Thread by @PKoppenburg

A #thread on penguins. The ones in particle processes.

What's the difference between these two drawings? [1/17]

Both are Feynman diagrams, a tool used to compute probabilities of particle processes.

In both cases a beauty (b) quark becomes a strange quark (s) and two muons (μ). The difference is inside, which is the stuff we do not see. [2/17]

The penguin is that one. It involves three among the heaviest particles we know: the top quark (t), the W and Z bosons, that are responsible for the weak interaction.

They appear for a very very short time (more on that later) thanks to Heisenberg's uncertainty principle. [3/17]

The heaviest particles we know do their business without being spotted. But maybe other particles we do not yet know are involved. There could be a heavy Z boson (Z') or a leptoquark (LQ). [4/17]

How can we tell? By looking at the direction in which the muons fly. We can parametrise this in quantities like P'₅, which we plot versus the squared mass of the two muons, q².
(About measuring masses, see ) [5/17]

The result @LHCbExperiment got in 2020 is this. Here q²=1 would be about the mass (squared) of a proton.

Ref: [6/17] arxiv.org/abs/2003.04831

The data points in black differ from the theory prediction in orange. That could either indicate
1. New particles in penguins
2. Something wrong in the experiment
3. Some effect not accounted for in theory. That's where the other diagram enters. [7/17]

The other diagram involves a charm quark (c) and its antiparticle, the anti-charm quark. They also appear for a short time and annihilate into a pair of muons. [8/17]

Both diagrams contribute to the overall b→sμμ process. Sometimes it's the one, sometimes the other, sometimes both at the same time. Like in the double-slit experiment. [9/17]

For theorists predicting the charm diagram is much harder because it's nonlocal: the charm loop is not at the same place as where the muon pair is produced, and thus happens later.

How much later? we are talking of less than 10⁻²⁰ seconds. That's long for particles. [10/17]

That nonlocality breaks the maths tricks theorists use (don't ask why).

Penguins on the other hand are local. All fine. [11/17]

But if nonlocal is hard to compute, maybe we can measure the effect. So instead of vetoing the regions of dimuon mass where the charm loops dominate, we use them. They are in grey in the 2020 plot. [12/17]

That's the @LHCbExperiment data: here are the number of processes versus dimuon mass squared. Note the weird vertical scale. There are many many dimuons at the mass of the J/ψ and ψ(2S) particles, that are made of a charm and and an anti-charm. [13/17]

Here's what we get in P'₅. In red the total fit and in pink only the local penguin contribution. The difference is mostly where the J/ψ and ψ(2S) are. [14/17]

Now we do a trick: let's replace the penguin parameters we determine by those we would expect from theory, while the nonlocal parameters are those from our fit. That's the blue band. It gets closer to theory. In absence of new particles the blue and red would overlap. [15/17]

But the blue band is still not overlapping with the (binned) theory prediction in orange by @DannyD82 and colleagues. That indicates that there may be nonlocal effects not fully accounted for. [16/17]

The paper posted today concludes that there's some mild tension between the theory expectation and the data.

But to me the most important is the input to theorists so they can work on their predictions. [17/17]