Assessing the reliability of the Raman peak counting method for the characterization of SWCNT diameter distributions: a cross-characterization with TEM

Resonant Raman spectroscopy is a widely used technique for single-walled carbon nanotube (SWCNT) characterization, in particular in the radial breathing mode (RBM) range which provides direct information on the structure of the nanotube in resonance. The RBM peak counting method, i.e. acquiring Raman spectrum grids on a substrate with a select set of discrete laser lines and counting RBM peaks as single nanotubes, is frequently used to characterize SWCNT growth samples, despite the many factors that can induce errors in the results. In this work, we cross-characterize the diameter distributions obtained through this methodology with diameter distributions obtained by counting SWCNT diameters in transmission electron microscopy (TEM) and discuss the different results and biases between the techniques. This study is performed on a broad diameter distribution sample, and on two chirality-enriched samples whose chirality distributions are determined by photoluminescence excitation spectroscopy (PLE) and statistical analysis of high resolution TEM (HRTEM) images. We show that the largest differences between the Raman peak counting and TEM diameter distributions stem from the chirality-dependence of SWCNT Raman cross-sections and the patchy vision offered by the use of only a few discrete excitation wavelengths. The effect of the substrate and TEM-related biases are also discussed.