10 Inverse Femtobarns: Celebration Time At Fermilab!
    By Tommaso Dorigo | December 21st 2010 03:12 PM | 26 comments | Print | E-mail | Track Comments
    About Tommaso

    I am an experimental particle physicist working with the CMS experiment at CERN. In my spare time I play chess, abuse the piano, and aim my dobson...

    View Tommaso's Profile
    As sure as death and taxes, and as timely as a Swiss watch, the Tevatron collider never ceases to awe us. Well into its twentysixth year of life, the aged and celebrated proton-antiproton collider sitting just a few meters underground in the west Chicago suburbs hit the mark of 10 inverse femtobarns of collisions delivered to the core of the CDF and DZERO detectors.

    10 inverse femtobarns! Ten inverse femtobarns of proton-antiproton collisions is a HELL of a lot of them. Plus, you should multiply that number by two, since the same number of collisions happened inside two different collision areas -those manned by the two competing collaborations.

    If you need a measure of what 20 inverse femtobarns of collisions mean, walk to the beach. A mile-long stretch of sand, extending 100 meters into the land and with a depth of one meter, contains roughly the same number of sand grains as that number of collisions. Absolutely awesome. On the other hand, if you were downplaying it, you could shrug your shoulders and claim that the total mass of annihilated protons would just make a picogram of matter. Protons are tiny, but we need lots of them to study subnuclear matter!

    The graph below shows very clearly the progress of Tevatron data collection in the course of the last ten years. In red you can see the integrated luminosity, which should be read off on the axis on the right; in green the thin bars show the weekly integrated luminosity, which describes how the accumulation rate has increased for many years, and now stabilized to roughly 55 inverse picobarns per week.

    Another way to picture what do 20 inverse femtobarns of data mean is to make simple calculations. Take the Z boson, the carrier of neutral weak currents that Carlo Rubbia in 1983 extracted with a lot of sweating from the lower energy and lower intensity SppS collider (five events in total, if I remember correctly, were all it took him to claim the discovery of that new boson, which together with the W boson signal allowed him to secure a Nobel Prize in Physics). The Z boson at the Tevatron has a production cross section  of about 8 nanobarns: since 2002 the Tevatron has therefore created in the core of CDF and DZERO a total ofsuch particles. 

    Another example ? By using a temporary suspension of disbelief, assume that the Standard Model is correct, and that the Higgs boson exists and has a mass of 120 GeV. In that case, as many as 20,000 such particles have already been created inside the detectors. Twenty thousand!! How can such a signal have escaped ? Unfortunately, the Higgs is a quite eclectic particle, and only few of its possible decay modes are suitable for its clear observation amidst huge backgrounds from less interesting processes capable of mimicking its disintegration...

    The Tevatron might end its operations in less than one year, or continue to run until the end of 2014. Its future is in the hands of philanthropists. If you are sitting on a pile of cash, consider donating 150 million dollars for a just cause.


    No claim is really outrageous at this point.   The Standard Model predicts how many times a year Fermi should expect to see the Higgs boson and how often you should see particle signals that can mimic it.

    At 10 inverse femtobarns, if the Higgs exists, it is in there somewhere waiting to be found.  Kudos to you all for predicting 18 months ago you would reach that number by the end of 2010.  It's good a lab is making projections and meeting them.
    Daniel de França MTd2
    If Higgs has a mass of 128GeV, what is the expected deviation from the background in sigma with this much of events?
    Hi Daniel,

    I am not sure whether you are asking what the Tevatron experiments see, or what they could see given data and analysis techniques. The word "expected" implies the latter. If so, please have a look at the graph below.

    The figure shows a statistical discriminant (on the y axis), the log-likelihood-ratio between the Background-only and Signal-plus-Background hypotheses, given the data and the searches, as a function of the Higgs boson mass.
    The dashed curves show the expected value that the LLR would take if the Higgs was not (black) or was (red) there. A "Brazilian belt" is constructed around the B-only hypothesis to show how much variation, at 1- and 2- sigma (green or yellow areas), can be expected from the data due to statistical fluctuations. The thick black line is what we observe in the data.

