Interface and electrical properties of Au catalysts (dots) on Ge(111)
H. Zitouni, N. Hakiki, K. Driss Khodja,
Laboratoire de Physique des Couches Minces et des Matériaux pour l’Electronique (LPCM2E)
Université d’Oran El’Mnaouer Oran 31100, Algérie
A. Mehdaoui, L. Josien, G. Garreau, J. L.
Bubendorff, C. Pirri*
Institut de Science des Matériaux de Mulhouse, CNRS- UMR7661,
Université de Haute Alsace 68057 Mulhouse, France
carmelo.pirri@uha.fr
Abstract
We investigate the interface between Au nanodots and a Ge(111) substrate. Au dots are achieved by dewetting an Au layer above the Au-Ge eutectic temperature TE (TE = 362°C). Surface and interface structural properties are analysed by using scanning tunnelling microscopy, transmission electron microscopy and scanning electron microscopy measurements while electrical characterizations are performed a by 4-points I-V probe.
Index Terms—Au-Ge contacts- eutectic-Au droplets-germanium
I. INTRODUCTION
The interaction of liquids phases with surfaces is a very broad subject, and the catalytic property of melt alloys clusters is one of these remarkable properties. These alloys are also of interest in the overall metallurgy and microelectronic areas as solder materials and in all technological area in which low temperature and corrosion resistance are required, such as space technology, gas sensor and medical devices. The ability of some alloys to form liquid eutectics at a temperature well below the melting temperature of each component separately, such as Au and Si but also of Au and Ge has thus been exploited. A very nice issue of the use of these catalysts is the growth of Si and Ge nanowires in epitaxy on Si(111) and Ge(111), but also the formation of ohmic contacts [1,4].
In this paper, we will show that Au droplets formed by melt above the eutectic temperature TE and cooled down to room temperature can be used as ohmic contacts with Ge(111). The electrical behavior is analyzed at a nanoscale by using conductive atomic force microscopy (AFM) measurements and more interestingly at a macroscopic scale by using 4-points I-V probe.
II. EXPERIMENTS
Clean Ge(111) is obtained in ultrahigh vacuum by degassing the sample at 400°C for overnight and flashes at 700°C for a few tenth of seconds. This procedure guarantees a clean Ge(111) surface with c(2x8) reconstruction checked by STM. An Au film (1.2 nm thick) is evaporated under ultrahigh vacuum conditions. The crystallites are obtained by dewetting the pure Au film by direct heating up to 400°C of a 1.2 nm Au deposit and they are cooled by shutting down the current abruptly to freeze the droplets into monocrystalline particles.
Owing to the very small Ge(111) sample size (5 x 10 x 0.3 mm) and the weak thermal contact with the sample holder (Omicron@), it experiences a (global) temperature decrease through the eutectic temperature TE (from 400°C to 300°C) of a few seconds only, as controlled by an optical pyrometer.
This Au amount gave us the opportunity to form Au- Ge droplets with a lateral size between 10 and 500 nm, then easily observable by SEM for the largest. The deposition rate is controlled by a water-cooled quartz crystal microbalance and the nominal Au thickness is given with a precision better than 10%. The annealing temperature is monitored with an accuracy of ± 20°C. Scanning tunneling microscope (STM) images were made in a room temperature operating microscope (Omicron STM-AFM microscope), in the constant-current mode. Transmission electron microscopy (TEM) measurements were performed with a JEOL 3010 microscope operating at 300 keV. Scanning electron microscope (SEM) images were acquired with a XL30-FEG Philips microscope. The structure of individual droplet was characterized using a Cs-corrected JEOL 2200FS scanning- transmission electron microscope (STEM) operating at 200 kV with a resolution close to 1 Å.
III. RESULTS
Room temperature Au deposition on clean Ge(111) results in a rather flat layer [5]. Surface dewetting starts at an annealing temperature below 300°C, thus below TE, and results in flat, pure Au crystallized islands, in epitaxy on Ge(111), as shown in Fig.1 [4,5].
Fig.1. STM image of a 1.2 nm Au deposit on Ge(111) annealed at 280°C, Upon increasing the annealing temperature above the bulk Au-Ge eutectic temperature TE (360°C), dome-like islands or droplets are observed when the sample is returned at room temperature. Their shape is now associated with their crystallization after melt. The “bare” surface exposes a wetting layer (Au-Ge layer) with a √3 x √3 R30° surface periodicity.
This layer involves one Au monolayer in the topmost Ge(111) planes but extends vertically over more than 1 atomic plane.
