Q&A with a LHCb expert

Questions / discussions specific to the LHCb experiment or science
Post Reply
User avatar
LHCPortal Guru
Posts: 76
Joined: Wed Nov 25, 2009 10:28 am

Q&A with a LHCb expert

Post by photino » Thu Dec 17, 2009 1:30 pm

From the chat...

[20:53] <photino> CMS has also rediscovered the pion http://cms.web.cern.ch/cms/Media/Public ... es/K0s.png
[20:53] <photino> lhcb was faster...
[20:54] <Anitusar> lhcb also found the Ks and lambda :)
[20:54] <photino> sorry that's actually the K_0 production in the plot above, from pairs of pions
[20:57] <photino> it's 1967 all over again at the CMS !! --
[20:55] <photino> anitusar: they should have a bingo where they mark off what's been seen
[20:55] <robcrazee> lol lhc bingo
[20:55] <robcrazee> need a higgs to win
[20:57] <Anitusar> ahh, the Higgs is bad for the lhcb
[20:57] <robcrazee> how so?
[20:57] <Anitusar> unless its between 125 and 140 GeV
[20:58] <Anitusar> CMS and ATLAS are build to find the Higgs, LHCb not really
[20:58] <photino> lhcb is optimised to detect kaons and pions, isn't it
[20:59] <Anitusar> and low momentum muons
[20:59] <sAxion> LHCb will measure CP violations = very important for early universe
[21:00] <Anitusar> we also search for new physics the indirect way, complementary to the big ones
[21:00] <photino> which detector grabs low energy muons? ie do they make it to the muon detectors at the end?
[21:01] <Anitusar> at least M1, which is before the calo
[21:01] <Anitusar> we also have less material, so muons are not stopped
[21:01] <photino> why don't electrons trigger the muon detectors?
[21:02] <bd_> I'd expect electrons to interact strongly with the rest of the machine and be lost before they get there...
[21:03] <Anitusar> electrons are stopped pretty fast when they see material
[21:04] <photino> but why wouldn't the same interactions apply to muons?
[21:04] <photino> ie in what way does the extra mass change things
[21:05] <Anitusar> the electrons make distinct shower in ECAL, also different rings in the RICH2, so one can separate electrons from muons pretty easy
[21:07] <sAxion> dont they bend more in mag?
[21:07] <Anitusar> the extra mass changes changes the energy loss per volume element.
[21:07] <Anitusar> the bending depends also on the momentum
[21:08] <photino> what's the main CP-violating decay lhcb is built to observe?
[21:09] <Anitusar> Bs -> J/Psi Phi is one important channel
[21:14] <photino> will lhcb have trouble if new physics comes into play because they can't capture everything produced in a collision?
[21:15] <Tau> photino: I thought ATLAS was supposed to catch everything.
[21:15] <photino> tau: sure, i'm asking about lhcb
[21:15] <Anitusar> no, what we test is the standard model predictions. So if there are not correct, we also "measure" new physics
[21:16] <Anitusar> we also look "only" for b-quarks, which fly always in forward direction
[21:17] <photino> but in order to determine the energy of the b-quarks, aren't you making standard model assumptions about their production?
[21:17] <photino> i suppose that's what you meant before :)
[21:18] <photino> or maybe you work backwards from the decay products?
[21:19] <Anitusar> yes, you always measure the decay products
[21:19] <Anitusar> b-quarks have the nice feature, that they fly for about a 1cm, so one can reconstruct them very well
[21:20] <Anitusar> and what we look for is the difference between the b-quark and anti-b
[21:21] <Tau> Anitusar: I almost get it. What difference is there between them? Energy?
[21:23] <Anitusar> concerning the difference, this is complicated, but you look for asymmetries in the decay of the both particles
[21:21] <photino> ie the CKM matrix?
[21:22] <Anitusar> we measure 4 entries in the CKM matrix
[21:23] <Tau> Yes, I looked up the CKM matrix. I can read the math, although I'm not familiar with it. A bit confusing, but I think I get it a bit.
[21:24] <photino> tau: basically the particles are vectors in a hilbert space, and there are 2 relevant bases which enter interactions. the CKM matrix translates between them
[21:25] <Anitusar> yes, the mass eigenstates are not equal to the weak eigenstates, if this helps ;)
[21:26] <Tau> Yes, I get an idea. So normally, you see mixtures of quarks, instead of exactly a top, strange and so on, as they like everybody to believe? I'll have to remove the idea from my brain that a particle consists always of a nice triple of quarks; there are all kind of mixtures (quantum interpolations, to use the right term).
[21:30] <Tau> With this new knowledge, I am going to try to read the Wikipedia page on CKM matrix. Wish me luck ;-)
[21:29] <Anitusar> all i care is that are 3 parameter and one phase, which leads to CP-violation
[21:29] <photino> anitusar: i still don't think i understand how you can get away with not knowing exactly how the b's were produced? (or maybe you do know but don't need to measure?)
