Experimental And Model Study On Surface BRDF For Several Typical Materials

With the development of the technologies of solar energy, remote sensing, radiationthermometry and rapid thermal of semiconductor, a series of surface radiation problemsappeared which dominated by thermal radiation. Bidirectional reflectance distributionfunction (BRDF) is a fundamental property of rough surface, which could accuratelydescribe the distribution of the surface reflected energy in the space, so knowledge of theBRDF is crucial to emissivity modeling and heat transfer analysis.First of all, a variety of existing BRDF modeling theories are analyzed andsummarized in this paper, an improved BRDF model which based on microfacet roughsurface theory was proposed. Comparing the improved BRDF model with other models, theimproved model not only meet the energy conservation law, but also avoid the distortion ofthe value of BRDF simulation in large incident angle. Therefore the model that proposed inthis paper, which can more accurately describes the reflection characteristics of the objectsurface.Secondly, surface topography distribution data of semiconductor silicon material wasmeasured by atomic force microscopy (AFM), used in the BRDF simulation which wereapplied in the method of finite difference time domain (FDTD) and geometrical optics. As aresult, in the two dimensional rough surface, when there are anisotropy or smooth surfaces,the results obtained by the method of FDTD can be better fitting the experimental data.Above all, the BRDF of several typical materials, such as corundum, zirconium oxideand cordierite et al, were measured by the NIMS-III type bidirectional reflectancedistribution function measurement. As a result, the mirror peak of these materials BRDFhas increases with the angle of incidence. As the incident wavelength range of380-2000nm,BRDF distribution of each material except the silicon carbide increases with the incidentwavelength. Because it is need to spend a lot of time that measured the BRDF of thematerial in all angles of hemisphere space. So in this paper we use genetic algorithms to getthe model parameters by optimize the simulation data with the experimental data, theresults show that the improved model which was proposed in this paper can be good fit with the experimental data.