Chemical growth routes to graphene and graphene applications 2

Chapter High-Throughput Synthesis of Graphene by Intercalation-Exfoliation of Graphite Oxide and Study of Ionic Screening in Graphene Transistor Abstract We report a high-throughput method of generating monolayer exfoliated graphene sheets (>90% yield) from weakly oxidised, poorly dispersed graphite oxide aggregates. These large-sized graphite oxide aggregates consist of multilayer graphite flakes which are oxidised on the outer layers, while the inner layers consist of pristine or mildly oxidised graphene sheets. Intercalation-exfoliation of these graphite oxide aggregates by tetrabutylammonium cations yielded large-sized conductive graphene sheets (mean sheet area of 330 ± 10 µm2) with a high monolayer yield. Thin-film field-effect transistors made from these graphene sheets exhibited high mobility upon nullifying Coulomb scattering by ionic screening. Ionic screening versus chemical doping effects of different anions such as chloride and fluoride on these graphene films were investigated with a combination of electrical transport measurements and in situ Raman spectroscopy. 4.1. Introduction Graphene-based nanoelectronic devices are of great interest because the active channel can be scaled down to a single crystalline sheet of sp2-bonded carbon. Such devices can exhibit ultrahigh carrier mobility and long range ballistic transport.1, Currently, the adhesive tape method used for producing graphene layers from HOPG 76    is not compatible with industrial production. The search is on for a high-yield production route toward high-purity, large-sized graphene sheets which can be deposited as a uniform film on a wafer substrate. Reported production methods are varied and range from the chemical exfoliation of graphite oxide,3-5 liquid-phase intercalation and exfoliation of graphite6-10 and epitaxial growth11, 12 to chemical vapor deposition.13 Of these, solution-processed graphene sheets offer low-cost and high throughput for printable device fabrication on flexible substrates. The most commonly used Hummer’s method produces aqueous solution of graphene oxide sheets, which are used as precursors to generate mildly conducting graphene films. However, the harsh oxidation method results in small sized insulating graphene oxide sheets with lateral dimension in the submicrometre range.3 For the ease of device fabrication, it is desirable to synthesise graphene sheets with a lateral size larger than 20 µm.5, 14 Several researchers have demonstrated exfoliation and intercalation of graphite to produce monolayer graphene sheets.6-10 However, these methods suffer from low production yield (~1%) of monolayer sheets and sometimes involve the use of hazardous exfoliating and reducing agents such as oleum9 and hydrazine,4, respectively. Hence, a solution-phase method which can achieve high monolayer yield of large-sized conductive graphene sheets under mild oxidation conditions is highly desirable. In this chapter, we present an efficient and highly reproducible one-step intercalation and exfoliation method to produce large-sized, conductive graphene sheets without the use of surfactants. By removing the ultrasonication step completely, we are able to obtain large-sized exfoliated graphene sheets (with lateral dimension > 20 µm) without sacrificing the high production yield of monolayer sheets. The principle of the method is based on the rich intercalation chemistry of graphite 77    oxide.15, 16 Large amounts of graphite oxide sediments are formed after a brief oxidation of natural graphite by applying the modified Hummer’s method (see Materials and methods section 4.2.1.). These sediments consist of weakly oxidised graphite which cannot be dispersed well in aqueous solution due to their hydrophobic nature and large size. Our hypothesis is that these large-sized graphite oxide aggregates consist of multilayer graphite flakes which are oxidised on the outer layers, while the inner layers consist of pristine or mildly oxidised graphene sheets with oxygen functionalities mainly decorated at the periphery. These aggregated graphite oxide sediments were usually discarded by researchers, but we reclaimed these mildly oxidised graphite oxide sediments for further processing with the aim of recovering the inner pristine or mildly oxidised graphene sheets by performing intercalationexfoliation chemistry. 4.2. Materials and methods 4.2.1. Oxidation of graphite and intercalation by tetrabutylammonium ions Graphite (1.5 g) (Asbury graphite flakes), NaNO3 (1.5 g), and H2SO4 (69 mL) were mixed and stirred in an ice bath. Next, g of KMnO4 was slowly added. The reaction mixture was then stirred in room temperature for hour. After which, 100 mL of water was added and the temperature was increased to 90 °C for 30 minutes. Finally, 300 mL of water was slowly added, followed by another slow addition of 10 mL of 30 % H2O2. The reaction mixture was filtered and washed with water until the pH was about 6. The graphite oxide precipitate was dispersed in water: methanol (1:5) mixture and purified with three repeated centrifugation steps at 12000 rpm for 30 minutes. The purified sample was then dispersed in deionised water and centrifuged at 78    2500 rpm to separate monolayer oxidised graphene sheets in the supernatant from graphite oxide and some unreacted graphite sediments found at the bottom. These sediments (approximately 0.5 g) were recovered, dried, and dispersed in 20 mL of DMF and mL of tetrabutylammonium (TBA) hydroxide solution (40% in water). The mixture was then heated under reflux at 80 °C for over days. A black suspension resulted, which remained stable for months without precipitation. To recover monolayer mildly oxidised graphene sheets, the reaction mixture was purified with repeated centrifugation in water/methanol (1:5) mixture as mentioned above, followed by a fresh addition of DMF at the last purification step. The final purified sample was dispersed in DMF. Centrifugations at 1500 rpm for 30 were used to reasonably separate highly oxidised graphene sheets (supernatant) from mildly oxidised ones (bottom). A good monolayer yield of ~90% was obtained, which was confirmed by tapping mode AFM. 4.2.2. Fabrication and electrical measurements of GFET Mildly oxidised graphene sheets were spin-coated on oxidised silicon substrates (285 nm SiO2 with prefabricated markers) and annealed at 1000 °C. The samples were identified and located by optical microscopy. 