Skip to main content

Seeking the Higgs boson

39 dipole magnets, each 15m long and weighing 35 tonnes, had to be replaced when the LHC was repaired in 2009 © CERN

39 dipole magnets, each 15m long and weighing 35 tonnes, had to be replaced when the LHC was repaired in 2009 © CERN

On 10 September 2008, the largest and most complicated scientific instrument ever built, the Large Hadron Collider at CERN, came to life. Nine days later, a catastrophic hardware failure forced the machine to be shut down for more than a year. CERN’s Director for Accelerators and Technology, Dr Steve Myers FREng, describes how engineers got the LHC back on track and operated it so successfully that last summer physicists were able to discover a new fundamental particle: the Higgs boson

Particle accelerators have revolutionised our understanding of nature’s fundamental constituents. They are also among the most complex and expensive instruments ever built, exploiting almost every cutting-edge technology available and spanning many engineering disciplines. The Large Hadron Collider (LHC) at CERN near Geneva is the world’s largest and most energetic particle collider – a 27 km circumference ring of superconducting magnets located 100m beneath the French-Swiss border that smashes protons into one another millions of times per second inside detectors the size of cathedrals. Proposed 30 years ago, the machine was designed to put the ’standard model’ of particle physics to its ultimate test.

By late summer 2008, construction of the LHC was complete (seeIngenia 35) and the machine’s eight sectors had been cooled down to their operating temperature, -271°C, using some 10,000 tonnes of liquid nitrogen and 150 tonnes of liquid helium. This is the temperature at which the niobium-titanium cables inside the LHC’s 1,232 dipole and numerous other magnets become superconducting, allowing them to carry the enormous currents necessary to guide high-energy protons around the ring. The LHC was being put through hardware tests in preparation for circulating beams.

Particle accelerators have been around for 80 years, but none has attracted such public attention before. At the scheduled start up of the LHC on the morning of 10 September 2008, some 300 journalists and broadcast crew were present at CERN. It put a lot of pressure on us as engineers because you never know what you’re going to get when you switch on a new machine. Thousands of magnets and power supplies had to have the right polarity, and we had to make sure there was nothing in the path of the beam. In the early 1990s, for instance, we discovered two beer bottles lodged inside the vacuum pipe of the Large Electron Positron Collider, whose tunnel the LHC now occupies.

The copper ‘bus bars’ surrounding the superconducting cables that carry the large currents through the LHC’s magnets © CERN

The copper ‘bus bars’ surrounding the superconducting cables that carry the large currents through the LHC’s magnets © CERN

With the LHC, we were already aware of a problem with the ‘plug-in modules’ connecting different magnets, which did not expand properly when a sector was warmed up to room temperature (the LHC’s circumference shrinks by up to 80m when the machine is cold). We fixed around 3,500 modules and developed a simple yet powerful method to check the vacuum pipe for obstructions, using a ping-pong ball fitted with a radio-frequency transmitter.

But had we caught them all?

The LHC switch-on was an overwhelming success. Both beams made a full machine turn within hours of the scheduled time, and we managed to capture one of the beams with the radio-frequency system, which allows the protons to be accelerated. It was the first of many steps towards the LHC’s physics program, but for engineers it was vital because once you have beam going round you can start to make measurements and learn how to operate the machine.

Events were to change course dramatically. Shortly after the switch-on, a technical hitch with an electrical transformer forced us to cease beam commissioning for a few days. During this period, it was decided to test the last sector of the machine up to 9.3kA, which is 10% above the dipole current required for 5TeV beams. When the current reached 8.7 kA on the morning of 19 September 2008, a resistive zone developed in one of the 10,000 superconducting interconnects between the magnets. The connection evaporated under the massive electrical load, causing an arc that punctured the helium enclosure and then the vacuum beam pipes. The sudden release of helium initiated a pressure wave that built up over a region of more than 400m, causing major damage to magnets, interconnects and polluting the ultra-high vacuum system with debris.

