Physics experiments are designed to test the very frontiers of human knowledge. Sometimes, if you are clever or lucky, you can think of an easy experiment to look for something that is easy and no one else has thought of it yet. But a lot of the time you need to construct expensive machines in order to look for tiny effects, and then convince yourself (and the funding agencies!) that you will actually be able see something meaningful. The size and sensitivity some of these machines require some truly impressive feats of engineering. This note is devoted to some of the heroic experiments where the practical considerations have seemed immense.

KATRIN

What the experiment is about

For physicists: this experiment is designed to at least put a bound on the absolute mass scale of neutrinos. If all the neutrinos are approximately the same mass (i.e. the neutrino masses are greater than the mass splittings) then when an atom undergoes β-decay energy conservation tells the emitted electron has a maximum kinetic energy lower than the one a massless neutrino would give. This sort of experiment has already been done to give a mass limit of around 2.2 eV, the goal of this experiment is to the bound down to ~ 1 eV (or a detection!).

For non-physicists: we have to detect a difference of maximum possible energy of around one part in a million in an interaction with a particle that almost never interacts. The energies associated with random thermal motion are roughly the same as the difference in energies we are looking for. To do this we need a detector that is really big and really sensitive. It needs to be really big so that we have lots of possible interactions (each possible interaction is very low probability). The very sensitive is because we still won’t get many interesting events and we have to make sure we get them all.

What’s the problem?

The detector has to be one solid piece to meet the sensitivity requirements. If the detector was multiple pieces there would be small holes where the pieces failed to match. It is already difficult to make a large detector in one piece; it is much harder when you make the detector in Luxembourg and have to transport it next to your tritium source 300 km away in Munich. To give you an idea of how large this detector is I could tell you it is a cylinder 23 metres long and 10 metres in diameter. Or I could show you the detector moving through the city to its final destination by the tritium plant:

This brings us to the next problem: how do you take a detector this large from Luxembourg to Munich anyway? It must be nearly impossible to get such a detector down the Autobahn, right? Turns out the physicists thought so too. So here is the route they took instead:

Reference:

http://www-ik.fzk.de/tritium/

http://www-ik.fzk.de/tritium/overview/index.html

GRAVITY WAVES

We need to know a few things to know why this is included on the list.

1. Gravity deflects light
2. The effect of light deflection is incredibly small

To give you an idea of how small, if the sun bends the light of a galaxy directly behind it, from Earth we will get the direction to the star wrong by a few milli-arcseconds. This is a tiny angle; if an astronaut stood on the moon with his arms outstretched a milli-arcsecond is roughly the change in angle we would have to look at from Earth to look at his right hand instead of his left. As the sun makes a much bigger angle on the sky, we would not be able to see the event anyway!

Of course, one of the famous early tests of General Relativity was Eddington’s measurement of light bending. Here the star was not directly behind the sun but closer to the edge. Eddington was also lucky that the stars were a lot closer than galaxies, so the deflection angle was 1/1000 of the angle the moon makes on the sky (1.75 arcseconds). As you probably know it is almost impossible to see stars in the daytime, let alone those that are very close to the sun. The last element that was important was that there was a solar eclipse, so the sky was dark enough to take pictures of the stars. Because of the sensitivity and the rarity of having stars in the right places while solar eclipses were happening, Einstein thought that lensing would never be useful.

The thing that allows us to do science with lensing is

1. There are lots of masses in the universe

By looking at galaxies instead we have enough events around that even “rare” events can be seen with a non-trivial frequency. A nice biological analogy is “for a culture of bacteria, billion-to-one odds against an event means it will occur 8 times before Tuesday”.

Gravitational lensing can be distinguished for just magnification or stretching. Gravitational lens distort the image in a very non-standard way. For example, the picture below shows what the Mona Lisa would look like with a large mass between you and it:

The trouble is that in the universe we only get to see the distorted picture — how do we tell the difference between an ugly painting like the one on the right and a painting like the one on the left that has passed a gravitational lens?

More on this later…..

Reference:

http://astronomyonline.org/Cosmology/GravitationalLensing.asp

SLAC

What the experiment is about

SLAC is the Stanford Linear Accelerator Center, which is not really an experiment but an place where multiple experiments are carried out. The goals of the different experiment involving the accelerator are varied, but in essence all of them are about smashing thing A into thing B to try and extract information about the subatomic world.

What was the problem?

The design goal for the SLAC accelerator is that it must move less than ¼&#148*; over the course of a year. That is, no part of the tunnel can move more more in a year than the distance a snail can cover in 5 seconds. That may not impress you — after all a snail is pretty small. On a larger scale, the continental United States drifts roughly 1” every year. The SLAC tunnel has to move at less than quarter of the speed of North America! Combine this with the fact that the tunnel is built up of a series of sections each approximately 20 feet long, and you see that the engineering tolerances are very tight.

[Before people are too picky, it does not matter if the entire tunnel moves along together so the fact the entire United States is moving does not contribute to the ¼” a year limit.]

At the time that the linear accelerator was being put together, CalTrans had plans to build the I-280 highway over the accelerator. Unfortunately construction of the overpass would cause vibrations that would ruin any physics experiments at SLAC. As a solution, SLAC administrators managed to convince CalTrans to build the overpass and a short section of road on either side years before I-280 was supposed to be built. The road extended far enough so that when CalTrans finally got around to finishing the road the construction work would be far enough away not to disturb the experiments.

While this shows the SLAC administrators to be very forward thinking (and again emphasises how sensitive some of these experiments are), it did mean that for several years you could see the Kafka-esque sight of a bridge and a road in the middle of the plain with both ends suddenly terminating!

Note:
* I am a little bit weary of the ¼”/year limit. Presumably it only applies as an effective speed during accelerator operation, as the accelerator is tuned for each use. What the limit probably means in practise is that during earthquakes or other large vibrations SLAC must simply throw out their accelerator data. The speed can be found in the reference below.

Reference:
http://www.balearntofly.com/links/bay-area-history.html [3rd section]

LHC

Reference: