In fact, it will be hard to distinguish dark-matter-induced signals from spin noise, so this is expected to be the eventual sensitivity limit for such measurements. doi.org/10.1088/2058....

I will note that ultra sensitive #NMR experiments being built to search for #Axion-like dark matter are expected to be sensitive enough to measure spin noise without the feedback enhancement. (CASPEr, the cosmic axion spin precession experiment)

So inductive #NMR can indeed measure the signal from 10^10 spins, but only if there are 10^20 identical ones around to help amplify the signal

This feedback mechanism (confusingly called “radiation damping”) amplifies the spin noise by driving the other 9.99999e19 spins in response to the tiny spin-noise signal. This is analogous to the (usually annoying) audio signals produced by improperly arranged audio equipment amplifying random noise

Turns out inductive (“normal”) NMR can only measure spin noise because it is amplified by radiation damping. Because NMR probes are high-Q tuned resonant circuits, any small nuclear magnetization signal induces a small current in the circuit, and this current then acts on all the spins in the sample

But wait, ~1mL of a 1M solution should give me ~10^20 spins, so measuring spin noise should mean measuring ~10^10 spins. If we could do that, shouldn’t #NMR be capable of measuring (10s of) picomolar concentrations? Why isn’t @brukercorporation.bsky.social selling us 32-bit ADCs?

In hyperpolarized samples, you can actually get *negative spin noise*, which broke my brain for a while (what does negative noise power mean? Turns out the spins are cooling the probe circuit a tiny bit in that scenario.)

Spin noise is actually independent of polarization, in principle. Typical inductive NMR measurements actually amplify the effect via radiation damping, so they do have a magnetization dependence. The nanoscale case is more like “pure” spin noise (sometimes called “statistical polarization”)

It’s measuring the inherent statistical variance in spin projection. If you have N spins, you would expect sqrt(N) of them to be precessing at any given time, even in a totally unpolarized sample. A dominant effect for small values of N (e.g. MRFM or shallow-NV-detected NMR).

Either that or chicken wings are neutrinos

Now that I check my abstract, apparently I also promised some quantum sensing with diamond… Might need to start making slides if I’m going to fit this all in to 20 minutes for non-experts 😅

They should probably just switch to helium

My wife doesn’t yet know how much she hates you for showing me this

Coming from the Pacific Northwest, I think I’m required to fight you now

Waaaay into the Zeeman-dominated regime. None of those fabled “chemical shifts” however

Yes please!!