Controlling the Crystalline Quality and the Purity of Single-walled Carbon Nanotubes Grown by Catalytic Chemical Vapor Deposition

It is frequently observed that as-grown single-walled carbon nanotubes (SWCNTs) contain defects. Controlling the defect density is a key issue for the control of nanotube properties. However, little is known about the influence of the growth conditions on the formation of nanotube defects. In addition, SWCNT samples frequently contain carbonaceous by-products which affect their ensemble properties. Raman spectroscopy is commonly used to characterize both features from the measurement of the defect-induced D band. However, the contribution of each carbonaceous species to the D band is usually not known making it difficult to separately extract the defect density and relative abundance of each. Here, we report on the correlated evolution of the D and G’ bands of SWCNT samples with increasing growth temperature. In the general case, three to four Lorentzian components are required to fit them. Coupled with HRTEM characterization, the low frequency components of the D and G’ can be attributed to the contribution of SWCNTs while high frequency components are associated with defective carbonaceous by-products. The nature of these defective by-products varies with the type of catalysts and with the growth conditions. INTRODUCTION Raman spectroscopy is commonly used to assess defects in graphene-related materials [1,2] through the measurement of the intensity ratio of the defect-induced D band to the graphenic G band [3,4]. The Raman D band is activated in the presence of defects through a double resonance process involving the elastic scattering of electrons by defects [5,6]. When the distance between defects (LD) is larger than a few nm, the IG/ID ratio is inversely proportional to the density of defects [7]. The D-band profile is also dependent on the type and structure of sp carbon material. For instance, Souza Filho et al. [8] showed that for SWCNTs, the frequency of the D band depends on the nanotube diameter following the general trend ωD = 1354.8-16.5/d for Elaser = 2.41 eV. Flat single-layer graphene with a small amount of defects displays a quite narrow D-band with a linewidth around 20 cm which broadens for LD<5 nm [7]. Graphitic sp carbon materials (graphite polyhedral crystal, graphite whiskers) can also display D-band linewidth around 20 cm [9] while less ordered materials (CVD multi-walled nanotubes [10], turbostratic carbon [9], short stacked graphene patches [11], polycrystalline graphite [12] amorphous carbon [13]...) display a broad D band with linewidths ranging from 45 cm to 100 cm. The D band of a SWCNT sample is generally considered as the sum of two contributions: a first one related to SWCNT defects and a second one related to defective carbonaceous impurities. However, the nature and abundance of the defective by-products (pyrolitic amorphous carbon, catalytically-grown defective carbon structures, carbon filaments, carbon shells, CxHy polymers...) is usually poorly known making it difficult to determine the defect density of SWCNTs as a function of the growth conditions. Here, we report on our investigations for identifying the relative contributions of the different carbon species in the D band and its second-order mode, the G’ (or 2D) band for CNT samples grown in different CCVD conditions.