Looking back through old lab books for this Behind the Paper blog, the first time this paper's headline dysprosium bis(amide)-alkene compound was made in our lab was April 2022. Even more remarkably the yttrium analogue was first isolated in January 2021, and the dysprosium bis(amide) allyl precursor compound (with just one synthetic step needed to make a high-temperature Single-Molecule Magnet, SMM) was the very last thing that we made before Covid struck, a mind-boggling March 2020. Here we provide a brief account of the meandering journey from a synthetic oddity to a surprisingly decent SMM, and all the distractions we had along the way.
Grand designs
Our groups have been making and studying lanthanide SMMs for a long while now. The prospect of making Dy3+ compounds with only two amide ligands has been a target for us since Nick Chilton predicted astounding SMM properties for such compounds based on a samarium bis(amide) compound made by Mills group alumnus Conrad Goodwin.1 That was a decade ago, not to make anyone feel old.
The amide ligand in question is bis(tri-isopropyl-silyl)amide (or N††, ‘en tee tee’ to its friends). We like this ligand; it’s pretty easy to make, it has a relatively localised negative charge at the N atom, and the enormous bulk of the six isopropyl groups take up a lot of space when it’s stuck to a metal. This bulk means that a maximum of two of these ligands can fit around a Dy3+ ion. The dream Dy3+ SMMs would have two highly charged ligands sitting perfectly opposite to each other, as this maximises the energy barrier to reversal of magnetisation.
Atoms go walk about
As it turned out, making [Dy(N††)2]+ was far more fiddly than swapping an atom on the computer; the N†† ligand is prone to some self-destructive reactivity when near a Dy3+ ion. We caught a break in early 2020 when Jack Emerson-King (Jek to his friends, pronounced how you would expect) found that boiling KN†† and DyI3 in benzene led to two of the ligands binding to the Dy3+ ion. The complex also lost three protons. And two electrons. And one of the iso-propyl carbon atoms had relocated itself to a different isopropyl group, leaving behind an ethyl group, and forming an anionic allyl group (compound 2-Dy in the paper for anyone following along at home).
Initially this Dy3+ bis(amide) allyl compound was written off from an SMM viewpoint, as Dy3+ compounds with three strongly donating groups are the antithesis of a desirable SMM. Jek was immediately distracted by the curious organometallic reactivity of this compound and focused on investigating that with the Y3+ analogue. We’re pretty sure we know how the rearrangement happens now – more on that in a future paper. These investigations ultimately led to us finding conditions that prevented this transformation, allowing us to prepare the initially targeted compound [Dy(N††)2][anion] – only takes five steps.2 Unfortunately, while a nice synthetic study that has now given us access to a wide range of new compounds, this compound is actually a garbage SMM as the cation is highly bent.
May as well protonate it, I suppose
In 2022 we found the time to react the Dy3+ bis(amide) allyl compound 2-Dy with a proton source to give the Dy3+ bis(amide) alkene compound that is the subject of our Nature paper (1-Dy). Our initial assumption was that it would be a bad SMM, and so purifying 1-Dy was shunted solidly to the back burner, and we did not prioritise studying its magnetic properties.
A whole year passed before Gemma Gransbury (a former measurements postdoc in the Mills and Chilton groups) had some spare magnetometer time to measure a sample of 1-Dy which had been gathering dust. The initial low temperature measurements were unremarkable, fractionally better than [Dy(N††)2][anion], but then at higher temperatures magnetic relaxation was much slower than the best-performing magnets at the time. We observed magnetic hysteresis (a measurement of magnetic memory) up to 100 K (-173 °C) – about the temperature of the dark side of the moon but a substantial improvement on the previous SMM record of 80 K (-193 °C).
They have magnets on computers now
We conducted several additional measurements over the coming year to convince ourselves the observed parameters were real, delayed by the reticence of the title compound to crystallise. Ben Atkinson in the Chilton group then simulated the rate of magnetic relaxation to explain the unique magnetic behaviour that we had observed. The crystal structure has four compounds in each unit cell, and to simulate the phonons (vibrations in the solid-state structure which drive magnetic relaxation). Given that we need to perform calculations on the whole unit cell, these calculations particularly challenging.
We were ready to submit in mid-2024. Reviews were initially polarised, but contained several interesting points that we could work into an improved version of the manuscript. Chief amongst these was a proposal to model the magnetic hysteresis computationally. This was a large amount of work, and something which had until then not been attempted for any compound from first principles, but Ben and Nick were able to develop a model that simulated hysteresis, and described the behaviour of 1-Dy relative to a previously record-breaking SMM. As with the relaxation rates, it confirmed our description that the low-temperature relaxation of 1-Dy is relatively fast, driven by the flexibility of the N†† ligand, but slow at high temperatures because N†† is so charge-dense.
Future reflections
That the alkene compound 1-Dy is such a good magnet, when [Dy(N††)2][anion] is not, is worth reflecting on. As we set out in the paper, the primary reason is that alkene coordination forces the N-Dy-N angle to be much closer to the ideal 180 degrees. This additional equatorial coordination (long thought to be bad for SMM properties), does not turn off the high-temperature SMM behaviour. This gives us and others huge scope to improve on SMM performance further, and it’s exciting to see how the research will progress from here.
See also
- Chilton, N. F., Goodwin, C. A. P., Mills, D. P. & Winpenny, R. E. P. The first near-linear bis(amide) f-block complex: a blueprint for a high temperature single molecule magnet. Commun. 51, 101–103 (2015).
- Emerson-King, J. et al. Isolation of a bent dysprosium bis(amide) single-molecule magnet. J. Am. Chem. Soc. 146, 3331–3342 (2024).