Article pubs.acs.org/JPCC
Work Function Tuning of Reduced Graphene Oxide Thin Films Lamprini Sygellou,*,† Georgios Paterakis,† Costas Galiotis,†,‡ and Dimitrios Tasis†,§ †
Institute of Chemical Engineering Sciences (ICE-HT), Foundation of Research and Technology, Hellas, P.O. Box 1414, 26504 Rio Patras, Greece ‡ Department of Chemical Engineering, University of Patras, 26504 Rio Patras, Greece § Department of Chemistry, University of Ioannina, 45110 Ioannina, Greece ABSTRACT: Graphene oxide (GO) has shown great potential as a component in various devices due to its excellent solution processability and two-dimensional structure. However, the oxygenated form of graphene has a moderate charge-transport capability. The latter parameter may be enhanced through controlled deoxygenation of GO with subsequent tuning of its work function (WF). Various reduction approaches were employed to investigate the effect of the oxygen content on the work function of GO derivatives as thin films on an indium tin oxide substrate. Such films were reduced by stepwise thermal annealing in ultrahigh vacuum up to 650 °C, by chemical reduction with hydrazine, or by a combination of chemical and thermal reduction processes. The effect of the GO film thickness and the flake size on the WF was also investigated. UV photoelectron spectroscopy and X-ray photoelectron spectroscopy were used to correlate the WF of GO derivatives with their oxygen content. The results showed that the WF is strongly dependent on the oxygen content, reaching a ∼1 eV difference between GO and highly reduced GO, under the specific reduction conditions. The film thickness affects the work function, since in thin films interaction with the substrate is pronounced. Finally, the WF of reduced GO after combination of chemical and thermal reduction reaches its lowest value of 4.20 eV, due to the presence of heteroatoms which doped the surface.
■
INTRODUCTION Graphene is a two-dimensional nanostructure that possesses unusual intrinsic properties. Due to its conjugated network of sp2-hybridized carbon atoms, it is considered a conducting material. This has generated unprecedented interest in the development of devices in applications related to the electronics industry. For such applications, modulation of graphene’s work function (WF) value is critical in achieving the desired properties. WF is a fundamental electronic property of any material and provides understanding of the relative position of the Fermi level. WF tuning of graphene-based thin films is a key requirement in organic electronic devices because the carbon nanostructures are used as interfaces between an electrode material and an organic semiconductor.1,2 By using surface-sensitive techniques, such as X-ray photoelectron spectroscopy (XPS) and/or UV photoelectron spectroscopy (UPS), significant progress has been made toward the understanding of the factors governing the WF tuning of graphene-based assemblies. The ability to control the WF of graphene-based nanostructures is a very important asset in applying them as electrode materials. Especially for cathode components, materials with relatively low WF values are needed. Regarding graphene oxide (GO)-based thin films, it has been shown that WF is highly dependent on the oxygen content as well as functionality speciation.3 Several approaches have been adopted for preparing deoxygenated GO flakes, which leads to relatively lower WF values than that of the © 2015 American Chemical Society
parent oxygenated form. These include hydrazine-based reduction,4 combination of chemical and thermal reduction,5 modification with Cs2CO3,6 electrochemical reduction,7 Au nanoparticle decoration,8 synchrotron soft X-ray irradiation,9 and photoreduction under visible or UV irradiation.10 A theoretical study with molecular dynamics and density functional theory (DFT) calculations revealed that WF values vary by up to 2.5 eV in GO-based structures by precisely controlling the oxygen-containing functional groups. It was demonstrated that carbonyl groups show the largest impact on graphene’s work function among the oxygen-containing groups, inducing a WF of 6.8 eV, whereas the corresponding values for epoxy and hydroxy groups were 5.6 and 4.95 eV, respectively.11 Experimentally, the WF of a GO-based gate electrode in a complementary metal oxide semiconductor (CMOS) device was shown to be modulated from 4.35 to 5.28 eV by sandwiching different thicknesses of reduced graphene oxide (rGO) layers between the top contact metals and the gate dielectric SiO2.12 The WF of the rGO-based gate electrode showed a strong dependence on the rGO thickness as well as the oxygen concentration of the graphitic lattice.12 Alternatively, graphene oxide and its derivatives have been used as interfacial layers in polymer solar cells.2 In such devices, the WF Received: September 22, 2015 Revised: December 8, 2015 Published: December 8, 2015 281
DOI: 10.1021/acs.jpcc.5b09234 J. Phys. Chem. C 2016, 120, 281−290
Article
The Journal of Physical Chemistry C
differences in the size of the flakes, (b) the GO film thickness, and (c) the reduction conditions.27
of the hole-transporting layer must be close to that of the donor’s HOMO (highest occupied molecular orbital), which should be more than 5 eV, whereas the WF of the electrontransporting layer must be close to that of the acceptor’s LUMO (lower unoccupied molecular orbital), which should be close to 4.5 eV.2 Very recently, interest has emerged in the development of GO-based thin films as efficient holetransporting layer (HTL) components for high-performance polymer solar cells (PSCs).13−17 Alkylsilane-doped rGO films showed wide tunability of their WF values, which could substantially enhance the performance of organic field-effect transistors12,18 or ambipolar flash memory devices.19 Concerning the latter application, an alternative dopant involved gold chloride, which was used to tune the WF of the rGO-based floating gate component. 