Introduction

Semi-conductive nature of graphene (G) quantified by work function has been studied herein. We expected that the thermodynamic work function of G surface can be modulated by doping a second element or generation of defects or hole at the graphitic carbon skeleton. Though the zero band gap of G indicates its ballistic transport behavior, some electronic application is necessarily associated with the opening of band gap with controlled electron conduction property and gated semiconducting behavior. In addition, the realm of organic electronics demands the development of flexible, transparent and conductive electrodes, which partly can be full filled by modification of G1. Even in the present decade, ITO ensures its application with high light transparency (>80%) and low resistance (10 to 30 Ω/cm), its inhomogeneous surface and instability in rough condition remains the major drawback for micro-electronic application2. Out of ITO, carbon nanotubes, metal nanowires and some organic conducting polymers including poly (3, 4-ethylenedioxythiophene)-poly(styrenesulfonate) have also been used in flexible electronics3. Though G is expected to be the best candidate in this realm, modulating the conductivity of G to some desired extent remains still tough. The difficulty of controlled chemical functionalization or doping on G surface is the major reason for this inconvenience. Though, ensuring some harsh condition, G can be chemically functionalization to engineer the band gap, it affects some other necessary properties including light transparency, stability and sheet-like nature4. To avoid this unwanted situation, G based electrodes generally are fabricated from some indirect routes, where graphene oxide (GO) obtained from graphite by various oxidation techniques are reduced to form reduced Graphene Oxide (rGO) by means of chemical or thermal treatment5. The rGO exhibits properties similar to G. The band in rGO is open due to the existence of some residual oxygenous sites and doped elements6. Conventionally, pristine G obtained from chemical vapor deposition is used in light-emitting diodes7, photovoltaics8, field-effect transistors9, and photo-detectors10. As elemental doping affects the band gap and conductivity of rGO or G, instead of a single synthetic process, various routes to obtain G or rGO is preferred to tune and adopt the semiconductivity to the desired applications. Besides opening the band gap, chemical modification to G results in the defect and vacancy creation, for which this G can be used for magnetism, fluorescent and gas storing systems. Hence, finding routes for doping G by a second element has attracted worldwide concern.

In this work, we have succeeded chlorination of G by liquid pulsed plasma method, where graphite electrode was submerged and exfoliated in various organochlorine solvents. The exfoliation process along with doping was aided by a periodic high voltage stimulation11 for very short time. The method has been preferred for some advantages including (i) the ratio of chlorination was possible to be controlled only by changing the solvent having varying percentages of chlorine, (ii) compared with some established method, the doping and exfoliation time has found to be much lower (iii) large-scale synthesis of doped G was possible. A series of doped G, with a varied amount of chlorine dopant and band gap, was obtained by using CH2Cl2, CHCl3 and CCl4 as the solvent.

The synthesis of halogen-doped G has been investigated in recent years. G modified by chloride or fluoride possess high electronegativity and was found to exist with vacancy and chemical reactivity. Generally, chlorination or fluorination of G is accomplished by exposing with Cl2 or F2 plasma, photochemical reaction and microwave assisted reaction. However, wide application of these methods is limited due to the necessity for exclusive experimental condition and failure in large-scale synthesis. Moreover, there still doesn’t exist any route to control the level of halogen doping on G. Bearing these issues, herein we have studied improved pulsed plasma-aided chlorination of G in organochlorine solvents. The generated G was modified by control over the doping level and tunability of thermodynamic work function. We propose that these improvements are highly desirable for practical applications of G in sensor and other devices.

Results and Discussion

Figure 1 represents the AFM image of G nanosheets collected from aqueous suspension. Displayed morphology confirms the exfoliation of graphite rod into tiny nanosheet, which is G. The exfoliation is aided by pulsed plasma. Such exfoliation of graphite is commonly known to be assisted by any type of ionic intercalation12. The nanosheets seem to have a submicrometer dimension with the thickness below 3 nm.

Figure 1
figure 1

AFM image for GG@H2O shows the presence of G nanosheet with varied dimensions (a,b). The thickness is below 3 nm (c).

