Investigation of structural and electronic properties of epitaxial graphene on 3C-SiC(100)/Si(100) substrates.

Graphene has been intensively studied in recent years in order to take advantage of its unique properties. Its synthesis on SiC substrates by solid-state graphitization appears a suitable option for graphene-based electronics. However, before developing devices based on epitaxial graphene, it is desirable to understand and finely control the synthesis of material with the most promising properties. To achieve these prerequisites, many studies are being conducted on various SiC substrates. Here, we review 3C-SiC(100) epilayers grown by chemical vapor deposition on Si(100) substrates for producing graphene by solid state graphitization under ultrahigh-vacuum conditions. Based on various characterization techniques, the structural and electrical properties of epitaxial graphene layer grown on 3C-SiC(100)/Si(100) are discussed. We establish that epitaxial graphene presents properties similar to those obtained using hexagonal SiC substrates, with the advantage of being compatible with current Si-processing technology.

Introduction

al">Graphene is a new class of material that is intensively studied, which points to its remarkable phy<span class="Chemical">sical properties. Today, graphene is presented as a promising material for the observation of novel quantum phenomena and for the development of next-generation electronic and photonic nanodevices.1,2 Recent applications, such as in photodetectors, transparent electrodes, or energy storage, have established the large capacities of this material. A high-mobility graphene field-effect transistor (FET) array was fabricated on a flexible substrate using Al2O3 or h-BN as a gate dielectric in a self-aligned device configuration.3

al">Graphene is an allotrope of <span class="Chemical">carbon that is a single sheet of sp2-bonded carbon atoms densely packed in a honeycomb crystal lattice. Several promising methods have been reported on the synthesis of graphene. Therefore, micromechanical and chemical exfoliation of graphite,4–7 thermal chemical vapor deposition (CVD) growth on transition-metal substrates,8–14 CVD growth on semiconductors,15 plasma-enhanced CVD,16–19 graphite oxide reduction,5,20–22 and thermal decomposition on silicon carbide (SiC) substrates23,24 have been investigated. While exfoliated graphene flakes are considered the purest graphene with the highest recorded mobility, the limited size of the flakes prevents their use in commercial devices. The approach associating the chemical and thermal reduction of graphene oxide appears as promising. However, graphene sheets present structural defects, which may affect the mechanical and electrical properties. In addition, they tend to form irreversible agglomerates that are damageable, as their unique properties are only associated with individual sheets. With regard to CVD procedures based on metal substrates, although good quality and large areas of graphene can be achieved, graphene sheets need to be transferred onto insulating substrates for device applications. Nevertheless, this transfer step, using an invasive chemical treatment, is inevitably damaging to graphene-sheet properties.

The al">epitaxial graphene (EG) layer on single-crystalline <span class="Chemical">SiC substrates is considered the most promising technology for graphene-based electronics. In fact, the graphene synthesis following this procedure shows electronic properties similar to isolated graphene sheets. In addition, for the semiconductor industry, this approach also presents the advantage of obtaining graphene directly on an insulating substrate (SiC), and thus does not require any transfer before processing devices.

Recently, the EG layer was largely synthesized and the intrinal">sic mechanisms of EG were highlighted on hexag onal <span class="Chemical">SiC bulk substrates.23,25,26 Economically, however, the rather-high price of bulk hexagonal SiC substrates and their limited size are an obstacle for large-scale fabrication of graphene devices. To solve this crucial drawback, synthesis of the EG layer on a thin crystalline SiC layer formed on silicone substrates has been studied.27–30 This graphene-on-silicon method enables the formation of a large layer of well-ordered sp2 carbon networks on Si substrates and to fabricate electronic devices.31 Therefore, graphene synthesis on Si(111) can be used to form a channel for digital field-effect transistor operation at the terahertz frequency,27 while graphene on Si(110) and Si(100) can be used for ultrahigh-speed optical communications.32,33 Among different applications, biomedical applications of graphene have attracted increasing interest over the last few years.1 Among studies performed on gra-phene for biomedical applications, a lot of interesting studies have been carried out to explore the use of graphene as an antibacterial material, for biological sensing and imaging, and as a biocompatible scaffold for cell culture.34