20 Apart from using physical adsorption strategies (doping) to tune the WF of GO-based thin films, various reduction protocols have been applied, giving rise to the desired electron transport properties.21 The effect of combined chemical reduction and thermal annealing of GO films a few nanometers thick on their conductivity was investigated in various gaseous environments (Ar, H2, N2, or air).22−24 It was shown that optimized sheet resistance values were obtained with the GO films chemically reduced with hydrazine and subsequently annealed at 1000 °C in an inert atmosphere.24 The hydrazine-assisted chemical reduction results in the decoration of the graphitic lattice with nitrogen-based functionalities, whereas the subsequent thermal treatment eliminates to a great extent the heteroatomcontaining labile groups of GO. Depending on various parameters, such as the rate and duration of heating, and the final temperature, a variety of heteroatom-based functionalities remain attached onto the basal plane and edges of the graphitic nanostructure. The type and concentration of such functionalities have a decisive influence on the WF of graphene. Experimentally, the study of WF tuning of GO as a function of the oxygen content was investigated by contact potential difference (CPD) using a scanning Kelvin probe method (SKPM).25 The WF values were indirectly estimated and could be converted to absolute values by an appropriate calibration. The aim of this work is to correlate the absolute WF values of GO-based thin films with the heteroatom concentration in different conditions. The graphitic assemblies were reduced with either thermal (annealing in ultrahigh vacuum (UHV)) or chemical (hydrazine-induced) treatment. Also, a combination of both processes was studied. In the case of thermal reduction, two parameters were studied: (a) the influence of the graphite oxidation protocol and (b) the effect of the GO/rGO film thickness. For the former parameter, two different approaches of GO production were used, leading to the formation of either small-sized (500 nm) or large-sized (a few micrometers) flakes. For a specific batch, three different thicknesses were investigated, namely, ∼4, ∼7, and ∼10 nm, which are the most commonly used in devices.26 To correlate the WF with the heteroatom content and speciation, XPS and UPS were used in each reduction step. In the chemically reduced films, the duration of hydrazine treatment was varied, whereas for the combined protocol of chemical and thermal reduction, a chemically reduced rGO/indium tin oxide (ITO) film was stepwise annealed in UHV. To our knowledge, this is the first experimental work in which the heteroatom content and speciation in GO-based films have been correlated with the WF values of the nanostructures. The modulation of the WF was studied in relation to (a) the GO synthetic protocol, leading to
■
EXPERIMENTAL SECTION GO sheets were prepared by two different oxidation protocols, starting from natural graphite (NGS Naturgraphit GmbH, Germany). The first involves a modified Hummers method,28 comprised of two steps, giving rise to small-sized GO flakes (average 500 nm). Details of the GO synthesis are given in a previous paper by our group.29 The second batch was prepared by a one-step oxidation reaction,30 by which large GO flakes were produced (a few micrometers in size). To isolate the large flakes, the graphite oxide aqueous suspension was put into a bath sonicator for 5 min. This suspension was subjected first to centrifugation at about 3000 rpm for 5 min to discard the large graphitic aggregates. Then the supernatant suspension was centrifuged at 5000 rpm for 5 min to separate the large flakes, which were precipitated on the bottom, and the smaller ones stayed suspended in solution. The precipitated material was redispersed by short-time sonication and used for further study. Glass/ITO substrates were purchased from Visiontek Systems Ltd. with dimensions of 500 × 400 mm2 and a sheet resistivity of 12 Ω/sq. The substrates were cut into smaller pieces (1 × 1 cm2) and cleaned gradually with soapy water, deionized water, acetone, and ethanol (one time each) and three times with deionized water in a bath sonicator for 10 min at each step. After these wet cleaning processes and just before GO film deposition, the substrates were subjected to oxygen plasma treatment (rf (radio frequency) = 10.5 W, 20 min) to remove carbon contamination and to improve surface wetting and hence the coating homogeneity. Finally, the GO films were coated on the top of ITO by spin coating from an aqueous GO dispersion. To investigate the influence of the thickness on the oxygen concentration and work function of rGO/ITO films, three different GO thicknesses of small flakes were prepared, namely, ∼4, ∼7, and ∼10 nm. Chemically reduced GO films were prepared following a previously published method,31 through the incubation of the GO film in vapors of hydrazine. After the fabrication of the GO film onto ITO/glass, the sample was put in a clean glass Petri dish, which contained a smaller glass Petri dish with 1 mL of hydrazine monohydrate (99+%, Alfa Aesar). The large Petri dish was covered by a glass lid, sealed with Parafilm, and placed over a hot plate at 40 °C for 18 h. After this process, the color of the film was changed from light brown to metallic gray, indicating reduction of GO. Finally, the sample was dried under nitrogen flow at 80 °C overnight to remove hydrazine residuals. Concerning the thermal reduction process, the samples were mounted onto a stainless steel sample holder able to be heated up to 650 °C within the XPS instrument. First, the effect of oxygen plasma treatment on the bare ITO substrate was checked, and it was found that carbon contamination was removed from the surface and the work function increased from 4.0 eV (contaminated surface) to 4.5 eV (plasma-treated surface). The GO/ITO samples were stepwise heated from room temperature to 650 °C in the analysis chamber of the XPS instrument. The heating steps were not constant for the temperature heating range, that is, 25 °C for temperatures up to 250 and 50 °C for temperatures from 250 to 650 °C, and the duration was 10 min. Moreover, bare ITO substrate was thermally treated up to 650 °C in UHV, and XPS/UPS measurements were recorded in each temperature step. 282
DOI: 10.1021/acs.jpcc.5b09234 J. Phys. Chem. C 2016, 120, 281−290
Article
The Journal of Physical Chemistry C
