Evidence for Electron Transfer between Graphene and Non‐Covalently Bound π‐Systems

Abstract Hybridizing graphene and molecules possess a high potential for developing materials for new applications. However, new methods to characterize such hybrids must be developed. Herein, the wet‐chemical non‐covalent functionalization of graphene with cationic π‐systems is presented and the interaction between graphene and the molecules is characterized in detail. A series of tricationic benzimidazolium salts with various steric demand and counterions was synthesized, characterized and used for the fabrication of graphene hybrids. Subsequently, the doping effects were studied. The molecules are adsorbed onto graphene and studied by Raman spectroscopy, XPS as well as ToF‐SIMS. The charged π‐systems show a p‐doping effect on the underlying graphene. Consequently, the tricationic molecules are reduced through a partial electron transfer process from graphene, a process which is accompanied by the loss of counterions. DFT calculations support this hypothesis and the strong p‐doping could be confirmed in fabricated monolayer graphene/hybrid FET devices. The results are the basis to develop sensor applications, which are based on analyte/molecule interactions and effects on doping.


Experimental Procedures
All starting materials were purchased as reagent grade from Sigma-Aldrich. The reagents were used as received. Dry acetonitrile was purchased from Acros and stored under N 2 -atmosphere. Octyltriflate [1] and 1,3,5-tris(benzimidazolyl)benzene 3 [2] and 4 [2] were prepared according to literature procedures. Thin layer chromatography (TLC) was performed on plates from Merck (silica gel 60, F254). Substances were visualized under UV light (wavelength λ = 254 nm). Solvents and Rf values are stated in the experimental part. Column chromatography was performed on silica gel from Merck (35-70 micron). Some purifications were performed via automated column chromatography on a Biotage IsoleraTM Spektra One flash chromatography system usingBiotage SNAP-50 g KP-sil columns. The CVD graphene was purchased from Graphenea. Its quality was checked by analysis of the G and 2D peaks via Raman spectroscopy. Air-sensitive reactions were carried out in flame-dried glassware and under an inert N2-atmosphere using Schlenk techniques. Melting points were determined on a Büchi B-545 melting-point apparatus in open capillaries and are reported uncorrected. "Decomp" refers to decomposition.

2D material sample preparation
The sample preparation of non-covalent functionalized graphene for HR-MS, Raman spectroscopy, XPS, AFM and electrical transport measurements was always the same. A silicon wafer with deposited graphene, either CVD graphene or r-oxo-G was incubated in a 12 mM methanol solution of the respective tricationic molecule for 2 hours at 4 °C. After incubation, the functionalized graphene wafer was removed from the incubation solution and rinsed with methanol to remove excess of the tricationic molecules. XPS samples and Raman spectroscopy samples of neat molecules/adsorbates without graphene were prepared by drop-casting 3 drops of a 12 mM methanol solution of the respective tricationic molecule onto a silicon wafer and allow the solvent to evaporate.
Nuclear magnetic resonance spectroscopy 1 H NMR, 13 C NMR and 19 F spectra were recorded on a Agilent 400 spectrometer (400.1 MHz for 1 H, 100.6 MHz for 13 C and 376.3 MHz for 19 F). Chemical shifts (d) are reported in ppm and were referenced to the residual solvent signal as an internal reference (DMSO-d6: 2.50 ppm for 1 H, 39.52 for 13 C; ACN-d3: 1.94 ppm for 1 H, 1.32, 118.26 for 13 C; CD2Cl2: 5.32 ppm for 1 H, 53.84 for 13 C). Coupling constants (J) are given in Hertz (Hz) and the apparent resonance multiplicity is reported as singlet (s), doublet (d), triplet (t), quartet (q) or multiplet (m).

High resolution mass spectroscopy
High resolution mass spectroscopy data was obtained on an Agilent 6520 QTOF LC/MS coupled with an Agilent 1290 Infinity LC system. The signal of the molecular ion [M] + is reported in m/z units. The submitted samples (1 mg) were diluted to ca. 10 μg/ml in acetonitrile. The sample was analyzed using an Agilent 1290 infinity LC system equipped with autosampler tandem. HRMS spectra were recorded with a 0.3 ml/min flow rate using an isocratic method (50% MPA/50% MPB). Mobile Phase A (MPA): Water with 0.04% formic acid. Mobile Phase B (MPB): MeOH with 0.04% Formic acid. All samples were initially analyzed using an ESI source in positive mode (scan range 100-1200 m/z). Samples were also analyzed in negative mode to detect anionic molecules and fragments (scan range 50-1200 m/z).