Numerous studies on the structural and electrical properties of al">graphene synthesized on a crystalline 3C–<span class="Chemical">SiC(111)/Si(111),35–41 3C–SiC(111)/Si(110),29 and 3C–SiC(111)/Si(100) substrates have been reported.38,42–44 It has been established that the graphitization process on SiC(111)/Si(111) is similar to that on the Si-terminated SiC bulk substrates with the presence of a buffer layer at the graphene–SiC(111) interface. In contrast, graphene formed on SiC(111)/Si(110) is characterized by the absence of a buffer layer.42,44 The fabrication of an EG layer on a 3C–SiC(100)/Si(100) wafer with the most promising properties is very challenging, since it requires the formation of high-quality, large, and homogeneous single and bilayers of graphene. Although attempts to grow EG on this epilayer substrate have been reported,5,38 this silicon-substrate orientation for graphene epitaxy remains under investigation. In this paper, our work dealing with the structural and electronic properties of EG layers grown on 3C–SiC(100)/Si(100) epi layers by solid-state graphitization in an ultrahigh vacuum (UHV) is reviewed.

Experimental details

The substrates consist of 3C–al">SiC epilayers grown on Si(100) substrates by CVD. Deposition occurs within a resistivity-heated hot-wall reactor using <span class="Chemical">silane and propane gases as precursors diluted in hydrogen. After a carbonization step performed at 1,100°C, leading to the formation of an SiC nucleation layer by the reaction of propane with the Si substrate, a 1,500 nm-thick SiC is grown by exposing the surface to propane and silane precursors diluted in hydrogen (∼0.01%–0.02%) at 1,350°C and at a reactor pressure of 200 mbar. The surface of these heteroepitaxial materials is characterized by antiphase domain (APD) boundaries. More details about the 3C–SiC growth method can be found in Portail et al.45

al">Graphene layers are prepared in a UHV by electron-bombardment heating of 3C–<span class="Chemical">SiC(100)/Si(100) substrates. Prior to graphene formation, the pseudosubstrates are first outgassed for several hours at 600°C under UHV conditions and then annealed at 900°C under a silicon flux (2 Å/minute) to remove the surface-contamination layer and the native oxide. Afterward, the temperature is increased between 1,150°C and 1,300°C for a few minutes in order to achieve surface graphitization. During this procedure, the substrate temperature is monitored with an infrared pyrometer, and the base pressure of the system is maintained at around 2.1−10 Torr, with a peak pressure of 10−9 Torr during the graphitization process. This method to form EG on a 3C–SiC(100)/Si(100) epilayer surface is similar to the one used on the 3C–SiC(111) pseudosubstrate.37

al">Graphene samples are further cooled to room temperature and transferred ex situ for characterization. By combining micro-low-energy electron diffraction (μ-LEED), low-energy electron microscopy (LEEM), atomic force microscopy (AFM), Raman spectroscopy, and surface-potential (<span class="Chemical">SP) and X-ray photoelectron spectroscopy (XPS), the structural and electrical properties of the EG layer grown on the 3C–SiC(100) pseudosubstrate are determined.

Structural properties of epitaxial graphene on 3C–SiC(100)/Si(100): results and discussion

The graphitization of the al">SiC surface is achieved by solid-state graphitization. It consists of Si sublimation from the <span class="Chemical">SiC surface, enabling carbon enrichment of the surface.23,46 This graphitization step is monitored by LEED. The LEED pattern is collected to investigate the structural properties and the graphitization process of the sample surface. Initially, during the annealing of the sample under Si flux at 900°C, the LEED pattern is characterized by Si-rich (3×2) and (2×1) surface reconstruction. Transition from the Si- to the C-rich surface due to the sublimation of Si-0surface atoms takes place in the 1,050°C–1,100°C temperature range, and the c(2×2) reconstruction begins to develop during annealing between 1,100°C and 1,150°C. This phase is usually interpreted as a C-terminated structure.47 Further annealing induces the coexistence of the graphene layers and the (2×1) surface reconstruction of 3C–SiC(100). Finally, under UHV conditions, annealing at a minimum temperature of 1,150°C is required to develop the (1×1) graphitic phase on the SiC pseudosubstrate.

