The present paper proposes to investigate the links between the microstructure of polyurethane foams and their sound absorbing efficiency, and more specifically the effect of membranes closing the cells. This study is based on the complete characterization of 15 polyurethane foam with various cell sizes and reticulation rates (i.e., open pore content): (i) characterization of the microstructure properties (cell size C s , strut thickness t, reticulation rate R w …) from SEM pictures, (ii) characterization of nonacoustic parameters (porosity U, airflow resistivity r, tortuosity a 1 …) from direct and indirect methods. Existing analytical links between microstructure properties and nonacoustic parameters are first applied to fully reticulated materials. Then, they are improved empirically to account for the presence of the closed pore content. The proposed expressions associated to the Johnson-Champoux-Allard porous model allow a good estimation of the sound absorbing behavior of all polyurethane foams, fully reticulated or not. This paper also demonstrates the important effect of the presence of cell membranes: increase of the airflow resistivity, tortuosity, and the ratio between the thermal and viscous characteristic lengths while decreasing these two characteristic lengths. Thus, the sound absorption efficiency at low frequencies is improved but can be worsened in some higher frequency bands.

This paper proposes simple semi-phenomenological models to predict the sound absorption efficiency of highly porous polyurethane foams from microstructure characterization. In a previous paper [J. Appl. Phys. 110, 064901 (2011)], the authors presented a 3-parameter semi-phenomenological model linking the microstructure properties of fully and partially reticulated isotropic polyurethane foams (i.e., strut length l, strut thickness t, and reticulation rate R w ) to the macroscopic non-acoustic parameters involved in the classical Johnson-Champoux-Allard model (i.e., porosity /, airflow resistivity r, tortuosity a / , viscous K, and thermal K 0 characteristic lengths). The model was based on existing scaling laws, validated for fully reticulated polyurethane foams, and improved using both geometrical and empirical approaches to account for the presence of membrane closing the pores. This 3-parameter model is applied to six polyurethane foams in this paper and is found highly sensitive to the microstructure characterization; particularly to strut's dimensions. A simplified micro-/macro model is then presented. It is based on the cell size C s and reticulation rate R w only, assuming that the geometric ratio between strut length l and strut thickness t is known. This simplified model, called the 2-parameter model, considerably simplifies the microstructure characterization procedure. A comparison of the two proposed semi-phenomenological models is presented using six polyurethane foams being either fully or partially reticulated, isotropic or anisotropic. It is shown that the 2-parameter model is less sensitive to measurement uncertainties compared to the original model and allows a better estimation of polyurethane foams sound absorption behavior. V C 2013 American Institute of Physics. [http://dx