Next stop: tth

This was a summer of brutally dashed hopes for a quick discovery of many fundamental particles that we were imagining. For the time being we need  to focus on the ones that actually exist, such as the Higgs boson. In the Run-1 of the LHC, the Higgs existence and identity were firmly established,  while its mass and basic properties were measured. The signal was observed with large significance in 4 different decay channels (γγ, ZZ*, WW*, ττ), and two different production modes (gluon fusion, vector-boson fusion) have been isolated.  Still, there remains many fine details to sort out. The realistic goal for the Run-2 is to pinpoint the following Higgs processes:

• (h→bb): Decays to b-quarks.
• (Vh): Associated production with W or Z boson. 
• (tth): Associated production with top quarks. 

It seems that the last objective may be achieved quicker than expected. The tth production process is very interesting theoretically, because its rate is proportional to the (square of the) Yukawa coupling between the Higgs boson and top quarks. Within the Standard Model, the value of this parameter is known to a good accuracy, as it is related to the mass of the top quark. But that relation can be  disrupted in models beyond the Standard Model, with the two-Higgs-doublet model and composite/little Higgs models serving as prominent examples. Thus, measurements of the top Yukawa coupling will provide a crucial piece of information about new physics.

In the Run-1, a not-so-small signal of tth production was observed by the ATLAS and CMS collaborations in several channels. Assuming that Higgs decays have the same branching fraction as in the Standard Model, the tth signal strength normalized to the Standard Model prediction was estimated as

At face value, a strong evidence for the tth production was obtained in the Run-1! This fact was not advertised by the collaborations because the measurement is not clean due to a large number of top quarks produced by other processes at the LHC. The tth signal is thus a small blip on top of a huge background, and it's not excluded that some unaccounted for systematic errors are skewing the measurements. The collaborations thus preferred to play it safe, and wait for more data to be collected.

In the Run-2 with 13 TeV collisions the tth production cross section is 4-times larger than in the Run-1, therefore the new data are coming at a fast pace. Both ATLAS and CMS presented their first Higgs results in early August, and the tth signal is only getting stronger.  ATLAS showed their measurements in the γγ, WW/ττ, and bb final states of Higgs decay, as well as their combination:

Most channels display a signal-like excess, which is reflected by the Run-2 combination being 2.5 sigma away from zero. A similar picture is emerging in CMS, with 2-sigma signals in the γγ and WW/ττ channels. Naively combining all Run-1 and and Run-2 results one then finds

At face value, this is a discovery! Of course, this number should be treated with some caution because, due to large systematic errors, a naive Gaussian combination may not represent very well the true likelihood. Nevertheless, it indicates that, if all goes well, the discovery of the tth production mode should be officially announced in the near future, maybe even this year.

Should we get excited that the measured tth rate is significantly larger than Standard Model one? Assuming  that the current central value remains, it would mean that  the top Yukawa coupling is 40% larger than that predicted by the Standard Model. This is not impossible, but very unlikely in practice. The reason is that the top Yukawa coupling also controls the gluon fusion - the main Higgs production channel at the LHC - whose rate is measured to be in perfect agreement with the Standard Model.  Therefore, a realistic model that explains the large tth rate would also have to provide negative contributions to the gluon fusion amplitude, so as to cancel the effect of the large top Yukawa coupling. It is possible to engineer such a cancellation in concrete models, but  I'm not aware of any construction where this conspiracy arises in a natural way. Most likely, the currently observed excess is  a statistical fluctuation (possibly in combination with  underestimated theoretical and/or  experimental errors), and the central value will drift toward μ=1 as more data is collected. 

After the hangover

The loss of the 750 GeV diphoton resonance is a big blow to the particle physics community. We are currently going through the 5 stages of grief, everyone at their own pace, as can be seen e.g. in this comments section. Nevertheless, it may already be a good moment to revisit the story one last time, so as  to understand what went wrong.