wrinkles are formed during deposition onto the transmission electron microscopy (TEM) grid. After deposition of GO thin films onto the ITO substrate, their average thickness was assessed by XPS-based calculations, supported by atomic force microscopy (AFM) imaging. The average thickness of as-deposited GO films studied was found to be ∼4, ∼7, and ∼10 nm, respectively. The thickness of each prepared GO film was the result of both the concentration of the suspension and the rotational speed, during the spincoating process. The chemical state of the samples in each reduction step was investigated by XPS. As mentioned in the Experimental Section, plasma pretreatment on the bare ITO surface leads to significant removal of carbon contamination, so the C 1s peak intensity arises only from the deposited GO. Survey scans (not shown) of all the GO/ITO samples reveal only the presence of C, O, In, and Sn peaks. Figure 2 shows the corresponding C 1s and O 1s core level spectra of ∼4 nm GO films on ITO (small and large flakes) as well as the ∼10 nm GO/ITO sample (small flakes). Each C 1s peak is analyzed into the following five components: C−C sp2 and defective sp3 bonds at binding energies of 284.7 ± 0.05 eV and 285.6 ± 0.05 eV, respectively, carbon−oxygen components at 287.0−286.4 eV assigned to both hydroxyl and epoxides (C−OH, C−O− C),33 carbonyls (CO) at 288.5−288.7 eV, and carboxyls (OCOH) at 289.8−289.2 eV.29 In an analogous manner, the O 1s peak is analyzed into four components: the peak at a binding energy of 530.65 ± 0.05 eV was assigned to In−O bonds from the ITO substrate, and another three components were assigned to carbon−oxygen bonds, namely, components at 531.0−531.5 eV (CO bonds) and 532.7−533.2 eV (C−O bonds) and a low-intensity component at 534.7−535.2 eV (adsorbed H2O).29,34 From Figure 2A, it is obvious that the intensity ratio of the epoxide/hydroxide component to the corresponding C−C component (sp2 + sp3) is enhanced in the case of the large-flake sample (thickness 4 nm) and for the 10 nm thickness small-flake sample. Specifically, the aforementioned intensity ratio is ∼1 for the 4 and 7 nm small-flake samples and ∼1.3 for the 10 nm small-flake sample as well as the 4 nm large-flake sample. From this result, we conclude that the thick films and large-sized flakes trapped more epoxides and hydroxides than the thin films and small-sized flakes. Concerning the O 1s peaks, the only difference is the absence of the In−O component in the thick films due to the detection depth of the XPS technique. The C 1s and O 1s photopeaks were deconvoluted in the above components after each thermal reduction step. The results for all samples showed that annealing at UHV conditions resulted in appreciable loss of oxygen concentration, in agreement with our previous studies.29 This is evident from the intensity loss of the O 1s peak. The total intensity of the C 1s peak is stable after each heating step up to 650 °C, while the intensity of the carbon− oxygen components (C−OH, C−O−C, CO, OCOH) decreased and that of the carbon−carbon components (sp2/ sp3) increased. This result demonstrates that the heating in ultrahigh vacuum conditions leads to oxygen atom desorption but not carbon atom desorption (CO or CO2). To investigate the effect of the GO thickness on the thermal reduction mechanism, the atomic oxygen concentration (%), derived from the ratio of the O 1s and C 1s intensities, as well as the partial fractions of both C−O and CO components derived from O 1s peak fitting, was plotted versus temperature (Figure 3). The experimental error in the atomic concentration values was estimated to be about 10%. From the batch of small
Photoelectron spectroscopy measurements were carried out in a UHV chamber with a SPECS LHS-10 hemispherical electron analyzer. An unmonochromatized Al Kα line at 1486.6 eV and an analyzer pass energy of 36 eV giving a full width at half-maximum (fwhm) of 0.9 eV for the Ag 3d5/2 peak were used. The analyzed area was 2.5 × 4.5 mm2. XPS core level spectra were analyzed using a fitting routine, which can decompose each spectrum into individual mixed Gaussian− Lorentzian peaks after a Shirley background subtraction. The O/C relative atomic concentration in the analyzed region was calculated by dividing the peak areas of O 1s and C 1s by the appropriate relative sensitivity factor (RSF) after correction for the experimentally determined EA10 analyzer transmission characteristics (effectively the average matrix RSF). To subtract from the total O 1s peak intensity the contribution of O 1s from ITO, a blank experiment was performed, where the ITO/glass sample was heated up to 650 °C and the In 3d and O 1s peak intensities were recorded. The GO average thickness was estimated by dividing the intensity ratio of C 1s and In 3d5/2 spectra and using the appropriate equations.32 The UPS spectra were obtained using HeI irradiation with hν = 21.22 eV produced by a UV source (model UVS 10/35). During UPS measurements, the analyzer worked in the constant retarding ratio (CRR) mode, with CRR = 10. A bias of −12.29 V was applied to the sample to avoid interference of the spectrometer threshold in the UPS spectra. The high and low binding energies and HOMO cutoff positions were assigned by fitting straight lines on the high and low energy cutoffs of the spectra and determining their intersections with the binding energy axis. Regarding measurement errors, it should be noted that an error of ±0.05 eV is assigned to the absolute values for the work function and other UPS spectral cutoff features.
■
RESULTS AND DISCUSSION The morphology and size distribution of as-synthesized GO flakes, prepared by both approaches, are shown in Figure 1. In
Figure 1. TEM images of (a) GOsmall and (b) GOlarge batches.
general, both the size distribution and oxygen atomic ratio in GO samples strongly depend on the severity of conditions taking place during the oxidation reaction. These include the oxidative character of the reactants as well as the extent of sonication for the exfoliation of graphite oxide multilayers toward the preparation of graphene oxide in solution. It is clearly seen that the two-step oxidation protocol results in relatively small GO flakes (average size 500 nm) (Figure 1a). By using a one-step oxidation protocol, combined with minimized sonication treatment, we were able to prepare large GO flakes on the micrometer scale (see Figure 1b). Due to the relatively large size of the GO sheets, it is observed that 283
DOI: 10.1021/acs.jpcc.5b09234 J. Phys. Chem. C 2016, 120, 281−290
Article
The Journal of Physical Chemistry C
Figure 2. Deconvoluted (A) C 1s and (B) O 1s XP core level spectra of (a) 10 nm GO small flakes, (b) ∼4 nm GO small flakes, and (c) ∼4 nm GO large flakes on ITO.
Figure 3. Atomic oxygen concentration (%), C−O bond concentration (%), and CO bond concentration (%) for (a) 7 nm small flakes, (b) 4 nm large flakes, (c) 10 nm small flakes, and (d) 4 nm small flakes. The error in the atomic concentration values is 10%.