100 µL of 3% PMMA (molecular mass, 950 K) chlorobenzene solution was spin-coated on SiO2/Si substrates at 6000 rpm using Spincoater Model P6700 Series (Specialty Coating Systems, Inc.) and baked at 120 °C for 15 min. The thickness of PMMA is about 200 nm. Electron beam lithography was done using a Philips XL30 FEGSEM at 30 kV with a Raith Elphy Plus controller, with an exposure dosage of 280 µA/cm2. The PMMA was then developed with a methyl isobutyl ketone (MIBK)/IPA (1:3) solution. 79    Electrical contacts composed of 10 nm chromium (Cr) and 100 nm gold (Au) were deposited by thermal evaporation. The films were then lifted off in acetone at room temperature and rinsed with IPA. The Cr/Au contacts were annealed at 350 °C in a vacuum furnace for 40 minutes to improve the contact adhesion between metal and graphene. The transport measurements for devices were obtained with a B1500A Semiconductor Device Analyser (Agilent Technologies) using the in-built R-I Kelvin measurement software. Electron and hole mobility can be extracted from the linear regime of the transfer characteristics using µ = [(ΔIds/Vds) • (L/W)] / Cox ΔVg where L and W are channel length and width, respectively, Cox is silicon oxide gate capacitance (which is 1.21 × 10-8 F/cm2 for a gate oxide thickness of 285 nm), Ids, Vds and Vg is drain-source current, drain-source voltage and gate voltage, respectively. To allow for electrolyte gating of the channel, the contacts were insulated by spin-coating the device with a layer of PMMA baked at 150 °C and selectively exposing the channel area by electron beam lithography. The graphene channel was exposed by developing with a MIBK/IPA (1:3) solution. 4.2.3. Raman and optical contrast spectroscopy The Raman spectra were obtained with a WITEC CRM200 Raman system. The excitation source is 532 nm laser (2.33 eV) with laser power below 0.1 mW to avoid laser-induced heating. The laser spot size at focus was around 500 nm in diameter with a 100× objective lens (NA = 0.95). Spectral resolution is cm-1 for frequency range of 900-4000 cm-1. Spectral resolution is cm-1 for frequency range of 1000-2000 cm-1. The Raman hysteresis is less than cm-1. All Raman G and D peaks were adequately fitted with a Lorentzian component of the Voigt profiles. All Raman 80    spectra exhibited a prominent G peak which relates to the E2g vibrational mode between sp2 carbons. D peak, which relates to an out-of-plane vibrational mode, indicative of sp3 carbons in the surroundings, was left out for clearer representation of G peak responses to NaF and KCl concentrations. The ratio of integrated intensity of D to G peak (ID/IG) for thermally annealed graphene sheets ranges from 0.81 to 1.1, thus indicating homogeneity and good restoration of π-conjugated structure.3 For optical contrast spectroscopy, the sample was placed on the x-y piezostage to perform contrast imaging across the active channel region. The contrast spectra of graphene film are obtained by C(λ) = (R0(λ) – R(λ)) / R0(λ), where R0(λ) is the reflection spectrum from SiO2/Si substrate with SiO2 thickness of 285 nm and R(λ) is the reflection spectrum from graphene sheet illuminated by normal white light source. 4.3. Results and discussion 4.3.1. Reaction monitoring of intercalation-exfoliation of graphite oxide Figure 4.1 shows the schematic representation of the intercalation-exfoliation process. These graphite oxide sediments collected after centrifugation were intercalated with TBA under reflux condition and heated at 80 oC for over two days. The quality of exfoliated graphene sheets was determined by UV-Vis spectroscopy, XPS and electrical conductivity. 81    Figure 4.1. Schematic representation depicting intercalation of tetrabutylammonium ions in large graphite oxide sediments and unreacted graphite particles to obtain monolayer mildly oxidised graphene sheets dispersed in DMF. These exfoliated graphene sheets were deposited onto SiO2/Si substrate to form graphene thin film FET. Effect of ionic screening and chemical doping effect of NaF and KCl were investigated with electrical transport and in situ Raman measurements. The colour change of the reaction mixture as the graphite oxide sediments were intercalated by TBA over time is shown in Figure 4.2. Although no chemical reducing agent was used, the intercalation and exfoliation by TBA caused the colour of the dispersion to change from pale-yellow to dark brown and finally, black (Figure 4.2a). The black dispersion was consistent with an overall increase in the UV-visible 82    absorption region, due to the presence of extended π-conjugated structure.4 The products were hydrophobic but could be dispersed to form a homogeneous suspension in either DMF or chloroform after a brief vortex (Figure 4.2b).17 The formation of a stable dispersion allowed the reaction to be monitored by UV-Visible absorption spectroscopy. As shown in Figure 4.2c, the graphite oxide sediments displayed an absorption maximum at ~231 nm which is due to the π  π* transition of aromatic C=C bonds and a shoulder at ~290-300 nm which corresponds to the n  π* transition of the C=O bond.18 As the reaction progressed, the π  π* (C=C) absorption peak at ~231 nm displayed a gradual bathochromic shift to ~253 nm while the shoulder at ~300 nm for n  π* (C=O) absorption peak decreased in intensity. The overall absorption in the entire spectral region increased with reaction time. These changes were comparable to hydrazine reduction of GO, in which the bathochromic shift of the 231 nm absorption peak and increase in background absorbance suggest the restoration of π-conjugated network within the RGO sheets.4 We suggest that the overall increase in absorbance is due to the exfoliation of pristine graphene sheets originated from the interior of graphite oxide sediments. The reaction mechanism during the intercalation and exfoliation of graphite oxide by TBA was further investigated by XPS. As shown in Figure 4.2d, the percentage of C-C bonds increased from 55% after hour to 81% after days. The increase in the C-C component and corresponding decrease in C-O (epoxide, ether and hydroxyl groups) and C=O (carbonyl and carboxyl) components with reaction time indicated the restoration of large domains of π-conjugated structures. By comparing the relative peak heights of the C-O versus C-C peak, we conclude that these TBA-intercalated graphene sheets have a smaller percentage of non- 83    stoichiometric (10% oxidised C) oxidised graphene sheets compared to GO sheets obtained by the modified Hummer’s method.19 (a) (b) GO hr hrs hrs day days Absorbance (a.u.) 3.5 3.0 2.5 2.0 (d) 1.5 1.0 0.5 300 400 500 600 Wavelength (nm) 700 800 Annealed at 1000 oC 81 % days 66 % C- C (284.5 eV) 0.0 200 92.1 % Intensity (a.u.) (c) 4.0 280 12 hours 55 % C- O(286.7 eV) C= O(287.8 eV) 284 288 Binding Energy (eV) hour 292 Figure 4.2. Reaction monitoring of TBA intercalation in large graphite oxide particles. (a) Colour change of reaction mixture in DMF monitored over days. Suspension was centrifuged at 10 000 rpm for 10 minutes to remove unreacted particles. (b) Precipitation of relatively hydrophobic mildly oxidised graphene sheets in deionised water after reaction for day (left) and days (centre) and re-dispersion in DMF (right). (c) UV-visible absorption spectra of GO dispersions as reaction proceeded for over days. (d) C 1s XPS spectra of GO dispersion with reaction time show gradual increase in the C-C bonding component from 55% to 81%. The chemical exfoliation of the large graphite oxide sediments is inherently selective. The outer layers of highly oxidised graphene sheets with a larger proportion of oxygen functionalities will be exfoliated first due to greater interlayer distance and weaker van der Waals interactions, which afford greater ease of TBA insertion.15, 16 This is followed by exfoliation of the less oxidised inner sheets in which the oxygen functionalities are situated mainly at the graphene edge planes. Due to different solubilities and lateral dimensions, the smaller and highly oxidised graphene sheets 84    (yellow-brown in colour) could be reasonably separated from the larger and less oxidised ones by centrifugation (1500 rpm for 30 minutes). As shown in Figure 4.3, an analysis of the XPS spectra revealed that there were ~20% more C-O groups in the highly oxidised graphene sheets found in the supernatant as compared to those mildly oxidised ones found in the precipitate. Subsequent purification by repeated centrifugation resulted in a homogeneous black dispersion of mildly oxidised graphene sheets (Figure 4.2b). Mildly oxidised GO (precipitate) Highly oxidised GO (supernatant) Intensity (a.u.) Intensity (a.u.) C-C (284.5eV) (66%) C(epoxy/ether) (286.5eV) C=O (287.8eV) 280 282 284 286 288 290 292 294 Binding energy (eV) 280 285 290 295 Binding energy (eV) 300 Figure 4.3. Comparison of XPS spectra revealed a greater percentage of C-O groups (~20%) for highly oxidised graphene sheets in the supernatant as compared to those mildly oxidized ones in the precipitate. Inset shows the deconvolution of precipitate with 66% C–C group. After spin-coating the mildly oxidised graphene sheets on the SiO2/Si substrate, a good monolayer yield of ~90% can be obtained from our graphene dispersions with a mean sheet area of 330 ± 10 µm2, as evident in the optical micrograph and AFM topography study in Figure 4.4. The average topographical height obtained using AFM was ~0.93 nm which was comparable to the reported height of graphene sheets possessing residual oxygen functionalities (Figure 4.4c-d).3, 4, 85    7.3.3.3. Flow-catch-release sensing of single malaria infected cell Figure 7.7. Parasite differentiation by graphene channel conductance. (a) Conductance-time plots for (early to mid) trophozoite-PE and schizont-PE measured at Vg = 0.1 V and corresponding DIC images on the right. Device channel length and width was µm and 15 µm, respectively. (b) Box plots of percentage conductance changes for trophozoite-PE and schizont-PE. The top and bottom of the box denote 75th and 25th percentiles of the population respectively, while the top and bottom whiskers denote 90th and 10th percentiles respectively. Maximum and minimum values are denoted by open squares. Gaussian distribution of the raw data points is shown. (c) Current stability of graphene FET. Conductance spikes upon sudden increase in flow pressure from Pa to Pa were observed at around 50s and 120 s. However, such percentage increase in conductance (2%) was smaller than that induced by the adhesion of trophozoite-PE (~5%) and schizont-PE (~8%). The flow-catch-release sensing of single PE was demonstrated when the flow pressure was adjusted to Pa. Figure 7.7a shows the discrete time-dependent changes in conductance corresponding to the adherence of a single trophozoite-PE or schizontPE as it rolled across the CD36-functionalised graphene. All transistors were gated at CNP to avoid sample-to-sample variation in the initial state of device. The different level of conductance rise measured at CNP (Vg = 0.1 V) showed the ability to 170    differentiate between