Si(001

Subscriber access provided by Washington University | Libraries

Article

Metal-free CVD Graphene synthesis on 200 mm Ge/Si (100) substrates Mindaugas Lukosius, Jaroslaw Dabrowski, Julia Kitzmann, Oksana Fursenko, Marco Lisker, Fatima Akhtar, Sebastian Schulze, Gunther Lippert, Yuji Yamamoto, Markus Andreas Schubert, Hans-Michael Krause, Andre Wolff, Andreas Mai, Thomas Schroeder, and Grzegorz Lupina ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.6b11397 • Publication Date (Web): 18 Nov 2016 Downloaded from http://pubs.acs.org on November 20, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Applied Materials & Interfaces is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Metal-free CVD Graphene Synthesis on 200 mm Ge/Si (001) Substrates M. Lukosius1*, J. Dabrowski1, J. Kitzmann1, O. Fursenko1, F. Akhtar1, M. Lisker1, G. Lippert1, S. Schulze1, Y. Yamamoto1, M. A. Schubert1, H. M. Krause1, A. Wolff1, A. Mai1,T. Schroeder1,2, G. Lupina1 1

2

IHP, Im Technologiepark 25, 15236 Frankfurt (Oder). Germany BTU Cottbus-Senftenberg, Konrad Zuse Str. 1, 03046 Cottbus, Germany

*[email protected] Keywords: Graphene synthesis, CVD, Ge(001), CMOS, 200 mm, Ab-initio DFT, Faceting, nucleation

Abstract Good quality, complementary-metal-oxide-semiconductor (CMOS) technology compatible, 200 mm graphene was obtained on Ge(001)/Si(001) wafers in this work. Chemical vapor depositions were carried out at the deposition temperatures of 885 °C using CH4 as carbon source on epitaxial Ge(100) layers, which were grown on Si(100), prior to the graphene synthesis. Graphene layer with the 2D/G ratio ~ 3 and low D mode (i.e., low concentration of defects) was measured over the entire 200 mm wafer by Raman spectroscopy. A typical full-width-at-half maximum value of 39 cm-1 was extracted for the 2D mode; further indicating that graphene of good structural quality was produced. The study also revealed that the lack of interfacial oxide correlates with superior properties of graphene. In order to evaluate electrical properties of graphene, its 2x2 cm² pieces were transferred onto SiO2/Si substrates from Ge/Si wafers. The extracted sheet resistance and mobility values of transferred graphene layers were ~1500 ±100 Ω/sq and µ ~400 ±20 cm2/V∙s, respectively. The transferred graphene was free of metallic contaminations or mechanical damage. On the basis of results of DFT calculations, we attribute the high structural quality of graphene grown by CVD on Ge to hydrogen-induced reduction of

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 22

nucleation probability, we explain the appearance of graphene-induced facets on Ge(001) as a kinetic effect caused by surface step pinning at linear graphene nuclei, and we clarify the orientation of graphene domains on Ge(001) as resulting from good lattice matching between Ge(001) and graphene nucleated on such nuclei.

Introduction High quality, wafer scale graphene synthesis on CMOS compatible materials (dielectrics or semiconductors) is of importance for the fabrication of various electronic or photonic graphenebased devices within the mainstream Silicon (Si) technology. During the last years, Germanium (Ge) has shown the potential as an alternative substrate for the graphene growth

1-3

, due to its

catalytic activity 4, extremely low solubility of carbon 5 and availability of large area Ge on Si. 6 Recently, it has been reported by Kiraly et. al., that the electronic as well as mechanical properties of the graphene are influenced by the orientation of the germanium crystal. 7 Ge(110) orientation was found to be preferable, in particular, graphene with a single orientation can be grown on Ge(110), in contrast to Ge(001) and Ge(111). The extracted carrier mobility values for the graphene grown on Ge(110) were around 7000 cm2/V∙s, whereas the mobility values of 2500 cm2/V∙s have been found by low temperature Hall measurements for the Graphene/Ge(001) system.