On 19 September 2008, an electrical interconnect failed while carrying a current of 8.7kA, puncturing the LHC’s cooling and vacuum systems and causing major collateral damage to the machine © CERN

On 19 September 2008, an electrical interconnect failed while carrying a current of 8.7kA, puncturing the LHC’s cooling and vacuum systems and causing major collateral damage to the machine © CERN

I was in an LHC planning meeting at CERN when we received the news. When my colleagues and I arrived at the CERN control centre shortly afterwards, there were alarms going off everywhere, and clearly the whole system had fallen apart. It was a sobering moment for accelerator physics. Fifteen hours previously I had been nominated CERN Director of Accelerators and Technology, and immediately it was clear that we had a major repair job on our hands.


A CERN inquiry identified several causes of the damage, including the absence of solder on the magnet interconnect; a poor electrical contact between the superconducting cable and the surrounding copper stabiliser; insufficiently large pressure relief ports and inadequate anchorage of the magnets to the tunnel floor, and most importantly, the insensitivity of the magnet protection system (see later). Four people had signed off on the faulty interconnect, but human error probably was to blame for the incident. At the time, technicians were being interrupted by a number of evacuation alarms caused, for example, by smoke from welding, forcing them to travel to the surface and back down again. We don’t know for sure, but perhaps after returning from such an evacuation the team accidentally skipped a magnet, leaving it unsoldered.

During 2013 and 2014, more than 10,000 interconnects between the LHC’s 3,000-plus superconducting magnets will be replaced with more robust ones © CERN

During 2013 and 2014, more than 10,000 interconnects between the LHC’s 3,000-plus superconducting magnets will be replaced with more robust ones © CERN

In total, 39 dipole magnets plus 14 smaller quadrupole magnets were replaced. More than 200 magnet interconnects needed full or partial repair and 4km of ultra-high vacuum beam tube had to be cleaned of small pieces of super-insulation and soot. Half the dipole magnets were equipped with larger and additional pressure relief ports (the remaining 50% will be completed during the present long shutdown of 2013-2014), and a new longitudinal restraining system was installed on 50 quadrupoles. The incident set the LHC programme back by roughly 15 months.

A crucial aspect of the repairs is a brand new magnet protection system that uses 6,500 detectors and over 250km of cable to monitor the resistance of all inter-magnet superconducting connections with sub-nano-ohm precision. When the LHC was designed, electronics were not advanced enough to build such a system. Now, when the slightest rise in resistance is detected, the beam is dumped and the current is deposited in resistor banks located in the tunnel. The energy abort process takes around 50 seconds, during which the surrounding copper stabilisers carry the current safely. It is unlikely that any accelerator has undergone such a quality assurance analysis as has been carried out for the LHC. Exhaustive investigations of the machine design revealed three or four things that we could have done better, but there were thousands of things that we got right.


By late 2009, we were ready to start ramping the beam to higher energies. When we first threaded the beam around the machine the previous year we did it at the injection energy: 0.45TeV, which is the energy of the protons when they emerge from the Super Proton Synchrotron (SPS) after being accelerated in a series of smaller CERN accelerators. On 29 November, the operations team succeeded in accelerating both beams to 1.18TeV, corresponding to a dipole current of 2kA, breaking the record for the energy of a collider held for the previous two decades by the Tevatron at Fermilab in the US. During this short test period, we also brought the beams into collision at low energies quietly, just to see what would happen. Everything was looking good for the start of the LHC’s physics programme.

That first part of the ramp was very interesting for us. Going from 0.45TeV to around 1TeV is tricky because there are persistent eddy currents in the superconductors that can change the machine ‘optics’ enormously. A phenomenon called snapback, for instance, alters the chromaticity of the magnets. Things get easier once you are above 2TeV. However, as the energy goes up, the beam also starts to gain destructive energy of its own. For many years, 2MJ was the world record for the energy stored in a particle beam, but we were about to operate a beam packing over 100MJ. That’s sufficient to melt around 130 kilos of copper, and the beam dimensions are so small that the energy density could cause significant damage to the hardware if it were to veer off course.

In my view, another major accident could have been the end of CERN. That was the pressure on us as engineers. The question hanging over the LHC was how high we could push the energy without risking a similar hardware failure. Although all but one of the superconducting interconnects had passed the high-current test, during 2009 we realised that there were many more bad connections in the copper stabilisers surrounding the superconducting cables. Around 15% of the connections had resistances five or six times higher than they should, which would cause a problem if the superconducting state is lost due to a magnet quench when running at high energy. For this reason, we decided to restrict the energy to 3.5 TeV per beam (half of the LHC’s design energy), corresponding to a current of 6kA.