In the recent years, physics beyond the Standard Model has seen 2 other flops of comparable impact: the faster-than-light neutrinos in OPERA, and the CMB tensor fluctuations in BICEP.  Much as the diphoton signal, both of the above triggered a binge of theoretical explanations, followed by a massive hangover. There was one big difference, however: the OPERA and BICEP signals were due to embarrassing errors on the experiments' side. This doesn't seem to be the case for the diphoton bump at the LHC. Some may wonder whether the Standard Model background may have been slightly underestimated,  or whether one experiment may have been biased by the result of the other... But, most likely, the 750 GeV bump was just due to a random fluctuation of the background at this particular energy. Regrettably, the resulting mess cannot be blamed on experimentalists, who were in fact downplaying the anomaly in their official communications. This time it's the theorists who  have some explaining to do.

Why did theorists write 500 papers about a statistical fluctuation?  One reason is that it didn't look like one at first sight. Back in December 2015, the local significance of the diphoton  bump in ATLAS run-2 data was 3.9 sigma, which means the probability of such a fluctuation was 1 in 10000. Combining available run-1 and run-2 diphoton data in ATLAS and CMS, the local significance was increased to 4.4 sigma.  All in all, it was a very unusual excess, a 1-in-100000 occurrence! Of course, this number should be interpreted with care. The point is that the LHC experiments perform gazillion different measurements, thus they are bound to observe seemingly unlikely outcomes in a small fraction of them. This can be partly taken into account by calculating the global significance, which is the probability of finding a background fluctuation of the observed size anywhere in the diphoton spectrum. The global significance of the 750 GeV bump quoted by ATLAS was only about two sigma, the fact strongly emphasized by the collaboration.  However, that number can be misleading too.  One problem with the global significance is that, unlike for the local one, it cannot be  easily combined in the presence of separate measurements of the same observable. For the diphoton final state we  have ATLAS and CMS measurements in run-1 and run-2,  thus 4 independent datasets, and their robust concordance was crucial  in creating the excitement.  Note also that what is really relevant here is the probability of a fluctuation of a given size in any of the  LHC measurement, and that is not captured by the global significance.  For these reasons, I find it more transparent work with the local significance, remembering that it should not be interpreted as the probability that the Standard Model is incorrect. By these standards, a 4.4 sigma fluctuation in a combined ATLAS and CMS dataset is still a very significant effect which deserves a special attention. What we learned the hard way is that such large fluctuations do happen at the LHC...   This lesson will certainly be taken into account next time we encounter a significant anomaly.

Another reason why the 750 GeV bump was exciting is that the measurement is rather straightforward.  Indeed, at the LHC we often see anomalies in complicated final states or poorly controlled differential distributions, and we treat those with much skepticism.  But a resonance in the diphoton spectrum is almost the simplest and cleanest observable that one can imagine (only a dilepton or 4-lepton resonance would be cleaner). We already successfully discovered one particle this way - that's how the Higgs boson first showed up in 2011. Thus, we have good reasons to believe that the collaborations control this measurement very well.

Finally, the diphoton bump was so attractive because theoretical explanations were  plausible.  It was trivial to write down a model fitting the data, there was no need to stretch or fine-tune the parameters, and it was quite natural that the particle first showed in as a diphoton resonance and not in other final states. This is in stark contrast to other recent anomalies which typically require a great deal of gymnastics to fit into a consistent picture.   The only thing to give you a pause was the tension with the LHC run-1 diphoton data, but even that became  mild after the Moriond update this year.

So we got a huge signal of a new particle in a clean channel with plausible theoretic models to explain it...  that was a really bad luck.  My conclusion may not be shared by everyone but I don't think that the theory community committed major missteps  in this case.  Given that for 30 years we have been looking for a clue about the fundamental theory beyond the Standard Model, our reaction was not disproportionate once a seemingly reliable one had arrived.  Excitement is an inherent part of physics research. And so is disappointment, apparently.

There remains a question whether we really needed 500 papers...   Well, of course not: many of  them fill an important gap.  Yet many are an interesting read, and I personally learned a lot of exciting physics from them.  Actually, I suspect that the fraction of useless papers among the 500 is lower than for regular daily topics.  On a more sociological side, these papers exacerbate the problem with our citation culture (mass-grave references), which undermines the citation count as a means to evaluate the research impact.  But that is a wider issue which I don't know how to address at the moment.

Time to move on. The ICHEP conference is coming next week, with loads of brand new results based on up to 16 inverse femtobarns of 13 TeV LHC data.  Although the rumor is that there is no new exciting  anomaly at this point, it will be interesting to see how much room is left for new physics. The hope lingers on, at least until the end of this year.

In the comments section you're welcome to lash out on the entire BSM community - we made a wrong call so we deserve it. Please, however, avoid personal attacks (unless on me). Alternatively, you can also give us a hug :) 

Game of Thrones: 750 GeV edition

The 750 GeV diphoton resonance has made a big impact on theoretical particle physics. The number of papers on the topic is already legendary, and they keep coming at the rate of order 10 per week. Given that the Backović model is falsified, there's no longer a theoretical upper limit.  Does this mean we are not dealing with the classical ambulance chasing scenario? The answer may be known in the next days.

So who's leading this race?  What kind of question is that, you may shout, of course it's Strumia! And you would be wrong, independently of the metric.  For this contest, I will consider two different metrics: the King Beyond the Wall that counts the number of papers on the topic, and the Iron Throne that counts how many times these papers have been cited.

In the first category,  the contest is much more fierce than one might expect: it takes 8 papers to be the leader, and 7 papers may not be enough to even get on the podium!  Among the 3 authors with 7 papers the final classification is decided by trial by combat the citation count.  The result is (drums):

Citations, tja...   Social dynamics of our community encourages referencing all previous work on the topic, rather than just the relevant ones, which in this particular case triggered a period of inflation. One day soon citation numbers will mean as much as authorship in experimental particle physics. But for now the size of the h-factor is still an important measure of virility for theorists. If the citation count rather the number of papers is the main criterion, the iron throne is taken by a Targaryen contender (trumpets):

This explains why the resonance is usually denoted by the letter S.

Update 09.08.2016. Now that the 750 GeV excess is officially dead, one can give the final classification. The race for the iron throne was tight till the end, but there could only be one winner:

As you can see, in this race the long-term strategy and persistence proved to be more important than pulling off a few early victories.  In the other category there have also been  changes in the final stretch: the winner added 3 papers in the period between the un-official and official announcement of the demise of the 750 GeV resonance. The final standing are:

Congratulations for all the winners.  For all the rest, wish you more luck and persistence in the next edition,  provided it will take place.

Black hole dark matter

The idea that dark matter is made of primordial black holes is very old but has always been in the backwater of particle physics. The WIMP or asymmetric dark matter paradigms are preferred for several reasons such as calculability, observational opportunities, and a more direct connection to cherished theories beyond the Standard Model. But in the recent months there has been more interest, triggered in part by the LIGO observations of black hole binary mergers. In the first observed event, the mass of each of the black holes was estimated at around 30 solar masses. While such a system may well be of boring astrophysical origin, it is somewhat unexpected because typical black holes we come across in everyday life are either a bit smaller (around one solar mass) or much larger (supermassive black hole in the galactic center). On the other hand, if the dark matter halo were made of black holes, scattering processes would sometimes create short-lived binary systems. Assuming a significant fraction of dark matter in the universe is made of primordial black holes, this paper estimated that the rate of merger processes is in the right ballpark to explain the LIGO events.

Primordial black holes can form from large density fluctuations in the early universe. On the largest observable scales the universe is incredibly homogenous, as witnessed by the uniform temperature of the Cosmic Microwave Background over the entire sky. However on smaller scales the primordial inhomogeneities could be much larger without contradicting observations.  From the fundamental point of view, large density fluctuations may be generated by several distinct mechanism, for example during the final stages of inflation in the waterfall phase in the hybrid inflation scenario. While it is rather generic that this or similar process may seed black hole formation in the radiation-dominated era, severe fine-tuning is required to produce the right amount of black holes and ensure that the resulting universe resembles the one we know.