the trophozoite and schizont parasite development stage. While the adherence of trophozoite-PE produced a conductance change of 5.1 ± 0.3%, schizont-PE induced a change of 8.4 ± 1.3% (Figure 7.7b). Even at a high flow pressure of Pa, the baseline conductance values remained relatively constant and any conductance change corresponding to a sudden increase in flow pressure was small (~2%) compared to that induced by PE adherence (Figure 7.7c). This shows good device reproducibility with minimum fouling of the graphene sensor platform. Figure 7.8. (a) Conductance-time plot for the adherence of schizont-PE when Vg = 0.2 V. This resulted in a conductance drop of ~10%. (b) Conductance-time plot for the occupation of schizont-PEs showed distinct conductance rise which corresponded to single schizont-PE. The ambipolar nature of graphene allows detection at negative gate voltage as well. A reverse trend in conductance drop was obtained in this case when timedependent conductance was measured at Vg = -0.2 V. The adherence of schizont-PE typically induced a conductance drop of ~10%, as shown in Figure 7.8a. The sensitive detection of single cell binding event was further shown in Figure 7.8b where the adherence of two schizont-PEs on graphene induced two sharp conductance rises when measured at Vg = 0.1 V. When a single schizont-PE rolled across graphene, the change in conductance was ~8%. When a second schizont-PE squeezed past a 171    restricted area occupied by the first schizont-PE, the resultant conductance change was ~14%. This slight decrease from the expected conductance change of ~16% was due to the slightly reduced surface interaction when two PEs squeeze into the channel. 7.3.3.4. Adhesive dynamics of malaria infected cells Figure 7.9. Adhesive dynamics and parasite differentiation. (a) PE containing early trophozoite rolled along the surface with tank-threading-like motion. (b) PE housing late trophozoite exhibited tumbling or flipping dynamics. (c) PE bearing schizont peeled off the surface under hydrodynamic flow, followed by crawling dynamics possibly due to strain of parasite body on the membrane. (d) Box plots of cell velocities for trophozoite-PE and schizont-PE velocity obtained from the time taken for PE to cross graphene channel of 15 µm. The top and bottom of the box denote 75th and 25th percentiles of the population respectively, while the top and bottom whiskers denote 90th and 10th percentiles respectively. Maximum and minimum values are denoted by open squares. Gaussian distribution of raw data points is shown. The simultaneous DIC visualisation and time-dependent conductance measurements also provide an avenue to elucidate the adhesive dynamics of PE, thereby allowing for parasite differentiation based on cell velocity. The rolling of PEs along the microfluidic channel lined with parallel graphene active channels is facilitated by fluid stresses and random bombardment with other flowing red blood cells.22 As the PE entered the graphene channel, conductance increased and remained 172    relatively constant for the time duration, t, and dropped to baseline value when the PE exited the graphene channel (Figure 7.7a). The time duration, t, is known as the conductance dwell time, and from it the cell velocity can be calculated. The cell velocity is closely related to the dynamics of adhesion. As the cell progresses from the trophozoite to schizont stage of infection, its membrane stiffness increases. With a larger surface area of contact and lower membrane stiffness, it was noticed that early trophozoite-PE crawled along the surface in a tank-threading-like motion (Figure 7.9a).8 As the intra-cell parasite developed to late trophozoite and schizont (Figure 7.9b-c), the dynamics of adhesion changed to fast flipping or tumbling from side to side, some rolling motions and occasional firm adhesion.9 The mean cell velocity calculated from the conductance dwell time for trophozoite-PE (13 µm/s) was lower than schizont-PE (19 µm/s), with a statistical difference of p = 0.02865 at the 0.05 level using the two-sample Student’s t-test. This increase in cell velocity as a function of membrane stiffness can be understood as such: a higher membrane stiffness results in a smaller contact area as the cell is less deformable and thus PE rolls faster across the graphene active channel. The relationship between the calculated cell velocities and membrane stiffness correlates very well with numerical simulations.8 This showed that the computation of conductance dwell time and resultant cell velocity could also be used as additional parameters for statistical differentiation of the diseased cells. 173    7.3.4. Effect of charge impurity density on cell-graphene interface Figure 7.10. Effect of charge impurity density on GFET. (a) Conductance versus carrier density, n, for protein-functionalised graphene in solution (blue curve based on experimental data in Figure 7.6 for protein-functionalised graphene) and under influence of charge impurity density with input parameters for black curve (ε = 1, z = nm, Nic = 1.75 ×1012/cm2), red curve ( ε = 1, z = nm, Nic = 3.5 ×1012/cm2), yellow curve (ε = 10, z = nm, Nic = ×1012/cm2) and green curve (ε = 10, z = nm, Nic = ×1012/cm2). For each dielectric constant and cell-graphene separation, the charge impurity densities arising from cells are chosen to generate the experimentally observed width of the voltage plateau. The density shifts in the position of conductivity minima are not shown (see Inset). (b) Conductance versus n near CNP for protein-functionalised graphene in solution. The blue curve is based on the experimental data for protein-functionalised graphene. Curves with increasing conductance represent graphene under the influence of charged cells with varying percentage of randomly-distributed cell area in proximity with graphene: 0% (blue), 5% (red), 10% (black), 15% (yellow) and 20% (green). Input parameters: ε = 10, z = nm and Nic = ×1012 cm-2 are chosen to simulate the local changes in graphene conductivity (c) Simulated changes in the minimum conductance of graphene upon adherence of cells of different knob density. Baseline is given by proteinfunctionalised graphene doped at CNP to match experimental conditions. (d) C-V response for protein-functionalised graphene upon PE adherence. 