On 30 March 2010, we achieved collisions at 7TeV, opening a new chapter in the exploration of nature. Getting the protons to collide is not technically very hard, but it took three attempts on that day. There are two very narrow beams in a relatively huge vacuum chamber – it’s a bit like firing two knitting needles across the Atlantic and getting them to meet halfway. We rely on beam instrumentation and corrector magnets to shift the beam up and down or left and right in small increments until we detect collisions, but the protons also need to strike at precisely the right time for this to happen. The LHC beam is not a continuous stream of protons but a series of bunches, each containing hundreds of billions of protons, so it’s necessary to make sure that the bunches cross right in the centre of the giant detectors, built at four points around the LHC, to study the collision debris.

From there onwards, it was a case of boosting the LHC’s collision rate (luminosity) to deliver as much data to the experiments as possible. This involves increasing the number of bunches, putting more protons in each bunch and “squeezing” the beam more tightly at the interaction points. Everything worked incredibly well, although initially things went slowly because we had established a dedicated machine protection team. Their job was to ensure that each step in increasing beam energy was well understood and not dangerous for the accelerator components.

The decay of a potential Higgs boson recorded in May 2012, evidenced by the presence of four leptons inside the Large Hadron Collider’s ATLAS experiment © CERN

The decay of a potential Higgs boson recorded in May 2012, evidenced by the presence of four leptons inside the Large Hadron Collider’s ATLAS experiment © CERN


By the end of 2011, we had exceeded the LHC’s design luminosity for 3.5TeV beams by some margin – achieving 3.9x1033 collisions per square centimetre per second – despite using only half the maximum possible number of bunches (1,380 of 2,760). Topping the year off, in December the LHC’s two giant general-purpose detectors ATLAS and CMS also spotted the first signs of what would turn out to be a brand new fundamental particle: the Higgs boson – see LHC physics. Our top priority going into 2012 was to run the machine to either find or rule out the Higgs. Since we had experienced no magnet quenches during two years of operation, a rigorous risk analysis allowed us to increase the energy to 4TeV per beam, boosting the chances that new particles such as the Higgs could be produced.

For the next few months, we delivered so much data to the detectors that physicists could barely keep up. ATLAS and CMS were designed to handle around 20 collision events simultaneously but soon found themselves dealing with 45 individual events – all coming in at a rate of 20MHz. That physicists could contemplate making a discovery from this data ‘pile-up’ is quite amazing.

The morning of 4 July 2012 was one of the most exciting moments in the history of particle physics. Certainly, I’ve witnessed nothing like it in my 41 years at CERN. The ATLAS and CMS teams each presented their latest search results for the Higgs, and the atmosphere in the auditorium was fantastic – people had been camping outside it for days to secure a seat. When CMS spokesperson Joe Incandela announced that his team had hit ‘5 sigma’ statistical significance, showing beyond doubt that a new particle had been discovered, he immediately came over and shook my hand, explaining that it had been the previous month of exceptional LHC performance that had sealed the discovery. At the time we weren’t certain that the new particle was a Higgs boson, but that question was largely settled this year with data taken during the remainder of 2012.

As engineers, we can make an accelerator as beautiful as we like and marvel at getting the beam around it, but it is the discoveries that make you realise how important these machines are. The world’s first hadron collider – CERN’s ISR (Intersecting Storage Rings) in the early 1970s – was one of the best machines of its time, but it discovered very little new physics.


Hopefully, the most exciting period in the LHC’s life is yet to come. In February 2013, the machine entered a long shutdown for repairs to allow it to operate at or close to its design energy of 7TeV per beam, with first collisions for physics expected in spring 2015. That will give the LHC the ability to produce much more massive particles than it has up until now, perhaps those predicted by ‘supersymmetry’ or other theories that attempt to go beyond the standard model. Such a discovery would revolutionize the field, yet so far the LHC has found nothing that cannot be explained by the standard model.

Our main task now is to replace all 10,170 of the LHC’s magnet interconnects, which have been completely redesigned to make them physically and mechanically more robust. The teams will start by opening up the 1,695 interconnections between each of the cryostats of the main magnets and will then carry out repairs on around 500 interconnections simultaneously , gradually working their way around the entire circumference of the machine. Major renovation work is also scheduled for CERN’s 54-year-old Proton Synchrotron (PS) and 37-year-old Super Proton Synchrotron (SPS), which once were the lab’s flagship machines but now are part of the LHC’s injection chain.