flakes, calculations have been performed for all three films with different thicknesses, whereas from the batch of large flakes, only one sample was studied (∼4 nm thickness). The total atomic oxygen concentration (%) curves may be divided into two different temperature windows. First, a pronounced loss of oxygen moieties was observed in the temperature range between 50 and 200 °C. This was ascribed to the elimination of trapped water as well as desorption of oxygen-containing labile groups from the graphitic lattice.27 Specifically, for each carbon−oxygen component, it is obvious that appreciable loss of C−O functionalities (hydroxides and epoxides) takes place, whereas the CO concentration (%) is constant up to 150 °C. At temperatures between 150 and 200 °C, the CO concentration (%) increases for all the investigated samples with variable thicknesses. This indicates carbonyl and/or carboxyl formation simultaneously with epoxide/hydroxyl loss. Formation of carbonyls with parallel epoxide loss has been observed by heating of single-layer GO at 175 °C in UHV
conditions.34 The profile of all four curves in this region showed that there is no dependence on either the GO film thickness or the GO flake size. In the second temperature window, between 200 and 650 °C, the rate of loss of atomic oxygen concentration seems to be highly dependent on the thickness of the GO film. Specifically, the total oxygen concentration continuously decreased for the 4 and 7 nm thickness GO samples, whereas the 10 nm GO film showed a thermal stability in the aforementioned temperature region. Concerning the partial components, the CO concentration (%) decreases with a variable thickness-dependent decay rate, whereas the C−O concentration (%) profile is somewhat similar to that of total oxygen. The thermal stability of the thick GO films (10 nm) could be explained by steric hindrance phenomena. Due to the small size of the GO flakes, it is anticipated that they are packed homogeneously during deposition, creating a confined environment for the oxygen-containing moieties. Thus, it was observed 284
DOI: 10.1021/acs.jpcc.5b09234 J. Phys. Chem. C 2016, 120, 281−290
Article
The Journal of Physical Chemistry C
°C under UHV (rGO/ITO, solid line). In Figure 4a, the high binding energy cutoff is shown, where the WF of the surface can be determined by subtracting the spectrum’s width (i.e., the energy difference between the Fermi level and the high binding energy cutoff) from the HeI excitation energy. In Figure 4b, the valence band region is shown, whereas the region near the spectrometer’s Fermi level is shown magnified in the inset. The GO valence band region is dominated by a peak at around 6 eV assigned to O 2p states35 and the absence of states in the Fermi level as shown in the inset. The estimated work function value is 5.4 eV. After thermal reduction at 400 °C (solid lines), several changes in the spectra occurred. In the valence band region, a new peak at ∼3 eV appeared which originates from the hybridization of the 2pπ state. In addition, the O 2p peak at ∼6 eV vanished, and a density of states in the Fermi level (0 eV) appeared.36 These changes were observed after heating of any sample at temperatures above 150 °C and become more pronounced as the temperature increases. Finally, the work function decreased to 4.5 eV, as shown in Figure 4a. The above changes in the UPS spectra resulted from GO thermal reduction. The effect of both the reduction temperature (Figure 5a) and the oxygen content (Figure 5b) on the variation of the WF values of GO derivatives was investigated. Separate curves are demonstrated for GO-based films possessing different thicknesses and GO flakes with different flake sizes. The work function of all the GO films at room temperature was about 5.3 ± 0.2 eV. In Figure 5a, the WF variation of plasma-pretreated bare ITO with temperature is also provided. Heating up to 350 °C resulted in a WF decrease by 0.3 eV, whereas a slight increase (0.15 eV) took place by subsequent heating to 650 °C. The WF decrease in the plasma-treated substrate was attributed to the desorption of excess hydroxyls from the surface, reaching the stoichiometric atomic ratio of In to O at a temperature of about 250 °C. The slight increase in the WF for temperatures greater than 350 °C is accompanied by changes in the valence band of the ITO substrate and is attributed to changes in the
that the GO film thickness is an important factor in the thermal deoxygenation of such assemblies. The final oxygen content was 20% for the 10 nm, 10% for the 7 nm, and 5% for the 4 nm thickness small-sized GO flakes. The corresponding oxygen content for the 4 nm large-sized GO flakes was 7%. This is in agreement with the findings of a recent study in which the authors suggested that, in multilayer GO films, the layers below the first layer are reduced less effectively.12 Concerning the effect of the flake size in the high-temperature zone (above 200 °C), there is no appreciable difference between the corresponding profiles. Figure 4 shows representative UPS curves of 4 nm GO/ITO at room temperature (dashed line) and after annealing at 400
Figure 4. HeI UPS spectra of ∼4 nm small-flake GO/ITO at room temperature (dashed lines) and after heating at 400 °C (rGO/ITO, solid lines). (a) Secondary electron cutoff where the work function was derived. (b) Valence band region. In the inset, the region near the spectrometer’s Fermi level (low binding energy side, where electrons from the highest molecular orbital are ejected) is shown magnified for clarity.
Figure 5. Work function versus (a) temperature and (b) oxygen content for 4, 7, and 10 nm GO/ITO (small flakes) and 4 nm GO/ITO (large flakes). The WF evolution of the plasma-pretreated bare ITO substrate at various temperature steps is also shown in panel a (bottom). In the inset of panel b, the thermal decay profiles of CO functions for all thicknesses are given. 285
DOI: 10.1021/acs.jpcc.5b09234 J. Phys. Chem. C 2016, 120, 281−290
Article
The Journal of Physical Chemistry C structure.37 It is obvious that the WF of the ITO substrate does not affect the WF values of deposited rGO since the latter are higher than that of ITO. In addition, the shapes of the curves are similar regardless of the rGO thickness. It is observed that the GO work function depends on the temperature, on the oxygen content and speciation, and on the film thickness and flake size. Considering the temperature dependence, the work function decreases rapidly in all samples by heating up to 200 °C (Figure 5a), at which temperature the total oxygen has decreased appreciably as shown in Figure 3. This behavior leads us to the conclusion that the WF is strongly dependent on the population of hydroxide and epoxide moieties, which are the dominant species at temperatures below 200 °C. At higher temperatures, and specifically in the range between 200 and 450 °C, approximately, a less steep slope of WF decrease was observed for the three samples with variable thickness (small flakes). Furthermore, for temperatures up to 650 °C, the WF is stable within experimental error. It is noted that the minimum WF value, being at 4.50 ± 0.05 eV for all samples, was reached at different reduction temperatures. The minimum WF value for the thick film (10 nm) was achieved at lower temperature than that of the thin film (4 nm). This implies a clear effect of the film thickness on the modulation of the WF values for the same flake size. Furthermore, for differently sized rGO flakes, the work function minimum is reached at similar temperatures. Table 1 shows the temperature, the total oxygen content in the film, and the oxygen loss (%) at which the lowest WF value was first recorded.
