I, and many of my palaeomagnetist colleagues, often use thermal demagnetization (potentially up to 800°C) in our studies. Usually a stepwise technique is employed, so each sample is heated -- often 20 or more times -- to progressively higher temperatures. This can become a problem with jointed or friable samples: the pieces have to be glued together in order to make a measurement, but there are very few adhesives which are both heatproof and non-magnetic. By ‘non-magnetic’ I mean that it should carry as little remanence as possible; while it would also be nice if it had a negligible magnetic susceptibility, this is a less important factor.
In my search for an adhesive which is heatproof as well as non-magnetic, I've looked at:
- UHU. Good bond, non-magnetic, but not very heatproof. Even the high-temperature version breaks down at about 250°C. Beyond that, every heating step leaks noxious burning-glue fumes into the lab and I have to re-glue all the samples before measuring them.
- Cyanoacrylate (a.k.a. superglue). Non-magnetic, and bond usually OK, at least on fine-grained rocks. Even less heat-tolerant than UHU, but the fumes are a bit less noxious at least.
- Aremco Ceramabond 571, a magnesium oxide based paste which easily fulfils the temperature requirement (it's rated to 1760°C) and was recommended to me as non-magnetic (albeit not by the manufacturer). Unfortunately it turned out to carry a significant mangetic remanence, and contaminated measurements unacceptably. (It might work reasonably for strongly-magnetized volcanic samples, though.)
- Sodium silicate in aqueous solution, a.k.a. water glass. Magnetically, this fits the bill. Kostadinova et al. (2004) reported that it doesn't appear to affect remanence measurements (although it does alter susceptibility). It's thermally stable up to the temperatures I care about and beyond. With clay-rich material I had some trouble with expansion (intumescence) of treated samples upon heating and have heard from other researchers that this is a known problem; Otaka et al. (2002) indicate that heating rate is also a factor. Another problem is that the adhesive properties are not ideal: sodium silicate works well as a binder for soaking and impregnating loose, friable material, but doesn't make a good contact adhesive for less porous rock.
- Automotive silicone adhesives: I have seen various silicone adhesives designed for work on automotive exhaust systems. They would probably fulfil the temperature and adhesive requirements, but I haven't yet done any pilot studies on their magnetic properties. Given the intended applications, there's no reason for the manufacturers to specifically exclude magnetic compounds from the mix. I'd be interested to hear if anyone's tried any adhesives of this sort.
I should add that the adhesive requirements are actually very modest: the join doesn't have to bear any load beyond the weight of the sample itself, which will not exceed 50g.
- Kostadinova, M., Jordanova, N., Jordanova, D., & Kovacheva, M. (2004). Preliminary study on the effect of water glass impregnation on the rock-magnetic properties of baked clay. Studia Geophysica et Geodaetica, 48(3), 637-646.
- Otaka, T., & Asako, Y. (2002). Thermal intumescent characteristics of heated sodium silicate. In ASME 2002 International Mechanical Engineering Congress and Exposition (pp. 205-210). American Society of Mechanical Engineers.