www.MaterialsViews.com
Graphene Quantum Dots
Is the Chain of Oxidation and Reduction Process Reversible in Luminescent Graphene Quantum Dots? Min-Ho Jang, Hyun Dong Ha, Eui-Sup Lee, Fei Liu, Yong-Hyun Kim, Tae Seok Seo,* and Yong-Hoon Cho*
Graphene-based
quantum dots (QDs) have received a tremendous amount of attention as a new type of light-emitting materials. However, their luminescence origins remain controversial due to extrinsic states of the impurities and disorder structures. Especially, the function of oxygen-contents should be understood and controlled as a crucial element for tuning the optical properties of graphene-based QDs. Herein, a series of graphene oxide QDs (GOQDs) with different amounts of oxygencontents are first synthesized via a direct oxidation route of graphite nanoparticle and thoroughly compared with a series of reduced GOQDs (rGOQDs) prepared by the conventional chemical reduction. Irreversible emission and different carrier dynamics are observed between the GOQDs and rGOQDs, although both routes show a similar tendency with regard to the variation of oxygen-functional components. Their luminescence mechanisms are closely associated with different atomic structures. The mechanism for the rGOQDs can be associated with a formation of small sp2 nanodomains as luminescent centers, whereas those of GOQDs may be composed of oxygen-islands with difference sizes depending on oxidation conditions surrounded by a large area of sp2 bonding. Important insights for understanding the optical properties of graphene-based QDs and how they are affected by oxygen-functional groups are shown.
M.-H. Jang, Prof. Y.-H. Cho Department of Physics and Graphene Research Center of KI for the NanoCentury Korea Advanced Institute of Science and Technology 291 Daehak-ro, Yuseong-gu Daejeon 305-701, South Korea E-mail: [email protected] H. D. Ha, Dr. F. Liu, Prof. T. S. Seo Department of Chemical and Biomolecular Engineering and Institute for the BioCentury Korea Advanced Institute of Science and Technology 291 Daehak-ro, Yuseong-gu, Daejeon 305-701, South Korea E-mail: [email protected] E.-S. Lee, Prof. Y.-H. Kim Graduate School of Nanoscience and Technology Korea Advanced Institute of Science and Technology and Center for Nanomaterials and Chemical Reactions Institute for Basic Science (IBS) 291 Daehak-ro, Yuseong-gu, Daejeon 305-701, South Korea DOI: 10.1002/smll.201500206 small 2015, DOI: 10.1002/smll.201500206
1. Introduction Since the first report of luminescence from graphene oxide (GO),[1,2] the field of graphene-based quantum dots (QDs), consisting of disc-like graphene with a few nanometers in size, has received a tremendous amount of attention, not only as a new type of light-emitting materials but also in relation to a wide variety of technical applications, especially those specific to graphene-based QDs in biological labeling, imaging, and functional nanoscale devices.[1–4] Currently, the luminescence origins are explained in terms of their size,[5–8] shape,[6] fraction of the sp2 domains in the sp3 matrix,[9,10] and/or various functional groups.[10–12] However, since photoluminescence (PL) is simultaneously observed by a combination of these factors, it is difficult to identify the main origin of the PL. For the size effect, both size-dependent[5,8] and size-independent[7,13] PL behaviors have been reported
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
1
full papers www.MaterialsViews.com
in a size range of 2–10 nm, and the calculated bandgap in this size range has shown much lower energy than that of the visible light. Therefore, the origins of the visible PL change due to the size-effect remain controversial, and it is difficult to exclude the effects of extrinsic states, which are derived from oxygen-functional groups (i.e., epoxy, carboxyl, and hydroxyl groups), the edge effect, and the disorder structure of vacancies. Hence, a full understanding of the optical properties from the defect states and judicious control of their PL properties would be essential for a variety of applications. In particular, oxygen-functional groups, a common type of defects, render graphene-based QDs to be helpful for the dissolution in water and to interact with a wide range of organic and inorganic materials, greatly widening the range of applications. Furthermore, they are more controllable than other types of defects when using oxidation and reduction methods. Therefore, the oxygen-functional groups were considered as a crucial element for tuning the optical properties of graphene-based QDs.[5,14–17] Thus far, graphene-based QDs have been mainly prepared with oxidation strategies by the treatment of various carbon sources.