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

Ultrasonication-induced extraction of inner shells from double-wall carbon nanotubes characterized via in situ spectroscopy after density gradient ultracentrifugation

“…From the HRTEM images we determine the diameter distributions and abundances of the various CNT species present in the ultrasonicated DWCNT sample and compare them with those present in the purified DWCNT sample. Contrarily to HRTEM, where despite the improved TEM grids the bias is typically towards larger diameter CNTs [51], in our optical spectroscopy measurements only CNTs with a diameter up to 1.8 nm are detected. Hence, to compare the statistical results of HRTEM, shown in Figure 9, with the earlier spectroscopic conclusions, the statistical data is deliberately limited to those CNTs that have an inner shell-diameter below 1.8 nm.…”

“…Although HRTEM is a powerful technique to estimate statistically the diameter distribution of a CNT sample, it can be subject to many biases and thus a well-chosen methodology that avoids these biases is required [51]. Generally, porous supporting grids are employed in TEM to deposit the CNTs on.…”

“…Besides an increase in outer shell SWCNTs, there are significantly more small-diameter SWCNTs observed after ultrasonication, especially within the diameter range below the cut-off limit of 1.8 nm where diameters as thin as 0.6 nm are detected. When taking into account that such small diameters are easily overlooked in HRTEM or even destroyed by the electron beam during image acquisition [51], this is even more impressive. Compared to normal SWCNT samples, such thin shells are more stable in DWCNTs they are shielded from the harsh outer environment during growth by the outer tube [4,5].…”

The role of the bile salt surfactant sodium deoxycholate in aqueous two-phase separation of single-wall carbon nanotubes revealed by systematic parameter variations

“…The 2D RRS and PLE data are fitted simultaneously for all samples within one systematic separation series by using previously developed models for single-wavelength Raman spectra [44][45][46] and wavelength-dependent PLE spectra [31,[47][48][49][50] of SWCNTs.…”

“…The excitation profile is fitted using both excitonic and band-to-band transitions, as well as a phonon side band in excitation. The number of fit parameters is reduced, by making assumptions on the position and relative amplitudes of these phonon sidebands, for which more details can be found in references [31,[47][48][49][50]. Importantly, all PLE maps are then fitted simultaneously, with shared peak positions and linewidths that are optimized numerically, but similarly as in the Raman maps, with varying amplitudes that are obtained through linear regression and thus not optimized numerically.…”