Ultra-thin NaCl films as protective layers for graphene

Abstract

The ageing of graphene is an important issue for technological applications. Capping layers are a good way of circumventing this problem. In this work, we propose the use of ultra-thin NaCl films as easily-removable protective layers. We have carried out a detailed characterization of the NaCl/graphene interface on metal substrates, namely Cu(111) and Ir(111), by means of complementary microscopies, electron diffraction and spectroscopic techniques. Interestingly, we show that NaCl neither interacts in a chemical way with graphene nor intercalates through it. We demonstrate that the NaCl film is stable under ambient conditions, protecting the graphene surface from oxidation. In addition, after removing the protective layer, graphene remains intact.

1. Introduction

Graphene (Gr) is one of the most studied materials in the scientific community [1]. The wide range of potential applications [2], due to its exceptional electrical, mechanical and optical properties, identifies graphene as one of the essential “future materials” in technology. Despite the fact that Gr is rather inert in atmospheric conditions, as a result of its sp² honeycomb structure, it has already been proven that the progressive accumulation of oxygen and epoxy species over time may lead to a progressive degradation of its outstanding properties [3]. Therefore, encapsulation is used to protect graphene from the environment and to avoid degradation of the carrier mobility [4][5]. So far, different materials (h-BN [4], PMMA [6], Perylene [7]….) have been used for this purpose.

There are two main requirements that protective layers should fulfil; on the one hand they have to protect the surface avoiding any chemical interactions with the environment (water, oxygen…), and on the other hand, the interaction with the substrate must be negligible. Therefore, a thorough study of the interface between graphene and the proposed protective layer is mandatory in order to discern whether a layer is suitable or not to be used as a capping. In this respect, an interesting material is NaCl since crystalline NaCl layers can be easily grown stoichiometrically and atomically flat on metal or semiconductor surfaces. Several works describe the NaCl thin films structure on top of these surfaces [8][9][10], pointing out the low interaction between the NaCl layers and the substrate. Moreover, a recent study has even addressed the growth of 2D NaCl crystals on graphene flakes [11].

We show hereafter that NaCl is an ideal capping layer as it provides both protection and low interaction. We have performed a thorough structural characterization of the NaCl/Gr interface by means of Scanning Tunnel Microscopy (STM), Atomic Force Microscopy (AFM), Low Energy Electron Microscopy (LEEM), and Low Energy Electron Diffraction (LEED). This multi-technique study shows that NaCl films adapt to the graphene surface without disrupting its morphology. Moreover, a spectroscopy study using X-Ray Photoemission Spectroscopy (XPS) indicates that there is no chemical interaction between the salt and graphene while Low-Energy Ion Scattering (LEIS) demonstrates that NaCl films do not intercalate through graphene at room temperature (RT). Importantly, we demonstrate that after several weeks in ambient conditions the capping layer can be easily removed by a soft annealing, recovering the pristine graphene.

2. Results and discussion

Figure 1a shows an STM image of a Gr/Cu(111) sample where a moiré pattern of 6.5 Å periodicity is seen. This kind of superstructures is very common in the graphene-metal systems being a direct consequence of the mismatch of both lattices [12]. In the inset, an overview of the sample shows the copper surface completely covered by a monolayer (ML) of graphene. The LEED pattern (Figure 1b) shows both the hexagonal spots of the Cu(111) surface (yellow arrow), and a modulated ring (red arrow), associated with a multi-domain graphene. Similar multi-domain structures have been found in other single crystals such as Pt(111) [12][13]. After NaCl is evaporated in-situ in UHV at RT on this surface, new features appear which can be seen in the LEED pattern of Figure 1c. Two inner rings arise between the (0,0) spot and the Cu hexagon spots. For better viewing of these features, the LEED pattern was taken off centre. The inner rings correspond to a thin film of NaCl grown on top of Gr with different rotational domains. Thus, instead of the usual diffraction pattern with 12 well-defined spots, corresponding to the three square domains rotated 60° appearing when NaCl is adsorbed on bare Cu(111) [14], a multi-domain square structure is found. It has been previously reported that the domain structure of NaCl films is directly influenced by the temperature of the substrate during deposition. Actually, the temperature might depend on the nature of the substrate [8][9][14]. The same NaCl multi-domain structure was found for NaCl on graphene for all deposition temperatures between RT and 523 K, providing further evidence of the low interaction between both layers.

