• John
  • Felde
  • University of Maryland
  • USA

Latest Posts

  • USA

Latest Posts

  • Flip
  • Tanedo
  • USA

Latest Posts

  • CERN
  • Geneva
  • Switzerland

Latest Posts

  • Aidan
  • Randle-Conde
  • Université Libre de Bruxelles
  • Belgium

Latest Posts

  • Laura
  • Gladstone
  • University of Wisconsin, Madison
  • USA

Latest Posts

  • Richard
  • Ruiz
  • Univ. of Pittsburgh
  • U.S.A.

Latest Posts

  • Seth
  • Zenz
  • Imperial College London
  • UK

Latest Posts

  • Michael
  • DuVernois
  • Wisconsin IceCube Particle Astrophysics Center
  • USA

Latest Posts

  • Jim
  • Rohlf
  • USA

Latest Posts

  • Emily
  • Thompson
  • Switzerland

Latest Posts

  • Ken
  • Bloom
  • USA

Latest Posts

Brian Dorney | USLHC | USA

View Blog | Read Bio

To B or not to Bbar: b-Jet Identification

Well its been longer then usual since my last post.  This past weekend was commencement at my host institution; and I spent it with my friends who graduated with their Master’s degrees (myself included) and family who came down to visit.

But today I’d like to talk about something that is crucial to my own research, B-Tagging.  Or, the experimental tools we (as high energy physicists) use to identify b-jets.  Here b stands for the “beauty” or “bottom” quark; these are two interchangeable terms for the same particle.

Importance of B-Tagging

But first, what’s the big deal about B-Tagging?

Well, many current Standard Model physics process under study at the LHC have a b and an anti-b quark (termed bbar or simply b) in the final state.

To understand what “final state” means, let’s look at an example.  Suppose in a proton-proton collision a top quark, t, and an anti-top quark, t, are produced (this is called top-quark pair, tt, production).  Well the top quark decays to a W boson, and another quark, q, ~10% of the time [1]. i.e.

t → W + ql+v q

Here: the W+ boson has decayed leptonically into a lepton, l, a neutrino, v; and a quark, q. The quark may be either a d, s, or b quark.  The case for anti-top quark decay, t, is shown here:

t → W - ql-v q

Where q = d, s, or b.

Then the final state for this event is two oppositely charged leptons, two neutrinos, a quark, and an anti-quark (These quarks will turn into jets as they cannot exist freely, more information on this below).  But when looking for top-quark pair production events in a collider, it’s easiest to find them if you look for events containing two b-jets along with the charge leptons and neutrinos.

In addition to top-quark pair production, the Higgs Boson is theorized to decay into a bb final state; which in turn form two b-jets (see one of my older posts here, or one of Flip Tanedo’s posts here for more details regarding the Higgs Boson).

But b-jets don’t just don’t come from Standard Model processes, many new physics searches (such as supersymmetric searches) have bb final states.

With this in mind, it is of paramount importance to be able to find and identify b-jets.  But to do this we first need to understand the properties of the b quark itself; or more importantly b-Hadrons as these are what we actually observe.

Beautiful Hadrons

B-Hadrons are rather unique in elementary particle physics.  They offer us a chance to study so much; we are able to use them to investigate topics from Quantum Chromodynamics, to CP-Violation, and even physics beyond the Standard Model.  Simply put, B-Hadrons have got it going on!

They are very heavy particles, with rest masses of approximately 5-10 GeV/c2; or roughly five to ten times the proton’s rest mass [2].

B-Hadrons are also very “long lived” particles, with mean life-times of approximately 1.6 pico-seconds, or 1.6·10-12 seconds.  For comparison, the π0 meson has a lifetime of roughly 8·10-17 seconds [2]; and the top quark has an even shorter lifetime of ~10-23 seconds.  So B-Hadrons are very long lived indeed.  In fact, because of this long life-time a B-Hadron has a of approximately 480-500 micro-meters (τ being the proper time, or the B-Hadron’s mean life-time; and c is the speed of light).  Putting this into more tangible terms, a B-Hadron will travel roughly half a millimeter before it decays.  For comparison, the π0 meson has a of only 25 nano-meters (2.5·10-8 meters).

