## Monday, May 17, 2010 ... /////

### D0 claims evidence for a new source of CP-violation!

The Tevatron would like to claim that it's not quite dead yet. Tonight, the Tevatron's D0 collaboration will publish a new paper on the arXiv that is already available on their server:

Evidence for an anomalous like-sign dimuon charge asymmetry (available now, old URL)
The hundreds of people have looked at the 6.1 inverse femtobarns of their 1.96 TeV proton-antiproton collisions and picked the collisions in which two muons or antimuons with the highest transverse momentum had the same sign of the charge - either positive ("N++" events) or negative ("N-—" events). If you think that one of the minuses in "N-—" is longer, you may always buy a new LCD panel.

These numbers should be nearly equal, so the asymmetry defined as the ratio of the difference and the sum
A = (N++ - N-—) / (N++ + N-—)
should be nearly zero because there's an approximate CP symmetry between the matter and antimatter, including muons and antimuons (which also exchanges left and right in space, because of the P, but "A" is insensitive to this spatial label).

More precisely, the hadrons containing bottom quarks decay under the influence of the "slightly" CP-violating CKM-matrix. The small complex phase in this 3-by-3 matrix is the only "established" source of CP-violation in all of physics that we know (so far). That's why the standard model predicts "A" to be -0.0002 (plus minus 25%).

However, the D0-measured value is -0.0096, fifty times larger than that. The statistical error - the uncertainty from the "noise" of having a limited amount of data - is (plus minus) 0.0025 and the systematic error - from the gadgets, that can always go in the same direction - is (plus minus) 0.0015.

These two types of errors should be imagined as being "independent" of one another, i.e. as two sides of a right-angled triangle; the hypotenuse is "sqrt(0.0025^2+0.0015^2) = 0.0029" and determines the "total error".

The deviation of the "centrally" measured value, -0.0096, from the predicted value, -0.0002, is -0.0094 which is (minus) 3.2 times bigger than 0.0029. So we say that the measured value differs by 3.2 standard deviations from the predictions.

It's slightly more than three standard deviations. Is it enough of a discrepancy to claim that something must be wrong? The corresponding confidence level is something like 99.9%.

The risk that the bold conclusion about a new source of CP-violation is wrong is just 1 in 1,000, about 50-100 times smaller than what climate science calls "near certainty, an overwhelming consensus justifying witch hunts against those who want to kill the baby Earth who has fever".

However, in science, the deviation of the measured and predicted value by 3 sigma is called "just evidence" - the first word in the paper above. It's just a hint that there could be a new effect. However, the accuracy would have to improve for the deviation to go to 5 sigma - or 99.99995% confidence level - for the particle physicists to claim a discovery of a new phenomenon.

This immense increase of the confidence level can be obtained just by doubling the amount of data (at least, the statistical error will decrease enough; the systematic error has to be reduced in other ways).

So while the observation is tantalizing, 3 sigma really doesn't seem enough to reliably claim that a hole in the Standard Model has been found. Many other 2-sigma and even 3-sigma announcements of this kind had to be scrapped in the past.

While hundreds of the "world's top scientists" offer us a 99.9% evidence of a new source of violation, the people who expect that this result will go away are no deniers. They're very sensible, indeed. My guess - unsupported by direct polls at this moment, just by extensive experience with particle physicists' opinions and ways of thinking - is that most particle physicists (theorists and phenomenologists, and possibly the LHC experimenters as well) actually believe that the result will go away. Just statistically, looking at the actual history of science, such things usually go away whether you like it or not. 99.9% is just not enough to change your opinions about the good old Standard Model.

Moreover, the Tevatron folks may be nervous about the ongoing de facto death of their collider, so they may be more eager to publish impressive results. That should slightly diminish your confidence in the paper, too. I have personally doubts that the detectors, made out of matter and not antimatter, are really able to treat muons and antimuons (and their energy) so completely fairly, so I may imagine that the real systematic error is much higher. ;-)

On the other hand, if this enhanced CP-violation is real, new physical phenomena have to exist. Supersymmetry offers various tools how to get new CP-violation. In particular, read a new preprint:
Dobrescu, Fox, Martin: CP violation in B_s mixing from heavy Higgs exchange
Symmetry magazine popularization
that shows how you can reproduce these observed data via "uplifted supersymmetry", a region of MSSM with two Higgs doublets where the down-ish quark masses come from loops and the up-type Higgs doublet.

But other models of the extended Higgs sector could do it, too. The most obvious detailed way to get the enhanced asymmetry of the dimuon events is to believe that there exists a new oscillation of B-mesons and B-bar-mesons, both of which can be created from the bottom-antibottom pairs of quarks.

To produce such an extra new oscillation, you need a new quartic interaction term in the effective Lagrangian proportional to
L to make D0 happy = #.(bγ0s) (bγ0s) + h.c.
I inserted the Dirac matrices because it's actually slightly easier for me to type them and the daggers in normal HTML than the bars. ;-) After I corrected Jester's typo and modified the term so that it really changes strangeness by two, I had to re-add his Hermitean conjugate term. :-)

You're expected to submit many preprints that calculate this term from your favorite theory or model and determine the right coefficient that matches the observations. ;-)

Hat tip: Resonaances