could be performed at a very fast rate.
So we looked for a cryostat that would be easy for users to use, reliable, flexible, and fast.
According to Dominique Brochier:
We based the design on the "anti-cryostat" idea of Air Liquide, with a central insulated tube through which we could insert and remove the sample without having to bring the whole of the cryostat back to temperature.
The idea moved things on at the time, but it wasn't perfect. The tube was not insulated from the liquid helium bath and the gas extracted to cool the sample circulated in the tube. We had to plug the arrival of cryogenic liquid when changing the sample. Thus ice clogging of the arrival of the cryogenic liquid could occur when changing the sample. This set the scene for Dominique and Serge's invention of the VTI (variable temperature insert). Dominique explains:
The annular space - "it was obvious!" ... and the VTI made it possible to change the temperature in the short periods of time ILL required. Serge was there all through the process, with his cryogenics and metal-worker's experience. He would say "That's not possible - but we could do this...!" The process of designing and building the prototype (the "36") was done on the sly, as Management was convinced of the quality of the commercial cryostats used around the world, and it had no plans for in-house development.
There's an amusing sequel to the story of after-hours work. Why the heck are the pins of the thermometer's Jaeger plug ordered 1-6-3-4 and not 1-2-3-4 still today? Because the first cryostat was assembled in haste one evening, in a badly-lit corner, and the welder made a mess of the no. 2 pin!
It would have been logical to choose ILL blue for the cryostat, but Dominique Brochier explains that painting aluminium was not a trivial exercise at the time. He consulted an expert from Guy Gobert's group, who told him:
"I've just what you need, except that it's orange".
The colour was a bit strange but Maurice de Palma quickly got used to it, referring to it from then on as the "orange cryostat".
The annual reports however show that it took some time for the name to catch on.
- 1975: "Full range (1.5-300 K) variable cryostat, top access, light, small, with very large holding time, easy to operate"
- 1976: "The 1.5 to 300 K variable temperature cryostat, with access to the sample from above, lightweight, compact and with a very long holding time"
- 1977: "ILL cryostat"
- 1980: "Standard cryostat"
- 1984: "Orange cryofurnace, covering continuously the 1.5 K to 500 K temperature range".
Serge Pujol remembers that ILL first produced 5 orange cryostats in-house, before ordering 5 others from SDMS and Bernusconi, although insisting on keeping the assembly work in-house. The question of patenting or publishing was raised, but Dominique Brochier settled the matter: "No time!". Even though the 1977 technical report  resembles the draft of a patent application that was never submitted
ILL offered the design to several different companies, before finally choosing AS Scientific (Abingdon, UK) who went on to build the orange cryostats under a non-exclusive ILL licence. The original deal provided ILL with a free cryostat for every ten devices built; this later became a percentage.
After ten years however ILL forgot to claim its dues. It was finally Gerhard Collet (head of the Finance and purchasing service) who noticed it and obtained a few cryostats from AS Scientific in compensation.
Then later, in 2010, it was Isabelle Petit (accounts) who noticed that the dossier had again been neglected. AS Scientific finally came up with what it owed.
In 2015 a contract was negotiated when SANE developed a new heat exchanger to speed up the temperature changes in the Orange cryostat: three times faster cooling and twice as fast heating.
The VTI idea has been copied by all the companies producing liquid helium cryostats and cryomagnets. Some of the details couldn't be reproduced however, and the VTIs sold by these companies are not always reliable when it comes to controlling the temperature of samples below 7 K.
The material of the cryostat vessel
- The vessel walls were in aluminium, which was a novel choice at the time. Cryostats were in stainless steel until then, but Dominique Brochier and Serge Pujol wanted to lose some weight, to facilitate the countless mounting and unmounting operations on the spectrometers.
The VTI — Variable Temperature Insert
- To reduce the cryostat's energy consumption, Dominique and Serge added a liquid-nitrogen cooled screen between the helium bath and the sample tube. This made it possible to work above -200°C without having to rapidly empty the helium bath. They also had the idea of making the helium gas flow around an annular space surrounding the sample tube. With this brilliant idea it became possible to dissociate the helium flow from the sample volume and drastically reduce the probability of blocking up the cold valve.
- This is one of the key components in the Orange cryostat. The helium is introduced, expanded with a pump. The gas cools the copper exchange block, which in turn cools the walls of the sample chamber and the exchange gas inside it. The challenge was to get the dimensions and geometry right, to absorb as many calories as possible, whilst still being able to control the processes... Domique Brochier explains: "I made a few rough calculations and we set up a rig, testing by trial and error. A filter, by definition, presents an obstruction; so things get blocked,and we couldn't let the pressure rise in the cryostat. All the work on the exchanger was linked to that."
The cold valve
- The cold valve is used to expand the liquid helium from 4.2 K (about -269°C), transforming it into a colder gas (about 1.3 K) which will then cool the exchanger and the sample. If the valve is opened too far, the cryostat uses too much helium; if opened too little the sample won't reach the necessary temperature. It's a nightmare for the inexperienced, and all sorts of effort were put into automating the process. For Serge Pujol, the secret was not to turn or move the valve; the trick was to just create a variable leak by pushing the pin in contact with the body of the valve more or less firmly. The choice of material for the body of the valve is critical, as is the geometry of the hole it rests in.
