Driving force of the interface migration, caused by changing the thermodynamic variables (such as composition, temperature, etc.) has been investigated. Three experiments have been performed to prove that the coherency strain energy is the driving force. First, by temperature change the migration of liquid films (LFM) and grain boundaries (DIGM) have been found in 95Mo-5Ni system. And by adding Fe as an alloying element at 1400℃, LFM and DIGM have also been found. The driving force can be varied by changing the temperature difference, or the amount of Fe added. The migration rate was varied parabolically with the coherency strain or the concentration difference of the solid. This result is consistent with the coherency strain energy scheme. In this experiment, lowering the temperature makes the thin diffusion layer strained compressively, but the process of adding Fe makes that strained tensily. This property has been used in the second experiment. In the second experiment, the condition that the coherency strain is zero while the chemical free energy difference due to composition difference remains finite, has been made using the ternary Mo-Ni-Fe system. On lowering temperature to 1400℃, LFM has occurred as in the first experiment. At the same time, Fe which makes the thin layer strained tensily has been added. With increasing Fe the migration rate has been decreased and become zero. At this point, the driving force of LFM is zero, which means that the compressive strain caused by lowering temperature has been matched perfectly with the tensile strain caused by Fe addition. Therefore, the coherency strain is zero, but the composition difference remains finite. Further Fe addition has made LFM occur again. In the third experiment, the anisotropic property of the coherency strain energy has been investigated in 80Co-20Cu system whose elastic constants are much anisotropic. By lowering temperature from 1300℃ to 1150℃, LFM has also occurred. Faceting of the liquid films have been found, which means that the migration rate depends on the orientation of retreating grain. Such a phenomenon cannot be explained by the chemical free energy difference.
DIGM and the discotinuous precipitation in Mo-Ni system at temperature lower than the peritectic point (1362℃) has also been investigated. Discontinuous precipitation in 95 Mo-5Ni and 85Mo-15Ni system has occurred easily by lowering temperature and its rate has become faster with temperature difference. By replacing the Ni-Mo matrix in the specimen sintered at 1400℃ with the liquid Cu-Fe at 1400℃, DIGM and surface recrystallization have occurred. In this case the coherency strain is tensile. The same experiment as previous one using the processes both of lowering temperature to 1350℃ and adding Cu-Fe, has been performed. The similar result has occurred. Therefore, the driving force in this case has also been proved to be the coherency energy.