174    The adherence of PE modifies the mobility, carrier density and width of the voltage plateau at the conductivity minimum of graphene. These physical processes underpin the PE detection and differentiation. The overall changes in transport properties point to the enhancement of disorder in graphene upon cell adherence. It is important to note that the cell adherence on protein-functionalised graphene represents a rather complex physical phenomenon, involving several sources of charge scattering and a rapidly changing dielectric environment in the direction perpendicular to the graphene surface. In such a case, an exact treatment is non-trivial and therefore the focus will be on identifying the broad concepts responsible for PE detection on graphene. Prior to cell adherence, the carrier mobility of graphene is strongly limited by factors such as defects and substrate impurities and quantifying their effect is beyond the scope of this work.6, 23 The discussion will therefore be focused on the changes in graphene conductance induced by cell adherence. To estimate the influence of Coulomb scattering from these cells requires knowledge of the dielectric function of the system and of the density of charged scatterers (Nic). The layer of protein atop graphene has a low dielectric constant. However its small thickness of ~1.5 nm suggests that the screening properties of graphene will also depend on the environment beyond the protein layer, which comprises of the ionic media (with high dielectric constant) and the cellular media.14, 24 For the relatively simple case of planar lipid bilayer atop graphene (as discussed in Chapter 6), the three-dielectric scattering problem was solved following the approach of Ponomarenko et al.24 to obtain an effective dielectric constant for that system. In the present case, the values for the top dielectric media () is varied between and 10. Based on the self-consistent charged-impurity scattering theory for graphene described within random phase approximation,25 the estimated density of 175    randomly distributed surface charges acting as Coulomb scatterers needed to explain the observed changes in the width of CNP voltage plateau upon cell adherence is between Nic ~2 - x 1012/cm2 (Figure 7.10a). Such surface charge densities are quite reasonable for planar charged cell membrane systems14 and these can readily explain the observed mobility decrease in our devices. However, simultaneous to the changes in the width of voltage plateau and mobility, a large carrier density shift in the position of conductivity minimum is expected from theory (see Inset of Figure 7.10a), but this was not observed experimentally. This discrepancy may be related to the fact that cell surfaces are far from being planar, but are highly undulated and irregular. Therefore, any model which accounts for cell-surface interaction should consider the cell as a highly charged and irregular entity. In this case, only positively charged knob sites that adhere to the thin protein layer can exert significant Coulomb potential on charge carriers in the graphene sheet.14 The scattering arising from a majority of the charged cell surface is exponentially screened by a factor exp(-q·z) where q is the momentum transfer upon scattering and z is the cell-graphene separation.25 For pristine graphene doped at CNP, the energy landscape is usually described by the formation of electron-hole puddles with typical energy E F  v F n0 , where n0 is the disorder-induced residual carrier density.23 However, in regions where the heavily-charged cell surface is in close proximity to the graphene sheet, the size of these local electron-doped charged puddles is expected to be much larger in energy space. Under the assumption that the typical spatial extent of such charged puddles is comparable or larger than the mean free path, the cell-graphene system can be modelled as a two-dimensional random two-resistor network comprising of graphene regions with and without cell-induced Coulomb disorder.26 For graphene areas which are in close proximity to the cell, an additional Coulomb 176    disorder term is introduced. The shift in Fermi level and change in mobility in these puddles are calculated based on the self-consistent theory for charge impurity scattering in graphene.25 A charge impurity density of Nic = ×1012 cm-2 and a cellgraphene separation of z = nm were assumed. The total conductivity of graphene is then estimated for various fractions of such randomly distributed puddles based on the theory for two-dimensional random two-resistor network.26 Figure 7.10b shows the conductivity changes near CNP arising from this inhomogeneous charge density configuration, considered for the case of low density knobs (5 - 20%). Key features of detection involving conductivity changes near CNP (simulated in Figure 7.10c) can be explained using this model, while remote charge scatterers can possibly contribute to the shift in the position of conductivity minimum. This is supported by the C-V plots in Figure 7.10d, which indicate a small increase in interfacial capacitance of the system after cell adherence. The enhancement in the total capacitance has been associated with the presence of increased Coulomb scatterers in the vicinity of graphene,27 given by CQ ~(|nG|+|no|)2, where CQ is the quantum capacitance of disordered graphene, nG and no is the carrier concentrations caused by gate potential and charged impurities, respectively. Therefore, the C-V measurements and different output signals associated with the adherence of trophozoite-PE and schizont-PE can be understood within the density inhomogeneity model and strongly suggest that capacitive coupling is responsible for the changes in graphene conductance during cell adherence. 