    Now, at 128 GeV, the data is very signal-like, since the black curve overlaps the red dashes. The background-only hypothesis, on the other hand, is only about 1.2 standard deviations away there.
    The separation between the black and red dashes illustrates how discriminating are the tevatron analyses at a particular mass point. The most discriminating point is 165 GeV: there, if the Higgs were there, the Tevatron would have already obtained a 3-sigma evidence, on average. 128 GeV, on the other hand, is about the least discriminting point -the hardest for the Tevatron to investigate. 10/fb will not be enough (the plot was made for 6/fb) to even reach a 2-sigma separation.

    Daniel de França MTd2
    128GeV was not a random value that I chose. 126-128 is the most expected value for the Higgs if gravity is renormalizable.
    So, how long will it take to see a sigma 3 and a sigma 5 signal on LHC?

    "If gravity is renormalizable"? It's not, so forget that prediction.

    Daniel de França MTd2
    Non perturbatively renormalizable. 
    So Tommaso, any idea of how my barns will be necessary or the forecast of when 128GeV that it will be achieved on CMS or ATLAS?
    "Non perturbatively renormalizable."? It's not, so forget that prediction.
    (Unless you like theoretical ideas of the kind "and-then-a-miracle-happens")

    Daniel de França MTd2
    Yes, I am of that kind. And I won`t answer any more of your posts on this thread because you are here just to troll.
    Please Tommaso, what is the luminosity and forecast for that.
    Sorry Daniel, I wrote two posts on CMS and ATLAS forecasts no more than two months ago. Please search them in this blog and you will find them.
    Daniel de França MTd2
    Your forecasts were for exclusion of higgs at LHC or the the best signal for Tevatron. But there was not an explicit discussion on how much luminosity would be required to find a sigma 5 for higgs at 14TeV

    Page 14, here:
    For Atlas, there is an asymptotic curve about 130GeV, so I cannot determine what is the luminosity.
    For CMS, almost all of them are above 10fb-1

    On page 29, several channels are above 30fb-1

    So, my question was when, not with how much luminosity...

    Perhaps, never?

    Sure, perhaps never. As if the LHC ends in a boom, or if you die, or if I die. This kind of exercise is kind of silly.
    I expect that a 128 GeV Higgs will be conclusively discovered at the LHC in March, 2013.
    Daniel de França MTd2
    I was serious when I said never I meant that it would require decades of collecting luminosity at 14TeV. Wasn't 1fb-1 all of what was to be collected next year at 7TeV?

    Daniel, the forecast is 6/fb at 8 TeV in 2011. This would be enough to understand where the Higgs is sitting.
    I hope this is not too stupid a question, but how many inverse barns would correspond to one proton-antiproton collision?
    Robert H. Olley / Quondam Physics Department / University of Reading / England
    I was just writing in my thesis that the total inelastic cross section at Tevatron (proton antiproton collisions at a center-of-mass energy of 1.96 TeV) is about 0,06 barns. The total number of inelastic collisions per second N is the cross section (sigma) times the instantaneous luminosity (L).

    N = 1, sigma = 0.06 barns, so L = 17 barn ^ -1.

    If we want to ask ourselves how many inelastic collisons happen per second at Tevatron when we are at peak instantaneous luminosity of 2x10^32 cm^-2s^-1, we transform this in 2x10^8 barn^-1 s^-1.

    sigma = 0.06 barn, L=2x10^8 barn-1, so N = 12 millions collisions per second. Now, this is not consistent with the number of about 2.5 million collisions per second that take place at Tevatron, both elastic and inalestic I did something wrong in the calculation. Could someone correct this?


