H S Green – författare

135 We first describe the thermodynamic theory of surface tension and adsorption, by the method of the dividing surface of GIBBS. The use of a dividing surface or its equivalent is indispensable for the treatment of a curved interface, as otherwise the concepts of the area and curvature of the interface, cannot be pre­ cisely defined. In the case of a plane interface, however, the concept of the dividing surface is not necessary and a valid alternative exposition has been proposed by GUGGEN­ HEIM [3J, [4J in treating the interface zone as a separate entity of some definite thickness bounded by two mathematical planes. We make, however, little mention of this method, since it seems to be of only minor importance in connec­ tion with the statistical treatment of an interface. To avoid any ambiguity, the treatment of a spherical interface given in this article is based not on the original method of GIBBS but on the method modified by HILL [8J and KONDO [9]. This method, however, is not applicable to non­ spherical interfaces, which will not be dealt with in this article. Although all the relations for a plane interface can be deduced from the cor­ responding ones for a spherical interface by putting the curvature equal to zero, the planar and the spherical cases are considered separately because of the prac­ tical importance and easy physical visualization of a plane interface.

135 We first describe the thermodynamic theory of surface tension and adsorption, by the method of the dividing surface of GIBBS. The use of a dividing surface or its equivalent is indispensable for the treatment of a curved interface, as otherwise the concepts of the area and curvature of the interface, cannot be pre­ cisely defined. In the case of a plane interface, however, the concept of the dividing surface is not necessary and a valid alternative exposition has been proposed by GUGGEN­ HEIM [3J, [4J in treating the interface zone as a separate entity of some definite thickness bounded by two mathematical planes. We make, however, little mention of this method, since it seems to be of only minor importance in connec­ tion with the statistical treatment of an interface. To avoid any ambiguity, the treatment of a spherical interface given in this article is based not on the original method of GIBBS but on the method modified by HILL [8J and KONDO [9]. This method, however, is not applicable to non­ spherical interfaces, which will not be dealt with in this article. Although all the relations for a plane interface can be deduced from the cor­ responding ones for a spherical interface by putting the curvature equal to zero, the planar and the spherical cases are considered separately because of the prac­ tical importance and easy physical visualization of a plane interface.