A generalized dimensionless formulation has been developed to predict the spatial distribution of microwave power and temperature. The 'dimensionless analysis' is mainly based on three numbers: wave number, N-w; free space wave number, N-w0; and penetration number, N-p, where N-w is the ratio of sample thickness to wavelength of microwaves within a material, N-w0 is based on wavelength within free space and Np is the ratio of sample thickness to penetration depth. The material dielectric properties and sample thicknesses form the basis of these dimensionless numbers. The volumetric heat source due to microwaves can be expressed as a combination of dimensionless numbers and electric field distributions. The spatial distributions of microwave power for uniform plane waves can be obtained from the combination of transmitted and reflected waves within a material. Microwave heating characteristics are obtained by solving energy balance equations where the dimensionless temperature is scaled with respect to incident microwave intensity. The generalized trends of microwave power absorption are illustrated via average power plots as a function of N-w, N-p and N-w0. The average power contours exhibit oscillatory behavior with N-w corresponding to smaller N-p for smaller values of N(w)0. The spatial distributions of dimensionless electric fields and power are obtained for various N-w, and N-p. The spatial resonance or maxima on microwave power is represented by zero phase difference between transmitted and reflected waves. It is observed that the number of spatial resonances increases with N-w for smaller N-p regimes whereas the spatial power follows the exponential decay law for higher N-p regimes irrespective of N-w and N-w0. These trends are observed for samples incident with microwaves at one face and both the faces. The heating characteristics are shown for various materials and generalized heating patterns are shown as functions of N-w, N-p and N-w0. The generalized heating characteristics involve either spatial temperature distributions or uniform temperature profiles based on both thermal parameters and dimensionless numbers (N-w, N-w0, N-p). (c) 2005 Elsevier Ltd. All rights reserved.