yuetingchen

Recently, I looked at the equilibrium constants of H2S(g) in the GWB database (thermo.tdat). I believe some of the values are not right, and I’d like to share my observation with everyone. Below is what extracted from “thermo.tdat”: ------------------------------------------------------------------------------------- H2S(g) mole wt.= 34.0758 g 2 species in reaction 1.000 H+ 1.000 HS- -8.0615 -7.9791 -7.9737 -8.0646 -7.1449 -4.6551 .4746 9.3231 ------------------------------------------------------------------------------------------ The above data indicate that H2S(g) becomes more soluble when temperature increases above 100oC. That is at odds with my understand of gases solubilities in water, i.e., in general gases become less soluble with temperature increase. I understand that “thermo.tdat” is based on EQ3/6 database data0, but I cannot tell which version of LLNL’s “data0” is used to generate “thermo.tdat”. I checked the YMP version of EQ3/6 database: “data0.ymp.R2”. Below is what I found for H2S(g) from this database: ------------------------------------------------------------------------ H2S(g) sp.type = gas * EQ3/6 = ymp.R2, ymp.R0, com, ree, alt, sup YMP Qualification status = Q * mol.wt. = 34.082 g/mol V0PrTr = 0.000 cm**3/mol [source: 78hel/del] **** 2 element(s): 2.0000 H 1.0000 S **** 3 species in aqueous dissociation reaction: -1.0000 H2S(g) 1.0000 H+ 1.0000 HS- * **** logK grid [0-25-60-100C @1bar; 150-200-250-300C @Psat-H2O]: -8.0781 -7.9759 -7.9295 -7.9572 -8.0759 -8.2750 -8.5671 -9.0074 * * gflag = 1 [25C,1bar: reported delG0f used] * P-T extrapolation algorithm: Cp(T), const-V(P) integration * P-T extrapolation alg. ref.: 82wag/eva, N/A * reference-state data source: 82wag/eva * delG0f = -8.021 kcal/mol * delH0f = -4.931 kcal/mol * S0PrTr = 49.185 cal/(mol*K) * Cp coefficients [source: 60kel ] * T**0 = 0.78100000E+01 * T**1 = 0.29600000E-02 * T**-2 = -0.46000000E+05 * Tlimit = 2026.85C ----------------------------------------------------------------------- The trend of temperature dependence of log K data appears making more sense.