All in all, it's fair to say that the scenario where all or a significant fraction of  dark matter  is made of primordial black holes is not completely absurd. Moreover, one typically expects the masses to span a fairly narrow range. Could it be that the LIGO events is the first indirect detection of dark matter made of O(10)-solar-mass black holes? One problem with this scenario is that it is excluded, as can be seen in the plot.  Black holes sloshing through the early dense universe accrete the surrounding matter and produce X-rays which could ionize atoms and disrupt the Cosmic Microwave Background. In the 10-100 solar mass range relevant for LIGO this effect currently gives the strongest constraint on primordial black holes: according to this paper they are allowed to constitute  not more than 0.01% of the total dark matter abundance. In astrophysics, however, not only signals but also constraints should be taken with a grain of salt.  In this particular case, the word in town is that the derivation contains a numerical error and that the corrected limit is 2 orders of magnitude less severe than what's shown in the plot. Moreover, this limit strongly depends on the model of accretion, and more favorable assumptions may buy another order of magnitude or two. All in all, the possibility of dark matter made of  primordial black hole in the 10-100 solar mass range should not be completely discarded yet. Another possibility is that black holes make only a small fraction of dark matter, but the merger rate is faster, closer to the estimate of this paper.

Assuming this is the true scenario, how will we know? Direct detection of black holes is discouraged, while the usual cosmic ray signals are absent. Instead, in most of the mass range, the best probes of primordial black holes are various lensing observations. For LIGO black holes, progress may be made via observations of fast radio bursts. These are strong radio signals of (probably) extragalactic origin and millisecond duration. The radio signal passing near a O(10)-solar-mass black hole could be strongly lensed, leading to repeated signals detected on Earth with an observable time delay. In the near future we should observe hundreds of such repeated bursts, or obtain new strong constraints on primordial black holes in the interesting mass ballpark. Gravitational wave astronomy may offer another way.  When more statistics is accumulated, we will be able to say something about the spatial distributions of the merger events. Primordial black holes should be distributed like dark matter halos, whereas astrophysical black holes should be correlated with luminous galaxies. Also, the typical eccentricity of the astrophysical black hole binaries should be different.  With some luck, the primordial black hole dark matter scenario may be vindicated or robustly excluded  in the near future.

See also these slides for more details. 

CMS: Higgs to mu tau is going away

One interesting anomaly in the LHC run-1 was a hint of Higgs boson decays to a muon and a tau lepton. Such process is forbidden in the Standard Model by the conservation of  muon and tau lepton numbers. Neutrino masses violate individual lepton numbers, but their effect is far too small to affect the Higgs decays in practice. On the other hand, new particles do not have to respect global symmetries of the Standard Model, and they could induce lepton flavor violating Higgs decays at an observable level. Surprisingly, CMS found a small excess in the Higgs to tau mu search in their 8 TeV data, with the measured branching fraction Br(h→τμ)=(0.84±0.37)%.  The analogous measurement in ATLAS is 1 sigma above the background-only hypothesis, Br(h→τμ)=(0.53±0.51)%. Together this merely corresponds to a 2.5 sigma excess, so it's not too exciting in itself. However, taken together with the B-meson anomalies in LHCb, it has raised hopes for lepton flavor violating new physics just around the corner.  For this reason, the CMS excess inspired a few dozen of theory papers, with Z' bosons, leptoquarks, and additional Higgs doublets pointed out as possible culprits.

Alas, the wind is changing. CMS made a search for h→τμ in their small stash of 13 TeV data collected in 2015. This time they were hit by a negative background fluctuation, and they found Br(h→τμ)=(-0.76±0.81)%. The accuracy of the new measurement is worse than that in run-1, but nevertheless it lowers the combined significance of the excess below 2 sigma. Statistically speaking, the situation hasn't changed much,  but psychologically this is very discouraging. A true signal is expected to grow when more data is added, and when it's the other way around it's usually a sign that we are dealing with a statistical fluctuation...