[5,14–17] Therefore, most graphene-based QDs are composed of a high level of oxygen contents, which can be called as graphene-oxide QDs (GOQDs). Hence, the GOQDs with different amount of oxygen contents have been commonly prepared by partially reducing the oxygen content through various reduction methods, such as the use of chemical agents,[18] high-temperature annealing,[19,20] and/or photothermal methods.[9,10] For reduced GOQDs (rGOQDs) prepared by a reduction process, the changes in the optical properties have been generally explained by a model of the isolated sp2 nanodomains, resulting in a bluecolored emission. Even though there are several researches about the PL properties of GOQDs depending on reduction degree,[9,10,18–20] the studies on the PL of the graphenebased QDs starting from the pure graphene quantum dot (GQDs) depending on oxidation degree has rarely been done. The PL properties of the graphene-based QDs have recently attracted huge attention, but the clarification for the PL mechanism is not satisfied so far simply by investigating the GOQD reduction process. Therefore, a series of GOQDs and rGOQDs should be prepared experimentally via both oxidation and reduction to understand the detailed whole PL mechanisms and to confirm whether the optical properties are the same or different depending on the degree of oxygen contents. Particularly, the oxidation approach will elucidate the fundamental PL mechanism of GOQDs with controlled oxygen content, and will overcome several drawbacks of the reduction process, such as the requirements of a high-power lamp source, long fabrication times, the use of harmful chemical reagents, and the generation of disordered structures like carbon mono and multivacancies. The limitations of previously reported oxidative cutting methods[5,14–17] will be overcome by exfoliation methods of graphite nanoparticle (GNP),[12,21] which result in the synthesis of single-layered graphene-based QDs with regular shapes and sizes regardless of the oxygen contents. The exfoliation methods of GNPs can generate a series of GOQDs with different amounts of oxygen by controlling chemical
2 www.small-journal.com
oxidation condition. Moreover, it is possible to comparatively analyze the optical properties of GOQDs and rGOQDs at similar levels of oxygen-functional groups. Herein, we suggest an easy route to obtain a series of GOQDs with tunable PL emissions controlled by the chemical oxidation of GNPs. To the best of our knowledge, this is the first report on the synthesis of oxygen-dependent GOQDs without a reduction process. Thus, we prepared all kinds of graphene-based QDs from the GQDs to GOQDs by oxidation as well as from the GOQDs to the rGOQDs by chemical reduction, and performed the in-depth comparative studies on the PL properties of the GQDs, rGOQDs, and GOQDs. The GOQDs have shown clear multicolor emissions depending on the amount of oxygen. We also found that the spectral emission of rGOQDs does not match that of GOQDs at all, in spite of similar oxygen contents. The effects of oxidation and reduction on the optical properties are discussed in details by means of PL, PL excitation (PLE), UV–vis absorbance, and time-resolved PL (TRPL) measurements as well as first-principles density-functional theory (DFT) calculations. This will provide important insights for understanding of how oxygen-functional groups affect the optical properties in graphene-based QDs and whether the successive process of oxidation and reduction is reversible or irreversible to the optical properties of them.
2. Results and Discussion 2.1. Preparation Figure 1 shows a schematic illustration of the proposed route in which we controlled the oxygen-functional groups in graphene-based QDs. Our studies focused on understanding how the optical properties of graphene-based QDs are affected when controlling the number of oxygen defects by means of chemical oxidation and reduction. In the gradual oxidation process, we employed the modified Hummer’s method using a sulfuric/nitric concentrated acid mixture (H2SO4/HNO3), and the number of the oxygen-functional groups on the surface could be controlled depending on the ratio between sulfuric acid and nitric acid.[22,23] The GNPs were chemically exfoliated to produce GOQDs, and the volume ratio of the H2SO4/HNO3 mixture was changed from 0/80 to 60/20. Note that 60/20 means the used volume ratio of the H2SO4 and HNO3 was 60 and 20 mL, respectively. Different oxidation degree resulted in the tunable number of the oxygen-functional groups in the GOQDs, which were monitored by the C1s core-level peaks in a high-resolution X-ray photoelectron spectroscopy (XPS) (Figure S1, Supporting Information). The C1s XPS data show a sp2 carbon peak at ≈284.5 eV, and sp3 carbon peaks including a carbonyl peak (C–O) at ≈286.6 eV, an epoxy (C–O–C) peak at ≈288.0 eV, and a carboxyl (–COOH) peak at ≈289.0 eV. Figure 2a shows a summarized proportion of the oxygen content for GOQDs fabricated with different oxidation, revealing a gradual increase under higher volume ratio of sulfuric acid to nitric acid. Then, we performed the partial reduction process using sodium borohydride (NaBH4), a well-known reducing
