THERMR without nin

[marquezj]

In the current implementation THERMR takes two ENDF-6 files: nin (processed from the neutron sublibrary) and nendf (a TSL ENDF-6 file). Depending on the application you might need the data from nin, or not. In a very common use case, when you process a library with THERMR and ACER, you only need MF=3 and MF=6 for MT=mtref and MT=mtref+1 (if it is a solid).

Maybe we could make nin optional. That would simplify the generation of thermal ACE files.

What do you think?

[marquezj]

Continuing on this, it is possible to generate an ACE file with from a TSL ENDF-6 file using any ENDF-6 neutron file as a "template". For instance, this input would process the aluminum TSL using the neutron-neutron library (MAT=25):

20 22 /
'Dummy ENDF-6 tape' /
25 0 /
.01 /
0/
broadr
20 22 23 /
25 1 /
0.01 /
293.6 /
0/
thermr
31 23 41 /
53 25 32 1 2 1 0 1 237  0 /
293.6 /
0.001 5.0 /
acer
20 41 0 51 61 /
 2  1  1  .00 0/ 
' Aluminium at 293.6K' /
 25 293.6 al-met 2 /
 13000 13027 /
 237 200 238 0 1 5.0 2/
stop

The only strange effect of this (and the only place where the neutron ENDF-6 file affects) is that the atomic weight ratio of the file would be off. In this case:

al-met.00t 1.000000 al-metal.ace 0 1 1 1783786 0 0 2.530E-08

At least in OpenMC this value is read when the library is loaded, and never used again.

From a physical perspective it makes sense, because the atomic weight only plays a role above the energy range where the S(a,b) are used, and in that range the atomic weight ratio of each isotope would be used. But, maybe older codes use that value for something, for historical reasons.