Figure 1.

a) STM image of 1ML of Gr/Cu(111) showing a moiré pattern (0.375 nA, 827.4 mV, (5.9×5.9) nm²). Inset: Zoom out of the STM image (100.2×100.2) nm² (0.099 nA, 826.5 mV). b) LEED pattern of a ML of Gr/Cu(111) at 135 eV. The modulated ring corresponds to Gr while the spots are stemming from the Cu(111). c) LEED pattern of NaCl/Gr/Cu(111) at 59 eV. The two inner rings belong to the NaCl film, while the outer one to the Gr monolayer and the spots to the Cu(111). For a better visualization, the (0,0) Cu spot is off center. d) AFM image (700.7×349.3) nm² of a Gr/Cu(111) sample partially covered with NaCl. On the right-hand side a wrinkle (typical feature of Gr on metals) is visible. The inset shows the height profile along the blue line.

After the in-situ characterization, the sample was taken out from UHV and characterized ex-situ with AFM. An AFM image of submonolayer NaCl coverage is shown in Figure 1d. A wrinkle, visible at the right of the image, is a characteristic signature of a ML Gr [12] which usually appears during the growth process in metal surfaces. The bright finger-like islands correspond to the NaCl film. In this case, the film growth was stopped in the submonolayer range in order to determine the thickness of the NaCl film. A profile in the image of Figure 1d (inset) shows the height of the NaCl film with respect to the graphene of around 2.5 Å. This height is assigned to the thickness of a NaCl bilayer 2.9 Å [15] within the typical error of the AFM.

It has been demonstrated that moirés in graphene can modify the adsorption properties [16]. However, NaCl films are not affected by the graphene corrugation due to the low interaction between the two surfaces. In order to prove it, we have grown Gr on a different metal substrate, Ir(111), with different moiré patterns. Figure 2a shows a STM image of one of the characteristic moiré patterns of 1ML Gr/Ir(111). In the inset, the STM image shows the complete coverage of the Ir surface. The microspot LEED pattern (see Figure 2b) shows the Ir (111) and the Gr spots as well as the typical pattern from the moiré of Gr/Ir(111), consisting on hexagons around the Ir integer spots. When NaCl is deposited at RT, two inner rings appear in the LEED, Figure 2c. The structure is thus the same as previously found for the Cu(111) surface, evidencing the weak NaCl-Gr interaction.

Figure 2.

a) STM image of 1ML of Gr/Ir(111) showing a typical moiré (13×13) nm², (1.4 nA, 300 mV). Inset: Zoom out of the STM image. b) Microspot LEED of 1ML of Gr/Ir(111) taken at 58 eV. c) Microspot LEED of NaCl/Gr/Ir(111) taken at 58 eV. The two inner rings belong to NaCl. d) and e) LEEM images, FOV 20 μm and electron energy 18 eV, of NaCl/Gr/Ir(111) acquired just after salt evaporation (d) and few seconds after (e). Electrons damage the NaCl film, as a result a change in the contrast is visible on e).

Figures 2d and 2e show two LEEM images with a field of view (FOV) of 20 μm, recorded in-situ right after NaCl deposition and a few seconds later, respectively. In Figure 2d the contrast is quite homogeneous because the surface, at this length scale, is fully covered by an ultra-thin film of salt. These LEEM images, together with ex-situ AFM images (not shown); show a continuous layer of NaCl without pits or cracks and low surface roughness, supporting its use as a protective layer. Nonetheless, after a few seconds under the electron beam irradiation the morphology of the surface changes. This damage can be readily seen as a change in the contrast of the LEEM image (Figure 2e) where about 50% of the salt has already been desorbed. At this point it is worth to mention that NaCl may experience photo-dissociation on Gr, leading to Na intercalation, as reported recently [17].

