Since March 30, when the LHC at CERN first collided protons at an unprecedented total energy of 7 TeV (7 trillion electron volts) the machine has been steadily moving from crawling to walking. Last Saturday, I'd say it took its first steps, and like any toddler, will soon be running.

The plot shows what we call "integrated luminosity" which is simply a measure of the number of collisions of protons in the interaction regions at the four experiments. In this case, it's my own experiment, CMS, the Compact Muon Solenoid experiment. CMS and ATLAS are the two large general-purpose detectors, each with thousands of physicists eager for real physics data.

As you can see, the vertical axis of the plot is labelled in units of “nb-1” or inverse nanobarns. The unit “barn” is a unit of area, a kind of joke from Enrico Fermi and friends who, despite the tiny size of a nucleus, said it was “as big as a barn” even though in cross sectional area it’s on the order of 10-28m2 (which is in fact the definition of one barn). If we think about the cross sectional area of the protons colliding in the LHC, they have a cross sectional area (or simply a total collision cross section) of about 0.12 barns.

So what's an inverse nanobarn? Well, if we try to collide lots of protons, we might ask "how many collisions per barn or cross sectional area did we make?" It's like throwing little paint blobs at a wall, one at a time. Eventually the wall is covered, and then covered again, and then covered many times over. We can ask "how many paint blobs per unit area of the wall did we cover?" The nano in nanobarns means one billionth of a barn, and so, now, the LHC has managed to produce its first inverse nanobarn: one collision per every billionth of a barn of cross section.

It's just a unit - all that matters is "how many collision events of my favorite kind should have been produced?" To get this, you multiply the number of inverse nanobarns by the production cross section for that kind of event, and also by the probability that you actually detect it. So for Z boson production, for example, the cross section is about 30 nanobarns, so we should have a few by now. (I am not at liberty to say whether we do or not...)

The plot has stair steps - the horizontal axis is real time, and the LHC machine is filled with protons, then brought to full energy, then collimators put in, then the experiment turns on and records data for some time until the accelerator folks decide to dump the beam out and refill. As you can see this cycle has been going like clockwork, with fill after fill of the machine. And the experiment has been recording a very large fraction of the delivered collisions, the losses being quite normal and due to end effects and the occasional glitch.

But then came the LHC baby's first real step last weekend: squeezing the beam. By raising the quadrupole beam focusing magnets to high field, the transverse size of proton bunches in the machine shrinks down and the probability of collisions goes up. In this case, the luminosity went up by an order of magnitude - it was a stunning success. Any imperfection in the focusing fields can send the beam right out of the machine, and, clearly, that did not happen.

The goal in the next year is to get to one inverse femtobarn - a million times more data. In the next week or so the plan, if all goes well, is to achieve another couple orders of magnitude in luminosity. Shit's about to get real, folks...