Figure 1

GM spectra under pressure from the two set of experiments using 1.96 ($G^{1.96}$ in black) and 2.41 eV ($G^{2.41}$ in gray) excitation energies. The positions of the most intense bands in each spectrum are marked at different pressures. A blueshift is observed with increasing pressure for all the bands of the two spectra ($G^{1.96}$ and $G^{2.41}$). At 1.9 GPa the intensity of the $G^{1.96}$ spectrum is collapsed. At 3.5 GPa, the intensity of the $G^{1.96}$ spectrum is recovered, however a loss of the lower-frequency band is recorded (at the point marked “X”). At higher pressures, the bands in both spectra are broadened.Reuse & Permissions

Figure 2

Pressure evolution of the frequency of the most intense peak $G1$ from both sets of experiments using 1.96 (filled triangles) and 2.41 eV (open circles). The inset shows the transition region with the linear fits to the triangles (black lines) and the circles (gray lines) before and after transition. The intersection of each corresponding linear fits (black or gray) results in different transition pressures at $1.0\pm 0.1\text{GPa}$ and $1.7\pm 0.1\text{GPa}$ for $G^{1.96}$ and $G^{2.41}$, respectively.Reuse & Permissions

Figure 3

Left panel: GM spectra (gray) recorded at ambient pressure under each excitation. (a) The $G^{2.41}$ spectrum is fitted using four Lorentzian peaks (black dotted). These peaks are associated with semiconducting tubes in resonance. (b) The $G^{1.96}$ spectrum is fitted using four Lorentzian peaks and a BWF line shape. Both the BWF peak $\sim 1540{\text{cm}}^{-1}$ and the Lorentzian “$G+$” $\sim 1580{\text{cm}}^{-1}$ (solid black) are a contribution of metallic tubes in resonance, while the other three Lorentzians (black dotted) are associated with the semiconducting tubes in resonance. Right panel: GM spectra (gray) recorded at 3.5 GPa under each excitation. (a) The $G^{2.41}$ spectrum fitted using three Lorentzians (black dotted: $G1$, $G2$, and $G4$ been shifted under pressure). This spectrum is still considered as a contribution of semiconducting tubes in resonance. (b) The $G^{1.96}$ spectrum fitted using only three Lorentzians (black dotted: $G1$, $G2$, and $G4$ have been shifted under pressure). As these Lorentzians are associated with semiconducting tubes in resonance, the absence of the BWF peak together with the $G+$ Lorentzian is associated to a loss of resonance from the metallic tubes.Reuse & Permissions

Figure 4

Optical interband transition energies $\left(E_{ii}\right)$ versus tube diameters $\left(d_t\right)$ for all the tubes in the diameter range of the used sample (a) at ambient pressure and (b) at 3.5 GPa. The two solid lines (black and gray) represent the laser excitation energies (1.96 and 2.41 eV, respectively), while the two dotted lines (black or gray), represent the resonance window within each of the excitation energies (1.96 or 2.41 eV, respectively). Filled triangles refer to “$E_{11}$” from metallic tubes, while open circles to “$E_{33}$” and “$E_{44}$” from semiconducting tubes. All data have been plotted at ambient pressure using Refs. 27, 28, 29, while we suggest a redshift of $\sim 35\text{meV}/\text{GPa}$ for plot “$b$.”Based on the proposed shift, all “$E_{11}$” from metallic tubes will be out of the resonance window, hence no resonance from metallic tubes is expected.Reuse & Permissions

Figure 5

RM collected using the 2.41 eV excitation at ambient pressure. Two clear peaks are presented at 150 and $175{\text{cm}}^{-1}$. According to Araujo et al. (Ref. 27), these peaks correspond to resonance from tubes centered $\sim 1.6$ and 1.4 nm, respectively, with the high contribution from those of $\sim 1.4\text{nm}$.Reuse & Permissions