The aluminium tail
- It is very important for scientists that this part of the cryostat, which is inside the neutron flux, produce as little neutron background and parasitic diffraction spots as possible. According to Dominique Brochier, this material was chosen out of the blue - sometimes pure aluminium, sometimes single-crystal aluminium. And he thinks the result was not that bad in the end as the aluminium used in more recent cryostats seems to have created more problems. Aluminium alloys have evolved over time and it's not uncommon to find different results in the same series.
The cryogenics service had a lot of faith in its Orange cryostat and wasted no time on the training of newbies. Alain Filhol remembers his first encounter with a cryostat:
My cryogenics training lasted no longer than the time it took for Klaus Gobrecht to perform an express helium transfer. I was then left alone for the night in the guide hall, staring at the cryostat (oh oh, is this thing going to explode?) and fiddling with the (ever-so-delicate) Taconis level meter (ohh... what's that strange noise....? Empty? Full?), strange and worrying noises (ouille, ouille, ouille, is that normal?), spurts of helium in all directions (if I waste too much I'm going to be in trouble!)... and me just clenching my buttocks !
The Orange cryostat turned out to be remarkably forgiving of errors and easy to use, even if some of us did manage to sent a few sample sticks (or even worse, a dilution cryostat!) into orbit.
Dominique Brochier thought he was redesigning the steam engine...
The Orange cryostat had an autonomy in normal circumstances of 60 hours for liquid helium, and 30 hours for liquid nitrogen. So there was a need to top up manually with liquid helium and nitrogen - far too often, in the users' opinion. Cryocoolers are less powerful and slower, but they had the big advantage of requiring no surveillance... until automatic transfers of liquid helium and nitrogen became possible and cryostat cold valves were automated.
And in any case, with the Orange cryostat, there's a wider accessible temperature range (mini dilution refrigerator, cryofurnace), greater refrigeration power, samples can be changed faster, and it's easier to combine low temperatures and high pressures.
And, finally, there's much less maintenance, as there are fewer moving parts. In November 2015 the record time for completely reconstructing an Orange cryostat was held by Jean-Paul Gonzales: he had the cryostat off the instrument by the end of the morning, re-installed it in the course of the afternoon, with the sample sitting cold in the beam before supper!
A few key dates, compiled by Serge Pujol and Eddy Lelièvre-Berna:
- Back in 1977 the success of the Orange cryostat encouraged attempts to automate temperature control, leading to the development of a temperature controller far more modern and efficient than the horrors being produced by industry at the time.
- 1980: ILL's aim is to produce 10 cryostats a year.
- 1981: AS Scientific (UK) produces the Orange cryostat under licence.
- 1982: Launch of the dilution insert for the Orange cryostat, capable of descending to 50 mK. The insert was developed by Karl Neumaier and Jean-Louis Ragazzoni.
- 1984: Arrival of the cryofurnace, a development of the Orange cryostat, capable of reaching 300°C whilst continuing to descend to 1.5 K.
- 1985: Development of an Orange dilution cryostat reaching 10 mK. It's automatically controlled, with generous sample space (Ø80 mm), and it can be accessed without removing the cryogens.
- 1987: Serge Pujol developed a variety of skills in cryogenics thanks to the Orange cryostat. He went on to build a very compact helium-flow cryostatfor a Eulerian cradle. The results were impressive (1.8 K to 450 K), but helium consumption was high (1 litre / hour).
- 1989: Given the success of the compact cryostat, Serge Pujol and Alain Benoît develop an ultra-compact helium-flow dilution cryostat for Eulerian cradle. It is insensitive to gravity and has reached 110 mK on the D10 diffractometer. This is the only cryostat of its type in the world.
- 2002: Serge Pujol and Xavier Tonon replace the compact helium flow cryostat with a 10 K cold head, to which they add a Joule-Thomson stage. The high consumption problem is solved, and the minimum temperature is 1.8 K.
- 2006: Jean-Paul Gonzales fits an exchange gas flow thermometer chamber enceinte thermomètre à circulation de gaz d'échange inside the cryofurnace, thus improving considerably the precision of the temperature measurement.
- 2010: Xavier Tonon and Philippe Camu improve the dilution cryostat insensitive to gravity by adding an isotopic separator. The functional autonomy of the dilution, which had been limited to a week, becomes infinite. The system is entirely automated.
- 2015: Xavier Tonon, Julien Gonthier and Eric Bourgeat-Lami modify the Orange cryostat exchanger, which now operates three times faster. Signature of an amendment to the licensing agreement with AS Scientific.
The future seems assured for the Orange cryostat, as helium gas jet systems are far too expensive (€10/litre of liquid helium) and dry cryostats are practically 6 times slower.
- Letter to Jacrot and Mössbauer, Dominique Brochier, ILL, 13 October 1972.
- Minutes of the cryogenics meeting of 22 March 1973, Dominique Brochier, ILL.
- ILL note: "Cryostats pour les expériences", M. Jacquemain and D. Brochier, 8 October 1973.
- "Cryostat à température variable pour mesures neutroniques ou optiques", Brochier, D., Tech. Report 77/74 (1977)