The Large Hadron Collider is not just an extraordinary science experiment, it is also a remarkable engineering undertaking. As BBC science reporter James Morgan relates, just getting it built is an astonishing story in itself.
The LHC took 10,000 scientists a total of 14 years to assemble
How do you build a "Big Bang Machine"? That was the challenge which scientists at Cern began to ponder in the early 1980s, when the idea for the Large Hadron Collider was born.
Cern's governing council wanted to build a kind of time machine that could open a window to how the Universe appeared in the first microseconds of its existence.
If it could recreate the fleeting moments 13.73 billion years ago, when the fundamental building blocks of the cosmos took shape, then the world we live in today would be brought into much sharper focus.
It could discover how matter prevailed over antimatter, learn how dark matter was formed, and catch our first glimpse of the elusive Higgs boson - a "missing jigsaw piece" in our model of the Universe.
We might even find evidence of the existence of other dimensions. But to conjure up these conditions, the Cern council knew it needed to perform an engineering miracle.
The 12-storey ATLAS detector weighs in at 7,000 tonnes
To generate the necessary high energies, the designers required a particle accelerator more magnificently complex than any machine ever built.
Beams of protons would be hurled together at 99.9999999% of the speed of light, in conditions colder than the space between the stars and each travelling with as much energy as a car at the speed of 1,600km/h.
And yet the fruits of these explosions - high-energy particles - would decay and disappear from view in less than a trillionth of a second.
To "photograph" these valuable prizes would require a detector as large as a five storey building, yet so precise, it could pinpoint a particle with an accuracy of 15 microns - 20 times thinner than a human hair.
How on earth do you build a machine like that? The journey has taken 15 years, more than 10,000 scientists, from 40 countries, and a financial injection anticipated at up to 6.2bn euros - four times the original budget. But it was achieved, on time. Well, almost.
THE LARGE HADRON COLLIDER
The last of the LHC's 1,700 dipole magnets is lowered into place
The plans for the Large Hadron Collider began to gather momentum in the early 1980s, inspired by the success of its predecessor at Cern, a collider known as the Large Electron Positron (LEP).
But it was not until 1994 that the formal proposal for the LHC was ratified by Cern's member states, and the engineering work began.
The accelerator would be housed in a near-circular 27km-long tunnel, buried 50m-175m underneath the Jura mountains, criss-crossing the Swiss-French border. The tunnel was already in place - having been occupied by LEP, which was eventually disassembled in 2000.
Inside the LHC vacuum pipe, two parallel beams of sub-atomic particles (protons) or lead ions would hurtle in opposite directions at record energies.
Crashing together at specially designated junctions, they would release unstable, high-energy particles - including, perhaps, the elusive Higgs Boson.
To generate a magnetic field powerful enough to steer the high-energy particles around the pipe requires 1,740 superconducting magnets, which together required some 40,000 leak-tight welds and 65,000 "splices" of superconducting cables.
If you added all the filaments of these strands together, they would stretch to the Sun and back five times, with enough left over for a few trips to the Moon.
In order to conduct, the magnets must be cooled to within a couple of degrees of "absolute zero", the theoretical limit for how cold anything can get. This requires a constant supply of liquid helium pumped down from eight over-ground refrigeration plants - about 400,000 litres per year in total.
Engineers excavating the cavern for CMS encountered serious difficulty
At the junctions where particles collide, four enormous detectors have been designed to observe the microscopic wreckage.
Between 1996 and 1998, approval was granted for four giant "experiments" - Alice, Atlas, CMS and LHCb - to be housed in four enormous underground caverns, dug strategically around the collider loop.
Excavating these caverns out of sand, gravel and rock was a considerable feat. In the case of the 7,000 tonne Atlas detector, it took two years to burrow a cavern large enough to hold a 12-storey building.
But while Atlas may be the largest cavern, it was CMS - 10km up the ring below the village of Cessy - which proved the most problematic at the excavation stage.
The cavern shaft had to be bored through a 50m layer of glacial deposits, including fast flowing water, which threatened to flood the shaft. Engineers repelled these underground rivers by piping super-chilled brine down the shaft, allowing a wall of ice 3m thick to form around the circumference.
It took six months to freeze the walls of the two CMS shafts. But while the barrier worked initially, the water eventually broke through, forcing engineers to first pump down liquid nitrogen to turn the area into "Siberian permafrost", in the words of Austin Ball, CMS technical coordinator.
LHC components were transported to Cern from all over the world
Building the components of both the accelerator and the detectors was a truly international effort.
In the case of the 12,500-tonne CMS detector, the coiled strands of its central solenoid magnet - all 50km of them - began their life in Finland, before travelling to factories in Grenoble, Neuchatel and Genoa, to be braided, coated, and welded.
After being shipped to Marseille, they went up the river to Macon, where they were unpacked and driven by lorry under the mountains to Cern.
In fact, the diameter of the magnet was restricted to ensure it was just narrow enough that components could squeeze through the tunnels. The clearance was a matter of centimetres.
The CMS magnet is the most powerful solenoid ever built - conducting a current of 12,000 amperes - to create a magnetic field 100,000 times stronger than the Earth's.
ASSEMBLING THE DETECTORS
The detector units of CMS were squeezed in with centimetres to spare
The next problem, of course, was how to get a 45m-long, 25m-high, 7,000-tonne detector, through a shaft hole 20m wide.
The answer of course is to do it in bits. Atlas was lowered piece-by-piece over several years, and assembled almost entirely in the subterranean cavern.
The largest piece - the barrel toroid magnet - fitted down the cavern shaft with only 10cm of clearance on either side.
But the building of the detectors is not all heavy engineering. Layer upon layer of electronic sensors had to be wired and connected by hand, which meant up to 300 people a day working in the cavern cramped against each other.
Squeezing each piece into place was "like solving a wooden puzzle" - there is only one possible way of doing it, according to Professor Andy Parker of Cambridge University, one of the founders of Atlas.
"Everything fits together like Russian dolls. I saw one design for Atlas which fitted together, but you couldn't assemble it, because there was no room to move the pieces past each other. Every single millimetre of space was fought over," he said.
The CMS detector, on the other hand, was largely assembled above ground, in several enormous units.
The largest, at 2,000 tonnes (the weight of five jumbo jets, or one-third of the weight of the Eiffel tower) took 10 hours to lower down a 100m shaft, with a clearance of 20cm either side. The world's largest electromagnet had to be handled with extreme care.
Its cylindrically arranged silicon wafer detectors contain a vast network of micro-circuitry - including 73,000 radiation-hard, low-noise microelectronic chips, almost 40,000 analogue optical links and 1,000 power supply units.
To manufacture these required an entirely new method of auto-assembly.
PROBLEMS DURING TESTING
Failure of a magnet in testing delayed the LHC start-up by almost a year
Though the LHC was originally slated to begin operations in late 2007, the entire project was set back after a failure in one of the quadrupole magnets used to focus the beam, which buckled during testing.
This meant all similar magnets would have to be redesigned and replaced.
Other, less serious problems arose due to leaky plumbing of liquid helium, and also when some copper "fingers" used to ensure electrical continuity between magnets buckled when the magnets were warmed up.
The final tab for the LHC is expected to come in at a colossal 6.4bn euros.
But that sum still represents good value for money, according to Dr Chris Parkes, of Glasgow University, UK, who works on the LHCb detector.
He said: "Tom Hanks is to appear in the movie of Dan Brown's Angels and Demons, which involves scientists at Cern making anti-matter. But the new experiment at the LHC to understand anti-matter cost less than Tom Hanks will earn from the movie."