Revised handling of ideal gas contribution

[prehner]

The PR rethinks how ideal gas models are used in feos. Main motivations are

• without providing an ideal gas model, users should not be able to evaluate caloric properties...
• ...while still being able to obtain Total results for properties that do not depend on the deBroglie wavelength (e.g., pressure and its derivatives)
• Users should be able to combine any residual Helmholtz energy model with any ideal gas model