Effects of aging processes at the surface of the superconductor $\beta \text{−}\mathrm{FeSe}$

This work reports a complete analysis of surface degradation in $\beta \text{−}\mathrm{FeSe}$ single crystals; which includes the dependence on aging and temperature of Raman spectra, electrical transport resistivity, and XPS depth profiling characterization. All measurements are consistent with the emergence of clusters of Se and the presence of ${\mathrm{FeO}}_x$ in the surface of the crystals. The vibrational modes of the Se clusters are coupled to the $\beta \text{−}\mathrm{FeSe}$ ones as seen in the structural transition. Moreover, this degraded surface can behave like a quasi-two-dimensional layer, as supported by a theoretical model and also consistent with Raman spectra. While the degradation of thin films of $\beta \text{−}\mathrm{FeSe}$ in air is widely known, with a corresponding loss of the superconducting properties, in single crystals with aged surfaces we observe an enhancement of the superconducting properties.

Figure 1

(a) X-ray $\theta \text{−}2\theta$ diffractogram showing only $00l$ reflexions. Inset: The rocking curve around the [001] peak is shown. (b) Magnetization ZFC-FC (cooling in zero field and in an applied field of 10 Oe). (c) Electrical resistivity vs temperature curves with 16 T and with no applied field. The arrow points out the nematic transition temperature. In the inset a unit cell of $\beta \text{−}\mathrm{FeSe}$ is shown. Green corresponds to Se and brown to Fe.

Figure 2

(a) Photo of a sample immediately after taken out of the furnace. (b) Photo of an aged sample where part of the surface was cleaved (homogeneous pale pink part) and some part was not (yellow-green color part). (c) AFM deflection image of the cleaved zone some days after being made, where multiple clusters with the same shape can be seen. (d) Zoom on one cluster from image (c).

Figure 3

Raman measurements of the cleaved and noncleaved samples (bottom and top panels) at 300 (red line) and 5 K (black line). Both measurements were made in parallel polarization configuration. The arrows indicate the bulk Se and Fe vibrational modes.

Figure 4

Raman measurements of the cleaved part of the sample as a function of temperature. Upper section shows measurements below room temperature and lower section measurements above room temperature. The spectra spacing is proportional to the temperature difference between them. The red dotted line indicates the peak positions as a guide to the eye. See more detailed analysis in Fig. 7.

Figure 5

Upper section: Raman measurements for the noncleaved part of the sample as a function of temperature below room temperature. The arrow indicates the beginning of the broad phonon tail below 150 ${\mathrm{cm}}^{-1}$. In the lower section spectra polarization dependence is shown. The legend indicates the laser and analyzer polarization angle; on the top of each peak we display its normal mode representation.

Figure 6

Evolution of noncleaved sample spectrum from 293 to 473 K. Inset: Measurement at room temperature in two different positions of the sample after heating. The black line shows the spectra taken from the data of Ref. [26] for ${\mathrm{FeSe}}_2$ and the dotted red line corresponds to an iron oxide (magnetite) Raman spectra [27].

Figure 7

Raman cross section versus temperature below room temperature for the cleaved and aged parts of the sample of modes (a) $A_{1g}$, (b) $B_{1g}$, and (c) 233 and $250{\mathrm{cm}}^{-1}$. (d) Evolution of the Raman cross section of $A_{1g}$ and $B_{1g}$ modes for freshly cleaved samples is shown below and above room temperature. The data below room temperature were vertically translated to match the one above room temperature. Pink circles above 473 K (indicated by a dashed vertical purple line) correspond to the fit of the peak that results from the convolution of both modes. In (e) the Raman cross section temperature dependence of 233 and $250{\mathrm{cm}}^{-1}$ modes are shown. The left axis corresponds to the results below room temperature and the right axis to the ones above room temperature, indicated by arrows. Different samples were used in the low and high temperature runs. (f) The peaks energies as a function of temperature. The dashed lines on top of each data set corresponds to a 5 point average smooth. The shadow light gray in each graph indicates the structural transition temperature ranging from 80 to 90 K.

Figure 8

Upper panel: Spectra taken at ${\mathrm{Fe}}_{2p}, {\mathrm{Se}}_{3p}$, and ${\mathrm{Se}}_{3d}$ energies, in an old sample surface A, which displays the mentioned Raman spectra (see Fig. 3 top panel) as a function of sputtering. Lower panel: Spectra taken from an old sample surface B, which does not display the new Raman feature as a function of sputtering. Brown curves in both panels correspond to spectra taken in a cleaved sample C. Vertical lines indicate the higher spin state binding energies for the different chemical species: ${\mathrm{FeO}}_x, {\mathrm{Fe}}^0, {\mathrm{Se}}^0, {\mathrm{Se}}^{2-}, {\mathrm{SeO}}_2$, and $\beta \text{−}\mathrm{FeSe}$. The boxes on the right indicate the composition on each stage of the sputtering depth profiling.

Figure 9

Evolution of the superconducting onset temperature and critical temperature with thermal cycling in a sample with an aged surface. The dark gray line shows the resistivity versus temperature dependence of an as-grown sample. In the inset a scheme (not to scale) of the prepared sample is shown. Upper panel shows the transition region measured after one year in the same sample.

Figure 10

Upper section: Fit of the experimental data with the theoretical calculations. $\lambda$ is the eigenvalue from the theoretical calculation. Middle section: Scheme of one sheet of $\beta \text{−}\mathrm{FeSe}$. The pink slashed line determines one arbitrary unit cell of the system and the black square shows the notation used for the Hamiltonian. Bottom section: Scheme of the vibrational modes along the $c$ axis for a $\beta \text{−}\mathrm{FeSe}$ single sheet.

Figure 11

Black curve: Typical $\beta \text{−}\mathrm{FeSe}$ spectrum taken at 6.5 K. Red curve: Spectrum from the same sample 2 months later at 140 K. We see how the layer-breathing mode appears along with the new peak at $250{\mathrm{cm}}^{-1}$. Black curve was vertically translated for clarity.

Figure 12

Red curve: Spectrum collected with laser and analyzer polarization parallel to each other at 300 K. Black curve: Same spectrum after depolarization ratio analysis. We observe how the peak at 233 ${\mathrm{cm}}^{-1}$ completely disappears.

Figure 13

Raman spectra of FeTe single crystal before and after cleaving at room temperature. The peak values in the main frame are obtained fitting Lorentzian shaped curves.

Figure 14

Experimental spectra of the 250 and 233 ${\mathrm{cm}}^{-1}$ peaks and their fits taken at (a) 300 K, (b) 200 K, (c) 115 K, (d) 100 K, (e) 95 K, (f) 77.5 K, (g) 51 K, (h) 19.9 K, and (i) 5.2 K. The light gray line shows $A_{1g}$ and $B_{1g}$ peaks which were subtracted to perform the fit.

Figure 15

Measured $A_{1g}$ and $B_{1g}$ peaks and their fits taken at (a) 303 K, (b) 373 K, (c) 413 K, (d) 473 K, (e) 488 K, (f) 503 K, (g) 573 K, (h) 663 K, and (i) 763 K.