8

On the other hand, due to the possible incompatibility of Ge(110) with the standard

CMOS Si(001) flow, Ge(001) orientation is more attractive for production purposes, particularly when graphene is intended to be grown directly on the target wafer. Therefore, in this study we have employed epitaxial, low threading dislocation density and low surface roughness Ge(001) layers, produced on Si (001) substrates by CVD. 6


Page 3 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


An important aspect and one of the main motivations of this work was to explore the possibility of growing graphene on 8 inch wafers. Up to the date when this report was written, graphene synthesis was performed on Ge substrates no larger than 2 inches in diameter. Detailed microscopic and macroscopic characterization was done to assess the quality of the grown graphene. Furthermore, an explanation of the faceting of Ge(001) during the graphene synthesis is proposed on the basis of ab initio density functional theory (DFT) calculations. Understanding this mechanism is not only of fundamental scientific interest, but also important for the further developments of graphene/Ge(001) synthesis. Finally, it should be also stressed that the growth was carried out in the standard BiCMOS pilot-line, making this study unique as its results might directly pave the way to the further graphene integration and graphene-based device prototyping in the mainstream Si technologies.

Approach Graphene synthesis has been performed on Ge/Si substrates using Aixtron’s Black Magic BM300T CVD tool in 200 mm wafer configuration. 2 micron thick Ge (001) layers have been grown by the CVD method prior to the deposition of graphene 6. Graphene synthesis was carried out at the deposition temperatures of 885 °C using CH4 as a source of carbon and Ar/H2 mixture as a carrier gas. The pressure of 700 mbar was kept during the deposition; the optimized deposition time was 60 minutes. Carbon content was determined by X-Ray Photoelectron Spectroscopy (PHI Electronics, Al Kα), whereas the quality of the graphene layers was monitored by the Renishaw inVia Raman spectroscopy using 514 nm excitation wavelength of the laser. The surfaces of the samples were investigated by scanning electron microscopy



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(Merlin, ZEISS). Roughness of the samples was investigated by Park Systems NX20 Atomic Force Microscopy (AFM). Ellipsometry (KLA Fx200) was employed for the measurements of the graphene thickness over entire 200 mm wafers. Finally, graphene was electrochemically delaminated from the Ge surface and transferred onto 100 nm SiO2/Si substrates. Electrical properties have been extracted using Lakeshore 7600 Hall system after soldering indium contacts to the transferred graphene. Contamination levels in the grown as well as transferred layers were measured by ToF-SIMS (ION-TOF-5). Further insight into the growth mechanism was obtained from pseudopotential DFT calculations by the Quantum Espresso package 9, using the PBE gradient-corrected functional

10

and periodically repeated Ge(001) slabs with eight (001) planes.

The slabs were separated by about 10 Å of vacuum, terminated on one side by H, and electrostatically decoupled by dipole correction. Van der Waals forces were accounted for within the DTG-D approach. 11,12

Results In the first step, the thickness of Germanium epi layer was optimized, since it is well known that Si diffuses into Ge, especially at higher temperatures.

6,13

This effect is undesired, since the

formation of Si-C compounds that can co-exist with graphene prevents the growth of high quality graphene.

14

A typical example of the TEM/EDX profile of 3 micron thick Ge (annealed

at 885 °C for 1h) is shown in Fig. 1, where the diffusion of Silicon into Germanium of ~1.2 µm was measured. In order to prevent any Si diffusion onto the Ge surface, a thickness of 2 microns was selected for further experiments in this work.


Page 4 of 22

Page 5 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Fig.1 Line scan of the TEM/EDX depth profile of the 3 micron Ge(001) on Si(001) substrate.