177    7.4. Conclusion In summary, we have constructed a GFET sensor, integrated with microfluidic flow cytometric assay for the flow-catch-release sensing of malaria-infected red blood cells. The GFET sensor is able to generate dynamic disease diagnostic patterns in terms of conductance changes and characteristic conductance dwell times. Coulomb impurity potential exerted by charged protrusions on cell surfaces induces local doping and produces the observed conductance changes which could be used as a parameter to distinguish the stage of infection. The ability of graphene to interface readily with biorecognition proteins, coupled with its optical transparency, sensitivity towards capacitively-induced conductance changes and easy integration with microfluidic flow cytometric assay, exhibit great promise in clinical diagnostic applications. 7.5. 1. References Bhagat, A. A.; Bow, H.; Hou, H. W.; Tan, S. J.; Han, J.; Lim, C. T., Microfluidics for cell separation. Med Biol Eng Comput 2010, 48, 999-1014. 2. Hou, H. W.; Bhagat, A. A.; Chong, A. G.; Mao, P.; Tan, K. S.; Han, J.; Lim, C. T., Deformability based cell margination--a simple microfluidic design for malaria-infected erythrocyte separation. Lab Chip 2010, 10, 2605-2613. 3. Wang, Z.; El-Ali, J.; Engelund, M.; Gotsaed, T.; Perch-Nielsen, I. R.; Mogensen, K. B.; Snakenborg, D.; Kutter, J. P.; Wolff, A., Measurements of scattered light on a microchip flow cytometer with integrated polymer based optical elements. Lab Chip 2004, 4, 372-377. 178    4. Kim, M.; Hwang, D. J.; Jeon, H.; Hiromatsu, K.; Grigoropoulos, C. P., Single cell detection using a glass-based optofluidic device fabricated by femtosecond laser pulses. Lab Chip 2009, 9, 311-318. 5. Patolsky, F.; Zheng, G.; Hayden, O.; Lakadamyali, M.; Zhuang, X.; Lieber, C. M., Electrical detection of single viruses. Proc Natl Acad Sci U S A 2004, 101, 14017-14022. 6. Castro Neto, A. H.; Guinea, F.; Peres, N. M. R.; Novoselov, K. S.; Geim, A. K., The electronic properties of graphene. Reviews of Modern Physics 2009, 81, 109. 7. Dong, X.; Shi, Y.; Huang, W.; Chen, P.; Li, L. J., Electrical detection of DNA hybridization with single-base specificity using transistors based on CVDgrown graphene sheets. Adv Mater 2010, 22, 1649-1653. 8. Fedosov, D. A.; Caswell, B.; Karniadakis, G. E., Wall shear stress-based model for adhesive dynamics of red blood cells in malaria. Biophys J 2011, 100, 2084-2093. 9. Fedosov, D. A.; Caswell, B.; Suresh, S.; Karniadakis, G. E., Quantifying the biophysical characteristics of Plasmodium-falciparum-parasitised red blood cells in microcirculation. Proc Natl Acad Sci U S A 2011, 108, 35-39. 10. de Koning-Ward, T. F.; Gilson, P. R.; Boddey, J. A.; Rug, M.; Smith, B. J.; Papenfuss, A. T.; Sanders, P. R.; Lundie, R. J.; Maier, A. G.; Cowman, A. F.; Crabb, B. S., A newly discovered protein export machine in malaria parasites. Nature 2009, 459, 945-949. 11. Aikawa, M.; Kamanura, K.; Shiraishi, S.; Matsumoto, Y.; Arwati, H.; Torii, M.; Ito, Y.; Takeuchi, T.; Tandler, B., Membrane knobs of unfixed Plasmodium falciparum infected erythrocytes: new findings as revealed by 179    atomic force microscopy and surface potential spectroscopy. Exp Parasitol 1996, 84, 339-343. 12. Li, A.; Mansoor, A. H.; Tan, K. S.; Lim, C. T., Observations on the internal and surface morphology of malaria infected blood cells using optical and atomic force microscopy. J Microbiol Methods 2006, 66, 434-439. 13. McCormick, C. J.; Newbold, C. I.; Berendt, A. R., Sulfated glycoconjugates enhance CD36-dependent adhesion of Plasmodium falciparum-infected erythrocytes to human microvascular endothelial cells. Blood 2000, 96, 327333. 14. Ang, P. K.; Jaiswal, M.; Lim, C. H.; Wang, Y.; Sankaran, J.; Li, A.; Lim, C. T.; Wohland, T.; Barbaros, O.; Loh, K. P., A bioelectronic platform using a graphene-lipid bilayer interface. ACS Nano 2010, 4, 7387-7394. 15. Trager, W.; Jensen, J. B., Human malaria parasites in continuous culture. Science 1976, 193, 673-675. 16. Lambros, C.; Vanderberg, J. P., Synchronization of Plasmodium falciparum erythrocytic stages in culture. J Parasitol 1979, 65, 418-420. 17. Baenziger, N. L.; Majerus, P. W., Isolation of human platelets and platelet surface membranes. Methods Enzymol 1974, 31, 149-155. 18. Lim, C. T.; Li, A., Mechanopathology of red blood cell diseases — Why mechanics matters. Theoretical and Applied Mechanics Letters 2011, 1, 014000. 19. Loh, K. P.; Bao, Q. L.; Ang, P. K.; Yang, J. X., The chemistry of graphene. Journal of Materials Chemistry 2010, 20, 2277-2289. 180    20. Cole, D. J.; Payne, M. C.; Ciacchi, L. C., Water structuring and collagen adsorption at hydrophilic and hydrophobic silicon surfaces. Phys Chem Chem Phys 2009, 11, 11395-11399. 21. Kim, J.; Cote, L. J.; Kim, F.; Huang, J., Visualizing graphene based sheets by fluorescence quenching microscopy. J Am Chem Soc 2010, 132, 260-267. 22. Cooke, B. M.; Berendt, A. R.; Craig, A. G.; MacGregor, J.; Newbold, C. I.; Nash, G. B., Rolling and stationary cytoadhesion of red blood cells parasitised by Plasmodium falciparum: separate roles for ICAM-1, CD36 and thrombospondin. Br J Haematol 1994, 87, 162-170. 23. Das Sarma, S.; Adam, S.; Hwang, E. H.; Rossi, E., Electronic transport in twodimensional graphene. Reviews of Modern Physics 2011, 83, 407. 24. Ponomarenko, L. A.; Yang, R.; Mohiuddin, T. M.; Katsnelson, M. I.; Novoselov, K. S.; Morozov, S. V.; Zhukov, A. A.; Schedin, F.; Hill, E. W.; Geim, A. K., Effect of a High-kappa Environment on Charge Carrier Mobility in Graphene. Physical Review Letters 2009, 102. 