Typical diffraction patterns of 2-monolayer (ML)-thick al">graphene layers synthesized by annealing sample at 1,150°C during 10 min are shown in Figure 1A. Two contributions can be clearly distinguished. The first contribution corresponds to the (2×1) cubic surface of the 3C–<span class="Chemical">SiC(100) epilayer (white boxes), while the second is due to the two-domain hexagonal (1×1) graphene layer, confirming the presence of the graphene layer on the 3C–SiC(100)/Si(100) (white and blue arrows). This double contribution can be understood as the result of any of the following two phenomena: 1) a graphene domain oriented with +15° to the [110[ direction of the inphase/anti-phase 3C–SiC(100) domains will appear at ±15° in the LEED images (Figure 1B); and 2) a possible alternative consists in the presence of two hexagonal lattices rotated by ±15° with respect to a single-domain square SiC lattice (Figure 1C).

Figure 1

(A) μ-Low-energy electron diffraction (LEED) images (Ep =46 eV) of epitaxial graphene (1.5 layers) on 3C–SiC(100) taken on different regions. The intense bright large spot is due to backscattered secondary electrons. The magnification shows the two sets of spots present in reciprocal space, corresponding to the misalignment between the graphene layer. (B and C) Crystallographic axes of the graphene layer and cubic substrates determined from the LEED images.

If we analyze finely the second contribution establishing the presence of EG on the surface, we observe that two sets of spots are present in reciprocal space. Other than the rotation by ±15° outlined earlier, an additional ±2° misalignment can be observed, indicated by the double white arrows on the magnification of Figure 1A. This rotation is assigned to the misalignment between the first layer and the second al">graphene layer constituting the analyzed sample.48 Therefore, by considering the double contribution and the misalignment of the <span class="Chemical">graphene contribution evidenced by the LEED image, we can conclude that the graphene layer consists mainly of two hexagonal lattices rotated by ±15° (±2°) with respect to the square 3C–SiC lattice. This would be in agreement with the recent scanning tunneling microscopy observations of Chaika et al49 consisting of two domains rotated by ±13.5° from the (110) axis and two domains rotated by ±13.5° from the (1–10) axis, which generates 24 LEED spots.

The al">graphene layer is further studied by LEEM in order to evaluate thickness spatial distribution.50,51 During the LEEM experiments (LEEM III; Elmitec, Clausthal-Zellerfeld, Germany), the electron gun and sample are biased at 20 kV. The bright-field (BF) and dark-field (DF) LEEM images (Figure 2) have been obtained by selecting the corresponding LEED spots via an aperture (contrast aperture) placed on the o<span class="Chemical">ptical path of the microscope column.

Figure 2

Low-energy electron microscopy images of sample covered by 1.5 layer of graphene.

Notes: (A) Bright-field image (VST [start voltages] =2.69 eV); (B) dark-field image (VST =7.97 eV) for the (01)SiC low-energy electron diffraction (LEED) spot; (C) dark-field image (VST =7.97 eV) for the (10)SiC LEED spot. The field of view is 25 μm.

A typical BF LEEM pattern from an EG layer achieved for electron energy of 2.69 eV is represented in Figure 2A. In different regions, the image-intensity levels change in different manners. The layer exhibits large domains separated by thin black lines. These fine structures can be caused by the thickness in homogeneity (mono/bilayer al">graphene) over the surface, this mono/bilayer distribution being characterized by two hexagonal lattices rotated by ±15° with respect to the square <span class="Chemical">SiC lattice.

Figure 2B and C present DF LEEM images from the two al">SiC LEED spots “(01)<span class="Chemical">SiC” and “(10)SiC” represented on the LEED pattern from Figure 1A. Depending on which LEED spot is considered, the intensity on the LEEM images is reversed. This result indicates that the domains of 3C–SiC(100) on Si(100) have twofold symmetry. Due to the alternating layer-stacking sequence in the [001[ crystallographic direction of 3C–SiC(100), the substrate surface possesses only a twofold symmetry, which would necessarily imply that only one of the domains should be present. The presence of both domains indicates that the 3C–SiC layer presents two distinct epitaxial orientations with respect to the Si(100) substrate: [100[//[100[ or [100[//[010[.41 Comparison of the BF LEEM image with the two DF LEEM images clearly indicates that the black lines between the large domains of graphene layers can be correlated with the domain boundaries of the 3C–SiC(100). We thus establish that the graphene areas are delimited by the domains of the substrate.

At this stage, one may wonder what the factors limiting the size of al">graphene domains are. Before graphitization, the <span class="Chemical">SiC(100)/Si(100) surface is characterized by APD with stepped terraces (Figure 3A).41 After graphitization, within a given APD, the graphene layer is constituted by atomically flat terraces, as evidenced by AFM imaging (Figure 3). This observation establishes that the graphene layers are not limited by the SiC terraces, but rather by their surface symmetry, and the APDs initially characterizing the 3C–SiC(100) surface. This result shows that graphitization achieved by Si sublimation from the SiC surface is strongly affected by the substrate morphology.

