The emergence of graphene as a 2D Dirac material has ignited a frenzy of study into its distinctive chemical, electronic, magnetic, optical and thermal properties, while broadly uniting scientific efforts across multiple disciplines . Due to the sheer magnitude of mobility exhibited by its ambipolar charge carriers, high speed electronics, integrated circuits, and memories are of particular interest.
Chemical modification of graphene presents a viable pathway for tailoring electronic properties such as band-gap and majority carrier type. Covalent functionalization also enables subsequent coupling, which is vital for molecular diagnostics and molecular electronics. Recent strategies include the photochemical decomposition of benzoyl peroxide, azide-modification followed by click-coupling reactions with alkynes, a Diels-Alder reaction of graphene with tetracyanoethylene, and plasma chlorination of graphene. Surface chemists continue to establish reliable protocols to prepare densely ordered, high quality functional monolayers on graphene. The controlled formation of densely ordered and stoichiometric derivatives of graphene remains difficult to attain. Examples of dense terminations include hydrogenation of graphene into fully sp3-hybridized graphane by exposure to atomic hydrogen and fluorination of graphene into fully sp3-hybridized fluorographene upon reaction with XeF2; each of which become wide-gap insulators at complete monolayer coverage.
Hyperthermal ion beams (1–100 eV particle kinetic energy) present a unique approach to modification, due to their capability of triggering a chemical reaction when their kinetic energy is converted when they come to a halt upon collision with a surface . Here we demonstrate the covalent modification of graphene via the hyperthermal impact of beams of 4,4’-azobis(pyridine) (AZP) .
Figure 1 illustrates the experiment. Intact, gas phase, mono-protonated beams of AZP are generated from solution using ESI. The beam is transferred into vacuum through a capillary, and then steered, focused and accelerated towards a sample of chemical vapor deposited (CVD) graphene transferred onto SiO2/n+ Si substrates. The collision energy is adjusted by a substrate bias VB. Time-of-flight mass spectrometry (TOF-MS) is used to verify the formation of azopyridinium cations (AZP+; m/z = 185 u/e), and in combination with a mass-selecting quadrupole, a narrow mass-to-charge window is specified so that the reagent is highly purified before exposure. During modification, a deposition current is monitored at the substrate (200–330 pA) and integrated over time such that the accumulated dose is precisely controlled. The total cation exposure has been set to 1 nAh (3.6 µCoulomb), which corresponds to a dose of 2·1014 AZP+/cm2 and is achieved in 3–5 h. On the exposed area this would correspond to a coverage on the order of a close-packed molecular monolayer assuming a sticking coefficient of unity.
Ex situ ambient confocal Raman scattering (Fig. 2) confirms the characteristic G-Band and 2D-Band of pristine single-layer graphene before deposition. A small D-peak is initially detectable, which has been shown to originate from wrinkles. Collisions at 5, 75, 100, and 125 eV produce the same Raman spectrum as that of pristine graphene (Fig. 2, lowest spectrum). These depositions can be considered as soft landing. Upon modification with 165 ± 3 eV ions, a dominant D-peak appears, which is consistent with sp3-hybridization induced disorder, while suppression of the 2D/G ratio and slight blue-shifted D-peak position further indicates the presence of disorder/doping.
To confirm the covalent binding of the molecular ions, the modified samples were subjected to thermal treatment. After heating in vacuum at 200°C for 1 hour, Raman spectroscopy revealed a significantly reduced D-peak intensity. This change clearly testifies desorption of chemisorbed molecules. It should be noted that if carbon vacancies created by the ion impact were responsible for the pronounced D-peak, such restoration of the carbon framework of graphene would not be possible owing to the lack of a suitable carbon source and sufficiently high temperature. Raman maps displaying the D-peak area over 100 µm2 regions show a uniform, large-area modification, which has further been verified over the ≈ 4 mm × 4 mm sample area.
From the Raman spectra, the average distance LD between the sp3-defect centers (i.e., between the covalently attached azopyridyl groups) can be estimated. The intensity ratio of the D and G modes (ID/IG) well above 1 in the spectra of the modified samples indicates the high defect density regime, with LD being 1.4 nm. This translates into a high functionalization degree of approximately 3%, corresponding to 50% of impacting ions being bound to the graphene. As the elevated collision energy gives rise to competition from elastic scattering, and fragmentation efficiencies close to unity, it is reasonable to attribute the remaining 50% of the collisions to elastic or dissociative projectile scattering and hence no covalent bond formation.
Ambient tapping-mode atomic force microscopy (AFM) shows that despite the significant collision energy supplied to the surface, the pristine topographic quality of CVD graphene is still preserved after exposure, producing highly smooth and flat surfaces.
The elemental composition and chemical binding at the surface before and after modification was probed by X-ray photoelectron spectroscopy (XPS). The carbon signal in the spectra change upon covalent modification and an additional nitrogen signal appears. From the relative abundance of azo-nitrogen to the combined neutral and protonated pyridinic-nitrogen, a ratio of 2:1 for azo- to pyridinic-nitrogen is calculated. This points to a binding mechanism wherein the diimide group cleaves upon impact at the surface at the protonated pyridyl-site (C5H4NH)-N=N-(C5H4NH+), producing a neutral azopyridyl-radical (·N=N-C5H4N) that binds to graphene through a C-N covalent bond, as illustrated in Fig. 3.
Hyperthermal ion chemistry presents a unique solution to the challenge of controllably producing dense or ordered monolayers on graphene. Electrospray ionization accommodates both an enormous molecular size range (1–106 Da) and practically unlimited choice of reagents to activate non-equilibrium hyperthermal reactions of graphene that would otherwise remain kinetically unfavorable. High vacuum deposition further minimizes the influence of an electrochemical water/oxygen redox couple at the SiO2 substrate which is responsible for doped hysteretic devices in ambient.
The present study demonstrates a one-step non-destructive route to covalently functionalize chemical-vapor-deposited graphene using the controlled deposition of hyperthermal molecular ion beams of azopyridine. Since the kinetic energy of the impinging reagents of 165 eV/ion is significantly larger than that of typical covalent bond-dissociation depths (1 ≤ D0 ≤ 10 eV/bond), the excess energy supply breaks chemical bonds, and overcomes the costly activation barrier (EA) that otherwise prohibits addition to the basal plane. Raman scattering in combination with XPS demonstrates the formation of a covalently azopyridyl-modified surface with a high functionalization degree of 3%, while AFM imaging shows the resulting graphene retains its topographic integrity.