Elsevier

Applied Surface Science

Volume 424, Part 3, 1 December 2017, Pages 269-274
Applied Surface Science

Band bending at magnetic Ni/Ge(001) interface investigated by X-ray photoelectron spectroscopy

https://doi.org/10.1016/j.apsusc.2017.04.168Get rights and content

Abstract

We report the molecular beam epitaxy growth of Ni on a clean Ge(001) surface with an intermediate NiGe layer forming at the interface at room temperature. The crystallinity of the substrate is lost after the deposition of more than 2 Ni monolayers. The Schottky barrier formation is investigated by X-ray photoelectron spectroscopy. The method allows us to infer a 0.39–0.45 eV band bending at the interface between the compound and Ge(001). Magneto-optical Kerr effect measurements were conclusive in detecting the ferromagnetic ordering of Ni outermost layers.

Introduction

Nowadays the tendency in nanotechnology is to create smaller devices with improved characteristics, which has increased the necessity of developing new devices with higher charge mobility and lower Schottky barriers. In the case of a ferromagnetic metal, these characteristics are useful in developing spintronic devices, as well. In this paper we investigate the band bending in germanium with metallic nickel grown by molecular beam epitaxy (MBE) using X-ray photoelectron spectroscopy (XPS). The method has been used in recent years due to its advantages involving the surface sensitivity and the possibility of obtaining highly clean samples. [1], [2]. We chose to investigate this system as an alternative for Schottky diodes based on other technologies than on silicon. Germanium has a better charge mobility than silicon [3], and nickel is a natural ferromagnet with 3 values of the work function, depending on the surface planes [4]. Ni–Ge contacts have been proven to have many possible applications, such as Schottky source/drain transistors (SSDT) [5] and thin film transistors (TFTs) [6].

Taking as a starting point the basic theory of Schottky barrier height (SBH) formation in metal-semiconductor interfaces [7] we may expect to observe a bend bending of ΔΦ = ΦNi  ΦGe  0.05–0.65 eV, considering the work function (WF) values reported in literature, ΦNi = 5.15–5.35 eV [8] and ΦGe = 4.7–5.1 eV [8], [9]. Thus, at first sight the nature of such a contact is spanned from nearly ohmic (flatband) to severely rectifying. The SBH at the interface between Ni and Ge has only been studied by using electric measurements, [10], [11], [12] and the results are similar for Ni/Ge and NiGe/Ge interfaces. This behavior was explained by taking into consideration a strong Fermi level pinning effect, with a direct consequence of a large value of the barrier height [13], [14]. There are many studies in the literature that report results on NiGe/Ge interfaces. Usually, NiGe is obtained by heating the as deposited films. It has been reported that the compound is stable for temperatures up to 1023 K [15], [16] and that it has a low resistivity [5], [15], [17]. Yan et. al. characterized the electric properties of NiGe and Ni5Ge3 nanowires obtained by chemical vapor deposition (CVD), concluding that the NiGe nanowires are highly conductive, with a resistivity four times smaller than the Ni5Ge3 nanowires [18].

Although these results are promising, the values of the Schottky barrier were dependent on the deposition method [19] as well as on the surface morphology of the germanium [20], and the epitaxial planes showed dependence on the temperature [21]. Using density functional theory (DFT) calculations, Niranjan et. al. obtained a 0.45 eV variation of the NiGe work function depending on the orientation of the planes [22]. Even though the number of variables is so high and that these materials are dependent to the chemical process, the barrier height is easily controlled by the dosage of the dopants [5], [23].

Since the issue is quite delicate, it is important to understand the interface mechanism. We report the characterization of the band bending effects using X-ray photoelectron spectroscopy (XPS). The method consists of the deposition of 10 nm of Ni on a clean substrate, in this case a p-Ge(001) surface, in 7 steps, with XPS data acquisition after every deposition step for C 1s, Ni 2p, Ge 3d and Ge 2p core level electrons. XPS is a surface sensitive technique that allows the calculation of the Schottky barrier height in a direct way by following the changes in the binding energies (BE) of both Ge 2p and Ge 3d core levels, because the depletion depth at the interface (lD=(2ϵΔΦ)/(e2ND))1/2, in the range 10 nm–1 μm, where ϵ is the permittivity of the dielectric, ΔΦ is the barrier height, e the elementary charge and ND the concentration of impurities, is lower than the inelastic mean free path of the electrons in the material λ  1–2 nm [24]. There are many advantages of the technique: it can detect Ohmic contacts, while the electrical measurements are not able to measure these types of metal/semiconductor junctions, as well as being a useful tool in the detection of the growth mechanism of the thin film [25]. Its sensitivity also allows the detection of other interface effects such as compound formation. By using deconvolutions as an additional tool for data analysis [26], in the case of band bending at Pb(Zr,Ti)O3/metal contacts, XPS data was enough to lead to the detection of the preferential deposition of the metal on areas with different out of plane polarizations on ferroelectric substrates [1], [2], and the effect of the initial contamination of the substrate could be evaluated [27], [28]. All these previous results along with the uncertainties about Ni/Ge interfaces in the literature are good arguments for characterizing the interface using an ultrahigh vacuum technique with the purpose of studying the interface phenomena.