Raman spectroscopy
Raman spectroscopy data was obtained on a WITec alpha300 R instrument using an excitation wavelength of 532 nm. The laser intensity was kept at 1.7 mW for all performed measurements unless stated otherwise. The integration time was 0.1 s. Single spectra were four times accumulated. For most measurements a grating of 600 g/mm was chosen and the spectral center was set to 1900 cm -1 . For a grating of 1800 g/mm, the spectral center was set to 1350 cm -1 . The sample was placed on a motorized x,y table and focused before each measurement. For larger areas it was ensured that the focus of the excitation laser was constant. The data analysis was performed with the Project Manager software from WITec and visualized with OriginPro 9. The G and the 2D peaks were fitted by one Lorentzian function each. The graphene samples were analyzed via Raman spectroscopy before and after non-covalent modification.

X-ray photoelectron spectroscopy
X-ray photoelectron spectroscopy (XPS) experiment was conducted in a PHI5000 VersaProbe III Scanning XPS Microprobe system. The X-ray source was a monochromated Al anode (E = 1486.6 eV) with the beam diameter 100 μm (Energy resolution: 0.646 eV). Dual charge compensation was conducted by using argon ion gun (+ve) and electron neutralizer (-ve) for non-conductive materials. The survey scan was performed in the range between 0 and 1250 eV (step size: 1.0 eV) for compositional analysis. The high resolution narrow scan for the selected regions was performed with the step size of 0.1 eV for chemical state analysis. The energy scale calibration (ISO 15472) was performed by aligning the core level peaks of Au4f7/2, Ag3d5/2 and Cu2p3/2 at 83.96 eV, 368.21 eV and 932.62 eV, respectively. The sample was adhered on the platen with double-sided tape, and the surface contaminant was removed by blowing with dry N2 gas. The analyzing chamber was kept under UHV condition with pressure lower than 3.0 x 10 -6 Pa. The data analysis was performed with MultiPak software (Ver. 9.7.0.1). To ease data analysis, three energy states, including C1s at 284.6 eV [3] for the sp 2hybridised carbon in graphene (i.e. C=C bond), Si2p at 103.6 -104.0 eV and O1s at 532.6 -533.0 eV for the native SiO2 layer on silicon wafer, are used accordingly as the reference positions in this study.The used CVD graphene just showed the elemental signals for carbon, oxygen and silicon (from the wafer) in the measurements (see Figure S27).

Imaging time-of-Flight Secondary Ion Mass Spectrometry
ToF-SIMS data were obtained with a TOF.SIMS 5 instrument (ION-TOF GmbH, Münster, Germany). The instrument is equipped with a 25 keV Bi3 + cluster ion gun as the primary ion source and a 10 keV C60 + ion source for sputtering and etching. The samples were analyzed using a pulsed primary ion beam (Bi3 ++ , 0.14 pA at 50 keV) at a field of view of 500 μm × 500 μm. All spectra were acquired and processed with the Surface Lab software (version 6.4, ION-TOF GmbH). Low-energy electrons were used for charge compensation during analysis.

Oxo-Graphene and electrical measurements
The electrical transport measurements were performed in a two-probe configuration at ambient conditions. Oxo-Graphene flakes were deposited onto the Si/SiO2 (300 nm) substrate by Langmuir-Blodgett technique (LB, Kibron µtrough). Reduction was performed by vapor of hydriodic acid and trifluoroacetic acid (1/1 mixture by volume) at 80 °C. Reduced oxo-Graphene was cleaned with doubly distilled water (Carl Roth). Patterning of the electrode structure was achieved by standard electron beam lithography processing (Raith PIONEER TWO). The 5/70 nm Cr/Au electrodes were deposited by thermal evaporation (Kurt J. Lesker NANO 36). All electrical measurements were carried out at ambient conditions using a two-probe station with micromanipulated probes and two source-measurement units (Keithley 2450). Oxo-graphene and r-oxo-G were synthesized according to a previously published literature. [4] Atomic Force microscopy AFM characterization was performed using a JPK Nanowizzard 4 atomic force microscope in tapping mode at room temperature.

Density functional theory calculations
For all the results presented in this work, a 9 × 9 graphene supercell (162 atoms) with the molecule M adsorbed on top was employed. All the results were obtained using the CASTEP program package within the Material Studio 2017 framework in conjunction with the PBE functional including the semi-empirical dispersion-corrected DFT approach (DFT-D2) [5] of Grimme. Core electrons were described by on-the-fly ultrasoft pseudopotentials and 517 eV cut-off energy. Atomic positions were optimized with the BFGS algorithm using delocalized internal coordinates. The evaluation of the partial atomic charges on the graphene and on the molecule was made based on Mulliken population analysis provided by the DFT calculations.

