We’d many more events over 2012 if the LHC simply kept running the way it already was at the end of the year. That’s because for most of the year, the luminosity was increasing over and over as the LHC folks added more proton bunches and focused them better. But we expect that the LHC will do better, starting close to last year’s peak, and then pushing to ever-higher luminosities. The worst-case we are preparing for is perhaps twice as much luminosity as we had at the end of last year.

But wait, why did I say “worst-case”?

Well, actually, it will give us the most interesting events we can get and the best shot at officially finding the Higgs this year. But increased luminosity also gives more events in every bunch crossing, most of which are boring, and most of which get in the way. This makes it a real challenge to prepare for 2012 if you’re working on the trigger, because have to sift quickly through events with more and more extra stuff (called “pileup”). As it happens, that’s exactly what I’m working on.

Let me explain a bit more of the challenge. One of the triggers I’m becoming responsible for is trying to find collisions containing a Higgs decaying to a bottom quark and anti-bottom quark and a W boson decaying to an electron and neutrino. If we just look for an electron — the easiest thing to trigger on — then we get too many events. The easy choice is to ask only for higher-energy electrons, but beyond a certain points we start missing the events we’re looking for! So instead, we ask for the other things in the event: the two jets from the Higgs, and the missing energy from the invisible neutrino. But now, with more and more extra collisions, we have random jets added in, and random fluctuations that contribute to the missing energy. We are more and more likely to get the extra jets and missing energy we ask for even though there isn’t much missing energy or a “Higgs-like” pair of jets in the core event! As a result, the event rate for the trigger we want can become too high.

How do we deal with this? Well, there are a few choices:

1. Increase the amount of momentum required for the electron (again!)
2. Increase the amount of missing energy required
3. Increase the minimum energy of the jets being required
4. Get smarter about how you count jets, by trying to be sure that they come from the main collision rather than one of the extras
5. Check specifically if the jets come from bottom quarks
6. Find some way to allocate more bandwidth to the trigger

There’s a cost for every option. Increasing energies means we lose some events we might have wanted to collect — which means that even though the LHC has produced more Higgs bosons, it’s counterbalanced by us seeing fewer of the ones that were there. Being “smarter” about the jets means more time spent by our trigger processing software on this trigger, when it has lots of other things to look at. Asking for bottom quarks not only takes more processing, it also means the trigger can’t be shared with as many other analyses. And allocating more bandwidth means we’d have to delay processing or cut elsewhere.

And for all the options, there’s simply more work. But we have to deal with the potential for extra collisions as well as we can. In the end, the LHC collecting much more data is really the best-case scenerio.

In any case, it used to be that when I wanted to work on my detector, I had only to go across the street. Now I have to drive out of Switzerland and several miles into France. Except, I don’t like driving. So I’ve been working on alternate means of transportation. A few months ago I walked. Last night I had to go to downtown Geneva, so I took the bus. It’s actually pretty good, although the bus stop is a mile away from CMS. There’s also the shift shuttle, which runs from the main CERN site to CMS every 8 hours via a rather roundabout route. And I can bike, once the weather gets better and I get myself a little more road-worthy. To be honest, every option for getting here is much slower than driving, but I enjoy figuring out ways to get places enough that I’m going to keep trying for a while.

I have plenty of chances to try, because I’ll be here in the CMS control room a lot of the time over the next few weeks. Right now, I’m learning and helping with the pixel detector calibration effort. (We’re changing the operating temperature, so all the settings have to be checked.) Soon I’ll be learning to take on-call shifts. So the more I stay here, the more I learn. I got here this morning, and I won’t leave tonight until about 11 pm. I could take the shift shuttle back — or maybe I’ll just get a ride.

That’s not exactly π, but it’s quite close. What I really should do is look up where it’s already defined in the math library I’m using, but this is more than accurate enough for any reasonable purpose. It’s too many digits, in fact, although I know a few more. So why do I always write out exactly that many places? Well, after thinking about it for a minute a little while ago, I remembered the answer: it’s the number of digits of π my dad taught me when I was a kid.