In addition to their long life-time and large mass, a B-Hadron will produce five charged particles per decay (on average)!  In comparison a Δ++ (a light baryon made of u quarks) will produce only 3 charge particles per decay; and a π+ (a light meson made of a u and d) will produce merely a single charged particle per decay.

Also, when a B-Hadrons decays, there is a 10% chance that a lepton will be produced during the decay process.

In summary, B-Hadrons have:

  1. Large mass
  2. A long life-time
  3. Large 
  4. High number of charged particles per decay
  5. Chance of leptonic decay


But what does this have to do with B-Tagging?  For this we must ask ourselves how these properties listed above would show themselves within our detectors.

Experimental Signatures: B-Tagging

Hadrons will exist in clustered collimated groups within a detector, known as jets.  Thus we say jets are due to “hadronic activity.”

But what causes this clustering/jet structure to occure?  Well when protons collide in the LHC, a quark or gluon may “escape” the proton it was originally found in.  But quarks/gluons cannot exist freely in nature!  Thus quarks/anti-quarks use some of their kinetic energy to pull other quarks/antiquarks out of the vacuum to form hadrons (this is called hadronization, see one of Flip’s old posts here for more info).  And gluons “split” into a quark an anti-quark pair.  The produce quark and anti-quark in this “gluon-splitting” will then in turn undergo hadronization process.

But this all occurs within a jet!

So these B-Hadrons that we have been talking about are going to be found within jets.  So b-jets must have B-Hadrons inside them (hence the name)!

Now here is where our B-Hadron properties start to come into play.

Since a B-Hadron has a long lifetime, and travels some distance before decaying, we are able to look for what’s called a secondary vertex (SV).  A SV is a spot that new particles spew from because of the decay of a heavier particle (creating  tracks if these new particles have an electric charge).  So if a jet has a SV it is much more likely to be a b-jet.

To get an idea what an SV might look like in our detectors take a look at this image uploaded by Anna Phan.  Here a B-Hadron (Bs) has decayed into a charmed hadron (D+) and a muon (μ); the charmed hadron has then decayed into three other particles (hence a total of five charged particles were produced due to the B-Hadron’s decay!).

Notice how these particles are collimated and clustered together, i.e. this is a jet; and in this case it is a b-jet!

Also, because many charged particles are produced by a decaying B-Hadron, a jet that has a large number of tracks within it is more likely to be a b-jet.  So these four tracks that are produced after the SV (image I linked above) gave experimentalists in the LHCb Collaboration the ability to determine that this was a b-jet.  Had this jet come from say, a pion or a Δ++, there may have only been one or two tracks, and it would not be tagged as a b-jet (and rightly so!).

Finally, this picture has also shown the B-Hadron having a lepton (the muon) in its decay chain.  As a result, experimentalists look for jets that have nearby leptons when looking for b-jets.

So in summary, we as experimentalists in our B-Tagging efforts combine as much of this information as possible when searching for b-jets.  When we want to determine if a jet is a b-jet, we look to see if it has one or more of the following things:

  1. A secondary vertex
  2. A large number of tracks within them
  3. A nearby lepton


So from our list of B-Hadron properties we were able to construct a list of what to look for when attempting to perform B-Tagging.


But that wraps up our discussion for today.  There are many more levels to this then I’ve illustrated here.  If you are interested in more details, simply post below and I’ll try to answer your questions!

One last piece of information, creating the tools to perform B-Tagging is not something that one person alone can do.  For example, CMS has a very large staff of researchers (over 100) who focus directly on developing techniques these  techniques that I spoke on briefly.  Needless to say, without the researchers who develop these B-Tagging techniques, my own research would be impossible!

But until next time,




[1] Particle Data Group, http://pdg.lbl.gov/2010/tables/rpp2010-sum-quarks.pdf, May 11th 2011.

