The K2K Long-Baseline Neutrino Oscillation Experiment
What, Where, Why, and How?

JimHill

Abstract:

This talk is aimed at an audience of college level science students. I assume some basic knowledge of the names of some things in particle physics, but not really of the topic itself. The first section is historical. The second is an overview of present works. To understand the basic points here, the only really important concepts other than these few names are how to interpret a graph and some sense of what it means for data to be consistent with a calculation.

Almost all the slides are either reproduced or scanned here and much of the (planned) narrative is filled in. A few pictures are left out in favor of links to pages with original pictures (and more of them!).

This cover slide actually shows many of the main points of our experiment. The bottom left is a schematic drawing of the accelerator at KEK (Japan's National Laboratory for High Energy Physics) which produces our neutrino beam. Bottom center is an event display showing something of our data taken at the lab and right is another showing data taken 250km (155 miles) away measuring the same beam! Don't worry if you don't understand all this jargon yet; that's part of the point of giving the seminar.

The name of the experiment is `` ${\cal K}{\mathfrak{2}}{\cal{K}}$'' for ${\cal K}$EK ``to'' ${\cal K}$amioka describing where our neutrino beam travels.

I have compiled a small set of links on neutrino physics on this page.
While it is far from inclusive, it might be a nice place to start looking for extra information.

The idea of the neutrino was first suggested when people studied nuclear beta decay in the 1930's. Beta decay is when an electron (then sometimes called a ``beta particle'') is ejected from a nucleus leaving a slightly different nucleus behind. The difference is that one neutron is missing and one more proton appears. The first logical assumption seemed to be that the neutron split into a proton and electron and the electron escaped. There were a couple problems with this idea, but the biggest was that if one thing decays into two things, those two things should always get the same momentum on the way out. People could measure the distribution of momenta of electrons from beta decay (the ``spectrum'') and found that each time the energy was different. This distribution had a characteristic shape that was continuous from zero to some maximum depending on what nucleus was decaying.
Here is a link to a scanned image of the drawing I made depicting this.

As the idea of the neutrino became accepted, people began to try to think about how to detect such a thing. It was almost impossible since they are not charged and hardly ever interact with other particles. The idea arose however that if we know neutrinos arise in a process like this: $\epsfbox {bdecay.eps}$ perhaps they could be detected by a process like this: $\epsfbox {inverse_bdecay.eps}$

In the 1950's an experiment to observe this was done by Reines and Cowan. They successfully observed neutrinos emanating from a nuclear reactor by observing the above interaction in their detector and carefully checking for a ``anti-coincidence'' with a surrounding detector. That is, the point was not only to see the products of this interaction, but to remove the possibility that the same products somehow arose from some particle other than a neutrino entering the detector. This technique is now fairly standard in neutrino physics and was one of the great innovations of the experiment that eventually got it the Nobel prize.

Over the years, people learned a lot more about neutrinos.

$\bullet\$1962: There is a second type of neutrino!
This type is now called the ``muon neutrino,'' or $\nu_\mu$ The original is the ``electron neutrino,'' $\rm\nu_e$.

$\bullet\$1977: ...and probably a third!
(The ``tau'' neutrino, or $\nu_\tau$.)

$\bullet\$1968: Neutrinos from the sun are detected.

$\bullet\$1970's: ``Atmospheric neutrinos'' from the sky are detected.

$\bullet\$1985: Atmospheric neutrinos are seeming to behave rather strangely...

$\bullet\$1987: Neutrinos from a supernova are detected!

Before I said there were three different kinds of neutrino. In the particle world, it's not always straightforward to distinguish two separate things from one thing with two aspects.

This is a little similar to how some early astronomers didn't know if the ``morning star'' and the ``evening star'' were the same thing. (They are both the planet Venus.)

It is possible that all neutrinos are different manifestations of one underlying particle with properties that we see as one type or another. If that is the case, it is also possible that in time the neutrino which appears to be one type will seem to transform into another type. Of course, it takes some deeper mathematics to describe this meaningfully in a correct way, but this might give the flavor.

If such a transition occurs we call it ``oscillation.'' The name is because if it does happen it must happen bilaterally and obeying certain laws of probability. So a large number of neutrinos of one type at some time would change back and forth between types as they traveled. If we have a case with two types and look at how many of one type there are as they move along, the plot of a sine wave would correctly show the expected number of that type as a function of distance.

So what do we know so far...

$\bullet\$are very light or maybe massless.

$\bullet\$hardly ever interact.

$\bullet\$are incredibly abundant everywhere.

$\bullet\$come in several types:

neutrino	``anti-neutrino''	...go with
$\nu _{\rm e}$	( $\overline{\nu}_e$ )	electron, protons, neutrons
$\nu_\mu$	( $\overline{\nu}_\mu$)	muons, ...
$\nu_\tau$	( $\overline{\nu}_\tau$)	tau leptons, ...

We do not yet know whether neutrinos oscillate.

Back to the atmospheric neutrinos and their alledged ``strange behavior.''

Whatever we use to do this should be:

Big because neutrinos are hard to catch,

Deep because many other things from the sky are confusing,

Clear so we can detect the particles in such a big device,

Accurate because we need to distinguish different types.

Rather than show the single picture I did in my talk, here is the Super-Kamiokande home page to look at for many more interesting photos and diagrams.

It made the observation that has convinced most particle physics to take the ``atmospheric neutrino anomaly'' seriously as evidence for neutrino oscillations.

