Plant biomass is a potential source of large-scale biofuel production as long as the cell wall polysaccharides can be efficiently broken down into sugars that can be fermented into ethanol. Little is known about the precise 3D cell wall molecular organization, although such knowledge would greatly facilitate improved deconstruction schemes as well as rational cell wall engineering. The primary objective of our research is to obtain a realistic model of plant cell walls at molecular resolution, by determining 3D architecture of the cell wall via electron tomography and correlated chemical composition via Raman imaging. 3D architectural and Raman data will be analyzed with the help of sophisticated algorithms, and integrated into a molecular model of the cell wall. As the first step, we have tested different sample preparation methods to obtain the best possible preservation of the cell wall structure and of the Raman signal, including 1) high-pressure freezing and freeze-substitution followed by resin embedding; 2) microwave-assisted TEM processing; and 3) cryo-sectioning of high-pressure frozen samples. We explored differential staining and labeling approaches to increase overall contrast and/or to identify specific cell wall components. Using widefield correlative TEM and optical imaging at high resolution, we have monitored differences and similarities arising from various parts/tissues of the same plant, different ages of the same plant and also different plant species. We discuss our preliminary findings on the different sample preparation methods, 2D widefield TEM and Raman imaging as well as cell wall 3D tomograms.