[21:30] <Anitusar> we do not care, how the b was formed. We measure from the point of production onward
[21:31] <photino> and you determine the energy/momentum of those b's entirely working backwards from the decay product? wow!!
[21:32] <Anitusar> an example: you get two muons
[21:32] <Anitusar> combine them to a J/Psi
[21:33] <Anitusar> next two Kaons and build a phi
[21:33] <Anitusar> then you build the bs by combinening those two
[21:34] <Anitusar> measure where the primary vertex and you know, how long the Bs lived
[21:34] <Anitusar> the position of the primary vertex
[21:35] <photino> the primary vertex being the J/psi phi production vertex?
[21:35] <Anitusar> no the bs decay is the secondary vertex
[21:35] <Anitusar> no primary vertex is constructed using all the other tracks
[21:38] <Tau> I've seen those tables with decay probabilities: you probably need these for the reconstruction. Does that imply you're never sure what happened?
[21:39] <Tau> Or is it almost always possible to figure it out?
[21:40] <Anitusar> as we go backwards, we do need the tables for reconstruction
[21:41] <Anitusar> but we get the Standard model prediction for a decay, and then we look if we see diviations
[21:42] <Anitusar> e.g. we should see 190 decays of a certain kind and find 400, we know the Standard model misses something
[21:42] <Tau> Anitusar: OK, so that's how they made the tables.
[21:44] <Anitusar> sorry, backwards misses a "not" in the sentence ;)
[21:48] <Tau> Well, I learned a lot this far. Thanks.
[21:49] <photino> yes thanks for your time anitusar!
[21:50] <Anitusar> no problem, but i am more of an electronic guy, so the theories are not my best topic
[21:52] <photino> i've never understood the CKM matrix. i mean, where did it come from? looks so ad-hoc...
[21:52] <photino> hope the lhc finds out for us :)
[21:53] <Anitusar> you need the CKM mechanism to get CP-violation, got the nobel prize 2008 for predicting the third family
[21:55] <Anitusar> we will just measure the elements of the matrix with more precision to see how large the CP-violating effects are in the Standard model
[21:57] <photino> oh, i meant to ask how does the VELO detect the vertex position?
[21:59] <Anitusar> the beams are tuned for lhcb in such a way, that we have only one interaction per collistion in average
[22:00] <Anitusar> so most of the track are coming from that collision, therefore the primary vertex is where the most tracks intersect
[22:02] <photino> interesting, does that mean later at high intensities they will squeeze the beem very finely to get single interactions?
[22:03] <photino> whereas the other experiments only care about luminosity?
[22:03] <Anitusar> the opposite, the beam is not as much focused as for the other experiments
[22:04] <Anitusar> thats why lhcb has no problem with low luminosity in the beginning
[22:04] <photino> so not only is the velo really close to the beam, but the beam is wider than for the other experiments as well
[22:04] <photino> if i understand you right, it is designed for relatively low luminosity
[22:05] <Anitusar> yeap
[22:05] <photino> still lots of collisions but over a wider area
[22:06] <Anitusar> atlas and cms have over 20 interactions per collision, wher ewe will have 1
[22:07] <photino> but they want to make quark-gluon plasma unlike you
[22:07] <photino> (or does the term only apply to lead ion collisions)
[22:08] <Anitusar> usually ion collisions
[22:08] <Anitusar> but they want to create as much particles as possible
[22:10] <photino> a tradeoff with vertrex tracking precision
[22:11] <Anitusar> yes, we need to the position of the primary vertex very well and also in which primary vertex the b was produced
[22:11] <Anitusar> so having only one makes things easier
[22:12] <photino> i think i'm beginning to see now why the design is the way it is :)
[22:12] <photino> thanks lots!
[22:15] <spencer> thx here too
[22:15] <Anitusar> glad i can help :)
[22:16] <photino> at higher energies, will you rename the experiment LHCt? ;)
[22:16] <spencer> now my universe is inside-out
[22:17] <Anitusar> No, the t is bad. It cannot commit to other quarks to form a meson. Wants to be alone ;)
[22:19] <Axion> Surely it cant be alone due to colour confinement
[22:20] <Anitusar> it decays before there is any chance to form a meson (quark anti-quark pair)
[22:20] <Axion> ok
[22:21] <photino> the top is VERY heavy for a quark
[22:22] <photino> it weighs as much as a whole heavy metal atom
[22:22] <Axion> top heavy ;)
[22:22] <photino> haha
[22:26] <Tau> I've been following the discussion. Thanks for the explanations.
[22:27] <homolupus> so have I, it has been most interesting :)
[22:28] <Axion> I liked it too
[23:03] <spencer> BTW did LHC get the 1 million events ?
[23:04] <Anitusar> lhcb was at 423000 yesterday
[23:06] <Anitusar> well everyone wants the J/Psi for which you will need about a million events