After the determination of the required Ge thickness, the attempts to synthesize graphene have been performed at the deposition temperature of 885 °C. The quality of the layers was investigated by Raman spectroscopy by performing a line scan (~100 points) across the entire wafer as depicted in Fig. 2a by the red dotted line. The measurements revealed that very similar Raman spectra could be obtained over the wafer, indicating that uniform graphene can be homogeneously grown over large areas. For simplicity, only several typical spectra of three selected areas, indicated as 1, 2 and 3, are shown in Fig. 2b. From the Raman spectra one can notice that graphene layers have small D mode (ID/IG ratio ~ 0.1), which indicates a low concentration of defects and therefore good quality of as-grown graphene. The 2D mode was found everywhere on the wafer. It has a ratio of ~ 2.8 to the G mode over the full 200 mm wafer, as shown in the histogram in the inset of the Fig. 2b).



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fig. 2. (a) Graphene grown on 200 mm Ge/Si wafer and (b) Raman spectra at the indicated places. The histogram of the 2D/G ratio over the entire wafer (~100 measured points) is depicted in the inset of Fig. 2 b).

In order to evaluate the properties of the graphene on the microscopic scale, a micro Raman mapping was performed over the area of 10×10 microns. A full-width-at-half-maximum (FWHM) map of the 2D mode is plotted in Fig. 3a). It is seen that the values varied within the range of 36 to 42 cm-1, which is a similar range to that reported in the literature.8 A histogram and a typical 2D curve of the graphene layer are presented in Fig. 3b, where an average value FWHM of 39 cm-1 was extracted. This is higher than the FWHM of perfect monolayer graphene and might in principle point to presence of bilayer islands, but no significant concentration of such islands was observed in SEM images taken after transfer of the graphene onto SiO2 (see Fig. 8a). In addition, it was recently reported that broadening of the Raman 2D line is also strongly related to the strain variations in graphene in the nanometer scale, that is, over distances smaller than the spot size of the Raman laser.15 In spite of small variation of strain averaged over the spot area, one can therefore tentatively attribute the FWHM broadening to strain variations (Fig. 3c), albeit in the nanoscale.


Page 6 of 22

Page 7 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Fig. 3. Raman analysis. (a) FWHM map of 2D mode. b) Histogram of the 2D mode. (c) 2D-G probability distribution plot for as-grown sample. Doping is given in cm-2. (d) The same after transfer to SiO2/Si. Doping and strain was estimated by plotting the probability distribution of the 2D mode vs. the measured probability distribution of the G mode.16 The strain is compressive and its magnitude is the same in as-grown graphene and after its transfer to SiO2/Si. If biaxial strain is assumed, it amounts to -0.15%. As-grown graphene is slightly doped (2×1013 cm-2) and the doping level did not change after the transfer, which may indicate that it stems mostly from molecules adsorbed during exposition to ambient air. Strain and doping of our graphene are comparable to those deduced from a 2D-G plot for MBE graphene on Ge(001) (-0.3% biaxial and 1013 cm-2).

17

There is however a significant difference to the graphene grown on Ge(001)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 22

from C2H4 by UHV CVD (+0.4% biaxial, no indications of an interfacial layer). This variation may be associated with different nucleation and growth mechanisms during both CVD processes. Even though in both cases two orientation domains of graphene were observed, much larger domain size (as evidenced by much weaker Raman D mode and by the LEED pattern) is obtained by the current approach. As explained below on the DFT basis, these differences can be attributed to physical conditions of the process (pressure and temperature of the gas) rather than to the choice of the precursor. During the UHV-CVD process, hydrogen coverage of the Ge(001) surface is expected to be considerably lower than during exposure to a hydrocarbon precursor at near-atmospheric pressure, as was the case in this work. Hydrogen is delivered to the surface as a decomposition product of the precursor, and it originates also from the H2 carrier gas. Higher H coverages decrease the nucleation probability of graphene by promoting desorption of the product of hydrocarbon polymerization reactions taking place on the surface.