25. Adam, S.; Hwang, E. H.; Galitski, V. M.; Das Sarma, S., A self-consistent theory for graphene transport. Proceedings of the National Academy of Sciences of the United States of America 2007, 104, 18392-18397. 26. Bianco, B.; Giordano, S., Electrical characterization of linear and non-linear random networks and mixtures. International Journal of Circuit Theory and Applications 2003, 31, 199-218. 27. Xia, J. L.; Chen, F.; Li, J. H.; Tao, N. J., Measurement of the quantum capacitance of graphene. Nature Nanotechnology 2009, 4, 505-509. 181    Chapter Conclusions 8.1. Uniqueness of graphene Graphene is uniquely different from its carbon allotropes such as CNT and diamond owing to its ambipolarity, optical transparency, ultrahigh carrier mobility at room temperature, mechanical robustness and high conformity to arbitrary substrates. In addition, CPG can be readily functionalised with different functional groups for the synthesis of novel organic-graphene hybrid materials. These hybrids can have wide ranging applications spanning catalysis to energy storage and chemical and biological sensing. 8.2. Summary and outlook The chemical route to graphene via chemical exfoliation of graphite and graphite derivatives is a viable route for the large scale production of solutionprocessable graphene sheets. However, GO sheets derived from the harsh oxidation of graphite yielded poor electrical performance. An alternative route which can reduce the collateral damage of oxidation is needed. In Chapter 4, we presented an efficient intercalation and exfoliation method of graphite oxide without the use of surfactant or ultrasonication to yield big-sized (with lateral dimension > 20 µm) graphene sheets with reduced oxygen groups on the basal plane. With larger lateral dimension, RGO thin films exhibited lower number of sheet-to-sheet junctions and higher conductivity. It was discovered that impurity scattering effects exerted by substrate ionised charges 182    limit the mobility of carriers in RGO films. Using ionic screening, we demonstrated that the impurity charge scattering effect exerted by underlying substrate impurities could be reduced. Consequently, a nearly one order improvement in the carrier mobility of the films can be achieved. While NaF is effective in screening underlying charged impurities, KCl displays a concentration-dependent interplay of ionic screening and chemical doping, which may arise from specific adsorption of polarisable Cl- ions at the graphene-water interface. In Chapter 5, we improved the oxidation pathway of graphite and applied careful gradient separation to isolate bigsized GO sheets with a significant reduction in oxygen groups. In addition, significant reduction in contact resistance and the asymmetric factor and improvement in carrier mobility were achieved by replacing metal drain/source contacts with multilayer RGO interconnects. This paves the way for printable all-carbon post-CMOS electronics built on flexible substrates. Graphene is emerging as a new bioelectronic platform for sensing biomolecules and cells owing to its chemical inertness, mechanical robustness, optical transparency and electrical sensitivity to changes in its local charge environment. Therefore, it is timely to study how charged biomolecules influence the transport properties of GFET. In Chapter 6, we interfaced lipid bilayer with graphene and found that its presence on graphene created charged impurities which exerted spatially inhomogeneous Coulomb potential, leading to changes in the finite minimum conductivity, carrier concentration and carrier mobility of graphene. Utilising the Coulomb potential exerted by lipid charges on graphene, we could monitor proteininduced changes to biomimetic membranes and elucidate the pathway of membrane disruption. This is especially pertinent to the disruption of bacterial cell membranes by antimicrobial peptides. By forming biomimetic membranes mimicking the 183    composition of bacterial cell membrane, we demonstrated that the bactericidal activity of antimicrobial peptides, Magainin 2, could be sensed electrically based on the interplay of biomolecular doping and ionic screening effect of mobile ions in the electrolyte. This sensing concept may be extended to other biorecognition systems such as ligand-receptor binding and gated control of ion channels embedded in membranes. Moving on from biomimetic systems, we presented in Chapter 7, a GFET array integrated with microfluidic flow cytometric assay for the simultaneous optical and electrical detection of malaria-infected red blood cells in a fluid stream with single-cell resolution. The morphological distinction between healthy and malaria infected red blood cells lies in the surface protrusion of positively charged knobs exhibited by the latter. The Coulomb impurity potential exerted by these charged knobs influences the conductance of graphene. When gated in the n-type regime, the momentarily capture of single infected cell as it rolled across the graphene surface functionalised with biorecognition proteins manifested as a sharp rise in conductance upon the entry of the infected cell followed by a sharp drop upon its exit. The ability to a statistical percentage count of the infected cells as a mixed population of infected and healthy cells flow through the GFET array in a microfluidic channel exhibits great promise for clinical diagnostic applications. This dissertation shows that the opportunities for graphene in flexible printed electronics and nanobioelectronics are vast. In particular, graphene is suited for biological applications owing to its biocompatibility and ultrasensitive nature to its environment. Graphene establishes a good biointerface between biotic and abiotic system because of its stable interaction with