Non-destructive depth profile reconstruction of single-layer graphene using angle-resolved X-ray photoelectron spectroscopy

Highlights

• Single-layer graphene grown on polycrystalline copper foil was identified by Raman spectroscopy

• Dominant C sp² bonding in graphene was verified by C 1s and X-ray induced C KVV Auger spectra

• The depth profile reconstructions of C, O and Cu concentrations and their bonding states were calculated by the Maximum Entropy Method

• The copper substrate was fully covered by the graphene

• The thickness of air-exposed graphene overlayer was reduced by cleaning from ~0.6 to ~0.3 nm.

Abstract

We analyze air-exposed and cleaned graphene samples grown by the chemical vapor deposition method on polycrystalline copper. Raman spectra verify their single-layer nature. C 1s and C KVV Auger spectra confirm the dominant C sp2 coordination in the films. We use angular-resolved C 1s, O 1s, and Cu 3p photoelectron spectra to acquire non-destructive concentration depth profiles and for in-depth distribution of resolved bonding states by the Maximum Entropy Method. The elemental distributions show that the air-exposed surfaces of the samples are enriched by carbon- and oxygen- bearing species, resulting in an overlayer 0.6 nm in thickness. The in-depth distributions of the resolved bonding states reveal that the oxygen bonded to carbon is located at the top surface and the oxygen bonded to copper is located at the graphene/copper interface. Almost no oxygen is present at the surface of the samples cleaned by annealing. The percentage of carbon falls by ~40%. The thickness of the carbon overlayer falls to about 0.3 nm, and the graphene layer completely covers the substrate. We emphasize that the results for the in-depth distribution of element concentrations and for resolved chemical bonding states are obtained nondestructively, i.e. without any modifications to surface composition and bonding.

Introduction

Graphene, a monolayer of sp² hybridized carbon atoms, was first isolated from a graphite crystal in 2004 [1]. Since then, graphene has been prepared by a number of methods [[2], [3], [4]] and has been studied for its attractive physical and chemical properties [1,[5], [6], [7]], leading to numerous applications [1,2,5,8].

Identification of graphene layers (i.e. distinguishing between single-layer graphene, few-layer graphene, and/or bulk graphite) is the key problem in the analysis of two-dimensional carbon films. There are several methods that can be used for characterizing them. The main analytical techniques are: Raman spectroscopy (RS) [[9], [10], [11]], scanning tunneling microscopy/spectroscopy [12], optical microscopy [11,12], transmission electron microscopy [13,14], low energy electron microscopy [15], and also surface-sensitive electron spectroscopy, e.g. photoelectron spectroscopy (XPS, UPS) [9,12,16], X-ray induced and electron induced Auger electron spectroscopy (XAES, AES) [11,17] and electron energy loss spectroscopy (EELS) [10,18]. RS is currently used as the standard fingerprint method for the identifying graphene layers. Core-level photoelectron spectroscopy seems to be less eligible, because the binding energy of the C 1s peak recorded from air-exposed graphene surfaces can be influenced by surface adsorbates and by the underlying substrate [19]. However, core-level peak areas can be used for estimating the thickness of an overlayer [20]. While XPS is currently used for evaluating surface composition and bonding, angle-resolved XPS (ARXPS) spectra are seldom recorded to estimate the thickness of carbon overlayer [20] or to reveal (qualitatively) the in-depth homogeneity in graphene samples [21,22].

The EELS spectra of single-layer graphene differ from those of graphite [23], as has been shown for a suspended graphene layer [18]. When the layer is grown on a substrate, the EELS spectra consist of a spectral signal both from the graphene and from the underlying substrate. Interpreting the spectral intensity is a complicated task. This is, of course, not valid for the transmission mode in an electron microscope equipped with the EELS technique, where a suspended graphene foil can be analyzed.

C KVV Auger spectra, as a self-fold of the valence band density of the states, are currently used for estimating sp² and sp³ hybridizations of carbon atoms in carbon-based materials [24]. The width of the C KVV spectrum reflects the density of the electronic states in the occupied valence bands, and contains contributions of sp² and sp³ hybridizations of carbon atoms. The C sp²/sp³ ratio can easily be evaluated from parameter Δ [25,26], which is defined as the energy difference between the maximum and the minimum of the first-derivative C KVV spectrum. The dependence of parameter Δ on the C sp² or C sp³ percentage is assumed to be linear between the Δ values of graphite (Δ = 23.1 eV) and diamond (Δ = 13.2 eV) [26]. The method is independent of surface charging and independent of the position of the Fermi level.

Generally, air-exposed graphene surfaces are covered by carbon- and oxygen- containing adsorbates that can overlap the inherent properties of clean single-layer graphene [9]. In addition, the oxygen can easily be intercalated beneath the graphene layer, due to the polycrystalline nature of graphene. Oxygen atoms were observed in the OC and OCu bonding states [27,28], the latter forming Cu₂O, the low-oxidation state of metallic Cu [29]. Lu et al. [30] have shown that the Cu₂O layer is mechanically and chemically weak, graphene films can be detached and transferred to arbitrary surfaces, and the Cu substrates can be re-used for graphene growth. Barinov et al. [27] have shown that epoxy groups are formed on perfect graphene, whereas oxygen atoms are present in ether and carbonyl configurations on the vacancies of defective graphene surfaces.

Several attempts to clean graphene surfaces have been implemented in the past, using sputter-cleaning [10], annealing under UHV conditions [9,12,13,15,16], and electron irradiation [11]. The most successful method seems to be annealing under ultra-high vacuum (UHV) conditions, resulting in a clean graphene surface accompanied, however, by some structural disorder, as revealed from the RS spectra [13,16].

In the present work, we analyze air-exposed and cleaned single-layer graphene layers grown on a copper foil by RS, by XAES C KVV spectra, and by high-energy resolved core-level photoelectron spectra. To see beneath the surface of the graphene, we recorded photoelectron spectra at various emission angles and therefore at various electron sampling depths (SD). The SD is approximated by λcosα, where λ is the inelastic mean free path (IMFP) of the photoelectrons and α is the electron emission angle. Up to now, no attempt has been made to reconstruct the in-depth distribution of the elements and the in-depth distribution of the discerned bonding states for graphene samples. We obtained the concentration depth profiles non-destructively from the angular dependent peak areas of C 1s, O 1s and Cu 3p lines using the Maximum Entropy Method (MEM). In addition, we were able to determine the in-depth location of the CO, CuO, C sp² and C sp³ bonding states.