So, if you have a cool model explaining the h→τμ  excess be sure to post it on arXiv before more run-2 data is analyzed ;)

Plot for Weekend: new limits on neutrino masses

This weekend's plot shows the new limits on neutrino masses from the KamLAND-Zen experiment:

KamLAND-Zen is a group of buddhist monks studying a balloon filled with the xenon isotope Xe136. That isotope has a very long lifetime, of order 10^21 years, and undergoes the lepton-number-conserving double beta decay Xe136 → Ba136 2e- 2νbar. What the monks hope to observe is the lepton violating neutrinoless double beta decay Xe136 → Ba136+2e, which would show as a peak in the invariant mass distribution of the electron pairs near 2.5 MeV. No such signal has been observed, which sets the limit on the half-life for this decay at T>1.1*10^26 years.

The neutrinoless decay is predicted to occur if neutrino masses are of Majorana type, and the rate can be characterized by the effective mass Majorana mββ (y-axis in the plot). That parameter is a function of the masses and mixing angles of the neutrinos. In particular it depends on the mass of the lightest neutrino (x-axis in the plot) which is currently unknown. Neutrino oscillations experiments have precisely measured the mass^2 differences between neutrinos, which  are roughly (0.05 eV)^2 and (0.01 eV)^2. But oscillations are not sensitive to the absolute mass scale; in particular, the lightest neutrino may well be massless for all we know.  If the heaviest neutrino has a small electron flavor component, then we expect that the mββ parameter is below 0.01 eV.  This so-called normal hierarchy case is shown as the red region in the plot, and is clearly out of experimental reach at the moment. On the other hand, in the inverted hierarchy scenario (green region in the plot), it is the two heaviest neutrinos that have a significant electron component. In this case,  the effective Majorana mass mββ is around 0.05 eV.  Finally, there is also the degenerate scenario (funnel region in the plot) where all 3 neutrinos have very similar masses with small splittings, however this scenario is now strongly disfavored by cosmological limits on the sum of the neutrino masses (e.g. the Planck limit Σmν < 0.16 eV).

As can be seen in the plot, the results from KamLAND-Zen, when translated into limits on the effective Majorana mass, almost touch the inverted hierarchy region. The strength of this limit depends on some poorly known nuclear matrix elements (hence the width of the blue band). But even in the least favorable scenario future, more sensitive experiments should be able to probe that region. Thus, there is a hope  that within the next few years we may prove the Majorana nature of neutrinos, or at least disfavor the inverted hierarchy scenario.

Off we go

The LHC is back in action since last weekend, again colliding protons with 13 TeV energy. The weasels' conspiracy was foiled, and the perpetrators were exemplarily electrocuted. PhD students have been deployed around the LHC perimeter to counter any further sabotage attempts (stoats are known to have been in league with weasels in the past). The period that begins now may prove to be the most exciting time for particle physics in this century.  Or the most disappointing.

The beam intensity is still a factor of 10 below the nominal one, so  the harvest of last weekend is meager 40 inverse picobarns. But the number of proton bunches in the beam is quickly increasing, and once it reaches O(2000), the data will stream at a rate of a femtobarn per week or more. For the nearest future, the plan is to have a few inverse femtobarns on tape by mid-July, which would roughly double the current 13 TeV dataset. The first analyses of this chunk of data  should be presented around the time of the  ICHEP conference in early August. At that point we will know whether the 750 GeV particle is real. Celebrations will begin if the significance of the diphoton peak increases after adding the new data, even if the statistics is not enough to officially announce  a discovery. In the best of all worlds, we may also get a hint of a matching 750 GeV peak in another decay channel (ZZ, Z-photon, dilepton, t-tbar,...) which would help focus our model building. On the other hand, if the significance of the diphoton peak drops in August, there will be a massive hangover...

By the end of October, when the 2016 proton collisions are scheduled to end, the LHC hopes to collect some 20 inverse femtobarns of data. This should already give us a rough feeling of new physics within the reach of the LHC. If a hint of another resonance is seen at that point, one will surely be able to confirm or refute it with the data collected in the following years.  If nothing is seen... then you should start telling yourself that condensed matter physics is also sort of fundamental,  or that systematic uncertainties in astrophysics are not so bad after all...  In any scenario, by December, when first analyses of the full  2016 dataset will be released,  we will know infinitely more than we do today.

So fasten your seat belts and get ready for a (hopefully) bumpy ride. Serious rumors should start showing up on blogs and twitter starting from July.