A further important condition that capping layers have to accomplish, is not to intercalate underneath the surface they are protecting. At this respect, is still unclear whether NaCl intercalates or adsorbs in graphene at ambient conditions. To shed some light on this open question we have characterized the samples with LEIS, a technique suited for this purpose due to its exceptional surface sensitivity as most of the signal comes from the outermost atomic layer of the system under study [18]. LEIS identifies elements by their masses, therefore the assignation of the main peaks based on their backscattering energies can be obtained following a simple equation derived from classical physics [18]. Figure 3a shows the LEIS spectra for: a clean Cu(111) sample (ochre spectrum), a Gr/Cu sample (grey spectrum) and a NaCl/Gr/Cu sample (blue spectrum). For the clean Cu(111) sample, only the copper peak is visible, which manifests the absence of contaminants. After the growth of a Gr ML, there are no peaks at all. The copper peak was not expected since the sample is completely covered by graphene and LEIS is only sensitive to the last atomic layer. One might expect the presence of carbon emission which should be visible as any other element with this technique. However, for graphitic carbon or graphene an unusual strong matrix effect has been observed [19], which explains the absence of its signal in the spectrum. This phenomenon is related with the electronic structure of graphene that causes a strong neutralization of the backscattered helium ions [19]. Hence, no peaks are observed. Finally, the LEIS spectrum of the NaCl/Gr/Cu sample displays two peaks that can be assigned to Na and Cl. Figure 3b compares the LEIS spectra of a NaCl/Cu sample (upper green spectrum) with a NaCl/G/Cu sample (lower blue spectrum). Both NaCl evaporations have been performed with the same experimental conditions (evaporation time and substrate temperature). Nonetheless for the metal substrate the Na signal intensity is lower compared to the graphene one. Besides, a Cu peak appears in the upper spectrum, indicating that the NaCl has not completely covered the whole metal surface. On the contrary and as also demonstrated by the LEEM image shown on Figure 2d, a continuous NaCl film covers the whole graphene surface, proving that the wetting of the graphene takes place faster than that of the metal substrates. This set of consecutive experiments, where step by step the different layers have been grown to build up the whole system, evinces that NaCl remains in the surface and does not intercalate through graphene at ambient conditions. Although in the LEIS technique the intensity of a peak cannot be easily quantified, the fact that the Na signal is much higher than that of Cl suggest a higher neutralization probability of the He ions impeaching on negative Cl atoms. Other structural reasons, as buckling or surface relaxation in the salt layer cannot be excluded.

Figure 3.

a) LEIS spectra taken at 0.5 KeV, 10 mA, 2×10^-7 mbar of He of clean Cu(111) (ochre), Gr/Cu(111) (grey), NaCl/Gr/Cu (blue) samples, respectively. b) LEIS spectra taken at 0.5 KeV, 10 mA, 2×10^-7 mbar of He of NaCl/Cu (green) and NaCl/Gr/Cu (blue) samples. c) Upper, medium and lower panels: Na1s, Cl2p and C1s core levels, respectively, of a NaCl/Gr/Ir(111) sample after several days in ambient conditions, before and after annealing at 823 K in UHV.

Finally, we have evaluated by XPS the chemical interaction at the NaCl/Gr interface and the stability of the ultra-thin NaCl film under ambient conditions as well as its protective capacity in a non-vacuum environment. Figure 3c (upper, middle and lower panels) shows the Na1s, Cl2p and C1s core levels, respectively, of a sample after several days in air (closed room, at RT and 30% of humidity), before and after an annealing treatment in UHV at about 823 K. As it can be seen, prior to the annealing, NaCl remains on the surface, showing its stability under ambient conditions. The measured binding energies of Na1s and Cl2p peaks are 1073.3 and 200.1 eV, respectively. Both peaks are shifted 2 eV towards higher binding energies as compared to a freshly cleaved NaCl single crystal [20]. This shift can be ascribed to –O and –OH groups adsorbed on top of the salt surface, meaning that the NaCl film behaves as a sponge capturing oxygen related contaminants and therefore avoiding the ageing of graphene by these very same groups. The lower panel of Figure 3c shows the C1s core level of the same sample after several days in air. Before the annealing treatment, the peak slightly widens with a tail at higher binding energies associated with the adsorbed species on top of graphene (mainly NaCl). The main component at 284.1 eV can be assigned to the C-sp² [21][22]. To remove the protective layer an annealing treatment in UHV at about 823 K was carried out, until Na1s and Cl2p signals completely disappear (see XPS spectra Figure 3c upper and middle panels). After the annealing, the high binding energy tail of the C1s peak completely vanished, together with the Na and Cl signals, and only the sp² graphene component remains (Figure 3d, lower panel). Even though, before annealing the C1s component has a tail that suggest some sort of interaction in between both interfaces, we would like to stress that after a mild annealing the tail completely disappears and the sp² component remains unaltered. This slight modification of the C1s lineshape and the recovery after annealing indicates that the chemical interaction of the salt and the graphene is very weak. Moreover, in a recent work we were able to characterize the electronic band of the NaCl/Gr system, demonstrating that there is no modification of the graphene band structure at the dirac-cone bands [17]. Although from the structural point of view the lattice parameter of the NaCl compresses about 10% in the very first layers (see LEED patterns of Figures 1c and 2c, and refs [15][23]), this modification remains on the NaCl film without affecting the graphene, as it has been reported even for more interacting systems [24].

