radsub06.f

Adds radiation, optionally, to the gas equation of state interpolated in the OPAL 2006 tables. The older versions were RADSUB and RADSUB01 (both removed since rp_1).

• Who calls it: ESAC06, only when its argument irad is 1, and it passes irad on (esac06.f, lines 481–485↗); so irad is always 1 here, and the branch for irad $= 0$ never runs (unused?).
• Modules: EOS06MISC_MOD, which merges the former /EEOS06/ (ESACT, EOS(MV)) and /BEOS06/ (ZZ, IRI, INDEX, NTA, NRA) with two other blocks; shared with ESAC06, OEQOS06, READCOEOS06, RHOOFP06, T6RINTEOS06, CALCAD and EQBOUND06. This routine uses EOS and IRI only.
• Output: EOS(IRI(1..9)) overwritten.

As in RADSUB01 (equations in RADSUB; both removed since rp_1), with the 2006 order of quantities: IRI(4) is $\partial E/\partial\rho$, IRI(5) is $c_V$, IRI(6) is $\chi_\rho$, IRI(7) is $\chi_T$, IRI(8..10) the $\Gamma$s. With irad $\ne 0$, the radiation terms are added, $\partial E/\partial\rho$ becomes $\partial E_g/\partial\rho - E_r/\rho$, and the differences (with minus without radiation) are added to $\Gamma_1$ and to GAM2PT. The line that would add the difference to eos(iri(10)) ($\Gamma_3$, line 102↗) is commented out, while RADSUB01 corrected all three $\Gamma$s; it has nothing to correct, since EOS(10) is never interpolated (the tables hold nine quantities, readcoeos06.f, lines 124–126↗) and nothing reads it (settled from the code, Nov 9, 2026). With irad $= 0$, the routine only rewrites the gas values, with $c_V$ multiplied by $Nk/M$ (molenak/tmass).