    Quite a legal question Robert.
    Proton-antiproton collisions at 2 TeV energy have a total cross section of about 80 millibarns. So one of them corresponds to 1/0.08=12.5 inverse barns.

    Assuming mH = 201 GeV/c2, how many Higgses shoud have been produced at the Tevatron by now with an integated luminosity of 10 inverse femtobarns? And how many H -> ZZ -> µµµµ would one expect to see?

    Ok this is a bit more complicated. I assume you are triggered by the fact that CMS saw a ZZ->4mu event with a mass of 201 GeV. Let's see what comes up for the Tevatron experiments.

    At the Tevatron, the total cross section for H production is about 247 femtobarns, as you gather from table 1 in this paper, for instance. This means that in 10/fb one experiment would have got NH = 2470 produced Higgs bosons, in total.

    At 200 GeV the H-->ZZ branching fraction is BHZZ = 0.2533 (also in table 1 of the reference above). On the other hand, the Z branching fraction into dimuon pairs is BZmm=0.03366 (see here).
    So the production times branching fraction is N(mmmm) = NH x BHZZ x BZmm^2 = 0.709 events.
    Now, you should be aware that most of these events will not leave a completely reconstructed signal in the detector. Both CDF and DZERO see muons in a central rapidity interval, which means that they do not cover the full solid angle. Let us say that the total coverage amounts to 80%, and let us also say that in that region the detectors "see" 90% of the muons from Z decay. These are rough numbers, but in any case one cannot make detailed calculations on a piece of paper: one really needs to run a simulation which correctly describes the angular distribution of the decay products, etcetera etcetera. In any case, you want to see all four muons: this is possible only in a fraction of times equal to 0.8^4 x 0.9^4 = 0.27. This means that CDF, or DZERO, might expect to detect 0.709x0.27= roughly 0.2 events of that kind in 10/fb.

    Now please note that the experiments have not yet analyzed such amount of data - they have collected less than 10/fb, due to downtimes etcetera, and they are currently analyzing datasets of about 7/fb each. In such data they expect about 0.15 H-->mmmm events if the Higgs has that mass.

    Now also note that the standard model production of ZZ pairs has a cross section which is about 1.5 picobarns. This is roughly 25 times larger a number than the productiuon cross section of H->ZZ events (247 fb x 0.2533). So you get two things. First, that in 7/fb our back-of-the-envelope calculation predicts that from SM we expect to see roughly 25x0.15=4 four-muon events, not far from what is actually being observed. Second, that there is no chance to detect a Higgs boson there in this final state.

    Let me finally tell you how probable it is that at least one H->ZZ->4m event has shown up in either CDF or DZERO or both, in 7/fb of data. We expect 0.15 events in each experiment: this makes the average 0.3 in total. The probability to observe at least one event, if the expected value is 0.3, is calculated with an integral of the Poisson distribution of mean mu=0.3, from 1 to infinity. This amounts to 1-exp(-mu)=26%.

    Hope this helps... i will make a separate post with this information later.

    Thank you very much indeed, Tommaso, for your comprehensive answer.

    Let's celebrate,

    And let's remain silent about the fact that the LHC has already collected data about any physics above 150 GeV that surpassed the Tevatron and that the Tevatron can't match even if it were running for decades. At 35/pb, the CMS has excluded SUSY at much more compact islands than the CDF and D0 did.

    Daniel, pure gravity is spherically non-renormalizable. It means that it is non-renormalizable whatever direction you look at it.
    Daniel de França MTd2
    I will side with the crackpots on this one. Gravity is asymptotic safe:

    And soon Goroff-Sagnotti will be shown to be not spoil this property.

    BTW, is 11d SUGRA more divergent than pure gravity in 4D or not? 
    Dear Daniel, you don't have to side with them - you're one of them. ;-)
    You would have to define the question more accurately. What is the "degree of divergence" so that it's universal for all dimensions? 