The √3 x √3 R30° superstructure induces a distortion in the deeper Ge layers [6-9].
Figure 2 shows a SEM image taken at grazing incidence, for Au crystallites annealed above the TE for 2 hours and cooled down slowly through TE. The crystallites show several facets, with a dominant (111) and (100) orientation, in agreement with data published in the literature [10]. The Au- Ge droplets formed on Ge(111) were analyzed in STM and SEM.
Fig.2. SEM images of Au-Ge crystallites with a grazing incidence.
They are analyzed by STM as long as they stay in UHV for annealing series and they are analyzed by SEM ex situ afterward. Note that the STM images are more difficult to measure for so high Au-Ge droplets but they allow an accurate droplet height measurement, in contrast with SEM.
Nevertheless, artifacts, such as multiple tip effects often present in STM images, are avoided by the use of SEM. The crystallite size is about 200 nm within such annealing conditions.
For an annealing temperature close to the eutectic temperature TE, one can observe some platelets still remaining on the surface. Figure 3 shows such a configuration. On this image, the crystallite on the right side is a platelet. The other, on the left side is an Au-Ge crystallite after melt above TE. One can see that the largest droplets are perched on a pedestal (indicated by an arrow) after annealing above TE. This pedestal is due to precipitated Ge, suggesting that quenching of the droplet composition is not completely efficient. The pedestal has a round form, whatever the nanocrystal shape, which is a reminiscence of the liquid droplet form. The contrast between the upper and the lower part of the droplets is due to large difference in atomic number Z between Au and Ge. The pedestal is a Ge rich phase while the upper part of the droplet is Au rich. This has been controlled by recent TEM measurements [5], as shown below.
Fig.3. Top-view SEM images showing the coexistence of an Au platelet and an Au-Ge crystallite. This image was taken after annealing at the eutectic temperature TE. The arrow indicated a weak contrast due Ge precipitation.
Fig.4. STM image showing Au droplets on Ge(111). In inset is shown a profile along the white line.
Thus, SEM images show that the interface between an Au droplet or crystallite and the Ge(111) substrate is formed by an extra Ge-rich layer, namely the pedestal.
Figure 4 shows an STM image measured on the Ge(111) surface after annealing at 400°C (thus above TE) for 2 hours and cooling down to room temperature. This image shows the surface morphology, with large areas between the Au crystallites. The Au has now completely reacted with Ge to form Au-Ge droplets. Note that the Ge pedestal is not visible below the Au crystallites, due to the lack of Z selectivity of STM for the present system. Upon cooling down to room temperature, droplets of various height and lateral size are observed. The size measurement on droplets is shown in inset.
These droplets are 15 nm height and 60 nm lateral, typically.
Figure 5 shows a cross-section TEM image of an Au droplet at room temperature, after melt at 400°C. This image shows the Ge precipitation in the form of a pedestal but also a strong interaction of the Au droplet with the Ge(111) substrate [5]. EDS measurements shown in the lower part of Fig.5 reveal that an alloyed zone is formed between pure Au and Ge(111) substrate.
Fig.5. Cross-section TEM images for a 1.2-nm Au deposit annealed at 400°C. This figure also shows the Ge L, Ge K, and Au Mα lines intensity across the line of the cross-section images at selected points (From reference 6).
The electric properties of this Au/Ge/Ge(111) interface is now investigated by conductive AFM as well as by 4-points I- V measurements. The purpose of the following paragraphs is to take a macroscopic characterization of the Aucrystal/Gepedestal/Ge(111) system. Upon using macroscopic tips
for I-V measurements on such dispersed Au crystallites, one have to be sure that the current passes through them and not directly from Ge(111) to the tips.
We have done conductive AFM measurements on that surface. These measurements consist in the measurement of a current I induced through an AFM tip in “contact” with the surface at a given bias voltage V. These measurements have been done on a sample exposed to air after growth, before macroscopic I(V) measurements. Within this procedure, one would expect that the “bare” surface between the droplet is oxidized enough and then non-conductive.
In conductive AFM, the tip radius is well below the mean Au crystallite diameter and then we can estimate the conductive or non conductive areas of the surface. The tip is scanned across the surface, as shown in Figure 6a and 6b, and we measure the surface topography as well as a current image, at a given bias voltage. Figure 6a shows a topographic z (x,y) image and Fig. 6b shows a current I(x,y) image.
Fig.6. (a) Topography AFM image of Au-Ge crystallites and (b) conductive AFM current image.
