Section VII Example simulations

[ramess101]

@mostafa-razavi

7.1 : The agreement with previous simulation results on ethane, dodecane and TIP4P/2005 water is good ; However, this article does not say how liquid densities have been computed with ITIC. It is likely that the liquid densities have been either (i) taken from previous work cited in reference, or (ii) that ITIC contains some undocumented process to evaluate liquid density and related uncertainties (see section 2.5).  In case (i) ITIC is not a method of coexistence calculation and relies on previous coexistence calculations such as Gibbs Ensemble Monte Carlo simulations. In case (ii) the method may be good but this manuscript is forgetting a key point.  

7.2 In figure 23 the caption should indicate the temperatures corresponding to each isotherm.  

7.3 Ethane and dodecane, the compounds investigated in Tables IV and V with TraPPE‐UA, have been already investigated by GEMC + thermodynamic integration [ Ungerer et al. , JCP, 112, (2000), 5499] in a similar temperature range, finding also growing overestimation of saturation pressure by TraPPE‐UA when temperature decreases.  

7.4 TIP4P/2005, the model investigated in Table VII , has been already investigated by Gibbs Duhem   thermodynamic integration [Vega et al. , JCP 125 (2006), 034503] in a similar temperature range, finding also a significant underestimation of saturation pressure with this model. It would be fine to discuss possible differences between both methods

[ramess101]

7.1 : The agreement with previous simulation results on ethane, dodecane and TIP4P/2005 water is good ; However, this article does not say how liquid densities have been computed with ITIC. It is likely that the liquid densities have been either (i) taken from previous work cited in reference, or (ii) that ITIC contains some undocumented process to evaluate liquid density and related uncertainties (see section 2.5). In case (i) ITIC is not a method of coexistence calculation and relies on previous coexistence calculations such as Gibbs Ensemble Monte Carlo simulations. In case (ii) the method may be good but this manuscript is forgetting a key point.

We are case (ii). So we just need to make it more clear that we fix rhol and solve for Tsat.

[ramess101]

7.2 In figure 23 the caption should indicate the temperatures corresponding to each isotherm.

Sounds reasonable.

[ramess101]

7.3 Ethane and dodecane, the compounds investigated in Tables IV and V with TraPPE‐UA, have been already investigated by GEMC + thermodynamic integration [ Ungerer et al. , JCP, 112, (2000), 5499] in a similar temperature range, finding also growing overestimation of saturation pressure by TraPPE‐UA when temperature decreases.

This form of thermodynamic integration is a lot different from what we are doing. Essentially, you could just fit your GEMC results to the Clapeyron equation to extrapolate to as low of Tr as desired. Our approach is much more refined than that.

The issue regarding overestimation of Psat with TraPPE-UA is not really relevant, since we are not trying to validate force fields. However, I think it would be good to discuss that the Potoff potential appears to extrapolate well even to very low Tsat, since no one has validated the Mie potential near the triple point.

[ramess101]

7.4 TIP4P/2005, the model investigated in Table VII , has been already investigated by Gibbs Duhem
thermodynamic integration [Vega et al. , JCP 125 (2006), 034503] in a similar temperature range, finding also a significant underestimation of saturation pressure with this model. It would be fine to discuss possible differences between both methods

We should look into this. It would be useful to include their results on the same figure.