Well, it’s hit Slashdot now, so it won’t take long to make the rounds. The Large Hadron Collider project at CERN suffered something of a setback on Tuesday, when, during a test of one of the quadrupole magnets, which focus the beams down to tiny size at the collision regions, the magnet failed catastrophically. Though saying it like that sounds bad, here is a photo:
The failure occured during a test which simulated what the magnet might experience during a “quench” which is when some part of the superconducting cable inside the magnet suddenly goes “normal” and then resists the flow of the huge current in it. This releases heat, of course, causing the rest of the superconducting material to go normal. Liquid helium boils rapidly, creating large asymmetric pressures inside the magnet cryostat. These pressures can reach 20 bar, and it was during a 20-bar test that this particular magnet failed. No one was in the LHC tunnel when it happened – it must have been quite a sight, and sound, though.
I heard about this on Thursday, as did many of my colleagues, and we waited with trepidataion to hear how bad a problem this was. Yesterday at Fermilab, Steve Holmes, an accelerator physicist, gave a talk at our weekly US CMS meeting, and sounded confident that a solution could be found quickly to the design flaw which gave rise to the problem. The magnets were designed and built at Fermilab, and delivered to CERN and installed last year. CERN is leading the effort to find and implement a fix, with Fermilab’s help. Personally, I have little doubt that these world experts will solve this thing rapidly, and hopefully it won’t affect the LHC schedule much, and perhaps not at all.
You can get more details in Fermilab’s statement.
March 31st, 2007 at 5:03 pm
I hope there won’t be too much in the way of finger pointing and this problem is solved rapidly. I had heard that the CERN tests ran the magnets in a mode not tested at Fermilab.
March 31st, 2007 at 5:51 pm
Just to quibble a bit:
“one of the quadrupole magnets, which focus the beams down to tiny size at the collision regions”
I thought most of the quads in an accelerator are used to correct the beam, and that only a few special purposed ones are used for the actual final focus. Of course I could be wrong. (IANAAP)
March 31st, 2007 at 6:15 pm
April 1st, 2007 at 4:59 pm
[...] On blogs about science, I thought that Resonaances’ post continuing the saga of the (real) construction problems at LHC was really funny. It reports on the resignation of the director general of CERN, and the quote from the new appointee is funny. See that post here (it riffs on an earlier post located here, which also has some good -if alarming- humour). [...]
April 1st, 2007 at 7:40 pm
It looks like the duct tape didn’t hold up.
April 2nd, 2007 at 2:30 pm
So, does Fermilab have a couple extra months to find the Higgs first?
April 2nd, 2007 at 6:57 pm
To reply to “njim”‘s quibble:
All quadrupoles in accelerators are used to focus the beam. The millions of particles in a bunch travelling along an evacuated beamline tend to spread out as they repel their same-charged bunch-mates. So, every so-often we put a quadrupole to focus them to reduce their spreading; these “regular” quads do not focus the beam to a very small spot, that is the purpose of the final focus quads.
The “triplet” of quads that suffered the accident at LHC were indeed a set of final focus quads that focus the beam to a very small spotsize at an interaction point where the 2 circulating-in-opposite directions proton beams collide.
These triplets are of a different design from the approximately 400 quads used to focus the LHC beams as they travel around their circular path.
I’m not sure what you meant by “correct the beam”? The term “correct” is usually applied to the function of dipole magnets which steer the particles in the beam so they change direction, and all move in one direction. Whereas a quadrupole affects the particles that are off-center, say left and right, and forces them towards the magnet’s center from both sides, this action we call focusing.
Hope this information will help your future quibbling about accelerator magnets.
April 3rd, 2007 at 2:55 pm
I overheard (or rather got an e-mail forward) that the support structure for the inner triplets would have to be re-engineered. Would anyone (such as Cherrill, or another magnet engineer) happen to know why one couldn’t just put 10-atmosphere relief valves on the cryo-lines, rather than re-engineering the support?
April 4th, 2007 at 11:47 am
Looks like a squirm can failure to me
April 4th, 2007 at 7:29 pm
So where did you get that photo, anyway? Did you take it? It’s a great shot, and I haven’t seen anything as good elsewhere. I notice that Nature borrowed it from us.
April 4th, 2007 at 7:36 pm
Answering my question (#8), via a suggestion from a colleague!:
Amount of stored energy in an LHC quadrupole ~=
7.8 x 10^5 J
(from
http://quench-analysis.web.cern.ch/quench-analysis/phd-fs-html/node12.html
– that is for an arc quad, but final triplet quads are probably of the same order of magnitude)
He heat of vaporization = 82.9 J/mol
.0224 m^3/mol (ideal gas @ STP, as it would quickly heat up to room temp)
=> ~200 m^3 of helium gas!
= a major personnel hazard in the tunnel (oxygen displacement) if released via relief valves in the event of a quench.
April 4th, 2007 at 8:00 pm
I just heard that the experts now have a proposed fix which is under review. Assuming this fix passes the reviews, the remaining magnets of this type can be fixed, in the tunnel without having to be removed. The magnet that blew up obviously has to be replaced. Don’t know if there are spares, or if Fermilab has to build another one, ASAP, from scratch. Nobody knows the impact on the schedule yet.
April 5th, 2007 at 12:24 pm
Looks like ‘only the 9′ it says final focus quads – but these would likely carry the most current (shorter focal lengths), probably resulting in a more serious quenching problem, that perhaps was not specificly tested for.
April 6th, 2007 at 10:36 pm
>>>>> Stephen Uitti on Apr 2nd, 2007 at 2:30 pm
>>>>> So, does Fermilab have a couple extra
>>>>>> months to find the Higgs first?
Applauses transforming in ovation.!
April 12th, 2007 at 4:31 pm
More pictures here:
http://8teslas.free.fr/zenphoto/index.php?album=Triplet+incident
April 12th, 2007 at 4:36 pm
By the way, we do have spares, at least for the most deteriorated magnet. About the other triplets – now know to all have the same shortcoming-, we are looking for an alternative to the replacement of the faulty piece – which would involve removing all triplets, bringing them to the surface and so on…