17

Lower

nucleation site density leads to increased domain size. Furthermore, decreased probability of polymerization processes on the surface due to reduced diffusivity and increased volatility of adsorbed species as well as increased precursor pressure enhances the role of polymerization in the gas phase. Being indicative of weak graphene-substrate interaction, these observations are compatible with small difference between strain in as-grown CVD films and after its transfer to SiO2. The uniformity of the grown graphene was also controlled by spectroscopic ellipsometry in the wavelength range of 200-800 nm using ~71° angle of incidence in visual (240-800 nm) and with ~63° angle of incidence in ultraviolet range (200-300 nm). The spot size was 28 x 14 µm. The optical constants of graphene were extracted by point-by-point fit of spectra measured for exfoliated graphene on SiO2/Si at each wavelength. The extracted data was then


Page 9 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


applied for the thickness determination of graphene deposited on Ge/Si substrates. The GOF (goodness of fit) value has been used for evaluation of measurement results. In addition, the developed optical model was also successfully tested on the absence of Graphene layer on blanket Ge/Si wafer. Fig. 4 shows the 121 points graphene thickness map over the 200 mm wafer. The measured average graphene thickness was ~0.27 nm with a standard deviation value of 4% over the full wafer. This indicates a homogenous distribution of graphene on the wafer scale and is in a good agreement with the Raman and other complementary techniques. The value 0.27 nm is less than the distance between carbon layers in graphite (0.34 nm) and also than the calculated distance from C to the average position of topmost Ge atoms (0.32±0.02 nm).18 On the other hand, it is not straightforward to interpret this “graphene thickness” as a distance between atomic layers, also because the optical constants for graphene/Ge/Si measurements were adapted from the Graphene/SiO2/Si system, and because the model is slightly simplified and does not fully take into account such effects like roughness of the substrate or the presence of surface hydride. In spite of this, spectroscopic ellipsometry turns out as a suitable tool for fast and non-destructive on-line control of the graphene deposition process. In general, it could be concluded from the Raman and ellipsometry data that it is possible to grow uniform, large area and good quality graphene on germanium/silicon substrates.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 22

Fig. 4. Thickness map of the graphene grown on Ge/Si over the entire 200 mm wafer, measured by spectroscopic ellipsometry. The marked sites correspond to the measurement positions.

Substrate roughness is another quantity that was measured in this study. Firstly, roughness of initial Ge and temperature treated samples were then monitored by optical profilometry on 1x1 mm scale. As-grown Ge surfaces were very smooth (typical Ra ~ 0.5 nm) and the exposure of the substrate to the graphene deposition temperature slightly increased the roughness, to Ra around 0.7 nm, indicating that no significant roughening of the surface occurred. A more detailed surface morphology was provided by SEM and AFM. A micrograph of a graphene sheet on Ge(001) obtained in this work by an optimized growth process is shown in Fig. 5. Graphene-induced Ge faceting is clearly recognizable. This observation is in a good agreement with other reports,8,19 where graphene was synthesized on Ge(001) by CVD under growth conditions similar to those used in this work. It was reported that during the graphene growth the Ge(001) surface broke up into hills and valleys: two families of (107) facets oriented 90° to each other appeared, which has been associated to the formation of two different graphene orientations on Ge(001).8 Ge faceting can be also clearly recognizable in the AFM micrograph,


Page 11 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


depicted in Fig. 5 b), where the typical heights of the facets of 2-5 nm and Ra values of 0.7-1.1 nm have been measured.

Fig. 5. SEM (a) and AFM (b) images of the graphene on Ge(001).

However, no facets have been observed after graphene growth on Ge(001) by atomic carbon molecular beam, where no H background was present 20 and by UHV CVD, where the H coverage of the surface was estimated to be relatively low. 17 This may indicate that the faceting is induced by adsorbed hydrogen. To verify this concept, we modelled the facet energetics by comparing the surface energy of Ge(001) and surface energies of Ge(108) surfaces with various reconstruction at its terrace edges, including the rebounded geometry considered to be the lowest energy structure of surfaces of this type.