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201500206
www.MaterialsViews.com
2.2. Luminescent Characteristics of GOQDs and rGOQDs We examined the luminescent characteristics for all samples under the same UV excitation with a 325 nm at room temperature. Figure 3a–c shows the PL spectra and photographic images of various types of GOQDs and rGOQDs. For the GOQDs, the emission colors of the PL spectra were gradually red-shifted from sky-blue (0.21, 0.24) of the GOQD(00/80) to greenishyellow (0.33, 0.43) of the GOQD(50/30) in the international commission on illumination map (Figure S6, Supporting Information). The highest PL peaks also move toward a lower energy from 2.80 eV of the GOQD(00/80) to 2.35 eV of the GOQD(50/30) as shown in Figure 3d. Hence, GOQDs controlled by the oxiFigure 1. Schematic illustration of the tunable luminescent graphene-based QDs through the dizing conditions are easily able to offer reversible oxidation and reduction processes. tunable PL. In contrast to the oxidation approach, there were nearly unchangeable agent.[24] Starting from the GOQDs which were produced by peak positions of rGOQDs with different amounts of oxygen the treatment of H2SO4/HNO3 (60/20), the GOQD(60/20) (Figure 3d). However, an interesting finding is that a lumineswas gradually reduced by adding a different amount of cence peak centered at 2.8 eV (440 nm) appears after chemNaBH4 into the 10 mL GOQD(60/20) solution. The reduc- ical reduction, and its relative intensity gradually increases tion process was investigated by XPS (Figure S2, Supporting as the amount of reducing agent increases from 1 to 50 mg Information), and summarized in Figure 2a. The XPS results (Figure S7, Supporting Information). As shown in Figure 3d, display that the sp2-carbon structure could be clearly recov- although the amount of oxygen-functional groups of both ered with an increased reducing agents from 0 to 50 mg, by GOQDs and rGOQDs shows similar variation as a function showing the decreased intensity of the oxygenous carbon of chemical oxidation and reduction, the spectral emission peaks. Also, the Fourier transform infrared spectroscopy shows a wholly irreversible emission tendency from oxida(FT-IR) spectra support the existence of oxygenous groups tion to reduction. Therefore, their luminescent characteristics by showing C=O (1725 cm−1), and O–H peak (3400 cm−1) imply that there exist different recombination mechanisms (Figure S3, Supporting Information). Similar to the XPS data, between GOQDs and rGOQDs with different structures due the peaks of oxygenous functional groups tended to increase to the chemical reactions. Furthermore, since it is well known according to the oxidation ratio, whereas those gradually that the oxygen-functional groups are very sensitive to the decreased in proportion to the reduction ratio. pH condition than sp2 carbon bonding,[11,26] we prepared graTo determine the size distribution, we conducted a high- phene-based QDs in the pH-dependent solution (Figures S8 resolution transmission electron microscopy (HRTEM) and and S9, Supporting Information). At pH of 1, the dominated an atomic force microscopy (AFM) analysis for the pristine- PL spectra clearly emit more bluish light than those of pH 7 graphene QDs (GQDs),[12] GOQD(40/40), GOQD(60/20),[21] due to the blue-shifted PL peak position or the increased and rGOQD (5 mg) (Figures S4 and S5, Supporting Infor- shoulder peak intensity of blue emission. On the other hand, mation). Due to the uniform size of the GNPs with average at pH of 14, the PL peak positions are similar to those of pH 7, diameters of 4 nm, all the produced graphene-based QD and but the full-width at half-maximum (FWHM) of the PL specits derivatives show a narrow size distribution between 2 and trum is broader than that of pH 7 due to the shoulder peak of 4 nm with an average diameter of ca. 3.37 nm (Figure 2b). the greenish emission. When the oxygen-functional groups are Thus, the optical properties of our graphene-based QDs were protonated under acidic condition, the greenish emission from dominantly depended on the number of oxygenous defects the oxygen-functional groups becomes inactive in the PL specrather than the size effect. In addition, the heights of the trum. Under alkaline condition, however, the oxygen-funcgraphene-based QDs display a slightly different distributions tional groups are deprotonated and the luminescence from among the categories of samples in shown Figure 2c. Inter- oxygen-functional groups can be maintained.[29] Especially, estingly, the individual sheets of GOQD (40/40) and GOQD in the case of rGOQDs, the degree of the blue-shift from pH (60/20) have an average thickness of 0.9–1 nm, significantly 7 to pH 1 is proportional to the strength of chemical reduchigher than that of pristine-GQDs and rGOQDs (about tion, 50 mg (143 meV) > 10 mg (129 meV) > 5 mg (81 meV). 0.7 nm). These findings could be rationally explained in terms This suggests that as reduction process is applied, more bluish of the oxygen-functional groups (e.g., epoxy) anchored on luminescence dominated than that of GOQDs. More detailed investigation of luminescence mechanisms is as follows. the basal plane of graphene.[25] small 2015, DOI: 10.1002/smll.201500206