eventually in the WF evolution. To understand why a thick sample reaches its WF minimum value at lower temperature than that of a thin sample, we have to analyze each parameter separately. It is noted that, when reaching the WF plateau at about 4.5 eV, the total oxygen content of the GO films continues to decrease. This is more pronounced for the 7 nm and 10 nm thick samples (see Figure 5b). This observation clearly suggests that the total oxygen concentration is not the main parameter controlling the evolution of the work function.11 As mentioned above, by comparing the decay profiles of “C−O-type” thin (4 nm) and thick (10 nm) films in the temperature range of 200−550 °C (Figure 3), we observed that the C−O concentration did not vary. On the contrary, monitoring the “CO-type” concentration decay curves showed some similarity to the corresponding curves of work function evolution. It is suggested that, under the specific conditions of our thermal annealing process, the CO-type decay profile may govern the evolution of the work function to a great extent. Specifically, we observe that, when the temperature-dependent decay rate of the CO-type concentration reaches an infinitesimal value approaching zero, the WF value reaches its plateau in the very same temperature zone as shown in the inset of Figure 5b, where the CO concentration versus temperature in shown. In addition to the aforementioned parameters, especially for the case of thin films, the WF depends on the substrate, where substrate polarization effects take place near the interface.39 Taking these into account, we came to the conclusion that, in the case of the small-flake samples, a decrease of the film thickness results in a change of the electron density at the rGO/ITO interface (unsymmetric potential thin film system). The thermal reduction profiles of the films with similar thicknesses (∼4 nm GO/ITO) but different flake sizes are similar (Figure 3); therefore, no differences in the chemical environment that could affect the WF values are likely to occur. Moreover, it has been shown that the work function depends on the perimeter size of the graphene nanoflakes because of the presence of the edges only when the perimeter size is smaller than 4 nm.40 Hence, in the present case, we do not expect the differences in the sample’s perimeters to affect the measured WF values. It is possible that the differences in the WF are due to the substrate contribution, which would have a larger effect on the small flakes. Of course, this is a not a conclusion but merely a suggestion and should be investigated in detail in a subsequent study. These results would be very useful for the device properties, i.e., field emission performance, since the oxygen content and film thickness affect the conductivity, whereas the work function affects the field emission characteristics.41 Therefore, depending on the desired work function value, one can choose the heating temperature and thickness and vice versa. Very interesting results were obtained by investigating the combined chemical and thermal reduction of a GO film (4 nm thickness) by XPS/UPS. The XPS data showed that, besides oxygen moieties, nitrogen-based species were present on the graphitic surface after hydrazine treatment at 40 °C and remained even after heating to 650 °C in UHV conditions. From the intensities of the N 1s, C 1s, and O 1s peaks, the atomic ratios (%) of N and O were calculated. Figure 6 shows the work function evolution (left axis) and the nitrogen and oxygen atom concentration (%) variation (right axis) after hydrazine treatment at 40 °C and subsequent gradual heating in UHV to 650 °C. In the same figure, the work function
Table 1. Annealing Temperature, Oxygen Concentration, and Oxygen Loss after Reduction for the Different GO Thicknesses and Flake Sizes Where the Minimum Value of WF = 4.50 ± 0.05 eV Is Reached GO flake thickness and size
annealing temp (±10 °C)
10 nm small flakes 7 nm small flakes 4 nm small flakes 4 nm large flakes
280 350 420 440
oxygen concn (%) 22 16 12 19
± ± ± ±
2 2 1 2
oxygen content loss (%) 48 55 68 55
± ± ± ±
5 6 7 6
Regarding the oxygen content dependence (Figure 5b), for the group of films comprised of small-sized GO sheets, the WF value of a thin film is larger than that of a thick film, for the same oxygen content. Concerning the flake size of the GO film, the WF of rGO large flakes is lower than the WF of small flakes for similar oxygen content values. To try to explain the variations of the WF values by thickness and flake size, certain assumptions have to be made. Generally, the work function of a material depends on both of the following parameters: (a) the chemical potential (such as doping concentrations, chemical changes, oxidation states, and defect sites)38 and (b) the surface dipole, which represents an additional electrostatic barrier to removing an electron from a solid surface.38 Therefore, changes in the chemical composition of the film are expected to lead to changes in the work function. To clarify the temperature effect, we suggest a plausible scenario which also incorporates the decay profiles of oxygencontaining moieties during gradual heating. Since the duration of each thermal annealing step is limited to 10 min, it is logical to assume that steric hindrance and confinement play a primary role in the decay profiles of oxygen functionalities and 286
DOI: 10.1021/acs.jpcc.5b09234 J. Phys. Chem. C 2016, 120, 281−290
Article
The Journal of Physical Chemistry C
To give insight into the understanding of WF evolution through the combined reduction process, deconvolution of N 1s peaks and O 1s peaks, analyzed in a variety of heteroatomcontaining species, was performed. Figure 7 shows the N 1s and O 1s peaks analyzed in various nitrogen-containing species and oxygen species at room temperature as well as at some representative elevated temperatures (250, 550, and 650 °C). Specifically, the O 1s components attributed to the rGO film are reduced due to thermal reduction. The N 1s peak is decomposed into three components at binding energies of 398.5 ± 0.1, 400.3 ± 0.1, and 402.1 ± 0.2 eV, which are assigned to pyridinic, amine/amide/pyrrolic, and nitrogen oxide moieties, respectively.42 This N 1s peak deconvolution is applied at every temperature, and the absolute percentage of each component versus temperature is shown in Figure 8. After
Figure 6. Work function (black line) and nitrogen (red line) and oxygen (blue line) atomic concentrations (%) of a 4 nm small-flake GO film chemically treated with hydrazine and stepwise heated in UHV. The WF of 4 nm small-flake thermally reduced GO (dashed black line) is shown for comparison.
evolution of 4 nm small-flake GO/ITO, reduced exclusively by gradual heating in UHV, is shown for comparison. The profiles of oxygen concentration decay are similar for 4 nm thick GO films, experiencing either exclusively thermal reduction or combined chemical/thermal reduction, whereas the same stands for the partial components, C−O and CO. Concerning the nitrogen content, a moderate decrease was observed up to ∼350 °C, while at higher temperatures, the nitrogen content (%) is relatively stable. The work function of hydrazine-treated GO at 40 °C is 4.45 eV and is the same for the corresponding small-flake 4 nm GO sample heated gradually to about 600 °C. Subsequent heating of the hydrazine-treated sample in UHV results in a WF decrease, reaching a minimum of 4.2 eV at 250 °C. Further heating gave rise to a WF increase to about 4.35 eV at 500 °C, and then the WF reached a plateau up to 650 °C. The evolution of the WF in the combined reduction approach seems to be a result of both speciations of functionalities as well as the absolute values of the heteroatom contents at each temperature.