More pictures, and commentary, on Google+…

www.youtube.com/watch?v=LXdrk6LC1eU

The video goes along with this Nature News article. You may also be interested in the recent combined ATLAS and CMS Higgs result, which uses only the first half of this year’s data.

By the way, when I talk about a “minimal Higgs, that only does the part we know that something like the Higgs has to do,” I’m referring to the so-called fermiphobic Higgs. It plays the usual role of the Standard Model Higgs boson in breaking electroweak symmetry, but doesn’t couple to quarks and leptons (i.e. fermions). We already know from the way the weak and electromagnetic forces work that the relationship between them has its origins in something like the Higgs — but we have less reason to be certain that the same particle takes care of quark and lepton masses too. This version of the Higgs boson is more difficult to find, but perfectly sensible, and we’ll probably hear a lot more about it in coming years if we don’t have a big discovery this year or next.

For Europeans, the Wild West seems to be one of the commonly referenced themes from American culture. Perhaps they think that it reflects our national character even in the present day, or perhaps it’s just a very striking setting. I can’t quite imagine the Fermilab cafeteria putting up the equivalent — French Revolution Week, perhaps, complete with guillotine — but it’s all in good fun.

But in my view of how experimental physics is done, history has shown us that once in a great while, something is discovered that nobody thought of and nobody can fit into the existing theoretical mold. The measurements that led to the discovery of GR in the first place provide a good example of this. Such shifts are extremely rare, but I don’t like the idea of ignoring a result because it doesn’t fit with the theories we have.

No, we have to address the measurement itself, and satisfy ourselves that there really was a mistake. There are many ideas for what might have gone wrong, and as far as I know, the discussion is ongoing. I’m not an expert on it, but I know enough to disagree with some of the blogosphere discussion lately that has pronounced that the case is closed. There seem to be two categories of claims going around:

1. Articles that point out that the OPERA result is inconsistent with other measurements, as in this piece by Tommaso Dorigo (who is, incidentally, my colleague now that I’ve joined CMS). These are of course correct within the context of GR or any straightforward modifications thereof, as I said right at the start of this post. The question is whether there’s some modification that can accomodate the results consistently, and that’s a very hard thing to exclude. (There is some good discussion in the comments of Tommaso’s post about this, in fact.)
2. Articles that the OPERA result has been refuted because someone posted an idea on the arXiv server. A current example is this preprint, which asserts that a 60 nanosecond delay might be explained by OPERA having made a relatively trivial mistake in their GPS calculations. Of course, it’s possible that a trivial mistake has been made. But I’m not inclined to consider it definitive, especially because the author has already partially backpedaled upon learning more about how GPS works.

It’s great that people are sending ideas for what might have gone wrong with the result, or how it might be explained. But let’s wait for the discussion to settle down — and, indeed, for OPERA to finalize their paper — before we conclude that the case is closed. I do expect the result to be disproven, but what I want to see is one of these things:

1. OPERA finds that there really was a problem with their measurement, revises it, and the “superluminal” effect goes away.
2. Another experiment makes the same measurement, and gets a result consistent with GR.

Either way, I’ll consider the case closed, but there’s no reason to get ahead of ourselves. Doing science usually doesn’t mean knowing the answer in time for tomorrow’s news.

But then I remembered all the acronyms on ATLAS, and learned that “PU” has gotten more common there too now that the increasing pileup is a frequent subject of discussion. (I really wasn’t paying attention generally to either ATLAS or CMS for the year where I did my analysis and wrote my thesis.) So although the culture of acronym use may be a bit different, it’s really just a matter of translating from one experiment’s terms to another.

