Raman Characterization of C4KHg

The motivation for the use of Raman scattering comes from previous Raman studies of the different phases of C₄KHg by Timp and coworkers.[247] An extensive transmission electron microscopy (TEM) and Raman study by Timp et al. found zone-folded Raman modes associated with in-plane ordering. The identification of Raman peaks with zone-folded modes can give a clue to the in-plane ordering of GIC's, as explained below.

Zone-folded modes are zone-center (q = 0) phonons found in superlatices. These modes originate from the folding of the graphite bands into the zone-center that occurs in superlattice phases that have a larger real-space lattice constant than graphite, and hence a smaller Brillouin zone. Only q=0 modes participate in Raman processes due to the requirement of wave-vector conservation and the very small momentum of a photon compared to that of a typical phonon. In principle,[6] and in practice,[247] the comparison of the positions of zone-folded Raman peaks with theoretical models allows identification of the corresponding in-plane superlattice.

A Raman study of C₄KHg was performed by Timp et al. in Ref. [247]. The essence of this paper is that the pink samples are identified primarily with a (2 × 2)R0° in-plane superlattice, while red samples are identified primarily with a (sqrt3 × sqrt3)R30° ordering.[247] Figure 1.1 shows the atomic positions corresponding to these in-plane arrangements. The gold samples show no evidence of zone-folded peaks at all.[247] Both the pink and red phases were found to contain as a minority constituent a (sqrt3 × 2)R(30°, 0°). The observation of the (2 × 2)R0° and (sqrt3 × 2)R(30°, 0°) phases is in good agreement with neutron diffraction,[123] x-ray diffraction,[70] and TEM[246] experiments. Kamitakahara[123] found using (10l) neutron diffraction spectra that the (2 × 2)R0° and (sqrt3 × 2)R(30°, 0°) phases correspond to the I_c = 10.24 Å and I_c = 10.83 Å repeat distances. The x-ray, neutron, TEM and Raman experiments together unequivocally identify the 10.24 Å alpha-phase with the (2 × 2)R0° in-plane structure, and the 10.83 Å (beta)-phase with the (sqrt3 × 2)R(30°, 0°) in-plane structure.

The (sqrt3 × sqrt3)R30° phase has only been observed in the TEM and Raman[246] experiments, though. TEM shows that the (sqrt3 × sqrt3)R30° phase appears to be associated with an anomalous (9.4 ± 0.01) Å repeat distance which is not seen in x-ray or neutron diffraction experiments, even by Timp. Timp[246] also finds a (sqrt3 × sqrt3)R30° phase in C₈KHg associated with an anomalous (12.8 ± 0.1) Å\ repeat distance. The usual lattice constant for C₈KHg is I_c = 13.6 Å.[124,246] Because TEM and Raman scattering are more surface-sensitive than x-ray and neutron diffraction, the observation of the (sqrt3 × sqrt3)R30° structure only in the TEM and Raman experiments suggests that this phase might be found only near the sample surface. Although the evidence for the existence of the (sqrt3 × sqrt3)R30° phase is compelling, none of this evidence proves that this structure is found in bulk KHg-GIC's.

New Raman spectra were taken by Dr. G. Doll on pink and gold C₄KHg samples at room temperature. The GIC's were HOPG-based and about 3 mm by 3mm by 1 mm in size. An argon ion laser's 4880 Å line was focussed to a spot on the surface. Care was taken to limit the laser power to less than 10 mW to avoid damaging the samples. Spectra taken on T_c = 1.53 K and a T_c = 0.719 K GIC's are shown in Figure . The spectra were taken over a much wider range of frequencies than is indicated in the Figure, but the only peak observed in either sample is the one displayed. This peak corresponds to the E_2g2 mode of graphite.[247] The frequency of this peak in the new work is about 10 cm^-1 higher than in the experiments of Timp and coworkers.[247] Raman spectra were taken on 6 other specimens whose data is not shown, but no zone-folded peaks were observed. X-ray diffraction scans taken after the Raman spectra showed no degradation in staging.

Figure 3.6: Raman spectra on gold and pink C₄KHg. a) Spectrum of a T_c = 0.719 K gold sample with a single I_c value = 10.14 Å. The peak frequency is 1597.1 cm^-1 and the HWHM is 13.2 cm^-1. b) Spectrum of a T_c = 1.53 K pink sample with a single I_c value = 10.22 Å. The peak frequency is 1593.9 cm^-1 and the HWHM is 15.5 cm^-1.

It was hoped that the observation of zone-folded Raman peaks in C₄KHg would allow a systematic study of the relationship between in-plane structure and superconductivity. As was shown in Section 3.4, the c-axis structure seemed to provide no information about the superconductivity since the (00l) scans of low- T_c and high- T_c specimens were identical. Unfortunately the new results on C₄KHg Raman spectra also show the low- T_c and high- T_c samples to be identical, in that no zone-folded peaks were seen in any spectrum. The lack of zone-folded peaks would seem to imply in-plane disorder at least at the surface in these samples. Since T_c measurements were taken on the same samples as the Raman spectra, and since there was no sample transfer in-between the T_c measurement and the Raman experiment, one can conclude that the Raman experiments are much more sensitive to the in-plane order than the superconductivity is. The reason may well be that, as mentioned, Raman is a fairly surface-sensitive probe of ordering, while the T_c measurements are sensitive to superconductivity anywhere in a sample's cross-section (see Figure ). Thus any exposure to air that might have occurred during the sample handling might have affected only the surface, which would show up in the Raman but not the superconductivity.

Hydrogenated C₄KHg specimens also showed only one peak in Raman scattering, a Lorentzian E_2g2 line at about 1598 cm^-1. The spectra of the hydrogenated samples were basically indistinguishable from those of the pink and gold samples. A low-temperature Raman study of these different types of C₄KHg might provide more information.

Like the (00l) x-ray data, the Raman data show almost no difference between the lower- T_c and higher- T_c C₄KHg specimens. Therefore they can provide little insight into the relationship between the in-plane structure and superconductivity. In the hopes of answering these questions it was decided to try neutron diffraction. Since neutrons are quite sensitive to hydrogen, neutron diffraction data taken before and after hydrogenation might offer important clues to the puzzle.