Surface-thermal capacity of from measurements made during steady-state evaporation

When D2O(l) evaporates into its vapor under steady-state conditions with the temperature field in the liquid arranged so that there is no buoyancy-driven convection and the Marangoni number is less than approximately 100, it is found that the interface is quiescent and thermal conduction to the interface supplies energy at a sufficient rate to evaporate the liquid. However, if the evaporation rate is raised so that the Marangoni number goes above approximately 100, the interface is transformed: a fluctuating thermocapillary flow occurs, and thermal conduction no longer supplies energy at a sufficient rate to evaporate the liquid. An energy analysis indicates conservation of energy can be satisfied only if thermocapillary convection is taken into account, and the surface-thermal capacity csigma is assigned a value of 32.5+/-0.8 kJ/(m2 K) when the temperature is in the range -10 degrees C< or =TLV< or =3.7 degrees C. This value is consistent with that found previously for H2O, and application of the Gibbs model gives a qualitative explanation for the value. Once the value of the surface-thermal capacity is known, the local heat flux along the interface can be calculated and statistical rate theory can be used to predict the local vapor-phase pressure on the interface. Since this theory introduces no adjustable parameters, the predicted pressure can be compared directly with that measured: this comparison indicates the mean of the pressures predicted to exist on the interface is in close agreement with those measured approximately 20 cm above the interface, and the small pressure gradient along the interface is consistent with the thermocapillary convection predicted from the interfacial temperature gradient.