Figure 3

3-D-flattened atomic force microscopy images (5×2.5 μm) of the 3C–SiC(100)/Si(100) surface before (A) and after annealing at 1,050°C (B), 1,150°C (C), and 1,200°C (D). (E) Step profile performed on samples S1 (red curve) and S3 (purple curve) indicated by the dashed lines in (B and D).

Synthesis of material with the most promising properties requires the formation of mono- and bilayer al">graphene with controllable characteristics, dimensions, and localization. It is known that graphitization begins from the step edge and progresses towards the terraces. The EG layers are formed during the early stages of thermal desor<span class="Chemical">ption of Si from the SiC surface. However, this does not necessary lead to a uniformly graphitized surface. Important parameters governing the formation of graphene by solid-state graphitization from the SiC surface are the annealing temperature at which the sublimation of Si atoms occurs and the duration of this annealing. In order to control the degree of graphitization, the evolution of surface graphitization was investigated as a function of synthesis temperature for a graphitization duration of 10 minutes. Then, three different samples (S1, S2, and S3) were analyzed, heated at 1,050°C, 1,150°C, and 1,200°C, respectively. The surface morphology at a microscopic scale of the 3C–SiC(100) surface was studied by AFM. The AFM imaging technique, performed in tapping mode with an Al-coated tip characterized by a tip radius of 7 nm, is an efficient technique in ambient conditions that enables rapid and convenient characterization of the surface. Figure 3 presents a series of 3-D topographic images of the semi-insulating 3C–SiC(100)/Si(100) surface before and after graphitization.

The surface of as-grown 3C–al">SiC heteroepitaxial substrates without any heat treatment is characterized by APD boundaries presenting stepped surfaces (Figure 3A). The sample heated at 1,050°C (S1) still shows domains, always delimited by APD. These domains present well-defined steps through energy minimization, resulting in a reconstructed surface (Figure 3B). This surface feature shows that sublimation of <span class="Disease">Si atoms leading to graphene formation does not occur in these growth conditions at 1,050°C. By contrast, smoother domains, with the surface appearing flatter as if covered with a continuous layer, characterized samples heated at higher temperatures, 1,150°C for sample S2 (Figure 3C) and 1,200°C for sample S3 (Figure 3D).

Interestingly, the continuity of the al">graphene layer observed here is in accord with results obtained for the growth of <span class="Chemical">graphene layers on 6H–SiC vicinal substrates, where a carpet-like growth mode covering several substrate terraces and steps is observed by scanning tunneling microscopy and AFM images.26,52 This behavior is confirmed by comparing the step profile of samples S1 and S3 (Figure 3E). This specific variation of surface morphology with temperature is the clear signature of graphene formation.35 The results displayed previously reveal that the sublimation of few atomic layers of Si out of the SiC epilayer starts by annealing of the surface at 1,150°C or higher. In these used conditions and depending on the temperature and duration of graphitization, the surface is covered by single or multiple layers of graphene. Graphitization of the 3C–SiC(100) surface leads to a remarkable modification of the surface morphology on a microscopic scale. These results indicate the possibility of controlling the growth of large-scale graphene on 3C–SiC(100) beyond the SiC terraces. However, before developing devices based on few-layers EG, the thickness as well as the uniformity of the graphene coverage must be controlled.

Raman spectroscopy is the technique of choice to access al">graphene-layer thickness and uniformity. Micro-Raman spectroscopy was performed at room temperature with a Jobin Yvon T64000 (Horiba, Kyoto, Japan) spectrometer in a backscattering confocal configuration. The excitation source was an Ar+ laser at a wavelength of 488 nm, and laser power was controlled at 5 mW on the sample surface. We used a 100× objective lens for focusing the laser beam on the surface and collecting the scattered light at room-temperature measurements from different local spots forming a pixel pattern. The spatial resolution of the image was 1 μm, while the spectral resolution was less than 0.35 cm−1.