Figure 8. Absolute atomic concentration (%) of nitrogen-based species of the 4 nm small-flake GO film reduced by the combined approach vs temperature.
hydrazine treatment at 40 °C for 18 h, among the nitrogenbased functionalities, amine/amide/pyrrolic moieties are the dominant species. A similar distribution of nitrogen-based functionalities was observed in the early stages of ammonia plasma treatment of monolayer GO sheets.43 The only
Figure 7. Deconvoluted XPS (A) N 1s and (B) O 1s core level peaks of 4 nm GO small flakes (a) chemically treated with hydrazine and (b−d) heated in UHV for 10 min at (b) 250 °C, (c) 550 °C, and (d) 650 °C. 287
DOI: 10.1021/acs.jpcc.5b09234 J. Phys. Chem. C 2016, 120, 281−290
Article
The Journal of Physical Chemistry C
Figure 9. (A) Work function (black line), nitrogen (red line) and oxygen (blue line) concentrations (%), and sum of the nitrogen and oxygen (green line) absolute atomic concentrations (%) of a 4 nm small-flake GO film chemically treated with hydrazine vapors at different times. (B) Deconvoluted XPS N 1s core level peaks of 4 nm GO small flakes chemically treated with hydrazine for (a) 2 h, (b) 4.5 h, and (c) 9 h. (C) Relative atomic concentration (%) of pyridinic and amine/amide/pyrrolic nitrogen and nitrogen oxides of the samples treated with hydrazine vapors at different times.
content of pyrrolic groups decreased appreciably, when compared with their initial concentration after hydrazine treatment. This resulted in a further increase of the work function to 4.35 eV. At higher temperatures (above 500 °C), there is no noticeable change in both the pyrrolic and pyridinic contents, which results in a constant WF value. To investigate the time dependence of the chemical reduction process itself, three additional samples of 4 nm GO/ITO were treated with hydrazine vapors for various times, namely, 2, 4.5, and 9 h, respectively. These were compared with the sample which was reduced chemically for a period of 18 h and was further reduced by gradual heating. After chemical treatment, the rGO samples were characterized by XPS/UPS measurements. Figure 9A shows the work function evolution and the nitrogen and oxygen absolute atomic concentrations (%) as well as the sum of both heteroatoms. Figure 9B shows the deconvoluted N 1s peaks of the 2, 4.5, and 9 h samples and Figure 9C the relative atomic concentrations (%) of pyridinic and amine/amide/pyrrolic nitrogen and nitrogen oxide moieties of all samples, including the 18 h one. The results show that the duration of hydrazine treatment affects the atomic nitrogen and oxygen concentrations (%) but not the sum, although the atomic carbon concentration is stable. As the reduction time increased, the atomic nitrogen concentration increased and the work function decreased, reaching its lowest value after 9 h of hydrazine treatment. However, it is noted that further hydrazine treatment up to 18 h leads to an additional atomic nitrogen concentration increase, while the work function actually remains stable at about 4.45 eV. From the N 1s intensity components (Figure 9C), the duration of hydrazine treatment does not affect the relative concentrations of nitrogen-containing components significantly; only a small increase of nitrogen oxides was observed.
difference in our work is that no graphitic nitrogen is expected to be produced under the specific conditions of the combined reduction approach. The pyrrolic species were found to decrease appreciably up to 500 °C and remained stable at 500−650 °C. On the contrary, a slight concentration decrease was observed for the nitrogen oxides from 40 to 650 °C. The third component, the pyridinic nitrogen, was found to be relatively stable at low content up to 250 °C, increased notably up to 400 °C, and reached a plateau up to 650 °C. It is seen that, for temperatures up to about 250 °C, pyrrolic N is the dominant species, a condition which seems to be partially responsible for the slight decrease (0.25 eV) of the WF. This is supported by the work of Singh et al.,43 in which an increase of the pyrrolic content in the initial stage of ammonia plasma treatment and its dominance over the other nitrogen species resulted in an appreciable WF decrease by about 1 eV. By trying to assess the partial contribution of oxygen species to the WF decrease, we made some rough calculations. The oxygen concentration shown in Figure 6 after hydrazine treatment and after heating at 250 °C is ∼15% and ∼10%, respectively. Figure 5b shows that, between these two oxygen concentrations of the 4 nm thermally reduced GO, a WF decrease of about 0.15 eV is observed. Thus, we suggest that both nitrogen (mainly pyrrolic) and oxygen species contribute to the 0.25 eV WF decrease after heating of the hydrazine-treated sample to 250 °C. The work function of N-doped reduced GO at 4.20 eV was recorded for hydrazine-pretreated reduced graphene films annealed in a H2/ NH3 atmosphere at 750 °C.44 In the present work, this value was achieved under mild conditions, a combination of ex situ hydrazine treatment and heating at 250 °C in UHV conditions. For temperatures between 300 and 400 °C, the contribution of pyridinic N increases significantly and becomes comparable with that of the pyrrolic moieties. In such a condition, the WF increased slightly (∼0.05 eV) from the minimum value of 4.2 eV at 250 °C. In the temperature window between 400 and 500 °C, pyridinic groups became the dominant species, whereas the
■
CONCLUSIONS A detailed study of the work function evolution of GO derivatives after a chemical and/or thermal reduction process 288
DOI: 10.1021/acs.jpcc.5b09234 J. Phys. Chem. C 2016, 120, 281−290
Article