In fact, we are carrying out as much repair work as we possibly can during the shutdown. Before the LHC came along, CERN’s accelerators were closed for around three months every year for maintenance, but since it takes four weeks to warm and then another four to re-cool the LHC magnets, this model no longer works and we now operate the accelerators on a three-year cycle with a series of shorter breaks during the cycle. The task facing us now is almost as complicated as the first installation of the machine because we are doing all the work in situ. All four main LHC experiments are also using the shutdown to improve their detectors’ performance.

We have decided to stick to 6.5TeV per beam rather than the design energy of 7TeV because of training issues with the magnets at high currents. The LHC is still at the beginning of its life. Starting in 2019, the machine will undergo a major upgrade to boost its collision rate by a factor of 10, further increasing the chances of discoveries. Such is the complexity of the LHC that its high-luminosity upgrade, HL-LHC, requires about 10 years to implement. It relies on new superconducting magnets, compact and ultra-precise superconducting cavities and 300-metre-long high-power superconducting links – plus a brand new linear accelerator that allows more protons to be fed into the proton synchrotron. In order for the detectors to handle the additional collisions and harder radiation environment, the ATLAS, CMS, LHCb and ALICE detectors will also require major refits, especially concerning the sensitive equipment located closest to the LHC beam pipe.

Then we will run for another two to three years before the LHC’s third shutdown starting around 2022, which will be used to change the insertion quadrupoles so that the beam can be squeezed further. This will also be a crucial period for the detectors, giving physicists the opportunity to replace components that have been degraded as a result of the constant bombardment of radiation. That will take the LHC to 2030 and beyond, and produce unprecedented volumes of data. What then?

Tomography (PET) scanner at the Hopital Cantonal Universitaire de Genève. Development of the multiwire proportional chamber at CERN was seen as a potential device for medical imaging © CERN

Tomography (PET) scanner at the Hopital Cantonal Universitaire de Genève. Development of the multiwire proportional chamber at CERN was seen as a potential device for medical imaging © CERN


Particle accelerators have become so large and complex that they demand technical and financial resources beyond the scope of a single nation, maybe even a single continent. They also must be planned decades in advance, often relying on technologies that are yet to be invented. While operating and upgrading the LHC is CERN’s top priority for the coming years, the laboratory is also working on new ideas for colliders – principally a linear collider that accelerates electrons and their antimatter counterparts, positrons. Whereas the LHC’s collisions are ‘messy’, protons being composite particles made up of quarks and gluons, electron-positron collisions are ‘cleaner’ and easier to interpret. The Large Electron Positron collider was built to study the properties of the W and Z bosons discovered at the SPS proton collider, the second-largest machine in CERN’s accelerator complex, and now the case is being made to build an even larger electron-positron collider to measure the properties of the Higgs and any other new particles the LHC finds.

CERN’s proposal, called CLIC (compact linear collider), would use a novel acceleration technique to fire electrons and positrons in opposite directions along a straight tunnel tens of kilometres long. A separate design adopted by the International Linear Collider project would do the same thing using proven superconducting accelerator technology, but would be limited to lower energies than CLIC (around 0.5TeV compared to 2-3TeV). That would be sufficient to produce Higgs bosons cleanly and in bulk, and Japan has expressed a strong interest in putting up the costs to host such a ‘Higgs factory’ machine.

In the meantime, CERN is exploring all options to keep it at the forefront of high-energy physics research. These include an audacious LHC energy upgrade that would use new magnets to reach energies of 16.5TeV per beam, and a feasibility study concerning a new 80-100 km circumference tunnel that could one day host a 100TeV collider. For me, the big break point will come after the LHC runs for two or three years at high energies. If the Higgs is still the only thing the LHC has found, all we can do is study the Higgs in great detail because we won’t know what energy a linear or any other collider should be in order to guarantee further discoveries. If that situation arises, and so far it looks reasonably likely, then we need a new discovery machine. That means building a proton -proton collider at the highest energies possible and setting engineers their next big accelerator challenge.

Keep up-to-date with Ingenia for free