Implementation of neutrinoless mu to e conversion

flav-io / flavio

A Python package for flavour physics phenomenology in the Standard model and beyond

http://flav-io.github.io/

MIT License

71 stars 62 forks source link

Implementation of neutrinoless mu to e conversion #86

Closed Broko55 closed 5 years ago

Broko55 commented 5 years ago

I implemented the conversion rate for neutrinoless mu to e conversion in Au and Al targets.

coveralls commented 5 years ago

Coverage increased (+0.02%) to 93.969% when pulling 54cb0eff4a2b57bf2580e506e74eb73db60faddb on Broko55:master into 0d277f637c4e2eb8cd1849c88d9844200ddfbc4f on flav-io:master.

coveralls commented 5 years ago

Coverage increased (+0.1%) to 93.986% when pulling 3779bc00543d9d6d3954a5b2b7d30d6019413072 on Broko55:master into 545ff8611f306ad31304a8beb91ccd8b7e8488de on flav-io:master.

DavidMStraub commented 5 years ago

That's brilliant, thanks a lot! The implementation looks almost perfect. Just one small thing I noticed: the nucleon charges (GuS etc.) should be much better known now than in 2001 (see e.g. appendix 10 of 1902.08191). In fact I have implemented them (in this notation) in a private branch. Let me push this file to the master branch, and then we can rebase your implementation to use these conventions (and numbers) before merging it.

By the way, the numerical parameters (overlap intergrals and GammaCapture Al) are currently constants. Can their uncertainties be estimated somehow?

Broko55 commented 5 years ago

Hi! In the HFLAG review I only see the axial, tensor and scalar charges. Except for the scalar charges, the related operators do not contribute to the coherent process, that's why I ignored them so far.

For the capture rate I found this reference https://doi.org/10.1103/PhysRevC.35.2212 giving (gaussian?) uncertainties. As for the overlap integrals, the only reference where they have been calculated seems to be https://arxiv.org/pdf/hep-ph/0203110.pdf in this scenario and there they do not give any uncertainty estimations. Perhaps the thorough thing to do would be implementing their calculation and propagate the nuclear input uncertainties.

At least the uncertainties on GammaCapture I can implement easily and also Ti as a target nucleus.

Furthermore, I noticed that I do not use the charges GxV but rather have them hardcoded, they should be removed from the parameter file.

DavidMStraub commented 5 years ago

Indeed the GxV can be removed as they are exact integers.

I now pushed the nucleon matrix element function in 545ff8611f306ad31304a8beb91ccd8b7e8488de. The relation to your GxS should be:

from flavio.physics.edms.common import proton_charges

pc = proton_charges(par, scale)
GuS = (pc['gS_u+d'] + pc['gS_u-d']) / 2
GdS = (pc['gS_u+d'] - pc['gS_u-d']) / 2
GsS = pc['gS_s']

Broko55 commented 5 years ago

Done!

DavidMStraub commented 5 years ago

Great!

The SM uncertainties are currently not too meaningful (since the overlap integrals are constants), but since this is a null test of the SM I think that's not a big problem (until it's actually discovered).

As soon as the next flavio version is released, these can be added to the smelli global likelihood.