Editor’s note: For those who missed it, here’s “So You Want to Discover a New Particle (Pt. 1)“
I went away for a bit to pursue some other extra-curriculars, but now I’m back — and the blog has a new home! Perhaps a new audience as well. Since this (my first) Quantum Diaries post is actually the second part in a series, and I am a shamefully sporadic poster, I suppose I owe you a synopsis:
• The Standard Model is boring, but new particles are not.
• If you’d like to discover a new particle, theorists have conveniently provided you with an expansive menu of possibilities.
• Choose your particle with great care. This involves physics-, politics-, and plushy-based considerations.
So, let’s assume you’ve done some research, sent some emails, formed (or joined) a research group, and selected the new particle you’d like to discover. Congrats. Now what? Well, your particle hasn’t been found yet probably because… it isn’t easy to find. There could be multiple reasons for this:
• It has a tiny production cross-section (the LHC produces it only once in a blue moon, as opposed to millions of times per second).
• It looks like other, already-discovered particles, so it blends right in with the (so boring!) Standard Model.
• It is produced and/or decays in such a way that our particle detectors have a tough time seeing it and measuring its properties.
• It is shielded from present detection by a future version of itself, traveling back in time to thwart your research plans (see here or here).
With the notable exception of that last one, the practical consequence is that you’re essentially looking for a needle in a haystack. Only sometimes the needle looks an awful lot like hay. And in some cases, the needle could actually be lying outside the haystack, in a place you aren’t even looking. We call this hay “background”; the needle, “signal.” Before you can even think about looking for your needle of interest, you first have to get to know the hay.
In general, the bulk of your background is made up of Standard Model processes that resemble your signal but are produced far more copiously. As an added complication, this can be split into two groups: “physics” backgrounds, which actually involve the same “final state” particles as your signal — that is, those that are measured by the detector and then used to reconstruct the event; and “instrumental” backgrounds, which involve a different final state that is mis-measured by the detector such that it “fakes” looking like your signal. To make matters worse, it’s possible that non-Standard Model (new!) particles similar to your own could interfere with your analysis, adding an entirely theoretical component to your total background. Insidious! You must be aware that yours isn’t necessarily the only needle in this haystack.
(… As is often the case, one person’s trash is another person’s treasure. :))
Assuming you did the research before choosing your particle, you already know the main backgrounds for your signal. Great! Now to study these background processes in a systematic, controlled manner, we use what are called “Monte Carlo” (MC) simulations, aka “fake data.” Although I won’t go into detail about how the MC is generated, I will say that it’s a sophisticated simulation representing our best guess at how the data should look based on both theoretical and experimental constraints. It involves random numbers, probability distributions, lots of computing power, and magic.
Most likely your backgrounds have already been simulated, and all you need to do is get them. Do so. Practically speaking, this is person- and experiment-specific, so I offer no details, but let’s just assume that you now have access to the MC simulated datasets for your main backgrounds — and so the fun can begin! Except it is now 2am, and my workweek is about to start. Nuts!
Next time: Examining your simulated haystack, comparing it to your simulated needle, and finding ways to effectively separate the two with an algorithmic baler.