    11D SUGRA is divergent because of the high dimensionality; the higher, the more divergent theory one gets. However, its compactification on torus, N=8 SUGRA in d=4, is probably perturbatively finite.

    4D pure gravity has a lower dimension which reduces some divergences, but it doesn't have the supersymmetric cancellations, which increases the divergences. At any rate, the non-renormalizable 2-loop Goroff-Sagnotti counterterm of pure 4D gravity has been known for decades.

    You can't hide this problem. There is no way to prove "pure gravity renormalizable" in any sense because you can't define pure quantum gravity at arbitrarily short distances to start with. You would have to find a "pure gravity" UV theory that flows to gravity in the IR, but this is impossible even at the level of the symmetries.

    Of course that there is a theory, namely string theory, that is defined in the UV and flows to GR at long distances. But it is not "pure gravity" in any known sense. One has to know more about its dynamics than Einstein's equations. String theory, as every consistent theory of quantum gravity (unrealistically imagining that there can exist others as well), also has to produce the right density of black hole microstates at very high masses. These scalings of entropy etc. are clearly incompatible with a local scale-invariant quantum field theory in the UV.
    Vladimir Kalitvianski
    Lubosh, you are so good at "short distances" as well as at "long ones". Then why are you so shy of commenting my post? I do not see where I am wrong ;-)  What is not "okeyable" in it?

    How can I progress without your friendly feedback? I appeal to all experts in "distances": please tell me what you mean while speaking of "distances", - elastic, inelastic, or inclusive picture? Because they are so different!
    Tommaso and Luboš

    Can I get you to comment on the views of Fred Alan Wolf?

    Sorry Henry, but I could only follow the first four minutes of the first video. I am opposed to people talking about quantum physics with a lot of hands waving around and funny faces, trying to hijack scientific truths.

    If that sounds too harsh on the guy, it's because I do not know him, it is the first time I hear him speak, and because to me he basically looked like trying to use things people do not understand (quantum physics) to foster his own religion and fit to his own agenda.

    Dear Henry,thanks for the videos - unlike Tommaso, I have watched both pieces in full (well, skipped a few minutes in the second part).

    I didn't know the guy either, and be sure that the typical theoretical physicists will tell you much more negative stuff about this sort of stuff - and about the movie. For example, Lisa Randall - who hated the movie as well - will dedicate a page of her 2011 new book, Knocking on Heaven's Door, to a friendly discussion with a producer of Bleep. She argues that he eventually changed his mind and agrees it is bullshit to make religion out of quantum physics.

    It's great that you told me because this Wolf seems to be a primary ideologue behind the movie.

    His view of quantum mechanics may have some remote similarity to the Copenhagen interpretation but most of the "details" are just plain wrong. First of all, there is no popping of the quiwf. The collapse is not a real process because the wave function is not a real object. It is just a collection of numbers meant to calculate properties of the real objects. Second, and this is related, there is no vagueness in the real world. If you ask the question what is the probability that the Moon or a cat has a particular position, A or B or C, you will find out that it is 100%. The probability that a cat is both dead and alive is zero. All such things were beautifully discussed - without any nonsense - by Sidney Coleman in Quantum Mechanics in Your Face which is also available as a video:

    Because there is no popping of the quiwf, there is no Godself because popping of the quiwf was mentioned as a necessary condition. ;-) Even if you interpret QM with popping of the quiwf, it's clear that conciousness doesn't have to play any role. Any sufficiently macroscopic object etc. has the right to decohere - one may imagine that the sharp reality is "settled" after an interaction with any macroscopic system. The size matters, the consciousness does not.

    In the latter parts, he gets into a general rant against science and everything he says is pretty much wrong - about epicycles and their alleged analogies in the current science. I don't want to run it  again. I have rigorously explained why all these people are totally deluded many times in my blog, and if you haven't noticed it in my blog or understood it, you wouldn't understand it here, either.

    Best wishes