lirate88.f

LIRATE88 computes the burning rates of ⁶Li, ⁷Li and ⁹Be by protons at every point, from the fits of Caughlan and Fowler (1988) (spelled "Caughlin" in the comment), scaled by user factors. Called with J = 1 it computes the rates for the model passed in, stored as the end-of-step set; with J = 2, for the saved model at the start of the step, stored as the start-of-step set (the convention its comment now states).

• Who calls it: the main program, main.f (with J = 2 before the structure iteration and J = 1 after it, then LIBURN), and GETW (J = 1, with its midpoint model) for rotating models.
• Inputs: with J = 1, the arguments HD ($\log\rho$), HT ($\log T$), HCOMP(1) ($X$); with J = 2, the old model HDO, HTO, HCOMPP from OLDMOD_MOD. With J = 2, HD and HCOMP are not read and HT serves only the temperature cut (line 65↗); at that call (main.f, line 815↗, before the structure iteration) the old model is on the current grid and is the model passed in: HPOINT splines HCOMPP onto the new points (hpoint.f, lines 656–663↗) and ends by copying its splined HTO, HDO into HT, HD (hpoint.f, lines 691–692↗, hpoint.f, lines 733–739↗), and without rezoning (the helium flash) main.f copies HT, HD into HTO, HDO (main.f, lines 774–780↗) (settled from the code, Nov 9, 2026). The scale factors SLI6, SLI7, SBE91, SBE92, SBE93 from NMLPAR_MOD (namelist values read by PARMIN; formerly /BURNSCS/, G. Somers 6/14). The constants (CLN, CC13, CC23) come from CONSTANTS_MOD.
• Outputs: RLI6, RLI7, RBE9 in NEWRAT_MOD (J = 1) or RLI60, RLI70, RBE90 in OLDRAT_MOD (J = 2), read by LIBURN, LIBURN2, BURSMIX and GETW.

With $T_9 = T/10^9$, each rate is the Caughlan–Fowler fit $N_A\langle\sigma v\rangle(T_9)$ for: ⁶Li(p,³He)α; ⁷Li(p,α)α; and the sum of ⁹Be(p,γ)¹⁰B, ⁹Be(p,d)2α and ⁹Be(p,α)⁶Li. For example ⁷Li(p,α)α, with $T_{9a} = T_9/(1 + 0.759\,T_9)$:

$$N_A\langle\sigma v\rangle = \frac{1.096\times10^{9}}{T_9^{2/3}}\,e^{-8.472/T_9^{1/3}} - 4.830\times10^{8}\,\frac{T_{9a}^{5/6}}{T_9^{3/2}}\,e^{-8.472/T_{9a}^{1/3}} + \frac{1.06\times10^{10}}{T_9^{3/2}}\,e^{-30.442/T_9}.$$

Each is multiplied by its scale factor and by $\rho X$, giving the rate per nucleus, $\lambda = \rho X\,N_A\langle\sigma v\rangle$ (per second, if the fits are in the usual cm³ mol⁻¹ s⁻¹; the code does not say). The loop runs from the centre outward and stops at the first point where both the new and the old $\log T$ are below 6 (TLIM); from there to the surface the set being filled (NEWRAT_MOD's with J = 1, OLDRAT_MOD's with J = 2) is set to zero.