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
3
full papers www.MaterialsViews.com
emitting states by losing energy through a relaxation process. Also, the red-shifted region in the 2D matrices may be formed due to the inhomogeneous distributions of graphene-based QDs.[14,17] Thus, we could extract the excitation wavelength (λex) and the invariant emission wavelength (λem) at the point of the starting spectral red-shift in the 2D matrices (Figure S11, Supporting Information). These points are organized according to the composition ratio of oxygenous-/sp2-carbon in the XPS data in Figure 4c,d. The values at both points of excitation and emission wavelength gradually decreased by increasing the fraction of oxygenous carbon (versus sp2 carbon). To further investigate the effect of oxygen-functional groups in a view of band-gaps in graphene-based QDs, we performed the first-principles DFT calculations for oxidized GQD models. The theoretical band-gap, defined as the energy difference between the highest occupied and lowest unoccupied molecular orbitals, was obtained by atomic structures of oxygen-functional groups (i.e., hydroxyl) that are serially conjugated on the nanodisk graphene (Figure S12, Supporting Information). The oxygen-functional groups split the degenerated energy levels in pure GQDs, causing that the band-gap of GQDs with oxygenous defects becomes smaller than that of pure GQDs. Accordingly, this is in good agreement with the optical properties of GOQDs, i.e., the color-tunable emission from blue to greenish-yellow and the increase of the absorption in the visible region (Figure 4e). Figure 2. a) The ratio of oxygenous-/sp2-carbons versus oxidation (red) or reduction Regarding the emission wavelength (blue) conditions. b) Diameter distribution analysis of the pristine-GQD, GOQD(40/40), of the invariant PL under similar oxygen GOQD(60/20), and rGOQD (5 mg). c) Height distribution analysis of the pristine-GQD (black), contents for GOQD and rGOQD in GOQD(40/40) (blue), GOQD(60/20) (green), and rGOQD (5 mg) (red). Figure 4c,d, the PL emission wavelength of GOQDs is shorter than that of the 2.3. Band-Gap Investigation of GOQDs and rGOQDs rGOQDs. This is because the reduction leads to other disordered structures (i.e., vacancies) on basal plane through the To gain insight into the luminescence features in details, removal of oxygen. Our DFT simulation indicates that the we measured the PL spectra of the samples in water sol- formation energy which represents the stability of atomic vent under different excitation wavelength (Figure S10, defects depends on the position of an oxygen atom on the Supporting Information). As shown in Figure 4a,b, the 2D nanodisk graphene, and an oxygen atom on the basal plane matrices were plotted by normalized PL peak intensity of has a higher formation energy than that of an oxygen atom excitation-wavelength-dependent PL (EWD-PL) spectrum. at the edge (Figure S13, Supporting Information). Therefore, Each 2D matrix could be divided into two regions based the reduction process preferentially removes the oxygen on their emission properties. One is mostly invariant with defects anchored on the basal plane of the GOQD. This increasing excitation wavelengths, whereas the other is red- strongly correlates with the increased sp2 fractions after shifted. The invariant PL with various excitation wavelengths chemical reduction and the decreased thickness by removing is closely related to the bandgap of the extrinsic states origi- oxygenous defects on the basal plane. For these reasons, nated from the oxygen-functional groups.[10,11] It reveals that rGOQDs show not only a small change of the luminescent the carriers excited by high energy photons arrive at the properties for extrinsic states, but also the generation of a
4 www.small-journal.com
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201500206
www.MaterialsViews.com
Figure 3. The photoluminescence spectra of a) GOQDs and b) rGOQDs irradiated by a cw 325 nm He–Cd laser at room temperature. c) The photographic images indicate tunable color emission by chemical oxidation (GOQDs) and reduction (rGOQDs), respectively. d) The peak positions of GOQDs (red circles) and rGOQDs (blue circles) were obtained from the photoluminescence spectra; they are plotted according to the oxygenous-/ sp2-carbon ratio.
