Synthetic Metals 103 (1999) 2.506-2507 Infrared Reflectance of Single-Walled Carbon Nanotubes U. Kuhlmann’, H. Jantoljak’, N. PF&nder 2, C. Journet3, P. Bernier3, C. Thomsen’ ’ Institut fiir Festkbkperphysik, Technische Universitiit Berlin, Be&n (Germany) 2 FHI der Max-Plan&=Gesellschaff, Berlin (Germany) 3 GDPC, University of Montpellier II, Montpelker (fiance) Abstract A mid-infrared reflection study on single-walled carbon nanotubes revealed weak absorption structures which we assign to the vibrational modes of the tubes. The results are in good agreement with theoretical predictions for nanotubes. In comparison to graphite phonons a small shift to higher frequencies was found. First measurements at different temperatures indicate no temperature dependence of the phonon frequencies. Keywords: Infrared and Raman spectroscopy; fiilerenes and derivatives 1 Introduction The vibrational spectra of single-walled carbon nanotubes (SWNT) con&t of 15-16 Raman and 7-9 infrared active vi- brational modes, their number depending on symmetry. [l] Whereas Raman spectroscopy on carbon nanotubes is a well established tool for investigation of the vibrational modes of SWNT 12, 31, there is no publication on the infrared active vibrations with results of comparable quality. The main ob- stacle is the poor sample quality of raw material produced by either laser-ablation or arc discharge techniques. Only small amounts of purified sample material were obtained by purification processes up to now. As an additional diffi- culty the modes, which are almost independent, of the tube diameter and therefore are expected to contribute construc- tively to detectable absorptions, lie close to the frequencies of graphite at 868 cm-’ and 1590 cm”. It may be difficult to distinguish them from SWNT, if the presence of graphitic particles cannot be excluded. 2 Sample Preparation and Experiment The samples were prepared by the arc discharge technique [4] and consist of soft soot bundles. The material is inhomo- geneousiy composed of SWNT (> 50 %), Ni nanoparticles (a catalyst for SWNT production) and amorphous carbon, as verified by transmission electron microscopy (TEM). A few graphitic particles were detected, too. The tube di- ameters were estimated from the TEM images to be 1.4 (zkO.1) nm. The as grown material was not suited for the reflection measurements, therefore it was pressed between polished copper plates until it turned into thin and almost flat sheets. It was verified by Raman spectroscopy that the sample composition did not change during this procedure. The reflection spectra were obtained employing a BRUKER IFS 66~ FT-IR spectrometer in combination with a liquid N, cooled MCT-detector. The reflectance was referred to a gold mirror. In addition, the temperature dependence of the spectra was measured by cooling the~mirror and the sample with a He-flow cryostat. Polycrystalline graphite and_higly oriented pyrolytic graphite (HOPG) samples were inyesti- gated for comparison. 3 Results and Discussion Figure 1 shows the reflectivity of theSWNT sample in com- parison to polycrystalline graphite. The weak structure in the SWNT spectrum at 1590 cm-l resembles a strongly - ’ . 0, 1~~~~~ 800 1000 12c4l 1400 1600 Frequency [cm-l] Fig. 1: Infrared reflectance spectra of polycrystalline graphite (upper curve) and SWNT. Graphite has two in- frared active phonons, the A,, mode at 868 cm“ and the E,, mode at 1590 cm-‘. The fist derivatives of the spectra around these positions are shown in the insets. broadened graphite mode. To enhance this structure and to determine its position, we took the first derivative-of the spectra, which is shown in the inset to Fig.1. Applying 0379-6779/99/$ - see front matter 0 1999 Elsevier Science S.A. All rights reserved. PII: SO379-6779(98)01077-7