For example, I recently learned that a JSON (“JavaScript something something”) file indicates which LumiSections (not an acronym, oddly) are good in a set of runs — in other words, for which times are the recorded data for all parts of CMS in good shape? On ATLAS, it would have been a GRL (“good run list”) indicating which LumiBlocks were good.

I still think that acronyms are thrown around in conversation a bit more on CMS than on ATLAS. Fortunately, there is a public list of CMS acronyms to help me. I’m sure I’ll figure them out eventually.

We’re just kidding, of course.

I’ve been asked more seriously, on a number of occasions, why I switched from working with ATLAS to working with CMS. There are several ways I can answer that one:

1. Why not? ATLAS and CMS both look for the same exciting things at the LHC: the Higgs boson, supersymmetry, and all sorts of other new physics. They have roughly similar capabilities and, for the most part, conceptually similar designs. So I should be happy to work on either one.

2. It came with the job. Being happy to work on either experiment means I applied to some groups working on ATLAS and some on CMS. The job I ended up with is with Princeton, and they have a CMS group, so…

3. It’s good for our field to exchange techniques and expertise between experiments.

4. It’s good for me to know people from both collaborations and learn different ways of doing things, and good to be forced into doing something completely different than what I did as a graduate student.

So why would switching be a bad idea? Well, mostly, it’s harder. There is more logistics to deal with to get started as a postdoc — on top of the logistics of starting a job — and a lot of time spent learning new software and new organization. And it will take me quite a bit longer to be in a position where I know enough and people have enough confidence in my work to give me significant responsibilities. But all of this, I hope, is transitory.

In the end, neither experiment is the dark side. They do compete with each other — as intended, to keep everyone working hard — but they’re more like opposing sports teams than opposite sides of the Force. You may despise the team across town much of the time, but without them you couldn’t play baseball. And once in a while, players get traded.

In the meantime, I’m cutting back on other things I enjoy working on but don’t have time for. That means no more regular volunteering for the Prison University Project, where I’ve been teaching basic math to prison inmates for most of the time I’ve been in grad school. It also means, a bit more slowly, handing off my organizational responsibilities for the math class there, which have been great practice for academic politics. Finally, it means giving up on blogging for a while; I won’t be able to write more about finishing my thesis, or the transition to my new job — at least not until it’s started and I know if I have time!

So you won’t see me here for several months at the very least. In the meantime, you can see my occasional professional opinions and random comments about coffee on my Twitter feed, @sethzenz. I hope to be back as soon as I can, but in any case, thanks for reading!

[Ahem] Magnets, how do they work?

Magnets are crucial to our entire endeavor at the LHC. They accelerate the protons and keep them on the right path, and they curve particles so we can measure their momenta in the experiments’ tracking detectors. Yet I actually know very little about the materials science behind building the advanced magnets my work relies on — probably an unavoidable consequence of the specialization required for modern science, but still a shame. I had actually considered using this quote for the experimental apparatus chapter(s) of my thesis, but I rather disapprove of the lines following:

And I don’t wanna talk to a scientist
Y’all [individuals of dubious moral character] lying, and getting me [upset]

I don’t disapprove because I have to put awkward euphemisms whenever they swear, but rather because, well, we aren’t lying. Every line of explanation in every textbook represents years of effort and careful verification, and we know how magnets work perfectly well. And knowing how magnets work makes them more wonderful, not less!

Anyway, the Insane Clown Posse’s question was recently repeated in a round of Twitter questions for New York Mayor Michael Bloomberg, who replied:

Well, as you probably know, everything is made up of atoms, and atoms have electrons, usually in pairs orbiting around them, and they create mini-magnetic fields. But the two electrons spin in orbit, as pairs, spin in opposite directions. They cancel out each other . . . But magnetic materials aren’t in pairs. So the spins don’t cancel out each other. And if there’s enough of them, it creates a magnetic field.

He left out a few details, of course, but that’s a good conceptual start. So, kudos to Mayor Bloomberg, for knowing how magnets work and taking the time to explain it a bit. The world is full of miracles, but with hard work we can understand a lot of them: that’s what being a scientist is all about.