CV/UV-Vis-NIR spectroelectrochemistry:
Cyclic voltammetry was carried out in 0.1 M Bu4NPF6 solution using a three-electrode configuration (glassy carbon working electrode, Pt counter electrode, Ag wire as pseudoreference) and PAR VersaSTAT 4 potentiostat. The sample concentration was 10 -4 M. The ferrocene/ferrocenium (Fc/Fc + ) couple served as internal reference. Spectroelectrochemical measurements were carried out in an optically transparent thin-layer electrochemical (OTTLE) [6] cell (CaF2 windows) with a platinum-mesh working electrode, a platinummesh counter electrode, and a silver-foil pseudoreference electrode. Anhydrous and degassed ACN with 0.1 M NBu4PF6 as electrolyte was used as the solvent. [2] (100 mg, 0.21 mmol, 1.0 equiv.) was suspended in 10 ml dry acetonitrile under a dry nitrogen atmosphere. Methyl trifluoromethansulfonate (114 mg, 0.70 mmol, 3.3 equiv.) was added and the reaction mixture was stirred at room temperature for 24 h.

HR-MS spectra:
This paragraph shows the HR-MS spectra from 8 3+ , 9 3+ , 11 3+ , 13 3+ . Compounds 7 3+ , 10 3+ and 12 3+ could be not protonated with ESI.     A significant change between the CVD graphene and the non-covalently functionalized graphene can be observed by analyzing the 2D peak for both samples. The CVD graphene give a 2D peak at 2682 cm -1 (median value, Figure S26, right side, black). The position of the 2D peak in the non-covalently functionalized graphene is at 2659 cm -1 (median value, Figure S26, right side, red). This corresponds to a shift of the 2D peak of 23 cm -1 . The FWHM Γ2D of the 2D peak of the graphene sample is 32.6 cm -1 (median value, Figure S26, right side in black) and broadens to 43.7 cm -1 (median value Figure S26, right side in red) for the non-covalently functionalized graphene sample. That leads to a shift of 11.2 cm -1 of Γ2D.           In as much as DFT predicts electron transfer from graphene to tricationic molecule 6 6+ , further signatures for such reduction of ionic charge, from 8 3+ to 8 2+ , is provided here. Thus, Raman spectra for 8 3+ , 8 2+ , and 8 1+ in the gas phase have been computed. Indeed, it is the assumed non-covalent nature of the bonding between cation and graphene that implies changes in the Raman spectra of the molecular ions in the gas phase to be relevant. The qualitative differences in computed spectra allow the further validation of the noncovalent bonding. Hence, support for single-electron transfer is arrived at by matching the experimental Raman spectra of tricationic molecule 8 3+ (Figure 4A) in solution to the computed Raman spectrum of the 8 3+ charged moiety 6 3+ in the gas phase (Figure S41), and tricationic molecule 8 3+ on CVD graphene 8 3+ (Figure 4B) to the computed Raman spectrum of 8 2+ , while excluding the 8 1+ . Figure S44. Calculated Raman spectra of A) trication molecule 8 3+ , b) the reduced species 8 2+ and C) the twice reduced species 8 1+ on graphene.

30
Cyclic Voltammetry characterization of 9 3+ : Figure S45. Cyclic voltammetry of 9 3+ in ACN (+0.1 M Bu4NPF6) at different scan rates. A) Overview scan of 9 3+ at 100 mVs -1 and B) overview scans of 9 3+ at different scan rates between 25 -1000 mVs -1 . C) Overview scans of 9 3+ at 100 mVs -1 showing the dependency of the oxidation peak from each reduction event. D) Overview scans of 9 3+ at 100 mVs -1 showing the reduction events without the oxidation event.  Transport measurements: Figure S48. Comparison of each two carrier mobilities at the maximum (green dash lines) and minimum slopes (blue dash lines) for r-oxo-G (black) and r-oxo-G/8 3+ (red, r-oxoG/SBr69).
We calculated each two carrier mobilities for r-oxo-G and r-oxo-G/8 3+ (Figure S48). The carrier mobilities were based on the minimum (blue dash lines) and the maximum slopes (green dash lines) of both curves. For the r-oxo-G, the fluctuation of its carrier mobilities is between 26.4 and 28.6 cm 2 V -1 s -1 . After functionalization with tricationic molecule 8 3+ , the carrier mobilities of r-oxo-G/8 3+ are between 17.8 and 20.7 cm 2 V -1 s -1 . Obviously, the effect of hysteresis on carrier mobilities is less than that of r-oxo-G/8 3+ . Based on this point, we can also confirm that the effect of air on the sample can be negligible.