Full size image

Figure 2 represents the Raman spectra of pristine graphite rod, GG@CH2Cl2, GG@CHCl3 and GG@CCl4 and some other G samples collected from various colloidal suspensions generated from similar pulsed plasma experimentation in different solvents. For pristine graphite sample, two typical peaks of carbon materials namely the ‘D band’ (~1350 cm−1) and the ‘G band’ (~1580 cm−1) are observed. These two peaks are reported to be associated with the breathing mode (A1g) and the in-plane bond stretching motion of C sp2 atoms (E2g), respectively13. The D band implies the sp3 carbon sites, with edge and defects. Raman spectra of G is very specific, where a 2D band (~2650 cm−1) is derived from a number of layer of G. The G band position shifts slightly from 1579 cm−1 to 1580 cm−1 for the GG@CH2Cl2 sample. This inverse softening is a result of the change in electronic structure during exfoliation as the van der Waals stacking force is relaxed due to the exfoliation process. However, this hardening process is reverse to the softening explained in some previous reports14. The ratio of peak height (ID/IG) for D and G-band was calculated as 0.337 and 0.641 for graphite rod and GG@CH2Cl2, respectively. Though this difference corresponds to a 33% increase, the ID/IG value clearly indicates the existence of G after the rod being exfoliated. Although electrochemical exfoliation with aqueous solvent is often utilized, some oxidation was observed in previous reports15. The ID/IG value increasing for GG@CHCl3 and GG@CCl4 as 0.942 and 0.993 indicates the increase in chlorine content of the samples. We propose that chlorination at some carbon sites and generation of defects are responsible for such increase in ID/IG value. Reported that the ID/IG value is inversely proportional to the percentage of sp2 domain16. For GG@CH2Cl2, GG@CCl3 and GG@CHCl3, we also measured FT-IR spectra (Figure S1). Each sample shows gradual peak for C-Cl bond between 700 and 800 nm.

Figure 2
figure 2

Raman spectra of Graphite and various G synthesized by submerging graphite rod in various solvents. Pulsed Plasma method.

Full size image

Therefore, the increase in this value is the result of decrease in sp2 domain and generation of some sp3 sites17. For observing the effect of solvent on exfoliation, we also accomplished the experiment using a series of solvents including EtOH, MeOH, H2O, THF, ACN and toluene. In every case, we observed the generation of colloidal dispersion with observable changes in Raman spectra. In a previous study, we have found that the increase in sp3 carbon site sometimes is associated with the generation of epoxy groups at G basement18,19. Though there exists no scope for such oxidation, to justify this possibility we have performed the chemical structure analysis by XPS study.

Figure 3(a,c,e) represents the C1s spectra of GG@CH2Cl2, GG@CHCl3 and GG@CCl4 in XPS. The spectra clearly indicate that the chemical structure of the nanosheet is graphene analogous. Possibilities for the formation of oxidized G was justified by de deconvolution of the envelop peak. The deconvolution spectrum for C=C (284.5 eV), C-C (285.0 eV), C-H (285.4 eV), C-OH (286.4 eV), C-O-C (287.2 eV), C=O (287.7 eV) and O=C-OH (288.7 eV) are shown in the figure, from which it is clear that the oxygenated functional groups have insignificant contribution to the envelop spectra. The spectra for C-Cl bonding in C1s spectra was overlapped with the spectra for C-O-C group. Hence, the deconvolution peak could be carried out with three features centered at 284.6, 285.8 and 287.6 eV corresponding to sp2 hybridized carbon, sp3 hybridized carbon and C–Cl respectively20. Figure 3(b,d,f) show the XPS Cl2p spectra of GG@CH2Cl2, GG@CHCl3 and GG@CCl4. The peak of Cl 2p3/2 and Cl 2p1/2 attributed to C-Cl bonding is observable around 200.4 and 202.2 eV. The peak attributed to Cl- anion (the peak of Cl 2p3/2 and Cl 2p1/2 is about 197.8 and 199.4 eV) is hardly present21,22. In addition, according to the semi-quantitative analysis of XPS, the ratio of C-Cl bonding in Cl2p spectra is almost equal the ratio of the peak at 286.4 eV (C-OH, C-Cl) in C1s spectra These results indicate that the peak at 286.4 eV in C1s spectra raises for the presence C-Cl bonding. The existence of C-Cl bond also is confirmed from elemental ratio obtained from XPS analysis. The ratio of Cl/(C + Cl) in GG@CH2Cl2, GG@CHCl3 and GG@CCl4 are 9.8, 29.7 and 12.6, respectively. The ratio of Cl/(C + Cl) for GG@CHCl3 is much than that for other samples. From the XPS survey spectra, GG@CHCl3 contained Cl 24 at.% (Figure S3b). This Cl contents is much higher compare to previous data23,24. Moreover, this Cl contents is one of the highest contents compare with other halides24,25,26.

Figure 3
figure 3

C1s (a,c,e) and Cl2p (b,d,f) spectra in XPS: GG@CH2Cl2 (a,b), GG@CHCl3 (c,d) and GG@CCl4 (e,f).

