Closed peterstangl closed 6 years ago
Thanks! I added a few comments.
By the way, the failing test is my fault. Will fix it soon, you can ignore it in the meantime.
@DavidMStraub I have added the B0
decay and implemented both the B+
and the B0
decays in such a way that it would now be very easy to also add the combined l1 l2
+l2 l1
LFV decays with combinations of leptons in the final state different from e,mu
as well as decays to kaons instead of pions. Would there be some use of already adding these other decays?
Good question. They are measured separately so it's not really needed, on the other hand since it's trivial to add them I guess it doesn't hurt.
You could also add the measurement from https://arxiv.org/pdf/hep-ex/0703018.pdf while you're at it :)
@DavidMStraub concerning the measurements, did you read my comment https://github.com/flav-io/flavio/pull/57#discussion_r210670958? Before adding them, it would be nice to understand what's actually going on there. Probably I'm just missing something.
No sorry, I had missed that comment.
The way they write it it's indeed confusing. I would understand it like this:
To quote a combined limit for the isospin modes, they assume there is no isospin asymmetry, which implies BR(B+ -> pi+ l+ l-) = 2 tau_B+/tau_B0 BR(B0 -> pi0 l+ l-)
and they call this BR(B -> p l+ l-)
. Note that the absence of isospin asymmetry is an assumption for the non-LFV modes which is violated at some level in the SM, but its certainly a valid assumption for the LFV modes.
The table of limits now gives three numbers for each decay mode: the charged decay, the neutral decay, and the limit assuming isospin symmetry to hold. The above relation of course does not apply to these limits. Let's take a simple counterexample: if your detector cannot see pi0s at all, the limit on the B0 mode will be completely gone, and the limit on the BR assuming isospin symmetry will be given directly by the limit on the B+ mode.
By the way, I fear the factor of 2 between the neutral and charged mode is missing in flavio :(
Missing factor of 1/2 added in 0d95eec5a72c2b19a7ce0a1c29e7d4fce0580397
The failing test should now also be fixed in 1dec447aa5748bb14bdc080854760ce7fc0523c0.
Thanks a lot for the explanation, this makes sense!
I have added the remaining decays and rebased on master.
I will also add the measurement from https://arxiv.org/pdf/hep-ex/0703018.pdf to this PR
@DavidMStraub I have also added the two measurements from https://arxiv.org/pdf/hep-ex/0703018.pdf. I have only included the LFV measurements since the non-LFV BR(B->pill)
is not included as observable at the moment (and the PDG would provide lower limits for these channels than hep-ex/0703018).
Very good!
Yes in fact B->pimumu has been observed and measured by LHCb in the meantime. It is not yet in flavio because the hadronic contributions are really tricky due to |Vud Vus| ~ |Vtd Vts| (but they don't enter the LFV decays fortunately).
@DavidMStraub, this PR defines the
Observable
instance for the combined emu+mue LFV BR(B->pi(emu+mue)) as used in https://arxiv.org/pdf/hep-ex/0703018.pdf. To this end, I have put everything used to define the LFV M->PllObservable
instances into the function_define_obs_B_Mll
and removed the currently unnecessary loopfor br in ['BR',]:
. I hope this fine.Please double-check that my assumption BR(B->pi(emu+mue)) = BR(B+->pi+ emu) + BR(B+->pi+ mue) is correct.
Note that like in the last two commits in master, the test introduced in 4056db5df4387a3aea32adbe833a369714ca2604 is failing (
KeyError: 'test_obs 1'
).