21

The major difference between Ge(108) and Ge(107)

that is usually reported as the facet orientation is the width of the terraces; the terrace edge reconstruction and the orientation of facet ridges with respect to Ge(001) is the same, i.e., (010). We considered surfaces with various H coverages, from 0 % (2×1-type reconstructions) to 150 % (3×1-type reconstructions) and found that Ge(108) is always unstable against the transformation back to Ge(001). We also did not observe any stabilization of Ge(108) by hydrocarbon polymers.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 22

However, we found that when a linear chain of C6Hn rings attaches itself to Ge(001)-p2×2 in the (010) direction equivalent to the direction of Ge(108) ridges (Fig. 6), the lattice mismatch between the substrate and the chain is reduced to below 0.5%.

Fig. 6. Graphene ribbons on flat Ge(001) and Ge(108) surfaces. Brown solid line indicates the direction of Ge dimer rows on Ge(001), whereas black dashed line – one of the three C-C bond directions in graphene hexagons. The blue arrows point to C-Ge bonds.

This is in contrast to chains attached to Ge(001) in the direction perpendicular to the dimer rows: such chains are strongly strained and become unstable already when longer than three rings. The unstable chains may eventually become detached from the surface by hydrogen and evaporate or, if they retain chemical bonds with the substrate, their further growth is inhibited by high strain. But the chains accidentally oriented along (010), as illustrated in Fig. 7, are more stable: they have an increased survival chance and can also grow faster. Appearance of such a chain is a rare event and the contribution of these chains to graphene growth becomes


Page 13 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


pronounced only when other nucleation mechanisms, like the dimer vacancy dominating at low H coverages, become irrelevant. This is consistent with all abovementioned experimental observations. Furthermore, the orientation of graphene nucleated on the (010) chains is rotated by about 4° with respect to the dimer rows, which is consistent with the orientation of graphene domains and graphene ribbons observed in CVD growth on Ge(001). 7 The angle depends on the balance between the strain energy in graphene and the energy of the Ge-C bonds and is affected as well by the presence of H underneath graphene; we estimate that it varies between about 3.5° and 4.5°. The (107) facets are thus deduced to form by a kinetic mechanism. The linear polymers attached with two bonds to each dimer row (Fig. 6) pin the direction of surface steps to that corresponding to the direction of steps between (107) terraces (Fig. 6). The steps appear on the surface because at high temperatures and particularly under the presence of H (which is an etching agent) Ge atoms sublime from Ge(001). This conclusion is in agreement with the interpretation of the faceting as a kinetic process, given in literature.7 In the next stage, graphene was transferred from the Ge/Si substrates onto the 100 nm SiO2/Si substrates in order to evaluate the electrical properties of the grown graphene. Typically, 5 – 10 pieces, 2x2 cm in size, were cut out from the 200 mm wafer and then transferred by electrochemical delamination method onto SiO2. The SEM image of the transferred graphene is shown in Fig. 7a): there is no evidence of cracks, holes or wrinkles. To the contrary, reference graphene, transferred from Cu foil, suffers from these defects (Fig. 7b). This difference in transferred graphene quite important, since access to clean graphene sheets, free of transfer damage, can make the fabrication of proof-of-concept devices efficient also in environments where direct graphene synthesis is not feasible. Nevertheless, a weak Raman D mode is induced



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 22

by the transfer (cf. the inset in Fig. 7a), which indicates that further optimization of the transfer process is needed. Additionally, dark spots can be distinguished in SEM images of graphene transferred from Ge (Fig. 7a). They are attributable to bilayer graphene; in analogy to such areas seen on graphene transferred from Cu foil (Fig. 7b). The estimated bilayer fraction is below 1% (approximately, 0.6%±0.1%) for the Graphene/Ge system.

Fig. 7. SEM images of the graphene transferred onto SiO2/Si substrates from a) Ge/Si and b) Cu. A representative Raman spectrum of the graphene after transfer from Ge/Si is shown in the inset of Fig. a).