The Journal of Physical Chemistry C was performed. The GO films were either gradually annealed in UHV or reduced in a combined manner, that is, hydrazine treatment and subsequent thermal reduction. For the thermally reduced films, the reduction mechanism in relation to the film thickness and GO flake size was also performed. The rate of loss of atomic oxygen concentration seems to be highly dependent on the thickness of the GO film for temperatures greater than 200 °C, whereas the flake size did not affect the reduction mechanism. The work function depends on the total oxygen concentration of the film and on the film thickness where in the thick films the WF is lower than that in the thin films for the same oxygen concentration. The films comprised of larger sized flakes have lower work function values than those comprised of smaller sized flakes. The combination of hydrazine treatment and subsequent thermal annealing leads to an even lower work function, with a minimum value at 250 °C. It is suggested that WF evolution is a result of both the absolute values of the heteroatom content at each temperature and speciation of the functionalities. This study clearly demonstrated that the WF of rGO thin films is a very complicated electronic property depending on multiple parameters such as the heteroatom concentration, the thickness, the GO flake size, and speciation. It has provided valuable information for designing materials with fine-tuned electronic and chemical properties for the optimization of device performance. We strongly believe that such materials may be used as either hole- or electron-transport layers in solar cells, but not in the role of photon-harvesting components. In the latter case, WF variations are to be expected due to the instability of the GO derivatives under illumination conditions.10
■
High-Performance Bulk Heterojunction Solar Cells. Adv. Mater. 2012, 24, 2228−2233. (7) Naito, K.; Yoshinaga, N.; Matake, S.; Akasaka, Y. Work-function decrease of transparent conducting films composed of hydrazinereduced graphene oxide and silver nanowire stacked layers by electrochemical treatment. Synth. Met. 2014, 195, 260−265. (8) Khoa, N. T.; Kim, S. W.; Yoo, D.-H.; Kim, E. J.; Hahn, S. H. Sizedependent work function and catalytic performance of gold nanoparticles decorated graphene oxide sheets. Appl. Catal., A 2014, 469, 159−164. (9) Lin, C.-Y.; Cheng, C.-E.; Wang, S.; Shiu, H. W.; Chang, L. Y.; Chen, C.-H.; Lin, T.-W.; Chang, C.-S.; Chien, F. S.-S. Synchrotron Radiation Soft X-ray Induced Reduction in Graphene Oxide Characterized by Time-Resolved Photoelectron Spectroscopy. J. Phys. Chem. C 2015, 119, 12910−12915. (10) Li, B.; Zhang, X.; Chen, P.; Li, X.; Wang, L.; Zhang, C.; Zheng, W.; Liu, Y. Waveband-dependent photochemical processing of graphene oxide in fabricating reduced graphene oxide film and graphene oxide−Ag nanoparticles film. RSC Adv. 2014, 4, 2404−2408. (11) Kumar, P. V.; Bernardi, M.; Grossman, J. C. The Impact of Functionalization on the Stability, Work Function, and Photoluminescence of Reduced Graphene Oxide. ACS Nano 2013, 7, 1638−1645. (12) Misra, A.; Kalita, H.; Kottantharayil, A. Work Function Modulation and Thermal Stability of Reduced Graphene Oxide Gate Electrodes in MOS Devices. ACS Appl. Mater. Interfaces 2014, 6, 786− 794. (13) Li, S.-S.; Tu, K.-H.; Lin, C.-C.; Chen, C.-W.; Chhowalla, M. Solution-Processable Graphene Oxide as an Efficient Hole Transport Layer in Polymer Solar Cells. ACS Nano 2010, 4, 3169−3174. (14) Gao, Y.; Yip, H. L.; Chen, K. S.; O’Malley, K. M.; Acton, O.; Sun, Y.; Ting, G.; Chen, H.; Jen, A. K. Y. Surface doping of conjugated polymers by graphene oxide and its application for organic electronic devices. Adv. Mater. 2011, 23, 1903−1908. (15) Ryu, M. S.; Jang, J. Effect of solution processed graphene oxide/ nickel oxide bi-layer on cell performance of bulk-heterojunction organic photovoltaic. Sol. Energy Mater. Sol. Cells 2011, 95, 2893− 2896. (16) Yun, J.-M.; Yeo, J.-S.; Kim, J.; Jeong, H.-G.; Kim, D.-Y.; Noh, Y.J.; Kim, S.-S.; Ku, B.-C.; Na, S.-I. Solution-Processable Reduced Graphene Oxide as a Novel Alternative to PEDOT:PSS Hole Transport Layers for Highly Efficient and Stable Polymer Solar Cells. Adv. Mater. 2011, 23, 4923−4928. (17) Stratakis, E.; Savva, K.; Konios, D.; Petridis, C.; Kymakis, E. Improving the efficiency of organic photovoltaics by tuning the work function of graphene oxide hole transporting layers. Nanoscale 2014, 6, 6925−6931. (18) Kang, B.; Lim, S.; Lee, W. H.; Jo, S. B.; Cho, K. Work-FunctionTuned Reduced Graphene Oxide via Direct Surface Functionalization as Source/Drain Electrodes in Bottom-Contact Organic Transistors. Adv. Mater. 2013, 25, 5856−5862. (19) Han, S.-T.; Zhou, Y.; Sonar, P.; Wei, H.; Zhou, L.; Yan, Y.; Lee, C.-S.; Roy, V. A. L. Surface Engineering of Reduced Graphene Oxide for Controllable Ambipolar Flash Memories. ACS Appl. Mater. Interfaces 2015, 7, 1699−1708. (20) Han, S.-T.; Zhou, Y.; Yang, Q. D.; Zhou, L.; Huang, L.-B.; Yan, Y.; Lee, C.-S.; Roy, V. A. Energy-Band Engineering for Tunable Memory Characteristics through Controlled Doping of Reduced Graphene Oxide. ACS Nano 2014, 8, 1923−1931. (21) Kim, S.; Zhou, S.; Hu, Y.; Acik, M.; Chabal, Y. J.; Berger, C.; de Heer, W.; Bongiorno, A.; Riedo, E. Room-temperature metastability of multilayer graphene oxide films. Nat. Mater. 2012, 11, 544−549. (22) Diez-Betriu, X.; Alvarez-Garcia, S.; Botas, C.; Alvarez, P.; Sanchez-Marcos, J.; Prieto, C.; Menendez, R.; de Andres, A. Raman spectroscopy for the study of reduction mechanisms and optimization of conductivity in graphene oxide thin films. J. Mater. Chem. C 2013, 1, 6905−6912. (23) Petersen, S.; He, Y.; Lang, J.; Pizzocchero, F.; Bovet, N.; Bøggild, P.; Hu, W.; Laursen, B. W. Stepwise Reduction of
AUTHOR INFORMATION
Corresponding Author
*Phone: 00302610965263. E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the Bilateral R&T cooperation program between Greece and Germany, Project INSOLCELL GSRT GER 2272: Innovative materials for solar cell design and demonstration, cofinanced by the Hellenic Republic and European Union-European Union Development Fund.
