In N2O5_Inorg_Org gamma function, cap A (alpha) factor at 3.2e-8

This commit addresses the issue in geoschem/geos-chem #907.

In the computation of the gamma (rxn probability) for N2O5, the
alpha factor (here called A) in routine N2O5_Inorg_Org can get
too large.  Viral Shah writes:

   The issue is that A scales as Volume/Surface area and becomes large
   for coarse-mode sea salt aerosols. The parameterization is based on
   the assumption that the N2O5 reaction is volume-limited, but Gaston &
   Thornton (2016) show that the reaction is largely diffusion-limited
   for particles larger than ~100 nm radius.

   I discussed this with Chris Holmes @cdholmes, and we propose to fix
   this by capping A to 3.2e-8, which is the value reported in
   Bertram & Thornton (2009). This sets the maximum gamma value
   to ~0.036, and would be consistent with the lab results.

Signed-off-by: Bob Yantosca <yantosca@seas.harvard.edu>

KPP/fullchem/fullchem_RateLawFuncs.F90

@@ -2857,6 +2857,10 @@ SUBROUTINE N2O5_InorgOrg( H,      volInorg, volOrg, H2Oinorg,              &
        ! Their paper suggested an approximated value of A = 3.2D-8
        A = ( ( 4.0_dp * volTotal ) / ( speed * areaTotal ) ) * KH
 
+       ! Cap A at 3.2D-8 to prevent high gamma values when vol/area is high.
+       ! See Github issue: https://github.com/geoschem/geos-chem/issues/907
+       A = MIN( A, 3.2e-8_dp )
+
        ! k2f - reaction rate constant of N2O5 with H2O
        ! From McDuffie (2018): k2f = 2.14D5 * H2O
        ! This linear water dependence is not accurate at large