Figure 4A represents a typical Raman spectrum obtained in the frequency range (1,000–3,000 cm−1). We observed the most prominent Raman features: D, G, 2D, and G’, character-izing al">graphene (see Table 1). The G band corresponds to the doubly degenerate (transverse o<span class="Chemical">ptical and longitudinal optical) phonon mode associated with E2 symmetry at the Brillouin zone center (the Γ-point), which gives evidence of carbon sp2 reorganization.53,54 Its peculiar position value (1,594 cm−1) indicates the existence of a strong level of doping and a compressive stress of the layer.55 The D band, at 1,372 cm−1, involves phonons near the K-points normally absent in defectless graphene.56 Its presence is the signature of structural disorder, including corrugation and distortions attributed to twisting, and step edges, such as finite-size domains, atomic-scale defects, and armchair-type edge defects.57–59 The G’ band corresponds to an intervalley process involving one transverse optical and one longitudinal optical phonon at the K-point.60,61 Finally, the 2D band is assigned to the second order of the D band, where it is associated with two phonons close to the K-point in the Brillouin zone.62 The (D + G) peak corresponds to 2LO (K) activated by disorder due to the presence of defects in the graphene layers.

Figure 4

Raman spectra of graphene layers grown on 3C–SiC(100) on Si(100) substrate.

Note: (A) unpolarized, (B) in p-polarization, (C) in s-polarization.

Table 1

Raman shift of the graphene modes grown on 3C–SiC(100)/Si(100) substrate

Raman modes and ratio	D (cm−1)	G (cm−1)	G′ (cm−1)	2D (cm−1)	D + G (cm−1)	IG/I2D	ID/IG unpolarized	ID/IG (s)	ID/IG (p)
Raman shift and ratio	1,372	1,593	2,467	2,733	2,948	1.2	0.75	1.02	0.53

Intensity mapping is performed over a surface of (300×300 μm2) in regular locations along x and y directions across the sample surface to ensure that all changes in al">graphene modes will be detected. The local mapping intensity of the G and 2D bands showed great similarity in <span class="Chemical">graphene distribution across the sample surface (see Figure 5A and B).

Figure 5

Mapping Raman intensity of G band (A), 2D band (B), and the ratio of the G band on the 2D band, ie, IG/I2D ratio (C) of graphene grown on 3C–SiC(100)/Si(100) substrate.

We observed black flakes nonuniformly distributed across the sample surface surrounded by a more brightly colored area corresponding to maximum al">graphene-layer numbers. The layer number was identified by investigating the intensity ratio of the G and 2D bands (IG/I2D). The false color cartography in Figure 5C represents the intensity ratios of the G and 2D bands. Seventy percent of the surface exhibits an intensity ratio IG/I2D centered on 1.2, which indicates multilayer-<span class="Chemical">graphene covering of the sample surface (3 layer graphene).63–66

From the intensity ratio of the D and G bands, we evaluated the domain size (La) of the al">graphene layer given by the following expression:67,68 where ID(IG) is the intensity of the D (G) Raman mode and E the excitation-laser energy (E =2.54 eV). La of EG layers grown on 3C–<span class="Chemical">SiC (100) epilayer is about 17.8 nm. The smaller domain size, normally absent for highly smooth EG, could be assigned to the pseudosubstrate-morphology effect. In fact, a possibility of Bernal stacking of EG similar to exfoliated graphene has been reported.69 This possible effect can be investigated by performing polarized Raman spectroscopy measurement.

The polarization behavior of al">graphene Raman modes is presented in Figure 4B and C. Backscattering geometry , where the incident laser-beam polarization is fixed in the x direction and the polarization of the scattered light by the sample is selected in parallel (p-polarization) or perpendicular (s-polarization) configurations, was used. Within this approach, <span class="Chemical">graphene surface-roughness dependence is investigated by analyzing the (ID/IG) intensity-ratio variation in both configurations. The intensity ratio (ID/IG) reaches high intensity in s-polarization (cf Table 1), reflecting the Raman graphene modes, since the relative intensities of each showed a particular behavior. The G band belongs to the D3h space group in the (Γ) point for three-layer graphene in AB stacking. The associated irreducible representation is given by the following expression:60 the G band being assigned to a combination of E′ and E″ representations, and also the representation is Ramanactive with a lower-frequency component and its mode has no polarization dependence (Figure 4), which makes it permitted in p- and s-configurations.55,70 With regard to the D peak (,) which is assigned to the breathing mode involving phonons near the K-zone boundary, its intensity is not affected by the layer number, whereas its mode shows a high dependence on the amount of disorder.71 This disorder can result from the graphene layer and/or from the angle between the incident polarizations and the graphene-flake edge.