We report that our data is consistent with previous UHV experiments and our results show that the formation of intermetallic Ni–Ge compounds is instantaneous, and the alloys occurring at the interface form with the deposition of the first Ni monolayers [29], [30] even for room temperature (RT) deposition.

Section snippets

Experimental

The sample was prepared and characterized in a surface science cluster (Specs) comprised of a molecular beam epitaxy (MBE) system equipped with an Auger/LEED (Low Energy Electron Diffraction) setup and an X-ray photoelectron spectroscopy (XPS) chamber. The base pressure in the whole setup was below 5 × 10−8 Pa. XPS was performed in an analysis chamber equipped with a 150 mm hemispherical electron energy analyzer (Phoibos), a dual anode (Mg/Al Kα) X-ray source, and a monochromatized (Al Kα/Ag Lα)

Low energy electron diffraction

The surface crystallinity of the clean Ge(001) substrate, as well as after the first two deposition steps, was verified using LEED. Diffraction patterns were only visible after the deposition of the first 2 Å of metal. Since the distance between 2 Ni layers in the fcc cell is 1.75 Å, we can conclude the surface crystallinity of Ge is lost after the deposition of more than two monolayers of Ni. Fig. 1 shows the images of the Ge (1 × 1) reconstruction and the diffraction spots corresponding to [010]

Magneto-optical Kerr effect

The surface magnetization of the resulting film was characterized using MOKE. In order to approximate the thickness of the layer responsible for the magnetic signal, Fig. 6 represents the magnetization curves of three Ni/Ge(001) samples obtained at room temperature with different thicknesses (5, 10 and 20 nm). The sample of 5 nm Ni shows no magnetic behavior, while the sample of 10 nm, which is object of this study, reaches saturation at 1.3 mdeg. Three additional thinner layers (2.5, 1 and 0.5 nm)

Conclusions

We have investigated the Schottky barrier height in a gradual deposition of Ni on Ge(001) by X-ray photoelectron spectroscopy. We infer that an interface compound forms instantaneously and that it contributes to the band bending by lowering the energy of the core levels in the interface region. The band bending has been determined to be in the range of 0.39–0.45 eV, in good agreement with values determined by other methods [10], [11], [22]. LEED patterns are lost after the first two stages of

Acknowledgements

This work is funded by the Romanian Ministry for Scientific Research and Innovation, Core Program 2016–2017, Contract No. PN16-480101. The authors thank Dr. Victor Kuncser for useful discussions on the MOKE measurements.

References (37)

  • L.C. Tanase et al.

    Growth mechanisms and band bending in Cu and Pt on Ge(001) investigated by LEED and photoelectron spectroscopy

    Surf. Sci.

    (2016)
  • C. Teodorescu et al.

    An approximation of the Voigt I profile for the fitting of experimental X-ray absorption data

    Nucl. Instrum. Methods Phys. Res. Sect. A: Accel. Spectrom. Detect. Assoc. Equip.

    (1994)
  • N.G. Apostol et al.

    Band bending at free Pb(Zr,Ti)O3 surfaces analyzed by X-ray photoelectron spectroscopy

    Mater. Sci. Eng. B: Solid-State Mater. Adv. Technol.

    (2013)
  • N.G. Apostol et al.

    Band bending in Au/Pb(Zr,Ti)O3 investigated by X-ray photoelectron spectroscopy: Dependence on the initial state of the film

    Thin Solid Films

    (2013)
  • A.E. Bocirnea et al.

    Structural and magnetic properties of Ni nanofilms on Ge(001) by molecular beam epitaxy

    Appl. Surf. Sci.

    (2017)
  • C. Jacoboni et al.

    Electron drift velocity and diffusivity in germanium

    Phys. Rev. B

    (1981)
  • T. Sadoh et al.

    Low-temperature formation (<500 °C) of poly-Ge thin-film transistor with NiGe Schottky source/drain

    Appl. Phys. Lett.

    (2006)
  • S.M. Sze et al.

    Physics of Semiconductor Devices

    (2006)
  • Cited by (0)

    View full text