■
REFERENCES
(1) Greiner, M. T.; Lu, Z.-H. Thin-film metal oxides in organic semiconductor devices: their electronic structures, work functions and interfaces. NPG Asia Mater. 2013, 5, e55. (2) Zeng, H.; Zhu, X.; Liang, Y.; Guo, X. Interfacial Layer Engineering for Performance Enhancement in Polymer Solar Cells. Polymers 2015, 7, 333−372. (3) Garg, R.; Dutta, N. K.; Choudhury, N. R. Work Function Engineering of Graphene. Nanomaterials 2014, 4, 267−300. (4) Choi, K. S.; Park, Y.; Kim, S. Y. Comparison of graphene oxide with reduced graphene oxide as hole extraction layer in organic photovoltaic cells. J. Nanosci. Nanotechnol. 2013, 13, 3282−3287. (5) Mattevi, C.; Eda, G.; Agnoli, S.; Miller, S.; Mkhoyan, K. A.; Celik, O.; Mastrogiovanni, D.; Granozzi, G.; Garfunkel, E.; Chhowalla, M. Evolution of electrical, chemical, and structural properties of transparent and conducting chemically derived graphene thin films. Adv. Funct. Mater. 2009, 19, 2577−2583. (6) Liu, J.; Xue, Y.; Gao, Y.; Yu, D.; Durstock, M.; Dai, L. Hole and Electron Extraction Layers Based on Graphene Oxide Derivatives for 289
DOI: 10.1021/acs.jpcc.5b09234 J. Phys. Chem. C 2016, 120, 281−290
Article
The Journal of Physical Chemistry C Immobilized Monolayer Graphene Oxides. Chem. Mater. 2013, 25, 4839−4848. (24) Kymakis, E.; Stratakis, E.; Stylianakis, M. M.; Koudoumas, E.; Fotakis, C. Spin coated graphene films as the transparent electrode in organic photovoltaic devices. Thin Solid Films 2011, 520, 1238−1241. (25) Mishra, M.; Joshi, R. K.; Ojha, S.; Kanjilal, D.; Mohanty, T. Role of Oxygen in the Work Function Modification at Various Stages of Chemically Synthesized Graphene. J. Phys. Chem. C 2013, 117, 19746−19750. (26) Wu, J.; Becerril, H. A.; Bao, Z.; Liu, Z.; Chen, Y.; Peumans, P. Organic solar cells with solution-processed graphene transparent electrodes. Appl. Phys. Lett. 2008, 92, 263302. (27) Pei, S.; Cheng, H.-M. The reduction of graphene oxide. Carbon 2012, 50, 3210−3228. (28) Kovtyukhova, N. I.; Ollivier, P. J.; Martin, B. R.; Mallouk, T. E.; Chizhik, S. A.; Buzaneva, E. V.; Gorchinskiy, A. D. Layer-by-Layer Assembly of Ultrathin Composite Films from Micron-Sized Graphite Oxide Sheets and Polycations. Chem. Mater. 1999, 11, 771−778. (29) Nikolakopoulou, A.; Tasis, D.; Sygellou, L.; Dracopoulos, V.; Galiotis, C.; Lianos, P. Study of the thermal reduction of graphene oxide and of its application as electrocatalyst in quasi-solid state dyesensitized solar cells in combination with PEDOT. Electrochim. Acta 2013, 111, 698−706. (30) Zhao, J.; Pei, S.; Ren, W.; Gao, L.; Cheng, H.-M. Efficient Preparation of Large-Area Graphene Oxide Sheets for Transparent Conductive Films. ACS Nano 2010, 4, 5245−5252. (31) Becerril, H. A.; Mao, J.; Liu, Z.; Stoltenberg, R. M.; Bao, Z.; Chen, Y. Evaluation of solution-processed reduced graphene oxide films as transparent conductors. ACS Nano 2008, 2, 463−470. (32) Sygellou, L.; Ladas, S.; Reading, M. A.; van den Berg, J. A.; Conard, T.; De Gendt, S. A comparative X-ray photoelectron spectroscopy and medium-energy ion-scattering study of ultra-thin, Hf-based high-k films. Surf. Interface Anal. 2010, 42, 1057−1060. (33) Kozlowski, C.; Sherwood, P. M. A. X-ray photoelectron spectroscopic studies of carbon-fibre surfaces. Part 4.-The effect of electrochemical treatment in nitric acid. J. Chem. Soc., Faraday Trans. 1 1984, 80, 2099−2107. (34) Bagri, A.; Mattevi, C.; Acik, M.; Chabal, Y. J.; Chhowalla, M.; Shenoy, V. B. Structural evolution during the reduction of chemically derived graphene oxide. Nat. Chem. 2010, 2, 581−587. (35) Sutar, D. S.; Singh, G.; Divakar Botcha, V. Electronic structure of graphene oxide and reduced graphene oxide monolayers. Appl. Phys. Lett. 2012, 101, 103103. (36) Ganguly, A.; Sharma, S.; Papakonstantinou, P.; Hamilton, J. Probing the thermal deoxygenation of graphene oxide using highresolution in situ X-ray-based spectroscopies. J. Phys. Chem. C 2011, 115, 17009−17019. (37) Ishida, T.; Kobayashi, H.; Nakato, Y. Structures and properties of electron-beam-evaporated indium tin oxide films as studied by X-ray photoelectron spectroscopy and work-function measurements. J. Appl. Phys. 1993, 73, 4344−4350. (38) Gaspar, D. J.; Polikarpov, E. OLED Fundamentals: Materials, Devices, and Processing of Organic Light-Emitting Diodes; CRC Press: New York, 2015. (39) Amy, F.; Chan, C.; Kahn, A. Polarization at the gold/pentacene interface. Org. Electron. 2005, 6, 85−91. (40) Kvashnin, D. G.; Sorokin, P. B.; Brüning, J.; Chernozatonskii, L. A. The impact of edges and dopants on the work function of graphene nanostructures: The way to high electronic emission from pure carbon medium. Appl. Phys. Lett. 2013, 102, 183112. (41) Sygellou, L.; Viskadouros, G.; Petridis, C.; Kymakis, E.; Galiotis, C.; Tasis, D.; Stratakis, E. Effect of the reduction process on the field emission performance of reduced graphene oxide cathodes. RSC Adv. 2015, 5, 53604−53610. (42) Some, S.; Bhunia, P.; Hwang, E.; Lee, K.; Yoon, Y.; Seo, S.; Lee, H. Can Commonly Used Hydrazine Produce n-Type Graphene? Chem. - Eur. J. 2012, 18, 7665−7670. (43) Singh, G.; Sutar, D. S.; Divakar Botcha, V.; Narayanam, P. K.; Talwar, S. S.; Srinivasa, R. S.; Major, S. S. Study of simultaneous
reduction and nitrogen doping of graphene oxide Langmuir−Blodgett monolayer sheets by ammonia plasma treatment. Nanotechnology 2013, 24, 355704. (44) Hwang, J. O.; Park, J. S.; Choi, D. S.; Kim, J. Y.; Lee, S. H.; Lee, K. E.; Kim, Y.-H.; Song, M. H.; Yoo, S.; Kim, S. O. WorkfunctionTunable, N-Doped Reduced Graphene Transparent Electrodes for High-Performance Polymer Light-Emitting Diodes. ACS Nano 2012, 6, 159−167.
290
DOI: 10.1021/acs.jpcc.5b09234 J. Phys. Chem. C 2016, 120, 281−290