Whatever the edge orientation and disorder, D intensity reaches the maximum for polarization parallel to the edge and the minimum when polarization is perpendicular to the edge, without reaching zero for light polarized perpendicular to the edge.71 The maximum intensity of the D band for parallel configuration proves that the al">graphene flake is parallel to the average edge direction, and the decrease of the D-band intensity in s-polarization indicates that the degree of disorder is not homogeneous in the two directions of the <span class="Chemical">graphene layer and the grains are dissymmetric.71 This defect variation could be related to the roughness of the substrate. However, investigation of the change of the growth axes of 3C–SiC(100) on the off-axis Si(100) substrate that could originate from the thermal effect occurring during the graphitization process has demonstrated that non-Bernal stacking can be related to pseudosubstrate morphology.72 By consequence, a domain size of the order of 17.8 nm cannot be assigned to the surface morphology of the substrate, but to the intrinsic formation of the hexagonal phase on the cubic folder symmetric surface, due to the presence of different interdistances between defects in the EG layer.59

Based on the structural properties discussed in this review, it is interesting to underline that the growth of al">graphene, and specifically of the first <span class="Chemical">graphene layer on 3C–SiC(100)/Si(100) substrate, is similar to that reported for the carbon-terminated 6H–SiC substrates. Surface coverage with single or multiple layers of graphene can be adjusted by modulating the graphitization parameters (temperature and duration), and the carbon enrichment of the surface is characterized by a specific surface structure. The specificity of EG layers grown on 3C–SiC(100)/Si(100) resides in the fact that graphene layers do not present a rotational disorder, in contrast to previous observations of few-layers graphene on SiC(000-1) substrate, which exhibits a high degree of disorder.38

Electrical properties of epitaxial graphene on 3C–SiC(100)/Si(100): results and discussion

One of the major challenges for EG-grown films is to understand the influence of graphitization conditions not only on the film structure but also on the electrical properties. The substrate, in addition to influencing the growth mechanism during synthesis, can significantly influence the electrical properties. In fact, according to the substrate nature, the degree of graphitization may be different for the same synthesis conditions. However, EG coverage affects the electrical properties of al">graphene.

In order to establish the al">graphene thickness–electrical properties relation, <span class="Chemical">SP measurements can be performed. The SP mode, also called the Kelvin mode, is an AFM technique offering a powerful tool for measuring the distribution of the electric potential of a surface with nanometer resolution. This technique creates the SP image by detecting the interactions between the conductive tip and the sample through long-range Coulomb forces. SP mapping of the surface was achieved with a Cr/Pt-coated tip characterized by a tip radius of 10 nm and with a lift-scan height of 20 nm (lift mode). Figure 6A–C present SP images performed in the same conditions on three different surfaces: the c(2×2)-3C–SiC(100) reconstructed, ie, sample annealed but not covered by the graphene layer, and two different stages of the EG-layer formation, corresponding to the sample covered by 1.5 ML and 2.8 ML of graphene. The SP profiles corresponding to the maps through the x-axis are shown in Figure 6D. The average SP extracted from these profiles as a function of the EG coverage are plotted on Figure 6E, with the error bar corresponding to the peak-to-valley value.

Figure 6

Surface-potential (SP) mapping of (A) c(2×2)-3C–SiC(100) reconstructed, ie, sample annealed, but not covered by graphene layer, (B) sample covered by 1.5 monolayer (ML) of the epitaxial graphene (EG), and (C) sample covered by 2.8 ML of the epitaxial graphene. (D) SP profiles through the x-axis corresponding to the SP mapping of (A–C). (E) Variation of the average SP, extracted from SP profiles as a function of the graphene coverage.

The al">SP, significant for each material, is related to the work function required to move an electron from the material to the vacuum. It can be defined as: where Wtip and Wsample are the work functions of the tip and the <span class="Chemical">graphene-sample surface, respectively, and e is the elementary charge.73,74 While Wtip is unknown, because the same tip with the same imaging conditions have been used to analyze all samples, Wtip remains constant and the different samples may be compared. In these conditions, the variation of the SP reflects the variation of Wsample.