Electrical properties of the transferred graphene were extracted from Hall measurements from typically 10 pieces, cut out of the full 8 inch wafer. The separation between the electrodes was 1 cm. Sheet resistance of Rs = 2000 ±100 Ω/sq and electron mobility of µ = 300 ±20 cm2/V∙s were measured. These values are lower than those reported in literature.2,8 We primarily attribute these differences to the polymer contamination and/or to the residual transfer damage (Fig. 8a, inset). Another reason for the inferior Rs and µ could be related to surface quality of Ge prior to the synthesis of graphene. Residual oxide and unhealed surface damage might result in the formation of defects deteriorating the electrical quality of graphene. At the same time, such defects could act as weak spots through which oxygen might leak through the graphene sheet to


Page 15 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the Ge surface, oxidizing it. In order to check the oxygen content in the samples, TEM/EDX and X-ray Photoelectron Spectroscopy measurements were performed. For TEM/EDX, graphene was capped with thin, thermally evaporated Al layer; this was not the case for the samples examined by XPS. In addition, the samples were annealed in UHV conditions at 500 °C for 1 hour, prior to the XPS measurements. As illustrated in Fig. 8a, the TEM/EDX profile reveals an oxygen-rich interfacial layer between graphene and the Ge substrate. Clear evidence of O bonded to Ge was also found also by XPS analysis (Fig. 8c)

Fig. 8. EDX depth profiles of a) initial and b) optimized graphene on Ge. The XPS profiles were measured after annealing the samples in UHV at 500 °C for 1h. The corresponding Ge 2p signals are given in c) and d). The insets of the graphs represent the XPS Oxygen 1s signal.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 22

To improve the quality of the starting surface, H2 bake time before graphene deposition was increased. In addition, the cooling rate after graphene deposition was also decreased. The result of the optimized growth is shown in Fig. 8b): no oxygen can be seen at the interface by EDX. To confirm this observation, XPS analysis was performed again. In this case, only elemental Ge (Fig. 8d) at the binding energy of 1218 eV is visible. The extracted electrical values for these optimized samples have then improved to Rs ~1500 ±100 Ω/sq and µ ~400 ±20 cm2/V∙s. Nevertheless, mobility values are still lower that the ones in reported literature and further optimizations of growth as well of the transfer procedures are required. Finally, we note that a another issue that might limit the integration of graphene into the standard CMOS technology is the metallic contamination arising from the graphene growth on catalytic metallic substrates and also from graphene transfer to the target substrates. 22 Therefore, graphene grown in this work was tested for the presence of Cu, Al, Fe and Ni contamination. An example of such measurements for Cu+ and Al+ on graphene, directly on Ge/Si and on transferred to SiO2 are illustrated in Fig.9 a)-b) respectively. The measurements revealed that no Cu, Al and other metallic contaminations could be detected by ToF-SIMS neither on the graphene grown directly on Ge nor after subsequent transfer to SiO2/Si. These results are also compared to the reference graphene, transferred from Cu foil. To the contrary, reference graphene suffers from these metallic contaminations as shown in Fig. 9a-b). The surface concentrations of Cu and Al ions always exceeded 1014 atoms/cm2 for the Cu-graphene samples, whereas the levels of contaminations in the graphene, grown on Ge(001) and on transferred ones, were below the detections limits.


Page 17 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Fig. 9. ToF-SIMS mass spectra images of a) Cu+ and b) Al+ regions, acquired on Graphene/Ge/Si (▲), graphene on SiO2, transferred from Ge/Si ( ), and graphene on SiO2, transferred from Cu foil (○).

Conclusions Large area graphene synthesis by CVD on 200 mm Ge/Si (001) wafers was performed in this work. Uniform and high quality graphene layers can be grown, as it was verified by Raman spectroscopy. Spectroscopic ellipsometry confirms the uniformity of the grown graphene on the wafer scale. The grown graphene was successfully transferred on to SiO2 and electrical properties were extracted from Hall measurements. Sheet resistance and mobility of Rs ~1500 ±100 Ω/sq and µ ~400 ±20 cm2/V∙s, respectively, were measured. According to DFT results, high quality of the CVD graphene is attributable to increased surface coverage by H, which suppresses the nucleation of graphene on Ge(001) dimer vacancies and reduces the nucleation probability on flat surface. The growth of graphene was accompanied by appearance of surface facets, which may also contribute to the low mobility values. This can be explained by Ge(001)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 22

step pinning on linear graphene nuclei bonded to Ge(001) along (010) directions. This is also in agreement with the typical orientation of graphene domains in CVD graphene on Ge(001). Finally, we verified that the transferred as well as grown graphene were clean of any metallic contaminations.