[2] Particle Data Group, http://pdg.lbl.gov/2010/tables/rpp2010-sum-mesons.pdf, Math 11th 2011.


Tags: , ,

15 Responses to “To B or not to Bbar: b-Jet Identification”

  1. This new venue of the US/LHC bloggers is very disappointing, way fewer graphics, the immediacy of the bloggers with the readers is totally lost.

    I hope it improves. Right now it is a boring monochromatic disappointment.

    • Quantum Diaries says:

      @Richard — Appreciate the feedback. The blog may have a new URL and a different background color, but we think you’ll find that it’s still the same US/LHC blog as it was before, with the same bloggers working hard to write something you’ll enjoy. Hope you’ll stick around.

  2. jal says:

    Thank you for sharing your knowledge. I’m becoming a better informed amateur.


  3. Tome says:

    Dear all,

    Really nice and informative blog, and I follow it regularly, but the new background color is trully offputting.

  4. Mark G. says:

    Please go back to the old US/LHC blog. The design of this website is very poor. I’m not at all intersted in input from the other bloggers you’ve merged with. Between the two of them, I don’t see any reason to keep coming back here.

  5. TB2 says:

    and while we’re at it, how on earth does one see the older posts? All I can see is what is published on the home page, but not the previous entries.

    • Quantum Diaries says:

      @TB2: We’re aware of the issue and will be fixing it shortly. We apologize for any inconvenience. You should, however, be able to scan the archives of the US/LHC blog from its homepage on Quantum Diaries.

  6. Doubledoge says:

    I sadly have to agree the move to this new blog is a real turn-off that gets in the way of the brilliant content we have enjoyed from the US>LHC blog in the past. I often used to print the blogs out as a pdf and I have a wonderful archive of really informative particle physics stuff now. At least Chrome still prints out the blog as black on white instead of the horrible colour scheme here but sadly all the files are called “Quantum Diaries” instead of automatically adopting the more useful blog post titles. Change it soon please and keep up the helpful diagrams and the great content.

    • Quantum Diaries says:

      We’re always looking at how to improve, so we’ll absolutely take your comments into consideration. Thanks for the helpful feedback.

  7. Fatih Salim says:

    Wow, some of these posts above are rather harsh. I think the US-LHC team is doing an outstanding job in creating non-technical blog posts about particle physics. I had a question about confinement. During the hadronization of the jets, are the particles appearing from the vacuum virtual particles? Furthermore, do the gluons decay into these mesonic pairs or what kind of process do they undergo to be transformed into these quark-antiquark pairs? Thank you for your time.

  8. Brian Dorney says:

    Dear Fatih,

    Thank you for your kind words! I, and my fellow US LHC Bloggers, appreciate it (we are actually all volunteers).

    In answer to your questions:

    The particles that appear from the vacuum are actual physical particles. Virtual particles are particles that are exchanged between real particles and mediate an interaction.

    This can get into quiet a philosophical debate about what is virtual and what is real. But this is usually a “good enough” descriptor: A physical particle is something that can be directly observed (say by a detector), whereas a virtual particle cannot be seen experimentally.

    These particles that come from the vacuum, if they are electrically charged they will leave a track in CMS’s Silicon Tracker, and possibly deposit energy in CMS’s Electromagnetic Calorimeter. However if they are electrically neutral we will catch them from their energy deposits within CMS’s Hadronic Calorimeter. So they can be observed directly, and thus are real!!!

    When gluons split into a quark & anti-quark pair it can happen in a few different ways. This can result in what’s called the “Hard Scattering,” or the two particles that actually participate in the proton-proton scattering. Here a gluon from one proton interacts with a gluon from another proton (protons have gluons constantly exchanging between the quarks that make up the proton, and these gluons can interact with gluons from the other proton). These gluons then fuse into a virtual gluon, which the splits into a real quark & anti-quark. This process is called “gluon fusion” as a result. Another case is when a particle from the hard scattering is very energetic and emits a real gluon, much like an accelerating electron emits a photon. This real gluon then splits into the quark & anti-quark pair.

