Measurements of the solubility of carbon dioxide in water date back to at least the beginning of the nineteenth century. However, the first accurate measurements to appear are those of Bunsen (1). Since that time there have been over 75 investigations of the solubility of CO₂ in water at low pressure.

      Carroll et al. (2) reviewed all the solubility data for CO₂ in water for pressures below 1 MPa and for temperatures from 273 to 433 K (0 °C to 160 °C). They developed a model for calculating the solubility of CO₂ in water over this range of pressure and temperature using a Henry's law approach.

      Most of the papers used in the review of Carroll et al. (2) are evaluated later in this volume. However, a few papers report values which Carroll et al. (2) concluded were accurate, but contain only one or two points. Thus, it was concluded that they did not warrant a thorough evaluation; however, they did deserve some mention. They are: Just (3), Christoff (4), von Hammel (5), Essery and Gane (6), Bartholome and Friz (7), Enders et al. (8), and Davies et al. (9).

     The model of Carroll et al. (2) was used to calculate the solubility of water when the CO₂ -partial pressure was equal to 101.325 kPa (1 atm). That is, the product of the mole fraction of CO₂ in vapor and the total pressure (y₁x P) equals 101.325 kPa. These values are tabulated in Table 1 (next page)and plotted on Figure 1 . The values of the solubilities were then regressed to obtain the following correlation:

where x₁ is the mole fraction of carbon dioxide in the liquid at a partial pressure of 1 atm, and τ = T/100 where T is in K. This equation is valid for temperatures between 273 and 433 K (0 °C and 160 °C). The average absolute deviation between the model and Eq. (1) is 6x10^-5 mo1 %. Values based on the correlation are also included in Table 1. From Table 1 and Fig. 1 it can be seen that the solubility, as defined here, becomes weak function of the temperatures for temperatures greater than about 373 K. In fact the solubility shows a shallow minimum at about 421 K.

     The solubilities calculated using the model of Carroll et al. (2) were then converted to Ostwald coefficients (volume of gas per unit volume of solvent, both at the specified temperature). For this conversion the density of carbon dioxide was calculated using the truncated virial equation. Second virial coefficients were taken from Angus et al. (10). The density of water was taken from The Steam Tables (11). The Ostwald coefficients are also listed in Table 1. The average difference between the Ostwald coefficients listed in Table 1 and those that would be obtained by assuming that CO₂ is an ideal gas is 0.33% and the maximum error is 0.68 %.

     It is interesting to note that the minimum solubility, when expressed in Ostwald coefficients, occurs at about 408 K (0.300 cm ³ CO₂ / 1 cm³ H₂O). The difference in the location of the minima for the two solubility units is largely due to the fact that the density of water is a function of temperature.

     Using the procedure of Wilhelm et al. (12), the change in enthalpy, entropy and heat capacity for solution were derived from Eq. (1). These values are included in Table 1 as well. The minimum in the solubility noted above is the point where the calculated enthalpy change on solution is zero. The change in the sign of the enthalpy change on solution more clearly reveals the minimum solubility than the solubility values themselves. From Table 1 it can be seen that mole fraction solubilities at temperatures from 413 K to 428 K are 1.72x10^-5 (to three significant figures) whereas the enthalpy change on solution clearly changes sign around 421 K. Finally, this approach predicts that the ΔC_p is a linear function of temperature with (∂[ΔC_p]/∂T = 0.08 kJ mo1^-1 K^-2.

     The effect of pressure on the solubility of CO₂ is shown in Figure 2 . The curves on this figure were calculated using the model of Carroll et al. (2) . A few data points from the literature are plotted on this figure.

Experimental Data:   (Notes on the Nomenclature)