bio-entities, high adsorption capacity for biorecognition proteins and target molecules and high electrical sensitivity since every carbon atom in graphene responds to charges induced by bio-entities, negating the 184    necessity for an oxide passivation layer. Besides, if graphene sensors are fabricated on flexible substrates, intimate connection with living tissues can be established. One promising future application of graphene in this aspect would be to construct ultrasensitive and flexible neural or coronary probes. As of today, material scientists, clinicians and engineers are exploring the possibilities of building implantable medical devices based on graphene, taking advantage of its flexibility, chemical inertness, sensitivity and versatility to act as components for sensing electrodes, memory and logic units, communication antennas and transmitters as well as energy sources such as supercapacitors, batteries and fuel cells. With the concerted efforts of multidisciplinary scientists, it is only a matter of time before graphene-based biosensors and smart medical devices emerge at the market place. 185    [...]... 0.10 10 mM KCl 0.05 Dry 60 66 1604 64 16 02 62 1600 60 1598 58 1596 56 1594 54 15 92 52 200 400 600 800 KCl Concentration (mM) 1000 1400 1500 1600 1700 -1 Raman Shift (cm ) 1800 1900 2. 4 -2 68 1606 (d) 70 1608 G Peak Position (cm-1) 1610 0 1300 80 2. 2 12 0 20 40 Gate Voltage (V) FWHM (G) (cm-1) (c) -20 Hole Concentration (x 10 cm ) 0.00 -40 2. 0 1.8 1.6 1.4 1 .2 0 20 0 400 600 800 KCl Concentration (mM) 1000... contacts and active channel area exposed by electron beam lithography Channel length and width was 29 .3 µm and 37.8 µm, respectively (b) Transport characteristics of graphene film with 1 to 4 layers of graphene sheets Hole and electron mobility obtained were 10.1 cm2/(V s) and 4.9 cm2/(V s), respectively As the channel length increased from 29 .3 µm to 100 µm, carrier mobilities decreased by 4 to 5 times... m 2 35 Counts 30 25 20 15 10 5 0 (c) 0 (d) 500 1000 1500 2 2000 GO sheet area (m ) 25 00 4 Height / nm 3 2 1 0 -1 0 2 4 6 8 L a te ra l D is ta n c e /  m 10 12 Figure 4.4 Optical micrographs and tapping mode AFM characterisation of mildly oxidised graphene sheets (a) Optical image of large-sized mildly oxidised graphene sheets (b) Size distribution of monolayer mildly oxidized graphene sheets (total... organic semiconductors Nature 20 05, 434, 194-199 25 Chen, F.; Qing, Q.; Xia, J L.; Li, J H.; Tao, N J., Electrochemical GateControlled Charge Transport in Graphene in Ionic Liquid and Aqueous Solution Journal of the American Chemical Society 20 09, 131, 9908-+ 26 Chen, F.; Xia, J L.; Tao, N J., Ionic Screening of Charged-Impurity Scattering in Graphene Nano Letters 20 09, 9, 1 621 -1 625 27 Chen, F.; Xia,... Dielectric Screening Enhanced Performance in Graphene FET Nano Letters 20 09, 9, 25 71 -25 74 28 Du, X.; Skachko, I.; Barker, A.; Andrei, E Y., Approaching ballistic transport in suspended graphene Nature Nanotechnology 20 08, 3, 491-495 29 Bard, A J.; Faulkner, L R., Electrochemical Methods: Fundamentals and Applications 2nd ed.; John Wiley & Sons, Inc.: New York, 20 01 30 Das, A.; Pisana, S.; Chakraborty,...   A K., Monitoring dopants by Raman scattering in an electrochemically topgated graphene transistor Nature Nanotechnology 20 08, 3, 21 0 -21 5 31 Berciaud, S.; Ryu, S.; Brus, L E.; Heinz, T F., Probing the Intrinsic Properties of Exfoliated Graphene: Raman Spectroscopy of Free-Standing Monolayers Nano Letters 20 09, 9, 346-3 52 32 Lazzeri, M.; Mauri, F., Nonadiabatic Kohn anomaly in a doped graphene monolayer... layer), 0.143 ± 0.0 12 (two layers), 0 .21 1 ± 0.013 (three layers) and 0 .28 2 ± 0.015 (four layers), which increased approximately linearly and showed a saturation after 10 layers 0.40 0.7 0.6 4 0.30 0.5 0.4 0.3 0 .25 Contrast (c) 0.8 0.35 3 0 .20 2 0.15 0.10 0 .2 0.1 1 2 3 4 5 6 7 8 9 10 Graphene Layers 1 Intensity (a.u.) (b) Contrast (a) 4 layers 3 layers 0.05 Height / nm 2 1 2 layers 0.00 3 1 layer 400 0 -1... suspended graphene (Figure 4.8b).31 In addition, based on the change in the slope of the σ-Vg plot, an increase in hole mobility from 59 cm2/(V s) to 460 cm2/(V s) and increase in electron mobility from 17 cm2/(V s) to 310 cm2/(V s) could be estimated There was also an apparent reduction in the asymmetry of the σ-Vg curve and an increase in the Ion/Ioff ratio from 1.5 to 10 2 Conductivity (mS) 0 .25 0 .20 Air... 700 Wavelength (nm) 800 1000 120 0 1400 1600 -1 Raman Shift (cm ) 1800 20 00 -2 -3 (d) 0 5 1 0 15 20 2 5 L a t e r a l D is ta n c e /  m 1 30 2 3 4 2 88    Figure 4.6 Characterisation of graphene film thickness and morphology (a) Tapping mode AFM characterisation of graphene film with film thickness of ~3 nm prior to annealing (b) Contrast spectra for 1 to 4 layers of graphene sheets Inset shows calibration... several aspects Figure 4.10a shows the σ-Vg characteristic of graphene film in different KCl concentration When graphene was exposed to 10 mM KCl, the slope of the σ-Vg increased sharply, similar to the case of NaF exposure The hole and electron mobility increased initially from 59 cm2/(V s) to 300 cm2/(V s) and from 17 cm2/(V s) to 130 cm2/(V s), respectively However, one interesting difference was . (supernatant) 28 0 28 2 28 4 28 6 28 8 29 0 29 2 29 4 C=O (28 7.8eV) Intensity (a.u.) Binding energy (eV) C(epoxy/ether) (28 6.5eV) C-C (28 4.5eV) (66%) 86  Figure 4.4. Optical micrographs and tapping. the graphene edge planes. Due to different solubilities and lateral dimensions, the smaller and highly oxidised graphene sheets 28 0 28 4 28 8 29 2 66 % 81 % 92. 1 % 55 % C = O (28 7.8 eV) C - O (28 6.7. 800 0.00 0.05 0.10 0.15 0 .20 0 .25 0.30 0.35 0.40 3 2 1 Contrast Wavelength (nm) 4 (a) (b) 1 2 3 4 2 (d) 123 45678910 0.1 0 .2 0.3 0.4 0.5 0.6 0.7 0.8 Contrast Graphene Layers 1000 120 0 1400 1600 1800 20 00 4