Margules constant volume heat capacity is implemented incorrectly

[speth]

Problem description

The values returned by cv_mole for the Margules phase are inconsistent with the values of du/dT and ds/dT that can be calculated for the phase.

Steps to reproduce

import cantera as ct
p = ct.Solution('test/data/LiKCl_liquid.yaml')
p.TPX = 900, ct.one_atm, (0.2, 0.8)
u1 = p.int_energy_mole
s1 = p.entropy_mole
cv1 = p.cv_mole
dT = 0.01
p.TP = p.T + dT, None # note density is always constant
u2 = p.int_energy_mole
s2 = p.entropy_mole
cv2 = p.cv_mole

print('from cv:', 0.5 * (cv1 + cv2))
print('from s:', (s2-s1)/dT * p.T)
print('from u:', (u2-u1)/dT)

Behavior

The three values should be approximately the same. Instead, they are:

from cv: 63534.365309023735
from s: 71849.22708415563
from u: 71848.82792830467

It is "interesting" that the difference between the value calculated by cv_mole and the other two is ~8314.4 J/mol, or approximately equal to the gas constant.

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.