Doping of few-layered graphene and carbon nanotubes using ion implantation

Doping of nanostructured materials using a clean, efficient, and site-selective route such as ion implantation would be hugely desirable for realization of large-scale production methods. Here, ion implantation is used to create uniform impurity-atom densities which are both dose and spatially controlled within multiwalled carbon nanotubes and graphene. The technique is demonstrated for a range of dopants, including silver, representing a likely candidate for optical enhancement, and boron, which is predicted to introduce a plasmon within the visible-frequency regime. Electron energy-loss spectroscopy performed within an aberration-corrected scanning transmission electron microscope, in combination with high-angle-annular-dark-field imaging, is used to pinpoint and identify the bonding configuration of single foreign species within the matrix.

Figure 1

[(a)–(c)] HREM images obtained in a Tecnai F30 for direct comparison of pristine few- and multiwalled carbon nanotubes (a) and B- and Ag-doped tubes, (b) and (c), respectively, the latter both ion implanted to a dose of $2\times {10}^{15}{\text{cm}}^{-2}$ at 200 eV. The structural integrity does not appear higher in the pristine tubes. [(d) and (e)] HREM images obtained in a Titan 80–300 of the same Ag-implanted tubes as in (c) at superior resolution to show details of the Ag clusters and also the effect of e-beam irradiation at prolonged time on small tubes. (f) HAADF image obtained in a Titan of a three-walled tube, showing Ag-atom aggregates from nanometer-sized clusters clinging to the tube exterior, to subnanometer clusters which have partially formed inside the tube, to intercalated single atoms (dashed circles).

Figure 2

(Color online) (a) Schematics of CNT ion implantation and (b) of STEM and EEL spectrum imaging. [(c) and (d)] SuperSTEM high-resolution BF and HAADF image of a multiple-walled CNT implanted with Co at 100 eV and ${10}^{14}\text{ions}{\text{cm}}^{-2}$. The region where an SI was taken is framed in white. The small solid frame (pale blue) contains a Co atom on the surface. (e) HAADF image after the SI acquisition. The Co atom is still within the same frame. (f) Top to bottom: raw EEL spectrum from the small, solid-line frame in (d) and (e) with overlaid background fit showing the $\text{Co-}L$ edge signal (i), raw spectrum from the dotted small frame without Co atom (pink) in (d) with overlaid background fit (ii), background-subtracted spectrum from the solid-line small frame with overlaid $\text{Co}L$-edge standard spectrum (iii), background-subtracted spectrum from the dotted frame in (d) (iv), spectrum from the dashed small frame (green) in (d) showing vacuum counts (v).

Figure 3

(Color online) [(a) and (b)] SuperSTEM high-resolution BF and HAADF image of a multiple-walled CNT implanted with Ag at 200 eV and $2\times {10}^{15}\text{ions}{\text{cm}}^{-2}$. Like in Fig. 2e Ag-atom aggregates in the form of nanocrystals as well as single Ag atoms [some are encircled in (b)] are observed. (c) Selection of images from a through-focus HAADF image series, showing a part of the CNT wall projection in the frame in (a) and (b). Numbers above the images denote the defocus (in nm) of the e beam with respect to the CNT-entrance surface. The bright feature encircled with a solid line represents a planar Ag-atom aggregate confined to the space between two walls. The feature encircled with a dotted line represents a couple of intercalated Ag atoms. The feature in the dashed circle is a 3D aggregate consisting of several planar intercalation aggregates interacting across the CNT walls.

Figure 4

(Color online) (a) SuperSTEM high-resolution BF image of a multiple-walled CNT ion implanted with nitrogen at 100 eV and ${10}^{14}\text{ions}{\text{cm}}^{-2}$. The distance between the graphitic planes is $3.35\text{Å}$. Superimposed on the image is the EEL spectrum acquisition grid. Light gray (green) pixels indicate a captured nitrogen signal. The variation in tint of the pixels indicates the integrated intensity in the energy window 400–415 eV of the nitrogen signal of raw, background-subtracted spectra [like in (c) below]: a stronger intensity has a stronger (brighter) tint. (b) Approximate geometry and dimensions of the CNT in (a). (c) Raw EEL intensity in the energy region around the $\text{N}K$ edge after power-law background subtraction, extracted from the bottom row of pixels in (a), indicated by the black arrow. (d) Same spectra after having undergone principal component analysis and normalization to the same noise bandwidth. (e) Top and middle spectra: DFT (WIEN2K) calculations of the $\text{N}K$ EEL edge of substitutional N with and without adjacent vacancy in graphene, the structure model is shown in the middle of the figure with carbon atoms in gray (blue) and nitrogen atoms in black, bottom spectrum is of the $\text{N}K$ edge in hexagonal BN (gray line; green) and nitrogen gas (black line). (f) SuperSTEM high-resolution BF image of a multiple-walled CNT implanted with boron at 100 eV and ${10}^{14}\text{ions}{\text{cm}}^{-2}$. Superimposed on the image is the EEL spectrum acquisition grid. Dark gray (pink) pixels indicate captured boron signals. The variation in tint of the pixels indicates the integrated intensity of the boron signal in the energy window 190–193 eV: a stronger gray (pink) intensity has a stronger tint. (g) Top to bottom: DFT (WIEN2K) calculations of the $\text{B}K$ EEL edge of substitutional B with and without adjacent vacancy in graphene, $\text{B}K$ EEL edge of a hexagonal BN standard (Ref. 32) (gray; green) and of nonspatially resolved measurements of a highly boron doped (25%) multiple-walled CNT (black). (h) PCA treated, noise-normalized spectra extracted from the row of pixels in (f), indicated by the black arrow. (i) Signal in the energy region around the $\text{N}K$ edge in the same spectra as in (h). (j) SuperSTEM high-resolution BF image of five-layered graphene with locations of ${\pi }^{\ast }$ signals in the spectra, as in (h), at 190–193 eV, marked as white squares (yellow). (k) SuperSTEM high-resolution BF image of staggered graphene sheets; leftmost is a single layer, followed (toward the right) by a second sheet with a rough edge, which, in turn, is followed by four further graphene sheets with partially straight edges. B-signal positions are marked as white circles (yellow). (l) Top to bottom: $\text{B}K$ EEL edge of amorphous B, of ${\text{B}}_4\text{C}$ and of two typical, overlaid spectra from white (yellow) positions in (k). These signals are significantly above noise level, thus have not been PCA treated. The ${\pi }^{\ast }$ peak is missing and comparison with the reference spectra suggests that small few-atom B clusters have formed attached to sheet edges or to monoatomic steps.

Figure 5

(Color online) (a) EEL intensity distribution extracted from SIs at 2.6–3.4 eV in a pristine, (b) a Ag-implanted and (c) a B-implanted nanotube (both implanted at 200 eV and $2\times {10}^{15}\text{ions}{\text{cm}}^{-2}$) with intensities displayed on the same scale. The implantation has created a plasmon at uv/visible energies, as predicted by the DFT EEL-spectra calculations, using WIEN2K, of (d) substitutional B and (e) substitutional and intercalated Ag in layered carbons. Experimental spectra are overlaid. (f) Plasmon-intensity distribution overlaid on a BF-NorthWestSTEM image of a B-implanted CNT network [intensity displayed on a reduced scale compared to (c)], showing “hot spots,” where tubes are in close proximity.