The measured al">SP, plotted in Figure 6, decreases as the EG thickness increases, which demonstrates a corresponding increase of Wsample (according to the definition of <span class="Chemical">SP given above). This increase of Wsample is in good agreement with results published in the literature.75–77 In our graphitization conditions, the increase in graphene coverage from 1.5 ML to 2.8 ML corresponds to a decrease of SP of around 161±15 mV, and thus to an increase of Wsample by ∼160 meV. With layer thickness taken into account, this result is in good agreement with those presented by Filleter et al,75 who determined that bilayer graphene grown epitaxially on 6H–SiC(0001) increases the work function in the order of 135±9 meV compared to single-layer films.

The thickness of the EG layer strongly influences the electrical properties. However, the nature of the substrate can also influence EG properties; in particular, it can potentially play a crucial role on the formation of a buffer layer. The buffer layer is a graphitic layer covalently bound to the substrate. At this stage, an important question appears: is there an interface layer between the EG layer and the 3C–al">SiC(100)/Si(100) epilayer substrate? To answer this question, <span class="Chemical">graphene samples were analyzed by XPS. XPS experiments were performed in UHV conditions on the Tempo beamline at the SOLEIL Synchrotron facility (France). The analyzing chamber was equipped with a Scienta 2000 electron hemispherical analyzer with a delay-line 2D detector (Scienta Scientific, Uppsala, Sweden), which optimized the detection linearity and signal:background ratio. Figure 7 displays C 1s spectra achieved at two different photon energies – 340 eV and 510 eV. A sharp C 1 s peak (blue curve), located at 284.5 eV in binding energy, indicates the presence of sp2 hybridized C–C bonds. The C 1s spectra showed another component at 283.2 eV in binding energy. These components corresponded to the SiC bulk (green curve).36,77,78

Figure 7

C 1s X-ray photoelectron spectroscopy spectra for epitaxial graphene on 3C–SiC(100)/Si(100) at two different photon energies.

Notes: Blue curve for graphene [C–C], and green curve for Si-C. Red dots are experimental data points. The black curve is the envelope of the fitted components.

It is interesting to note that for a al">graphene coverage of only 1.5 ML, the bulk <span class="Chemical">SiC signal disappears in the C 1s spectrum collected at a photon energy of 340 eV. This graphene-related component (G) – located at 284.5 eV, ie, shifted by 1.6 eV toward higher binding energy with respect to the bulk SiC – indicates sp2 hybridized C–C bonds, which are a signature of the graphene layers. At the C 1s core level, the average graphene-film thicknesses can be determined by measuring the attenuation of the SiC contribution (282.8 eV) with respect to the graphene signal (284.5 eV).79 When increasing the incident energy, the line shape and position of the graphene peak remain unaffected. The signature of the bulk SiC substrate appears. By contrast, any third peak, corresponding to the buffer layer present at the graphene–substrate interface, is observed. These results demonstrate the absence of an interfacial graphitic layer covalently bound to 3C–SiC(100).35,44,80,81 We can conclude that the first graphitic layer grows on 3C–SiC(100) without an interface layer (buffer layer). This result is a major departure with respect to graphene on hexagonal SiC(0001) and graphene on cubic SiC(111),15,25,36,56 where the first graphene layer, while easy to grow uniformly, is nonconductive. Finally, in the knowledge that the intermediate C-rich layer (buffer layer) degrades the intrinsic mobility of EG layer,23 its absence is an important property that opens new perspectives for graphene-based electronic devices.

Conclusion

The properties of EG layers discussed in this review evidence the interest in the 3C–al">SiC(100)/Si(100) pseudosubstrate as a base for <span class="Chemical">graphene synthesis. In fact, the possible control of graphene coverage by controlling the graphitization parameters (temperature and duration) and the absence of an interfacial layer covalently bound to 3C–SiC(100) are important properties that constitute fundamental steps toward the development of Si-based technologies for the mass production of graphene. Although there are many challenges that need to be overcome, such as improvement of the surface morphology of the SiC(100) epilayer characterized by antiphase domain boundaries, which limits the graphene-domain size, this approach will likely open new perspectives for industrial-scale fabrication and mass production. In fact, the EG on SiC/Si(100) substrates presents the advantage of being able to be produced on substrates significantly larger than commercial hexagonal SiC substrates, and of being compatible with current Si-processing techniques, thus offering the opportunity to directly integrate graphene into standard low-cost Si technology.44 These encouraging and promising results clearly demonstrate that graphene can be an ideal material for future silicon electronic devices. The ability to grow a high-quality graphene layer on silicon substrate will open up new opportunities for semiconductor nanoelectronic technology, making the devices faster, cheaper, and more reliable.