High temperatures

It was clear, very early on, that high-temperature and very high-temperature studies (>2000ºC) into the structure and dynamics of soft or solid materials would be promising veins of research for the ILL. The ability of neutrons to penetrate most metals was an obvious advantage in the design of high-tech furnaces.

It led to furnaces of all sorts being designed, first by the users, and then by ILL. All of these devices had interesting characteristics, but we will dwell on some of the more emblematic models. We've also provided some extracts from the Annual reports which help understand how high-temperature technology progressed at the ILL.

The ILL's furnaces group started with Raymond Serve (1972-1988), who also worked on high pressures, soon to be joined by Pierre Andant (1974-2003) and draughtsman Antoine Valenti. Dominique Brochier took over in 1975, allowing the group, in his own words, "a lot of liberty". Paul Martin didn't arrive until 1990. The group was initially supported by an "animateur scientifique" (first Albert Wright, and then Winfried Petry in 1988) before the service was taken over by scientists (Claude Zeyen, Anton Heidemann, Eddy Lelièvre-Berna).

Raymond Serve at 91 ©2015, Françoise Serve	Pierre Andant

with the help of Dominique Brochier, Raymond Serve and Jens-Boie Suck.

When ILL started, Guy Gobert (Gérald Guy Gobert to be precise), head of ILL's Mechanical service, and Pierre Tardif built 3 furnaces one of which was published in 1977. According to Dominique Brochier, these machines failed for lack of experience in heat technology. Raymond Serve is rather less direct. He considers the design of the furnaces acceptable (two concentric resistors) but that the electrical supply and control systems were catastrophic, and they were almost impossible to assemble and disassemble. Pierre Andant adds:

We owed a lot to Serve in the early days; he had a lot of good ideas. He was responsible for the early electrics, and for introducing saturable inductors. He borrowed the idea from submarines, where they were used to control high-power transformers.

At the time the directors preferred that users bring their own furnaces with them. For example:

• The 1973 Annual report says that D4 had an operational 800ºC furnace. In fact it was a furnace built by Walter Knoll and Hans Egger (Max-Planck Institut, Stuttgart) capable of reaching 1600ºC.
• The 1974 report says that the D2 furnace was rebuilt and reached 1000ºC.
• It was Jens-Boie Suck who was the first to perform a high-temperature experiment (1400 K) with neutrons at ILL. It was in 1975 on the D1A powder diffractometer, and the furnace came from Karlsruhe [1].

In 1975, Rudolf Mossbauer regrouped high and low temperatures under the responsibility of Dominique Brochier. At the time there was promising research being done on refractory metals and materials (for the technology, space and nuclear fields). Two teams therefore tackled the difficult problem of high (HT) and very high temperatures (VHT), although for very different neutron spectrometries:

• VHT and solid materials (1975 onwards): Pierre Aldebert, Dominique Brochier and Raymond Serve, with neutron powder and quasi-elastic diffraction.
• HT and molten materials (1980 onwards): a team from ENSEEG/LTPCM, with wide-angle and small-angle neutron scattering.

At the same time the Annual reports describe the multiplication of less ambitious furnaces with a variety of designs. This diversity of units was to generate a lot of work for the ILL's high-temperature group. The series of "blue furnaces" (the name was only to appear in 1986) was the result of a desire to standardise efforts and reduce maintenance.

With help from Pierre Aldebert.

Work started seriously when Pierre Aldebert arrived at ILL in 1975 from the Fond Romeu (Odeillo) Solar Furnace, the nec plus ultra at the time as regards high temperatures in France, indeed around the world.

Jean Pierre Traverse, Pierre's thesis supervisor, had got him studying La₂O₃, ZrO₂ and α-Al₂O₃ at vey high temperatures, with a view to studying ThO₂ and UO₂, refractory materials of interest in the nuclear sector. He'd pushed him straight into neutrons, because he knew that neutrons could localise the oxygen atoms in metallic oxides, something which couldn't be done with the X-ray sources available at that time.

Pierre had built an initial version of a resistance furnace for neutrons at Font-Romeu and a second version at the CEA/DN; he built his third with Dominique Brochier at ILL. He had managed 2500ºC in 1975 at Mélusine but it was at ILL in 1978 that he produced quasielastic scattering measurements using neutrons on La₂O₃ at 2400ºC [2,3]. He was able to observe for the very first time the liquefaction of an oxygen sub-lattice in a ceramic powder at very high temperature.

At the end of his thesis in 1980 his research was continued by a British team of scientists, who reached a record temperature of ~2660ºC (2930 K) in 1984 [4] with a furnace design based on that of Pierre.

Pierre Aldebert's main innovations for his very high temperature neutron furnaces were the use of very thin concentric tubes made of tungsten (a heat-resistant metal) as both heat elements and heat reflectors, the use of zirconium (ZrO₂) and carbon fibre as refractory insulation transparent to neutrons, zirconium at 900ºC as an oxygen trap, and, finally, precise optical measurement of the temperature using the pseudo black body technique.

The UK team's furnace [4] sports a double concentric resistor system. Its top-mounted electrical supply is very practical and has been continued in the ILL's more recent furnaces, which don't exceed 1600ºC however.