    This quark & anti-quark that are created will be of the same type, or flavor. They may form what is called a resonance, which is a special type of meson, made of only one type of flavor (whereas a Pion is made of two types, up and down). Or this quark & anti-quark may go their separate ways, and hadronize independently.

    Hopefully this has helped!


  9. Fatih Salim says:

    Brian, I understand the hard scattering and resonance, but I am still perplexed about the virtual particles. The force mediators are the gauge bosons, right? If they confer information to other particles, in what sense are they virtual [are they a mathematical conveniences, kind of like renormalization, or do they possess some physical bearing in reality]? I realize this is not the aim of your post, but I truly enjoy these blogs, for they do not contain an abundance of jargon, so I really appreciate you and your team’s efforts. Moreover, how does, say, a virtual photon affect other fermions more so than real photons? Thank you for your time.

  10. Raincloud says:

    Hi, first of all thanks for all the good content. It is very helpful for me as an amateur. What confuses me about this B-Hadrons is the this: The decay of the top produces for example a b-quark, but how does this turn into a B-Hadron. Is a B-Hadron a hadron with a bottom quark? Or asked differently, what exactly is the lower case b and the upper case B?

  11. Brian Dorney says:


    Let’s start with Example 1, in this link: http://teachers.web.cern.ch/teachers/archiv/HST2002/feynman/examples.htm

    Here a neutron is undergoing beta decay. It decays into a proton (p, at right), an electron, and an anti-neutrino. What happened is one of the d quarks in the neutron emitted a virtual W boson, and became a u quark. Then this virtual W “split/decayed” in an electron and the anti-neutrino. This reaction happens all the time in nuclear physics.

    Suppose we view this reaction in the neutron’s rest frame (image a block of cesium sitting on a table, one of the neutrons in one of the atoms of cesium). The three particles are created in this decay will have some kinetic energy; the sum of which cannot be more then ~0.78 MeV as this is the difference in the rest mass of the neutron and the three outgoing particles.

    Let’s say I then observed thousands upon thousands of interactions, and I was able to detect the electron and the anti-neutrino from each decay. If I made a plot of the invariant mass of these two particles, It would not peak at the W’s mass.

    The reason for this is that in a Feynman Diagram, conservation laws must be upheld. Virtual particles will then take ANY mass or momentum that maintains conservation laws. In this sense we say a virtual particle exists “off mass shell.” The virtual W here does not have a mass of ~80 GeV/c^2 (its true mass). This is a sense of how they are “virtual.”

    Because of this “off mass shell” behavior a virtual photon could even have a non-zero mass!

    As another example, photons have polarization vectors. This has been known since the 19th century. This also carries over into quantum field theory (as it must). We also know that for real photons the polarization vector is perpendicular to the direction of the photon’s propogation. But a virtual photon may have a polarization vector that is in this direction!

    Also, virtual & real photons affect and interact with real fermions in the same way. But where you draw the dividing line between virtual & real can become a grey area in practice. In theory, anything that is an internal line in a Feynman diagram is virtual, anything that is an external line is real.

    For a more detailed description at an advanced undergraduate level I would recommend “Introduction to Elementary Particles,” by David Griffiths.

  12. Brian Dorney says:


    For your first question, I think Flip Tanedo has a good post here:


    That may help visualize the hadronization process. This is how any quark(s)/anti-quark(s) become hadrons.

    Heavy flavor hadrons (made out of c & b quark/anti-quarks) are called C-Hadrons or B-Hadrons because they have a b quark as one of their constituents. So yes a B-Hadron does have a b-quark within it!

    In particle physics these days we usually associate the lower case letters with quarks & upper case letters with hadrons. So when I write b or bbar, I’m talking about a bottom quark and an anti-bottom quark. But if I write B or Bbar, I’m talking about a bottom Hadron, or anti-Bottom-Hadron.

    Hope this helps!

Leave a Reply

Commenting Policy