I have a decent idea of how colliders like the LHC work, but I'm not sure what the value is in continuing to run them.

Once you've smashed all of the (not very many) light elements together at a range of speeds and angles and collected the data, what else is there to do apart from look at the data for one or another theoretical effect?

If someone has a new idea to test, why can't they just use old data? Isn't there petabytes of the stuff?

I understand that colliders are extended or replaced over time to reach higher energies but, while the maximum energy is held at any given limit, what changes between new experimental configurations, and why do those changes matter?

  • $\begingroup$ Once you've smashed all of the (not very many) light elements together... The LHC doesn’t smash light elements together. $\endgroup$ – G. Smith May 14 '20 at 17:16
  • $\begingroup$ From Wiki: "The LHC primarily collides proton beams"... are you saying Hydrogen nucleii aren't light elements? $\endgroup$ – spraff May 15 '20 at 7:22

Particle physics obeys the law of conservation of energy and the law of mass-energy equivalence. If you want to see a particle with a mass $m$ then you need to put in energy equal to at least $E = mc^2$. So if, for example, you'd like to find out if there's a particle with a mass of $125 \text{ GeV/c}^2$ but you only have data where the particles were smashed together with a maximum energy of $90 \text{ GeV}$ then you're out of luck.

So the primary reason for doing new experiments is it allows us to find out if there are more massive particles that we've never seen before.

People do look through old data though. It's a not uncommon practice, though it has its pitfalls. Every particle physics experiment is much more complicated than "smash some stuff together and see what comes out" and the raw data that comes out are far too numerous to actually store. So what actually gets stored is pruned and selected in order to look for certain things. You can use that data to look for other things, but you have to be very careful that you're not being fooled by the selection and exclusion effects.

For a hypothetical example, imagine you wanted to look for a new fermion that for some reason doesn't couple to the Z boson. You could look in old data for it, but you cannot use data from either the LEP experiment or the SLAC experiment because both of those were designed to produce almost entirely Z bosons. But your fermion doesn't couple to the Z boson so you would expect to never see it in that data.


You should be aware of the observation in nature that at different length scales different types of phenomena become apparent. At macroscopic scale matter appears to be completely continuous, whereas at around 1Angstrom atoms and molecules are found. At an even smaller scale of around $1fm = 10^{-15}$cm quarks were found as constituents of protons and neutrons and many other particles, observation that was not all to be expected at larger length scales. So physicists in general expect that at even smaller length scale -- and there is no limit (at least up to the Planck scale which out of any imaginable reach) in the sense of small -- physicists expect again new phenomena.

The LHC is the first accelerator providing a centre-of-mass energy of 13TeV to protons, so the highest energy ever achieved by human means. By going to further higher energy even smaller length scales can be explored corresponding to the inverse dependence of the length scale to the applied energy of the used particles:

$$ \lambda = \frac{hc}{E}$$

where $\lambda$ is the wave-length, $E$ is the energy of the used particles and $h$ and $c$ are Planck's constant and speed of light respectively. A wave-length which allows to sample structures of matter of the size of the wave-length (like at a much larger length scale ($\sim$ nm) electron waves in an electron microscope for instance).

Particular burning topics of this exploration is the understanding of dark matter(DM), which was discovered 60 years ago (DM was even already suspected in thirties of last century, but without clear proof) and still nobody has any idea what it could be. It is hoped that a particle could be found that could explain dark matter.

In the meantime physicists have a established a well-known model, the Standard model (SM). Although it answers a lot of questions already, it also leaves a lot of questions open, for instance why we have 4 interactions and whether they could be merged in one unique interaction, and why the particles in the SM have so different masses etc....

Finally physicists would like to build up a theory which is able to explain nature with a unique picture. And actual physics is far away from this state. A physics which at its full glory had governed the first instances of the universe shortly after the big-bang. So one can also understand the smashing of particles at the LHC as a far-distant (as we are still very far away from it) attempt to simulate processes which happened at the big-bang. But anyway, physicists already plan new accelerators even more powerful than the LHC with even more energy to be able to study even smaller length scales in the universe.


First of all, new particle accelerators get built to reach higher energy levels than their predecessors. This allows them to explore new "territory".

Second, new accelerators get built to work with different projectile types. For example, the SLAC at Stanford was originally built specifically to accelerate electrons only.

Third, new accelerators are rebuilt from old ones to increase their capabilities. For example, the SLAC was heavily modified over the years to increase its beam energy and to convert it from a linear accelerator into a collider.

  • $\begingroup$ So are the energies continuously increasing? I was imagining that the maximum energy of a collider stays about the same for many experiments, and I'm wondering what the value of additional experiments at the same maximum energy is. $\endgroup$ – spraff May 15 '20 at 7:21
  • $\begingroup$ present colliders are close to the practical limit for energies. The value of additional experiments at the same energy is that of decreasing statistical uncertainty in the measurements. $\endgroup$ – niels nielsen May 15 '20 at 17:20
  • 1
    $\begingroup$ The amount of data continuously increases when it's running. The energy is much more difficult to increase, so that just happens very now and then and requires significant work to do. See e.g. home.cern/sites/home.web.cern.ch/files/2018-07/LHCRep_11.07.png $\endgroup$ – rfl May 15 '20 at 18:07

The main change scientists at the LHC are working on now is the "High Luminosity" upgrade. The purpose is to drastically increase the number of proton-proton collisions. With more collisions available to analyze, there is a chance to find rarer processes.

As is the case so often these days in particle and astrophysics, the amount of data collected in a year can be more than the combined data in all the years prior, i.e. an exponential increase of data. That's why this new data, in its significant amount, is more useful than just archival data (which of course can still be analyzed when people come up with new ideas)

  • $\begingroup$ Okay, so there is a greater quantity of data, but what is the qualitative difference? They're working on the HL upgrade now, but until that happens what is the qualitative difference between this year/month's data and that of the previous year/month? $\endgroup$ – spraff May 15 '20 at 17:14
  • $\begingroup$ Haha, actually, LHC isn't running right now ;) $\endgroup$ – rfl May 15 '20 at 18:03
  • $\begingroup$ Once the LHC and its detectors are upgraded, a given day's data is similar to another day's data (unless they change e.g. whether they collide protons or lead nuclei etc). But from one upgrade stage to the other, many things change. The detectors are improved by a lot, for example. $\endgroup$ – rfl May 15 '20 at 18:04
  • $\begingroup$ @spraff are you forgetting that particle physics is a quantum domain, and in quantum mechanics the predictions are probabilistic, which means the larger the sample of events the more accurate the results of the study? and the more discrimination between proposed models? $\endgroup$ – anna v May 16 '20 at 6:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.