Unit conversion 1.602E-10 in DOSRZnrc, DOSXYZnrc etc is off by 0.011% vs current data

[dworogers]

Various RZ and other codes need to convert from scored MeV/g to Gy. Using latest codata values the conversion is 1.6021766E-10 rather than 1.602E-10 used in the codes, i.e. 0.011% higher. Not a big deal but might as well have it exact.

[ftessier]

Agreed. Since this value is exact in the SI, I would go all the way and define e = 1.602 176 634 x 10^-19 C.

[ftessier]

Found this in src/EEMF_macros_beamnrc.mortran: REPLACE{$EC_em}WITH{1.60217662E-19}; "electron charge". I suggest we simply make such a constant available throughout, along with conversion factors such as eVToMeV, MeVPerGramToGray, etc., as suggested by @MartinMartinov in #701.

[ftessier]

Something like this, in a new header egs_constants.h:

// Physical constants

const EGS_Float     e_charge = 1.602176634e-19;    // Coulomb, exact, CODATA 2018

// Conversion factors

const EGS_Float     eV_to_MeV           = 1e-6;
const EGS_Float     eV_to_Joule         = e_charge;
const EGS_Float     MeV_to_Joule        = 1e6 * eV_to_Joule;
const EGS_Float     MeVPerGram_to_Gray  = 1e3 * MeV_to_Joule;

// (and more...)

[mchamberland]

Just bumped into something where two codes were using slightly different values for the electron charge, which made it trickier than it should have been to tell if there was any difference between the two codes.

So, what I'm saying is: I hope this gets addressed soon!

[ftessier]

Yes, the EGSnrc rewrite we are considering will use the exact value of $\large e$ as a constant, and perhaps a solid unit conversion library (or at the very least conversion factors as pointed out above). No more hard coded numbers!