Full size image

Work functions of the sample surfaces were studied by Kelvin Probe Force Microscopy (KPFM)27,28,29. We have investigated the change in work function of G due to the variation in the extent of chlorination. KPFM measurement resulted in the CPD with respect to the height profile of the films. CPD quantifies the difference of work functions between the sample surface and the probe tip when they are in thermodynamic equilibrium. Hence, CPD analysis is very useful to confirm semiconducting properties of nanomaterials. A lower CPD value signifies higher work functions. We accomplished simultaneous imaging of topography and contact potential difference. Figure 4 shows the surface profiles, potential diagrams, superimposing images and CPD profile for each of GG@CH2Cl2, GG@CHCl3 and GG@CCl4. The surface profiles of the samples show that their morphologies are same as was estimated by TEM image analysis. The image for two-dimensional height profiles in Fig. 4(a,e,i) shows the nanometer ranged thickness of GG@CH2Cl2, GG@CHCl3 and GG@CCl4. Respective, potential energy mappings presented in Fig. 4(b,f,j) clearly indicate the topographical dependence of CPD throughout the surfaces. The superimposed mapping for height profile and CPD are presented in Fig. 4(c,g,k). Figure 4(d,h,l) display the superimposed curves for variation in both the sample heights and CPD as a function of scanned locations. CPD recorded for GG@CH2Cl2, GG@CHCl3 and GG@CCl4, lie within the range of 390.03-469.86, 69.95-130.09 and 100.20-201.19 mV, respectively. The CPD recorded for GG@CHCl3 is lower than CPD of other samples and hence the work function of GG@CCl3 is assigned to be the maximum. These results might be originated from the variation in the extent of chloride-modified domains in G basement.

Figure 4
figure 4

Study of Kelvin Probe Force Microscopy (KPFM) image and contact potential difference (CPD) profiles of the samples. Height profiles as (a,e,i); potential energy mappings as (b,f,j); superimposed topography for heights and CPD as (c,g,k) and superimposed curves for heights profiles and CPD as (d,h,l) are presented for GG@CH2Cl2, GG@CHCl3 and GG@CCl4, respectively.

Full size image

G is composed of sp2 hybrid carbon atoms with perpendicularly lying pz orbitals engaged in π bond network system. The large π system can participate in molecular charge-transfer process with both the electron donor and acceptor molecules/domains30,31. Such charge-transfer phenomena assist the formation of C-Cl bond during the pulsed plasma treatment of graphite in solvent. We propose that the bond formation is accomplished by simultaneous exfoliation of graphite with the formation of carbon and chlorine free radicals, which initiate the C-Cl bond formation. Therefore, remarkable changes in the Raman spectrum of G is noticed. Though the scope of oxidation during the reaction is very limited, the ID/IG ratio increases due to the effect of doping and vacancy generation. The pulsed plasma-aided vacancy generation also could function for modulating the work functions of the samples. The addition of chloride is analogues to the previously reported hydrogenation of G, where sp3 C–H bond formation in a reversible way was utilized for the chemical storage system of hydrogen gas32. In a similar study, reversible fluorination of G was also found to be convenient for gas storage33,34,35. Chlorination of G by 56 wt.% (30 at.%) in liquid chlorine medium was passible to carry out by UV light irradiation28. The chlorinated sample decomposes on heating or on laser irradiation, releasing all the chlorine. Almost similar results were observed for the bromination of G.

Any type of structural interruption can affect the band structure of G. Therefore, the theoretical zero band gap of G becomes opened due to elemental doping, distortion on edges or simply the creation of vacancies. The intrinsic electron mobility in G is greatly affected by its interaction with the environment, dopant and even with the substrate29. In every case the formation of ‘disorder potential’ results in the reduction of mobilities and limiting energy gaps in nanoribbons and bilayer G26. The doping of chlorine on G has revealed such usual change in work functions. In addition, the possibility of the existence of some substrate-induced electron-hole puddles, doping domains and local quatum capacitances may have some miner effects31. KPFM experimentation with single and few-layer graphene was reported previously32. The work function of single and bilayer G can be modulated with respect to the carrier density and doping-induced carrier concentration. Shifting of Fermi level reveals such effects.

The work function is a fixed characteristic of surface and signifies optimized thermodynamic energy/work necessary for releasing an electron from a spot anywhere of the solid surface to a vicinity outside the surface by atomic scale distance. This surface property is associated with the estimation of semiconducting behavior as in semiconductors, the conduction bands are adjacent enough to the Fermi level with having the option for thermal energy driven population of electrons or holes. Therefore, the samples seem to have bandgap of semiconductor due to the doping of chlorine at the carbon skeleton of G.