From the perspective of readers here, though, I’m not going anywhere fast. I will be switching universities, and (gasp!) switching experiments, but in neither case am I going far enough that I fall outside the subject of this blog. I didn’t mean “far enough” too literally, but the distances are about 2500 miles and 5.3 miles respectively.

This gives me lots to write about, so let me start with my thesis. It’s not that much fun. It has two parts: background stuff I have to look up, and describing my analysis in more detail. But it is, I have to admit, probably all wortwhile. The former part is all stuff that’s relevant to my work and I ought to be able to describe off the top of my head — and in fact, I usually can, but not with all the numbers and equations and details just right. The latter part is a good chance to really document what I did in my analysis, which is information that might not be public elsewhere. Maybe someone, someday, will want to look up what I did. Maybe I’ll look it up myself out of nostalgia. Once I get my thesis written at last, though, one thing I’m sure of: I won’t look at it again for a while. I’m already ready to move on!

While I’m working on that, here are some goodies from my thesis. The output of a helpful script I wrote:

Seth-Zenzs-MacBook-Pro:~ sethzenz$ python scripts/thesistimeleft.py

You have 75 days left to file your thesis!!!

It recalculates every day. I could make it send me automatic emails, if I really want to make myself nervous.

And here’s a bit of my (first draft) introduction, which tries to explain how my work fits into the overall context of the LHC program:

The Large Hadron Collider (LHC) was built to produce new particles and rare interactions at a high rate, but its first and foremost byproduct is sprays of low-energy hadrons. At its full design capacity, the LHC will cross proton bunches 31.6 million times per second at each interaction point, with an average of about 20 proton-proton collisions per crossing. Most of these collisions will be “soft” interactions, with relatively little energy exchanged and the outgoing hadrons having relatively little momentum perpendicular to the beam axis. These interactions are described in principle by Quantum Chromodynamics (QCD), the quantum field theory of the strong interactions. In practice, however, they are the most difficult to understand, because the theory becomes non-perturbative at low energies. Predictions can only be made via approximations and phenomenological models. This difficulty with low-energy strong interactions appears even in interactions that are initially well-described by perturbation theory. Outgoing high-energy quarks and gluons quickly “clothe” their strong color charge by evolving into jets of lower-energy hadrons, a process that again requires approximation and modelling.

The LHC’s general-purpose experiments, ATLAS and CMS, are equipped with multi-stage trigger systems that select against these common processes, for example by identifying leptons and missing energy produced in electroweak interactions. However, low-energy QCD still has a significant impact on the physics program in several areas. With so many collisions in each crossing, the most interesting collisions will have many low-energy collisions whose signals in the detector overlap with the objects of interest. In order for their effects to be subtracted, these features of these pileup collisions must be known quantitatively. The evolution of high-energy hadronic jets must also be well-understood. This is partially to account for their contribution as pileup events, but their energy must also be calibrated so they can be studied in their own right. Although even very high-energy jets are relatively common at the LHC, they can also serve as signatures of the decay of new particles.

The quantitative investigation of low-energy QCD is thus a foundational element of the LHC program, which will inform the studies and discoveries of the coming years. Initial low-energy QCD measurements have divided the problem between low-energy events and the study of higher-energy jet properties. In the former case, inclusive charged particle distributions are produced from events identified using a “minimum bias” trigger. In the latter case, higher-energy jets are triggered and studied using the calorimeter system built for the purpose.

This work focuses on the additional information to be gained in the case that the two issues are not-so easily factorized, by studying the emergence of low-energy jets from soft interactions. Particles are identified using the methods of the lowest-energy measurements, but grouped together into jets according to the algorithms used to study jets at higher energies. Low-momentum jets and their properties are measured using the ATLAS Inner Detector, the component of the ATLAS experiment that tracks charged particles, in events identified using the ATLAS Minimum Bias Trigger Scintilators.