User avatar
LHCPortal Guru
Posts: 403
Joined: Fri Nov 20, 2009 1:19 pm
Location: Stuttgart Germany

Re: Q&A with a LHCb expert

Post by chriwi » Thu Dec 17, 2009 3:01 pm

I onlyread this later, also many thanks from my side.
Its a real good Idea to keep such interesting IRC-catchs and preserve them in the forum. :)


Posts: 48
Joined: Fri Dec 11, 2009 5:26 pm

Re: Q&A with a LHCb expert

Post by Anitusar » Thu Dec 17, 2009 6:22 pm

WOW, I am an expert :D

Okay, i tried to summarise the informations. But first i have to say that i am not an expert on the theory or the final results LHCb wants to measure. Nevertheless i picked up some informations from various talks given about these subjects, which i will try to explain in the next posts. So if some LHCb expert on this stuff thinks that i have made things to simple, please fell free to enlighten me and the other readers :).

By the way: My field of expertise is the electronic which does the actual measurement.

Posts: 48
Joined: Fri Dec 11, 2009 5:26 pm

Re: Q&A with a LHCb expert

Post by Anitusar » Thu Dec 17, 2009 6:23 pm


It all starts with the biggest asymmetry one can measure: our universe :) . When the universe was cooling down, photons (radiation) started to converted into particle and antiparticle. And particles and antiparticles annihilated into photons. So one would expect some kind of equilibrium between these two processes and you end up with a certain amount of photons, matter and antimatter. As one creates always a matter particle together with an anti-matter particle, there should be as much matter as antimatter in our universe.

But when you look around in the universe, we so far only found matter. Where is the anti-matter? One explanation would be, that the anti-matter is in a different region of the universe, which we did not found yet (People keep looking for this, but nothing so far). A second explanation would be, that there is no antimatter in the universe anymore. Somehow when all the annihilation took place there was some leftover matter. How much? Well, for every antiparticles you would have particles.

But wait a minute, you created a particle always together with an antiparticle. How do end up with more particles than antiparticles? The answer is: There must be a difference in the decay of these things. So you start with 50:50 and after everything has decayed you got more matter than antimatter.

Such decays would be CP-violating. CP-what? Okay, imagine a particle X decaying into a final state y. If you want to compare this with the behaviour of the antiparticles, you first have to figure out, which process you would have to measure. So first, you have to change all particles in your decay into antiparticles and vis versa. This is done in Quantum mechanics by the C (charge) operator. It just changes the charge of all particles involved. But you are not done yet. You also have to take into account the handiness of the particles. For this you need the P (parity) operator. It acts like a mirror, so all x coordinates are transformed into -x. After this, you have your antiparticle aX decaying into the "anti" final state ay.

In an experiment you now count how often both of these processes take place. And if the rate differs, you have a process, which does not respect the CP transformation. It's CP-violating! Such a process was first found in 1964. Until today several of these processes were found.