Acknowledgments. Computing time support from the Jülich Supercomputing Center of the John von Neumann Institute for Computing (project hfo06) is gratefully acknowledged.


Page 19 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


References (1) Lee, J. H.; Lee, E. K.; Joo, W. J.; Jang, Y.; Kim, B. S.; Lim, J. Y.; Choi, S. H.; Ahn, S. J.; Park, M. H.; Yang, C. W.; Choi, B. L.; Hwang, S. W.; Whang, D. Wafer-Scale Growth of Single-Crystal Monolayer Graphene on Reusable Hydrogen-Terminated Germanium. Science Express 2014, 344, 286-289. (2) Wang, G.; Zhang, M.; Zhu, Y.; Ding, G.; Guo, Qinglei.; Liu, S.; Xie, X.; Chu, P. K.; Di. Z.; Wang, X. Direct Growth of Graphene Film on Germanium Substrate. Sci. Rep. 2013, 3, 2465. (3) Pasternak, I.; Weselowski, M.; Jozwik, I.; Lukosius, M.; Lupina, G.; Dabrowski, P.; Baranowski, J. M.; Strupinski, W. Graphene Growth on Ge(100)/Si(100) Substrates by CVD method. Sci. Rep. 2016, 6, 21773. (4) Lascutoff, P. W.; Bent, S. F. Reactivity of the Germanium Surface: Chemical Passivation and Functionalization. Annu. Rev. Phys. Chem. 2006, 57, 467-495. (5) Scace, R. I.; Slack, G. A. Solubility of Carbon in Silicon and Germanium. J. Chem. Phys. 1959, 30, 1551–1555. (6) Yamamoto, Y.; Zaumseil. P.; Arguirov, T.; Kittler, M.; Tillack, B. Low Threading Dislocation Density Ge deposited on Si (100) using RPCVD. Solid State Electronics 2011, 60, 26. (7) Kiraly, B.; Jacobberger, R. M.; Mannix, A. J.; Campbell, G. P.; Bedzyk, M. J.; Arnold, M. S.; Hersam, M. C.; Guisinger, N. P. Electronic and Mechanical Properties of Graphene-Germanium Interfaces by Chemical Vapor Deposition. Nano Lett. 2015, 15, 7414-7420. (8) Pasternak, I.; Dabrowski, P.; Ciepielewski, P.; Kolkovsky, V.; Klusek, Z.; Baranowski, J. M.; Strupinski, W. Large-area high Quality Graphene on Ge(001)/Si(001) substrates. Nanoscale 2016, 8, 11241-11247.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 22