That’s very unlikely to be final, but in any case that gives you a picture of the sort of thing I’m working on.

The ATLAS and CMS experiments play to the strength of their detectors, which both have very powerful and hermetic energy measuring capability. This allows them to measure jets of particles that emerge from collisions. Jets are formed as the basic constituents of nuclear matter, quarks and gluons, fly away from the collision point. In proton collisions, jets usually appear in pairs, emerging back to back. However, in heavy ion collisions the jets interact in the tumultuous conditions of the hot dense medium. This leads to a very characteristic signal, known as jet quenching, in which the energy of the jets can be severely degraded, signalling interactions with the medium more intense than ever seen before. Jet quenching is a powerful tool for studying the behaviour of the plasma in detail.

“ATLAS is the first experiment to report direct observation of jet quenching,” said ATLAS Spokesperson Fabiola Gianotti. “The excellent capabilities of ATLAS to determine jet energies enabled us to observe a striking imbalance in energies of pairs of jets, where one jet is almost completely absorbed by the medium. It’s a very exciting result of which the Collaboration is proud, obtained in a very short time thanks in particular to the dedication and enthusiasm of young scientists.”

I worked on this paper for only 4 or 5 days, meeting with a few other enthusiastic “young scientists” on Skype almost continuously in order to do a cross-check of the results using track jets. (I explained a bit about track jets when I explained my thesis topic a while back. This work wasn’t part of my thesis, but it’s an application of the same technique.) Our work ended up becoming three sentences on page four of the paper:

The analysis was independently corroborated by a study of “track jets”, reconstructed with [Inner Detector] tracks of p_T > 4 GeV using the same jet algorithms. . . . A similar asymmetry effect is also observed with track jets. The jet energy scale and underlying event subtraction were also validated by correlating calorimeter and track-based jet measurements.

My immediate partners on this part of the work were David Miller and Zach Marshall, but I don’t want you to think just because I’m blogging about it that we were the critical people working on this paper. Obviously there were many other cross-checks of this result, all done through equally hard work in an equally short time, and then of course there are the Heavy Ion experts who did the core measurement itself! We only did a small part of it all, but we worked hard and we’re proud.

ATLAS as a whole is proud. We have won some and lost some this year in the competition with our colleagues at the CMS detector to make new measurements and discoveries first. This time, we won, and it’s a great feeling to end the year with.

Experimental particle physicists like me mostly work on building our equipment, keeping it working, and simulating how it responds to (relatively) ordinary particles so we can be sure that our measurements work at a basic level. Once we’re confident what we’ve measured, turning it into the answer to a question about fundamental physics — have we seen a new particle, or are we sure it isn’t there? — is the conceptual “last step.”  Meanwhile, theorists spend a lot of their time working out the details of things that can’t be measured directly; their “last step” is to figure out how their ideas can be observed.  We all have calculations to do, but theorists often work on beautiful math, while experimentalists usually do ugly statistics.

But even if we think and work differently, at some point we have to interact: we experimentalists look for the things the theorists invent, and they change their theories based on what we find.  How does that usually work?  Through journal articles and online preprints.  Theorists publish their multitude of ideas.  Experiments search for a set of physics signatures that correspond to those ideas, and publish what we find.  (By “physics signature” I mean some particular combination of particles; for example, one of several signatures of the top quark would be a bottom quark jet, an electron or muon, and missing energy from an escaping neutrino.)  Assuming we don’t find anything in an experimental search, there are two categories of things that experimentalists can publish: one is a precise specification of the physics signature we looked for, and the other is a limit on some particular theory that we tried to find.

If we put a limit on a theory we searched for, that is very helpful information for whoever worked on that particular theory, but what about someone with a different theory?  What about somebody who comes up with a new theory later?  How can he or she figure out whether that theory has been excluded by previous searches?   What we discussed at the Theory-Experiment lunch meeting yesterday was some of the ways to answer that question.