Figure 6b clearly shows that, upon applying a bias voltage of 100 mV, the current through the Au crystallites is more than two orders of magnitude larger than that through the “bare”
surface. Thus, one can expect that the I-V curves measured with macroscopic tips also reflect the contribution of the Au/Geped/Ge(111) interface.
I(V) measurements performed on the Au droplets are shown in Figure 7a. This figure shows raw data in the voltage gap [-1V; 1V]. The current is shown to vary linearly with bias voltage, as expected for a perfect ohmic contact. Also shown in Figure 6 is the current variation versus voltage for droplets with the same mean height, mean diameter and dispersion over a Ge(111) surface, but containing 9% Mn (Fig.7b). For these Mn-doped droplets, the I(V) curves deviates significantly from linearity, due to the formation of a this interfacial Ge(Mn) alloy. The asymmetric I(V) curve is close to that expected for a diode. Note that this thin alloyed Ge(Mn) layer has also magnetic properties, which are not described in the present paper. For both Au and Au(Mn) droplets, the mean height, the mean diameter and dispersion over the substrate is quite the same and thus the modification in I(V) is only due to interface
electrical properties. As to their surface composition, a small Ge segregation occurs, whatever the Mn content. The surface is Ge-rich and this extra Ge on top of the droplets seems to stabilize them. The linear variation of I versus V for pure deposited Au is in line with recent measurements performed by Kishore et al [11].
Fig.7. (a) Room temperature measurement of 4-points I-V curves for Au droplets formed at 500°C and for Au-Mn droplets formed at 500°C.
IV. CONCLUSIONS
We have investigated the structural and electrical properties of Au droplets formed on Ge(111) upon dewetting an Au layer.
These droplets usually serve as catalysts for Ge nanowires growth but they can also serve for electrical contacts in electrical characterization of nanoscale one-dimensional systems in contact with Ge. Indeed, they exhibit a clear linear I(V) behavior, in a wide bias voltage range.
V. COPYRIGHT FORMS
You must submit the IEEE Electronic Copyright Form (ECF) as described in your author-kit message. THIS FORM MUST BE SUBMITTED IN ORDER TO PUBLISH YOUR PAPER.
REFERENCES
[1] Binary Alloy Phase Diagrams, 2nd ed.; ASM International:
Materials Park, OH, 1990; Vol. 1.
[2] The microstructure and electrical properties of nonalloyed epitaxial Au-Ge ohmic contacts to n-GaAs, J. Appl. Phys. Vol.
72, pp 4773, 1992
[3] C. Renard, L. Vincent, C. Gardès, R. Boukhicha, E. Oliviero, G.
Patriarche, F. Fossard, S. Hajjar, J.L. Bubendorff, C. Pirri, D.
Bouchier, Physica Status Solidi Vol. 11, pp 315–319, 2014.
[4] P. Farzinpour , A. Sundar , K. D. Gilroy , Z. E. Eskin , R. A.
Hughes and S. Neretina, Nanoscale, Vol5, pp 1929-1938, 2013 [5] S. Hajjar, G. Garreau, L. Josien, J. L. Bubendorff, D. Berling, A.
Mehdaoui, C. Pirri, T. Maroutian, C. Renard, D. Bouchier, M.
Petit, A. Spiesser, M. T. Dau, L. Michez, V. Le Thanh, T. O.
Mentes, M. A. Nino, and A. Locatelli, Phys. Rev. B Vol. 84, pp 125325, 2011.
[6] G. Le Lay, M. Manneville, J.J. Métois, Surf. Sci., Vol. 123(1) pp 117, 1982
[7] P. B. Howes, C. Norris, M. S. Finney, E. Vlieg and R. G. van Silfhout, Phys. Rev. B Vol. 48, pp 1632, 1993
[8] 37. M. Göthelid, M. Hammar, M. Björkqvist, U. O. Karlsson, and S. A. Flodström, C. Wigren, G. Le Lay, Phys. Rev. B Vol.
50, pp 4470, 1994
[9] H. Over, C. P., Wang, F. Jona, Phys. Rev. B Vol. 51, pp 4231, 1995.
[10] W.D. Kaplan, D. Chatain, P. Wynblatt, W. C. Carter, J Mater Sci Vol. 48, pp 5681–5717, 2013
[11] V. P. Kishore, P. Paramahans, S. Sadana, U. Ganguly, and S.
Lodha, Appl Phys Lett. Vol. 100, pp 142107, 2012.
