get code
git clone ssh://
cd mu2eccfc 
make Mu2eCCFCNoNtuples

sample job:

Mu2eCCFCNoNtuples cd3fcl2mergedConv cd3fcl2mergedDIO rpcFilesBkgFrames crFiles 1 >& results.out  &

There is also a that contains an ntuple version of the code; other versions of the code optimize the start time of the measurement period. No one but experts should use either of those options now. The Mu2eCCFCNoNtuples version uses only cutAndCount histograms and optimizes only the momentum window.

These instructions work on mu2egpvm0*. No guarantees for other machines. In order to get the code to compile you need BOOST and ROOT. My advice is to use whatever version of the Offline you are using and setup mu2e, source from that. Otherwise you will be chasing versions. If you want to do it yourself, you don't need my advice. As of this writing Mu2eCCFCNoNtuples compiles cleanly: no warnings.


The code combines signal and background and uses Feldman-Cousins to find an optimized 90% CL momentum window. The original code is from Markus Roehrken and was used for the optimization of the stopping target. I have adapted it for our purpose and added bugs. Various pieces of Markus' original code are left but not used for later cannibalization. It also uses his version of the Czarnecki spectrum, not what is in the Offline. They differ very slightly and the differences only change the efficiency estimate, not the result.

The code also calculates the 5 sigma discovery R_{\mu e} under the assumption that the efficiency is known exactly, using the F-C construction.

Some subroutines from Louisville are at the front that can be used to check the RPC normalizations and document my use of them.


The estimator uses standard cutAndCount histograms for conversions, DIOs, and RPCs. Cosmic Ray files don't exist yet but a dummy file is used. More background sources can be added in a straightforward manner. Antiproton backgrounds are purely based on scaling from very old results described in Chapter 3 of the TDR.

The code can be invoked with four strings and a double for inputs:

Mu2eCCFCNoNtuples conversionFiles dioFiles rpcFiles crFiles POTScale >& results.out

where the program takes the strings, appends .txt, and then gets the file. POTscale is a double that is the ratio to 3.6E20. It could be used for deadtime.

It creates a ROOT output with histograms, and results.out prints the relevant results.

Input Files (background mixing for CE and DIO)

Conversion Files

Before May 2017:

Were created from Analyses/test/Mu2eCCFC_conversion.fcl running on the standard CD3 samples:

the output .root files were ls'd to conversionFiles.txt and then fed to the code.

May 2017:

Come from Andrei G.

DIO Files

Were created from Analyses/test/Mu2eCCFC_DIO.fcl running on the standard CD3 samples:

(so 4003 only); otherwise the same as conversions. Weights for DIOs are from EventGenerator/src/ .

May 2017:

Come from Andrei G.

RPC Files

From Louisville, and they have to be in this order:


since there are internal constants that give relative normalizations. An infelicity, but this will go away. These samples were not run with background frames. An extinction of 1e-10 is assumed. It's set in the code, could be an input.

May 2017: see doc-db 8923.

Cosmic Rays

No files exist yet. I have made a flat histogram with a normalization that yields 0.4 events from 103.5 to 105 MeV/c. Do not believe the error for this sample, it is forced small until a real sample exists.

PBars, etc.

Nothing exists yet but adding is straightforward.

27 May: first tag v_0_0 committed, contains stable release and input files. Still much work to do.

2 June 2016: added 5 sigma discovery code for 90%CL window and for finding optimum discovery window.

5 May 2017: added This does not use any ntuples but is the "latest" set of CD-3 runs. See doc-db 9999 for details. uses TrkAna ntuples and does t0 optimization as described in doc-db 7952.

19 May 2017: added four new text files:

cd3fcl2mergedConv.txt are the conversion files used for blessing, results documented in doc-db 9999.
cd3fcl2mergedDIO.txt are the same for DIO.
rpcFilesBkgFrames.txt are the RPC files described in doc-db 8923.
crFiles.txt is a one-line file that is a placeholder.

The data files have PID cuts. A sample run would be (after setup mu2e, source

Mu2eCCFCNoNtuples cd3fcl2mergedConv cd3fcl2mergedDIO rpcFilesBkgFrames crFiles 1 >& results.out &

and the results files contains the answers. Mu2eCCFC.root contains some histograms I use for debugging.