For those who aren't savvy, particle accelerators work by more or less swinging tiny bits of mass around in a circle and smacking them into each other at near-light speed so that we can see what happens. With a little bit of know-how and a whole lot of math, superheroes called "physicists" then decipher the results in research papers which must then withstand peer review. If you think that the superheroic designation is at all hyperbolic, then you clearly don't understand just how grueling peer review is.
For a little bit of perspective, there's enough energy at work in the LHC's twin beams to melt about a ton of copper. That's not a very rough estimate: "According to CERN LHC accelerator physicist Rüdiger Schmidt, who is in charge of machine protection systems, each unimpeded beam [of which there are two - D] is capable of melting a 500-kilogram block of copper." An exposure to such a beam just eighty-six millionths of a second long will bore a hole forty meters into a block of solid copper! So how do they stop it from doing just that?
Carbon, that's how.
After colliding for a while, the particle beams are shunted off into a straight 700-meter "runway" where they are spread out to be more like shotgun fire and less like machine-gun fire (the beam is magnetically widened from .2mm to 1.5mm). Then they're scanned like a CRT image onto a ten-ton graphite cylinder (or "dump") 8m long and .7m wide, housed in a thousand tons of concrete. These guys seriously aren't fucking around:
The upside is that this system is completely reusable. Though the beams need to be refreshed from time to time (increased usage makes them unstable), the dump cools off in a couple of hours and only needs to be used once every ten hours (which is how often the proton beams are purged per failsafe protocols). Without diffusing the beam or scanning it onto the dump, however, direct exposure would melt the graphite as well.
Bottom line: physics is hot.