Incorrect implementation of partialMolarEntropies in SurfPhase

[speth]

Problem description

The implementation of partialMolarEntropies in SurfPhase does not give thermodynamically consistent results.

Steps to reproduce

>>> import cantera as ct
>>> s = ct.Solution('test/data/surface-phases.yaml', 'Pt-surf')
>>> s.TPX = 300, ct.one_atm, 'Pt(s): 0.7, H(s): 0.1, O(s): 0.2'

# Check g_k = h_k - T * s_k
>>> s.chemical_potentials
array([-9.06035518e+05, -4.22973366e+07, -1.14433845e+08])

>>> s.partial_molar_enthalpies - s.T * s.partial_molar_entropies
array([ 0.00000000e+00, -3.65375419e+07, -1.10402994e+08])  # no agreement

# These pairs should be equal
>>> sum(s.partial_molar_entropies * s.X), s.entropy_mole
(581.0752089855705, 7302.323496843512)

>>> sum(s.partial_molar_enthalpies * s.X), s.enthalpy_mole
(-25560030.399718467, -25560030.39971847)

>>> sum(s.chemical_potentials * s.X), s.gibbs_mole
(-27750727.448771518, -27750727.448771525)

>>> s.enthalpy_mole - s.T * s.entropy_mole
-27750727.448771525  # Equals gibbs_mole

System information

• Cantera version: 2.6.0 or main at 1ab81ac.

Additional context

This was discovered as part of #1299, which implements Cantera/enhancements#114.