Tip-induced mobilization upon cooling of Ni monolayers on Re(0001)

Usually, cooling a metal sample down to cryogenic temperatures leads to immobilization of the surface atoms. In this study, we demonstrate a movement of Ni adatoms at $4.6\mathrm{K}$ on Ni films grown on a Re(0001) single crystal, while the surface is rigid at room temperature. The mobility is observed from 2 to 20 atomic-layer-thick films. Measurements at intermediate temperatures reveal an increasing mobility with decreasing temperature. The observed velocity of advancing steps is consistent with a model considering a sudden release of material, eventually triggered by the tip, followed by free diffusion. According to the model, an increasing length scale for the mobilization event with decreasing temperature, tentatively explained by epitaxial strain, is responsible for the inverse mobilization behavior.

Figure 1

(a) Constant-current image of cleaned Re(0001) ($V=100\mathrm{mV}, I=2\mathrm{nA}$), the blue arrows mark point defects appearing as dark spots due to a lower LDOS. The inlay shows an enlarged defect on the Re surface ($V=-250\mathrm{mV}, I=3\mathrm{nA}$). The elliptical shape with a size of $2\times 1\AA$ is likely caused by adsorbed residual gas molecules. (b) 4.2 monolayers of Ni deposited on Re(0001) before annealing ($V=1000\mathrm{mV}, I=0.1\mathrm{nA}$) at $4.6\mathrm{K}$. The island height varies over 7 monolayer steps. (c) LEED pattern of a $0.7\text{−}\mathrm{ML}$ Re/Ni surface before annealing, measured at $70\mathrm{eV}$. The pattern shows the six main reflexes and superstructure spots. The weak reflexes marked by yellow arrows, indicate an additional $2\times 2$ superstructure. (d) The sketch illustrates the crystallographic relation of Ni(111) on Re(0001). Ni atoms (blue) grow in the fcc lattice structure on hcp Re(0001) (red).

Figure 2

Topographical STM images showing four series of Ni/Re surfaces. All measurements are recorded with the same parameters of $I=2\mathrm{nA}$ and $V=0.1\mathrm{V}$ with a size of $(200\times 200){\mathrm{nm}}^2$ and $5\mathrm{ML}$ Ni for (a)–(l) and $3.5\mathrm{ML}$ for (m)–(x). (a)–(f) Show a series at $T=4.6\mathrm{K}$, (g)–(l) at $T=12.9\mathrm{K}$, (m)–(r) at $T=78.3\mathrm{K}$, and (s)–(x) at room temperature. While the first two rows illustrate a movement of the Ni surface with speed decreasing over time, the third row reveals a stable topography as well as the fourth row [(s)–(x)]. The thermal drift causes the image area to move, but the relative position of step edges and islands remains the same.

Figure 3

This figure displays two series of measurements that illustrate some particular behaviors of the mobilized Ni surface. (a)–(f) ($I=2\mathrm{nA}, T=4.6\mathrm{K}, 8\mathrm{ML}$ Ni) show a measurement where the first two images are taken in $9\min$ per image, which is a quarter of the time of a normal measurement. Ni atoms move at high speed. In addition, the influence of fault lines can be seen very clearly here. (g)–(l) ($I=1\mathrm{nA}, T=4.6\mathrm{K}, 4\mathrm{ML}$) show a rather fluctuating movement, in which the Ni atoms move in different directions and close a hole. The moiré pattern is depicted by larger contrast in the Supplemental Material [40], Fig. 13 showing a magnified version of image (l).

Figure 4

(a)–(f) STM images measured at $I=0.1\mathrm{nA}, T=4.6\mathrm{K}$ with different voltages indicated in each image. The thickness is $4\mathrm{ML}$ Ni. The series shows an example of increased mobility over time. (g)–(l) Difference images, where two consecutively scanned images are subtracted. (g) Results from the difference of an image measured before (a) and (a) itself, and (h) results from (b) minus (a). The brighter color indicates increased terrace areas, while the darker color represents decreasing terrace areas. In Table 1, the fractions of the moved Ni areas are listed.

Figure 5

The relative step-edge positions for different image series are plotted as a function of time. The black data points result from the movement of the center step edge, marked with a green arrow in Fig. 2, in Figs. 2–2 where five different positions of the step edge were averaged. The red points describe the movement of the island in Figs. 2–2, marked with three black arrows in Fig. 2, and the blue data points originate from the movement of the three different step edges, marked with three blue arrows in Fig. 3 of the series shown in Figs. 3–3. The data ares fitted with a square-root proportionality $\left[a{t}^{1/2}\right]$, where $a$ is the only free parameter. The error bars indicate the fluctuation of the movement along the step edge.

Figure 6

(a) The difference image of Figs. 3 and 3; areas of the same level are in yellow. The zones with an increase of one layer (51.15%) are in red, with two layers (3.87%) in purple, with three layers (1.35%) in blue, and the turquoise spot at the bottom of the picture is with four layers (0.1%) of increase. The three green spots on the upper side of the picture (0.28%) are the only areas where the Ni disappeared. (b) A difference picture of forward and backward measurement of Fig. 3. Apart from the brighter step edges, everything is at the same height.

Figure 7

The atomistic dynamics is determined by a set of diffusion coefficients: $D_0$ for the diffusion of adatoms on the terrace, $D_E$ for the diffusion of edge atoms along the step edge, and $D_K\approx D_E^2/D_0$ for the dynamics of kinks. In the kinetic model, the speed of the step edge is $v=awk$ with kink density $k\approx 2\sqrt{D_K/D_E}{a}^{-1}$ and kink velocity $w$ [50].