As remembered by Jean Blétry, Pierre Chieux, Claude Sénillou

The ENSEEG/LTPCM [5] had been working on high temperatures since the sixties, but with mass spectrometry [6]. In 1980 the team of P. Desré, Claude Sénillou, etc., decided to tackle the challenge of building high temperature furnaces for research into molten materials. This required very low background noise and no parasitic Bragg peaks [7,8]. They were targeting the following neutron techniques:

• large-angle neutron scattering (LANS), up to 2000ºC (~2300K). The furnace was to be used on D4, initially for studying liquid Ni-V alloys,
• small-angle neutron scattering (SANS), up to 1900ºC (~2200K). The furnace was to be used on D11 and D17, initially to study the liquid alloys Ag-Ge and Au-Si

The particularity of the furnaces was that the heat resistor and the refractory heat screening were pierced to let the neutrons through; this can be done with liquids because their internal motion homogenize the temperature. In addition, the sample-holder crucible made of monocrystalline aluminium oxide (sapphire) can be oriented using a positioning goniometer made in molybdenium to avoid parasitic Bragg peaks.

Strange as it may seem, these furnaces were built completely independently of Aldebert's work; neither team was aware of the other's developments. And yet they were the only two teams in the world working on high or very high temperatures with neutrons at the time - in the very same place!
Around 1980 the l'ENSEEG/LTPCM team managed to reach fusion temperature for aluminium oxide (T_F=2054ºC), a record for that type of experiment. In fact, in 1975 Pierre Aldebert had studied alumina at Mélusine to temperatures almost as high, but powder diffraction does not imply the same experimental constraints.

The LTPCM furnaces designed for D4, D11 and D17 helped throw light on local chemical phenomena (the formation of aggregates) and demixing (spontaneous separation of a liquid into two immiscible phases) in molten metallic alloys.

The LTPCM gave its two furnaces to the ILL later.

With the help of Dominique Brochier, Pierre Andant and Paul Martin.

The importance of high temperatures at ILL in 1980 is proven by the 83 experiments requiring a furnace (664 beamdays), although generally at much less extreme temperatures than those mentioned above.

It was in 1980 that Raymond Serve launched an ILL furnace project based on Pierre Aldebert's design. The furnace reached 2500ºC in 1981 and 2600ºC in its third version in 1985 - even though, according to the Annual report, anything over 2000° was strictly for specialists! Raymond Serve says that nine out of ten experiments failed, generally with a melt-down of the sample. Precise temperature measurement was really a problem. According to Pierre Andant:

"Thermocouple measurements at high temperature were quite haphazard at that time. We had quite a few failures and it took a while to find our way. It was quite poetical in pyrometric terms! Measurements at that time with vanishing filament pyrometers involved assessing the colour across a number of filters with the naked eye. Nobody came up with the same temperature. We used Ircon pyrometers later and they were more reliable."

After 1986 there was less interest in very high temperatures. It seems that there was less demand and ILL started using standard furnaces. These were easier to use and cheaper in terms of manpower, but they didn't generate such high temperatures.

In the eighties Raymond Serve and Pierre Andant, followed by Paul Martin after 1990, devoted themselves to the production of a line of standard furnaces, therefore, which were not only cheaper but much easier to maintain than the record-breaking furnaces mentioned above. This was the blue series - a name that seemed to emerge in 1984. Paul Martin adds:

"Before the blue series arrived we would have to repair the furnace after every experiment; now it's only once a year. It's true that the temperatures were lower, but extremely high temperatures were no longer in demand - there was even a marked drop in the level of temperature required."

Around 1988/89 Pierre Andant innovated with a monolithic resistor using niobium foil 50 µ thick, bead welded using a technique copied from the production of radio tubes which also used heat-resistant metals. The technique was taken up under licence by AS Scientific (Abingdon, UK). On 5 April 1993 the contract set up for the sale of orange cryostats was modified to include blue furnaces. On 7 May Pierre Andant was asked to transfer his know-how and wrote to Colin Hillier, the head of AS Scientific:

"As regards the screening, there's an unbreakable rule: there must be the least possible 'actual contact' (thermal short-circuits) between the layers. The screens must be insulated from the 'resistor' and well-centred, whilst remaining free to float one relatively to the others."

Twenty years later Pierre Andant (nicknamed Pierrot) observes that:

"I'd perfected a method for producing the resistors which made the work easier and reduced the weight of the heat element. AS Scientific's copies were often poor in quality, because they were never as good as Pierrot on the welding point (It's all about getting the hang of it!)."

before adding:

"When Paul Martin arrived in the group in 1990 we moved from inspired and inventive DIY to a serious production operation."

With help from Claude Zeyen, head of the Development Branch (1993-1998). Claude headed the sample environment service after Dominique Brochier and before Anton Heidemann.

During the long reactor shutdown of 1991-1995, Claude went to Cadarache. It was then that he learnt that the nuclear experts were more than curious about the behaviour at very high temperatures of UO₂, the fuel used in most nuclear reactors. He decided to set up a research programme with Euratom (JRC-IRMM, Geel, Belgium) and started work on the construction of a furnace capable of reaching 3000ºC. There were four people in the project: Bernard Guerra (an engineer and specialist in very high temperatures, but with little neutronic experience unfortunately), Peter Ziegler (a physicist at Geel), Pierre Andant and Paul Martin (ILL). Claude, however, fell ill. He had to resign before the furnace was ready and his successor didn't continue with the project.

There's an interesting anecdote that Claude has passed on...:

"It was during the period of perestroika, and Russia was opening its doors to the world. I decided to buy up as much helium-3 as I could, and I billed it on at western market prices. That helped us finance the VHT furnace and the hard X-ray source of the ILL." 

Note that ILL now has access to very high temperatures thanks to sample heating using laser and levitation technology over a gas jet (Louis Hennet et al. [9]).