Here is a copy of a slide presented at a conference ``Neutrino '98'' showing the results of Super-Kamiokande's investigations of atmospheric neutrinos.

The point of this set of plots is that the data clearly do not match the expectations for where neutrinos are coming from. The data are the points; the shaded boxes are the output of calculations for what would be expected if the phenomenon of oscillations does not occur. The X-axis is direction: upward is on the left and downward on the right. The top plot is for electron type neutrinos and the bottom for muon type. Super-Kamiokande cannot easily detect tau type neutrinos.

The notes in the margins give several ways to quantitatively analyze this, but the point is that the points clearly do not match the prediction. In fact, even if one is allowed to ``re-scale'' the prediction representing our possible lack of knowledge of the total rate expected, the shape is still clearly different. This is evidence that the prediction is based on a wrong or incomplete way of looking at the problem. Further work shows that including the possibility of oscillations gives a prediction that does match the data well. (not shown in this talk)

$\bullet\$ ``Observation'' is different from ``Experiment''
and generally we prefer the latter before fully accepting new phenomena.

$\bullet\$ Reproducibility is absolutely essential, and
Confirmation is necessary for important results.

Confirmation is much more general than just reproducing a result. To reproduce a result by doing exactly the same experiment shows that the result is somehow real. It does not show how the result occurred. To do that, in general some slightly different experiment is needed.

Consider the example at the beginning of this talk: nuclear beta decay. People had shown easily and quickly that the distribution of energies of electrons was as depicted above. The question was still open if this was because of the underlying physics (suggesting the existence of the neutrino!) or because the detectors could not measure so well and ``smeared'' the distribution by random errors. The existence of the neutrino (while generally accepted) was not really confirmed until Reines and Cowan did the inverse beta decay experiments in the 50's.

$\bullet\$We can choose one neutrino type.

$\bullet\$We can choose an energy range.

$\bullet\$We can know the direction well.

$\bullet\$We can generally understand the beam as well as the detectors.

If we can make a beam of neutrinos somewhere on the earth and measure it as it travels past two places, we can say something really conclusive about the oscillation hypothesis.

We can measure neutrino energy accurately using the product particle energy if we know these things:
$\rightarrow\$The mass, energy, and direction of that particle
$\rightarrow\$The mass of the resting particle it came from
$\rightarrow\$The mass of the other leftover particle
$\rightarrow\$The direction of the incident neutrino

This is only exactly true for the case when there are two particles before and two after the collision. Fortunately, that case is relatively common and usually we can separate cases when that is true from those when it is not. In general for that case, we are looking at a neutron in the nucleus interacting with the neutrino to make a proton and a muon.

Again, rather than pick individual photos, as I had to in my limited time talk, please browse through the K2K home page for information on the experiment designed to check the Super-Kamiokande result. Note particularly that some nice photos are on the ``Introduction'' link as well as the ``photos'' link.

Here is a page describing one specific project of mine within the experiment. It has more photos and a general introduction to this particular piece of equipment. This project was designing and building a small detector for part of the charged particle beam that gives us our neutrinos.

My more recent work is more closely related to computer modeling of the experiment. Unfortunately I do not have a public page for this work yet.

K2K started taking data last year and is running right now. While it will take some time to get fully meaningful results, we already have some sense of what is going on and what we expect to see for the next few years that we plan to continue this experiment.

This graph showing the predicted and measured distribution of particle energies at our front detector gives us confidence we understand our beam and essentially know what to expect at the far detector if oscillations do not occur.

1999 Apr-Nov totals

Category	Obs.	Expected
		No Osc.	$\rm 3\cdot 10^{-3}eV^2$	$\rm 5\cdot 10^{-3}eV^2$	$\rm 7\cdot 10^{-3}eV^2$
Fully contained:
in fid. vol.	3	$12.2^{+1.7}_{-1.9}$	$8.0$	$5.4$	$4.6$
out of fid.	3	$5.5^{+1.1}_{-1.2}$	$3.7$	$2.4$	$2.1$
With OD activity:
ID+OD ``crossing''	2	$4.2\pm 1.6$	$3.2$	$2.0$	$1.3$
OD contained	4	$8.7\pm 3.3$	$5.6$	$3.6$	$2.8$
TOTAL:	12	$31\pm 4$	$20$	$14$	$11$

The question now becomes: Is 12 closer to 31 or 14 or 11?
--Is it so far from either that we must have just done something wrong?

This second half is quite a serious consideration and of course really needs more detailed input (which is available, but not within the time of this talk). The general issue of how well a result matches a prediction is crucial to experimental science. The answer for this case pivots around the large errors quoted. Of course 12 is closer to 11 than to 31, but at this point it is not conclusive that this is exactly the right answer or even that we have made no large mistakes. We must be very careful against prematurely drawing any conclusions based on such small numbers, but it seems we might already have a strong hint at what our outcome might be....

One might wonder why I have been so very (perhaps overly?) cautious about the interpretation of our results. I have been asked if we have some underlying worry that we have in fact made a mistake at some point and need to cover for that. The truth if that our group (and I personally) does not believe there are any mistakes in what we chose to present so far. We think we are on the right track. It is part of a general lesson about how to do science to show sufficient caution about early interpretation of results, especially in cases like this where there are not so many data.

Our experiment is still running and taking more data. By the summer of 2000 we should have enough data taken and analysed to make more concrete statements about its interpretation. Our current plans are to run until sometime around the end of 2003 to the beginning of 2004 to aquire enough data to answer the question we set out to answer: Is the atmospheric neutrino result due to oscillations with a certain set of parameters we can probe?