Posts: 48
Joined: Fri Dec 11, 2009 5:26 pm

Re: Q&A with a LHCb expert

Post by Anitusar » Thu Dec 17, 2009 6:24 pm

A tiny bit of theory

Within the standard model (SM) there is only one place, where you can generate such processes. It is the CKM matrix (Cabibbo, Kobayashi, Maskawa). This matrix describes, how you can change a downtyp quark (d,s,b) into an uptyp quark (u,c,t) via the weak interaction (emitting a W boson). This matrix has 9 entries: one entry for d -> u, one for d -> c and so on. These entries are complex numbers, so you end up with 2x9 = 18 parameters. Due to physics constrains you can reduce this to 3 real numbers and one phase. And this phase is the source of CP-violation.

So there you have your explanation on how end up with more matter, so everything is good, right? Well, no. The effect of the CP-violation within the SM is too small to explain our universe. In order to fix this, you can postulate CPT-violation. The T (time) operator reverses time in your process (t into -t). By doing C,P and T you basically should end up where you started. So if you find a CPT-violating process, you can fly directly to Stockholm and receive the Nobel prize, because you destroyed about every theory in particle physics ;) .

The not so radical solution is that there must be additional sources of CP-violation. Maybe a b can not only change into u,c and t via the W boson, but via something else (H ?). In order to measure this, you have to calculate the prediction of the SM for a specific process and then measure it. If this deviates, you found something new. And by looking at which processes you find something new, you can extract the properties of the new stuff.

And that is what we do at the LHCb experiment.

Posts: 48
Joined: Fri Dec 11, 2009 5:26 pm

Re: Q&A with a LHCb expert

Post by Anitusar » Thu Dec 17, 2009 6:25 pm

How do you measure it?

In an analyse you look at a specific process (see below). Lets take Bs -> J/Psi (mu+ mu-) Phi(K+ K-) as an example (In this decay the final state y and ay are the same, so look at b going into this state and anti-b going into this state).

In a first step the event is reconstructed. This means, you try to connect all the measurements in the tracking detectors to build a track. Then you try to combine the track with measurements from the RICH, calorimeter and Muon stations. With these informations you build hypothesis about what caused this track. So what you have after this stage is a bunch of muon candidates, kaon candidates, electron candidates and so on.

Now you build your decay back wards. So you take a mu+ candidate and a mu- one and try to intersect the tracks. If you can do that, you declare this point as the point, were the J/Psi decayed. So you combine the candidates and get a new particle with a certain mass and momentum. If the candidates truly originated from a J/Psi, the mass you get will be close to the mass of the J/Psi (remember, its a measurement). Otherwise you can get any value (also by accident the J/Psi mass).

You do the same for the K+ and K- Candidates to get your Phi candidates. The you combine your J/Psi candidate with the Phi candidate and end up with a Bs candidate. From all this it is clear that you can not tell for a specific event if contained really this decay or not. However you can look at many events and plot the i.e. the mass of all the Bs candidates. So for a each mass value you count how many candidates you found in your sample. You will get a peak where the real Bs mass is. The better you can suppress the wrong combinations the more clear becomes the peak.

In order to get the number of real decays you observed, you do the following: The number of background you get from the sidebands. This means you look at the entries, which are not near the Bs mass. From them you can determine the shape of the background and extrapolate this to the entries near the Bs mass. This tells how many wrong combinations are part of your Bs mass peak, so all entries - background entries gives you the number of observed decays.

What you are left with is to find out, whether the a b-quark or an anti-b decayed. When you have this, you got your to rates and therefore the size of the CP-violation effect.
Last edited by Anitusar on Thu Dec 17, 2009 6:31 pm, edited 1 time in total.

Posts: 48
Joined: Fri Dec 11, 2009 5:26 pm

Re: Q&A with a LHCb expert

Post by Anitusar » Thu Dec 17, 2009 6:26 pm

At what processes do you look?

Of course there are lot of things various people look for (even directly for the Higgs). But there are a few major studies:

* The Tree-level determination of gamma
* Charmless charged two-body B decays
* Measurement of mixing-induced CP violation in B_s -> J/psi phi
* Analysis of the decay B_s -> mu mu
* Analysis of the decay B0 -> K* mu mu
* Analysis of B_s -> phi gamma and other radiative B decays

I can try to answer (simple) questions if you want to know more about one of these studies, but this is clearly not my area of expertise.

Basically you look for a process with
* a good theoretical prediction for the SM value
* a (large) contribution from new physics which modifies this value
* a decay which you can actually measure (look how often the mu is mentioned)

Post Reply