(9) Giannozzi, P.; Baroni, S.; Bonini, N.; Calandra, M.; Car, R.; Cavazzoni, C.; Ceresoli, D.; Chiarotti, G.; Cococcioni, M.; Dabo, I.; Dal Corso, A.; de Gironcoli, S.; Fabris, S.; Fratesi, G.; Gebauer, R.; Gerstmann, U.; Gougoussi, C.; Kokalj, A.; Lazzeri, M.; Martin-Samos, L.; Marzari N.; Mauri, F.; Mazzarello, R.;Paolini, S.; Pasquarello, A.; Paulatto, L.; Sbraccia, C.; Sclauzero, G.; Seitsonen, A.P.; Smogunov , A.; Umari, P.; Wentzcovitch, R. M. QUANTUM ESPRESSO: a Modular and Open-source Software Project for Quantum Simulations of Materials. J. Phys. Cond. Mat. 2009, 21, 395502. (10) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865. (11) Grimme, S. Semiempirical GGA-type Density Functional Constructed with a Long-range Dispersion DCrrection. J. Comput. Chem 2006, 27, 1787–1799. (12) Barone, V.; Casarin, M.; Forrer, D.; Pavone, M.; Sambi, M.; Vittadini, A. Role and Effective Treatment of Dispersive Forces in Materials: Polyethylene and Graphite Crystals as Test Cases. J. Comput. Chem. 2009, 30, 934-939. (13) Lukosius, M.; Lippert, G.; Dabrowski, J.; Kitzmann, J.; Lisker, M.; Kulse, P.; Krueger, A.; Fursenko, O.; Costina, I.; Trucsch, A.; Yamamoto, Y.; Wollf, A.; Mai, A.; Schroeder, T.; Lupina, G. Graphene Synthesis and Processing on Ge Substrates. ECS Transactions 2016, 75 (8), 533-540. (14) Son, Y. H.; Park, J. H.; Kwon, S.; Park, S.; Rummeli, M. H.; Bachmatiuk, A.; Song, H. J.; Ku, J.; Choi, J. W. Choi, J. M.; Seok-Gwang Doo, S. G.; Chang, H. Silicon Carbide-free Graphene Growth on Silicon for Lithium-ion Battery with High Volumetric Energy Density. Nat. Commun. 2015, 6, 7393.


Page 21 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(15) Neumann, C.; Reichardt, S.; Venezuela, P.; Drögeler, M.; Banszerus, L.; Schmitz, M.; Watanabe, K.; Taniguchi, T.; Mauri, F.; Beschoten, B.; Rotkin, S.V.; Stampfer, C. Raman Spectroscopy as probe of Nanometer-scale Strain Variations in Graphene. Nat. Commun. 2015, 6, 8429. (16) Lee, J. E.; Ahn, G.; Shim, J.; Lee, Y. S.; Ryu, S. Optical Separation of Mechanical Strain from Charge Doping in Graphene. Nat. Commun. 2012, 3, 1024. (17) Dabrowski, J.; Lippert, G.; Avila. J.; Baringaus, J.; Colambo, I.; Dedkov, Y. S.; Herziger, F.; Lupina, G.; Maultzsch, J.; Schaffus, T.; Schroeder, T.; Kot, M.; Tegenkamp, C.; Vignaug, D.; Asensio, M. C. Understandting the Growth Mechanism of Graphene on Ge/Si (001) Surfaces. Sci. Rep. 2016, 6, 31639. (18) The calculations were done for a periodic (C56H10)n ribbon hovering over a (108) facet. (19) McElhinny, K. M.; Jacobberger, R. M.; Zaug, A. J.; Arnold, M. S.; Evans, P. G. Grapheneinduced Ge (001) Surface Faceting. Surf. Sci. 2016, 647, 90-95. (20) Lippert, G.; Dabrowski, J.; Schroeder, T.; Schubert, M. A.; Yamamoto, Y.; Herzinger, F.; Maultzsch, J.; Baringaus, J.; Tegenkamp, C.; Asensio, M. C.; Avila. J.; Lupina, G. Graphene Grown on Ge(0 0 1) from Atomic Source. Carbon 2014, 75, 104-112. (21) Hashimoto, T.; Morikawa, Y.; Fujikawa, Y.; Sakurai, T.; Lagally, M. G.; Terakura, K. Rebonded SB Step Model of Ge/Si(105)1×2: A first-principles theoretical study. Surf. Sci. 2002, 513, L445-L450. (22) Lupina, G.; Kitzmann, J.; Costina, I.; Lukosius, M.; Wenger, Ch.; Wollf, A.; Vaziri, S.; Oestling, M.; Pasternak, I.; Krajewska, A.; Strupinski, W.; Kataria, S.; Gahoi, A.; Lemme, M.; Ruhl, G.; Zoth, G.; Luxenhofer, O.; M. Wolfgang. Residual Metallic Contamination of Transferred Chemical Vapor Deposited Graphene. ACS Nano 2015, 9, 4776- 4785.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Graphical Abstract


Page 22 of 22

Si(001

Recommend Documents