diff --git a/QA/tests/cosmo_h2cco2_opt/cosmo_h2cco2_opt.nw b/QA/tests/cosmo_h2cco2_opt/cosmo_h2cco2_opt.nw
index c4dcfe90cb..c8edfd8e0b 100644
--- a/QA/tests/cosmo_h2cco2_opt/cosmo_h2cco2_opt.nw
+++ b/QA/tests/cosmo_h2cco2_opt/cosmo_h2cco2_opt.nw
@@ -69,6 +69,7 @@ end
 
 cosmo
  do_gasphase f
+ do_cosmo_ks
  cavity ses
  ndiv xfine
 end
diff --git a/src/solvation/cosmo_initialize.F b/src/solvation/cosmo_initialize.F
index 1dc0e794b9..e6ae156de9 100644
--- a/src/solvation/cosmo_initialize.F
+++ b/src/solvation/cosmo_initialize.F
@@ -761,12 +761,12 @@ c
 c
 c <-- MN solvation models
 c
- 9973 format(' charge correction applied to potential')
+ 9973 format(' charge correction applied to potential, '
+     $       'doi: 10.1021/acs.jctc.5c01368')
  9972 format(' surface charge correction     = ',a)
  9971 format(' Cosmo: York-Karplus, doi: 10.1021/jp992097l')
  9970 format(' Cosmo: Klamt-Schuurmann, doi: 10.1039/P29930000799')
-c TODO: describe the GEPOL modifications and publish a new article
- 8870 format(' SES surface, doi: 10.1002/jcc.540151009, ??')
+ 8870 format(' SES surface, doi: 10.1021/acs.jctc.5c01368')
 
  9969 format(' Using -Lebedev- grid')
  9968 format(' -iangleb',I4)
diff --git a/src/solvation/cosmo_input.F b/src/solvation/cosmo_input.F
index 2796d07b3b..5ae16a020b 100644
--- a/src/solvation/cosmo_input.F
+++ b/src/solvation/cosmo_input.F
@@ -378,8 +378,6 @@ c
           cosmo_cavity = COSMO_CAVITY_VDW
         else if (inp_compare(.false.,'ses',token)) then
           cosmo_cavity = COSMO_CAVITY_SES
-c     use KS model by default when using SES cavity
-          do_cosmo_model=DO_COSMO_KS
 c     correct potential when using SES
         cosmo_ptcor=.true.
         else
diff --git a/src/solvation/hnd_coschg.F b/src/solvation/hnd_coschg.F
index a24de7caff..38107e4050 100644
--- a/src/solvation/hnd_coschg.F
+++ b/src/solvation/hnd_coschg.F
@@ -457,6 +457,12 @@ C>   "Treatment of the outlying charge in continuum solvation models",
 C>   <i>J. Chem. Phys.</i> (1996) <b>105</b>, pp 9972–9981, DOI:
 C>   <a href="https://doi.org/10.1063/1.472829">
 C>   10.1063/1.472829</a>.
+C>
+C>   [12] R. de P. Soares, D. Mej&iacute;a-Rodriguez, E. Apr&agrave;,
+C>   "Recent Improvements to the NWChem COSMO Module",
+C>   <i>J. Chem. Theory Comput.</i> (2025), DOI:
+C>   <a href="https://doi.org/10.1021/acs.jctc.5c01368">
+C>   10.1021/acs.jctc.5c01368</a>.
 C>
       subroutine hnd_coschg(g_dens,ndens,rtdb,geom,basis,nat,nefc,
      &                      efcc,efcs,msrf,efcz,efczz,efciat,ratm,
@@ -1144,7 +1150,7 @@ c     scaling factors or Lagrangian multipliers then additional
 c     correction terms are needed to ensure the expressions agree.
 c
 c     A consistent correction is to recalculate the total potential
-c     with the corrected charges (see Ref. [11] and cosmo_ptcor above).
+c     with the corrected charges (see Ref. [11,12] and cosmo_ptcor above).
 c     After that, adjust the elec potential accordingly, with a
 c     constant nuclear potential. This makes both formulations
 c     agree and also makes scale and Lagrangian corrections