Conclusions

A series of chlorine doped G namely GG@CH2Cl2, GG@CHCl3 and GG@CCl4 has been synthesized by pulsed plasma treatment of graphite electrodes submerged in the organochlorine solvents including CH2Cl2, CHCl3 and CCl4, respectively. The samples were characterized by AFM image, which confirmed the sheet-like structure. The Raman spectra confirmed the increase in ID/IG value of the chlorine doped G samples. This increment confirmed the destruction of sp2 carbon site and formation of sp3 carbon atoms due to the formation of the C-Cl bond. Finally, the XPS spectra confirmed the formation of G, with some C-Cl bonds, where the Cl/(C + Cl) ratio in GG@CH2Cl2, GG@CHCl3 and GG@CCl4 were 9.8, 29.7 and 12.6, respectively. The Thermodynamic work functions of the samples were measured by Kelvin Probe Force Microscopy (KPFM). Measurement of contact potential difference (CPD) by KPFM shows that CPD values for GG@CH2Cl2, GG@CHCl3 and GG@CCl4, lie within the range of 390.03–469.86, 69.95–130.09 and 100.20–201.19 mV, respectively. The work function clearly indicates that all the samples behave like semiconductor. GG@CHCl3 exhibits the lowest value of work function. We propose that this report represents a new route for tuning the semiconductivity of G. Besides, the doping level of halogen on G is expected to be modulated by pulsed plasma treatment reported herein.

Methods

Simultaneous chlorination and exfoliation of G from graphite rod were accomplished by liquid pulsed plasma method in a glass made electrochemical cell, which was filled with the organochlorine solvent. The experimental setup is presented in Fig. 5. Highly pure graphite rod (diameter: 6 mm, 99.9998%) was submerged in the solvent and was treated by applying pulsed plasma (60 V, 2.5 A, 30 kHz, pulse duration: 10 µs) for 10 min. After the pulsed plasma treatment, the color of the clear solvent changed from limpid to black, which implied the formation of G colloid in the organochlorine solvent. Apparently, G seems to be exfoliated the graphite rod. A series of such suspension was prepared using CH2Cl2, CHCl3 and CCl4 as solvents. The G products were named as GG@CH2Cl2, GG@CHCl3 and GG@CCl4. In addition, a variety of solvents such as pure water, Ethanol (EtOH), Methanol (MeOH), Tetrahydrofurane (THF), Acetonitrile (ACN) and toluene were used also for observing the effect of solvent on exfoliation.

Figure 5
figure 5

Photograph (a) and schematic (b) for the experimental setup of pulsed plasma treatment. Schematic illustration for the KPFM model used for investigating GG@CH2Cl2, GG@CHCl3 and GG@CCl4. VT presents the recorded contact potential. Brighter region indicates higher recorded contact potential difference. A lower VT represents a higher work function.

Full size image

The morphology and thickness of the samples were studied by AFM (Bruker, Digital Instruments Nanoscope V). Raman spectroscopic investigations were performed on a micro Raman spectrometer (NRS-3100, Jasco, Japan) with a 532 nm excitation source at room temperature. An XPS instrument of Thermo Scientific, Sigma Probe was used for studying the chemical structure. A monochromatized X-ray source (Al Kα, hν = 1486.6 eV) and a discharge source (He I, hν = 21.2 eV) were used. Pt substrate was used to determine the Fermi level in Chlorinated G/Pt film. Vacuums better than 10−7 Pa was confirmed for the measurements. A hemispherical energy analyzer equipped with six channeltrons was used to detect the emitted electron. The KPFM measurements were accomplished at ambient conditions using a multimode microscope (Nanocute) operated by a Scanning Probe Microscopy controller. In the first step, along every scan line, the topography was recorded by controlling distance in tapping mode. In the second scan, the KFM measurement was performed using amplitude modulation (AM-KFM) with maintaining a constant distance from the sample (lift height = 5–50 nm). The applied electrical modulation frequency, VA, was adjusted to the eigenfrequency of the cantilever. Assumed that the Kelvin controller cancels electrostatic forces by adjusting the tip bias voltage, VT, until the amplitude becomes zero. The VT reimburses the potential differences between sample spots and the probe tip. For a known value of tip work function ϴtp, the work function of the sample can be calculated as ϴse = ϴtp − e ϴtp. The variation in work function can be calculated to be independent of ϴtp as ΔϴS = −eΔVT. Figure 5(c) represents the schematic for the measurement of VT expressed in color codes with respect to the topography mapped along the z-axis. The VT in fact represents the contact potential difference (CPD).