In summary, it can be confirmed that NaCl does not chemically interact with Gr. In addition, upon annealing the sample, which removes the protective layer, the graphene surface displays the same structural perfection of an untreated graphene without showing any structural or chemical damage.

3. Conclusions

We have carried out a thorough characterization of the NaCl/Gr interface, proving that ultra-thin NaCl films fully cover the graphene surface protecting it towards oxidation while presenting negligible interaction with the underlying graphene. Structural characterization performed by the combination of STM, LEEM and AFM, shows that NaCl grows as a bilayer which completely covers the graphene surface. LEED shows that NaCl films present a multi-domain structure due to the low interaction between the Gr and NaCl surfaces whereas photoemission results corroborate the stability of the NaCl film after several days in ambient conditions as well as its negligible chemical interaction with the graphene. Moreover, XPS also demonstrates that after removing the ultra-thin NaCl film, graphene remains undamaged. Finally, LEIS studies confirm that NaCl does not intercalate through graphene under ambient conditions. Taking all these results together, we can conclude that NaCl ultra-thin films can be considered as a technologically relevant option to be used as protective layers to prevent the ageing of graphene.

4. Experimental section

The growth and in-situ characterization of the samples have been carried out in ultra-high vacuum (UHV) chambers, with a base pressure of 1×10^-10 mbar. Ir(111) surfaces were cleaned by repeated cycles of argon ion sputtering and annealing in an oxygen atmosphere (T = 1373 K and P_oxygen= 2×10^-8 mbar). In order to avoid any residual oxygen on the surface, the last cleaning cycle was carried out without oxygen. Cu(111) samples were cleaned with cycles of sputtering and annealing (873 K). Graphene on Cu(111) was grown by thermal decomposition of C₆₀ (Sigma Aldrich, 98% purity). These fullerenes were evaporated directly in the UHV system from a home-made evaporator based on a tantalum crucible and a type-K thermocouple spot-welded to it. The samples were kept at 1173K during the evaporation [25]. Graphene on Ir(111) was grown by decomposition of ethene following the recipe described elsewhere [26]. NaCl (99.999% Sigma Aldrich) was sublimated from a home-made Ta crucible annealed at 803 K controlled by a type-K thermocouple spot-welded to it, keeping the sample at RT. Structural in-situ characterization has been carried out with an Omicron RT-STM and an Omicron LEED. WSxM software was used for data acquisition and analysis [27]. XPS was used to characterize the chemical composition of the samples. XPS spectra were acquired in an ultrahigh vacuum (UHV) chamber with a base pressure of 1×10^-9 mbar using a hemispherical electron energy analyser (SPECS Phoibos 150 spectrometer) and a monochromatic AlK_α X-ray source (1489.74 eV). XPS spectra were recorded at the normal emission take-off angle, using an energy step of 0.05 eV and a pass-energy of 10 eV for high resolution data, which provides an overall instrumental peak broadening of 0.4 eV. LEIS was measured using a hemispherical analyser SPECS Phoibos 100 MCD-5; using the analyser operating at CRR mode. He⁺ 0.5 keV have been used as incoming ions. LEEM and microspot LEED were measured at the LEEM-PEEM experimental station of the CIRCE beamline of the ALBA Synchrotron [28]. AFM characterization was done with an Agilent 5500 PicoPlus equipment operating in dynamic mode under ambient conditions. Silicon cantilevers (from AppNano) with a nominal radius smaller than 10 nm were employed. The images were formed by 1024×1024 pixels.