##
So, what’s wrong with the CDF measurement of Lambda_b lifetime? *June 19, 2007*

*Posted by apetrov in Particle Physics, Physics, Science.*

trackback

trackback

Today a new preprint from D0 collaboration appeared on ArXiv. They report on a new measurement of Lambda_b-baryon lifetime, working with semileptonic channels of Lambda_b decay. There is something about that baryon.

First, for a number of years (I’d say about 10 years) there was a sizable discrepancy between theoretical predictions and experimental measurements of this lifetime (now, realistically people compared ratios of lifetimes of this baryon and neutral B0 meson — the reason being that many theoretical and experimental uncertainties cancel out in that ratio). This problem seemed to be sorted out a couple of years ago after this paper, which included up-to-date perturbative and non-perturbative corrections to the ratios of Lambda_b and B0 lifetimes and included all previous theoretical updates. The final prediction is 0.86 +- 0.05, which overlaps nicely with the world average of experimental results of 0.80 +- 0.05. This corresponds to the current world-averaged lifetime of Lambda_b-baryons of 1.230 +- 0.074 picoseconds — that’s a really short lifetime on a human scale of things!

Then, as nicely reported by Tomaso here, CDF had a new number for this lifetime, 1.593 + 0.083 – 0.078 +- 0.033 ps, which is clearly quite a bit above the world-average. Moreover, it is also larger than the D0 number in the same decay channel. Now, today, D0 provided a new number, in a different decay channel (semileptonic), 1.290 + 0.119 – 0.110 (stat) + 0.087 – 0.091 (sys), which is more consistent with the old expt world average (and theoretical predictions) than the CDF’s large number… even given the larger error bars (which is a manifestation of a more challenging measurement when your decay channel involves a neutrino). So, can there be something wrong with the CDF result (or it’s just a statistical fluctuation)… or all other results are fluctuations and CDF is right? 🙂

Now, of course they are all consistent at some level… yet it’s interesting that CDF comes out on the upper side of things…

Hmmm. Taken at face value, 1.593-0.085 and 1.290+147 aren’t too different – compatible at the level of a bit less than 2-sigma.

I would tend to think that it is harder for a measurement with smaller systematics to be wrong. Assuming one has to inflate the systematic components alone to bring the two results to within 1-sigma, one would need to multiply the CDF systematics sixfold, while a twofold increase in the D0 systematics obtains the same result.

What I mean to say is that while statistical uncertainties cannot be wrong unless something really bad happened to some analysis code (bugs are always there, true), systematics can indeed be underestimated by honest experimenters. Two sigmas are not a notitia criminis, but it seems to me as if the CDF result is stronger, not by virtue of the overall smaller error bars, but by virtue of the small systematics.

Cheers,

T.

Hi Tommaso,

So would you put your tenure against CDF systematics estimates in this measurement? 🙂 At any rate, the uncertainty of the CDF result is still statistics-dominated, so I’d like to wait and see when CDF publishes an update…

Regards,

–Alexey.

Yes, the fact that the statistical error bar is larger is an additional guarantee. On one side, it shows confidence in the methodology – experimentalists would not spend too much brain power to squeeze the systematical error bar to that level given the overall error does not gain that much – and on the other side, it is a promise that we will see a better result out soon.

Cheers,

T.