A note to my colleagues in theory and experiment: if there are any ideas I have misrepresented or omitted, it is probably due to my own ignorance.  Please leave comments and I will update my post.  Oh, and while I’m giving caveats: any opinions on any of these options are my own.

And now, on to the ways we might answer our hypothetical theorist’s question: I’ve invented a new theory, what do I do now?

Option 1. Publish your theory, then see if anyone on an experiment is interested in looking for it.  This is the current default system.  The problem is that doing a new analysis from scratch is hard, and there are way too many theories to look for!

Option 2. Realize that your theory “looks kinda like” another theory that has been searched for.  Suggest to experimentalists that they rerun their analysis, but look for your theory instead.  So they have to rerun the parts of the analysis that simulate what they’re looking for, but they don’t have to rerun any of the backgrounds or change any of the details of the signature they’re searching for.  A background is something ordinary from the Standard Model of Particle Physics that manages to mimic the signal you’re hoping to get from your new theory — and figuring them out is actually the bulk of the work, because they often depend on very rare mismeasurements in the detector that are hard to simulate.  If experimenters can avoid redoing that by reusing most of the analysis, then they might have time to look, even if it probably won’t be as effective a search as if it had been custom-designed for your theory.

A framework called RECAST was released this past week, to provide a means of streamlining requests to “recast” analyses in this manner.  I don’t think the LHC experiments are likely to use it in the short term, for a few reasons.  First, we’re really rushing around to find new stuff, rather than apply existing measurements to old theories.  Second, we’re still figuring out our procedures for completing and in-house review and approval ordinary analyses; figuring out how to organize and approve recast analyses would be more on top of a process that’s already tough.  Older experiments like the ones at the Tevatron accelerator at Fermilab, where things are settled down and there’s an effort to take the years of accumulated data as far as possible, are more likely to have time to take a close look at such a system.

Option 3. You could run a version of the analysis yourself, using something like Pretty Good Simulation in place of having an experiment redo the analysis.  The problem is that rough simulation, while much less computationally intensive, doesn’t have all the details of the detector and all the experimental knowledge that went into understanding them.  That makes getting the backgrounds right hard, so although this could be useful for getting a general idea of whether your new theory should have been detected by now, it’s not going to give an exact answer.

Option 4.  If an experimental measurement did a good job of explaining exactly the signature they saw, in a way that’s fully corrected for detector effects, then maybe nobody has to redo the analysis at all! Just simulate collisions in your theory, count how many there are fitting the signature, and compare to the measurement. For example, in a search for Supersymmetry, ATLAS might publish the number of events we see with four jets above some momentum and a certain amount of missing energy. Your theory doesn’t have to be a kind of Supersymmetry for you to count jets and compare to that number. If your theory were right, would there have been a lot more events than that? The problem with this method is that experimentalists don’t always give such straightforward conditions for theorists to compare with — sometimes we can’t, because in some cases we have to assume things about what we looking for in order to correct our data. Still, this is definitely something to strive for.

Option 5.  Encourage experimentalists to put limits on more general models. When experimentalists compare theory to data, one of the problems that arises is that we pick some theory and limit what the masses of its particles could be — but that theory is too specific. But if there were simple models that could be translated into many different theories, then potentially even new theories could be tied back to limits that had already been set. The LHC New Physics Working Group is working to define a list of simple models. The problem with this is that theorists are good at coming up with new ideas that don’t fit nearly in existing lists of models: in fact, some folks specialize in inventing things that are hard to look for!

In conclusion, I think a lot of us are hoping that all these details are premature for the LHC, and that new physics will jump right out at us! If that happens, it will give us a lot more clarity on what to look for, how to look for it, and what kind of theories to compare with what kind of data — and, of course, a lot of excitement! Until then, I don’t think there’s a perfect answer to our question.