Figure 4. Tunable luminescent mechanism. 2D matrices via the normalized excitation-wavelength-dependent PL behavior of various types of GOQDs and rGOQDs. The white-circles in the 2D matrices express the peaks of the PL spectra. a) The diagrams show the GOQDs fabricated by 10 mL/70 mL, 30 mL/50 mL, and 50 mL/30 mL (H2SO4/NHO3). b) The diagrams exhibit the cases of 50 mg, 5 mg, and GOQD (0 mg) according to the reducing agent of NaBH4. c,d) The points of the excitation (red-circle) and emission (blue-circle) energy of GOQDs and rGOODs, respectively. The error bar of the x-axis relate to the XPS analysis by several measurement, whereas that of y-axis (blue-circle) is standard deviation of PL peak from 2D matrices. e,f) Normalized UV–vis absorbance spectra of GOQDs and rGOQDs, respectively. The inset in f) shows that a shoulder peak at 285 nm in UV–vis absorbance corresponds to the peak of the PLE spectrum. g) The intensity ratio compared with 285 and 325 nm in PLE spectrum of the detection wavelength at 450 nm. The ratio is calculated by intensity of PLE at 280 nm/intensity of PLE at 320 nm. small 2015, DOI: 10.1002/smll.201500206
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
5
full papers www.MaterialsViews.com
luminescence centered at 2.8 eV (440 nm) due to isolated sp2 domains. The PLE measurement of the rGOQDs shows the distinct peak at 285 nm, corresponding to the shoulder peak at 285 nm as shown in the UV–vis absorbance, as well as broad PLE spectra with a center at approximately 325 nm (Figure 4f). The distinct PLE peak of the rGOQDs is similar to the previously reported optical properties of the intrinsic state in the pristine-GQDs, which is responsible for the blue-colored emission.[12] In addition, the fraction of the intensities of the two parts in the PLE spectrum, a peak at 285 nm and shoulder at 325 nm in Figure S14, Supporting Information, was summarized in Figure 4g, indicating the increase in the strength of the intrinsic states of the rGOQDs as a function of reduction degree. On the other hand, the GOQD(60/20), a fully oxidized sample, shows only broad PLE spectra with a center of about 325 nm (Figure S15, Supporting Information). Therefore, the broad PLE spectra can be attributed to optical transitions from oxygen-functional groups. Interestingly, all GOQD samples have broad PLE spectra regardless of the ratio of oxygen contents (Figure S14, Supporting Information). This means that the emission of GOQDs is mainly associated with oxygen-functional groups instead of the intrinsic states originated from the sp2 carbon.
2.4. In-Depth Study of Carrier Dynamics and Recombination An in-depth study of carrier dynamics and recombination is necessary for accurately understanding the PL phenomenon. The TRPL was carried out using a femtosecond Ti:Sapphire laser and a streak camera detector at room temperature. Figure 5 exhibits the temporal profile of time-integral PL
(TIPL) spectra obtained from streak images. Interestingly, the TIPL of GOQDs and rGOQDs showed very different spectral evolution characteristics as time passed after the pump event. As shown in Figure 5a–c, upon a time delay, the spectral signals of the respective GOQDs retain an invariable emission wavelength due to the presence of only single luminescence origin derived from the extrinsic states of the oxygen-functional groups. All of the rGOQDs, in contrast, show red-shifted spectral migrations as time proceeds, as shown in the Figure 5d–f. These spectra show a fixed peak position at 440 nm in a range of 0–1 ns, and a slight blueshift of the peak position from 515 to 480 nm as a function of the reducing agent in a range of 15–31 ns. This difference between GOQDs and rGOQDs is attributed to different recombination dynamics. According to TIPL spectra in a time range of 15–31 ns, both results of the red-shift from GOQD(10/70) to GOQD(50/30) and the blue-shift from GOQD(60/20) to rGOQD(50 mg) are associated with the change of the bandgap with the extrinsic states originated from oxygen-functional groups.[12] The emission with 440 nm of rGOQDs has significantly shorter PL lifetimes than that of the extrinsic state (Figure S17, Supporting Information). These properties are in good agreement with previously reported results,[12] corresponding to the fast decay of the blue emission in the pristine-GQDs and the slow decay of green emission in the GOQDs (Figure S16, Supporting Information).The blue emission with fast recombination in the rGOQDs originates from the intrinsic states of the sp2 nanodomains. Their dynamical spectral migration clearly supports that the luminescent origin of GOQDs is only the extrinsic states derived from oxygen-functional groups, whereas that of the rGOQDs include the extrinsic states as well as the intrinsic states formed by chemical reduction.
Figure 5. The time-dependent integral PL spectra from the GOQDs (blue) and rGOQDs (red) groups under excitation at 266 nm by a thirdharmonic-generation femtosecond pulse laser. The diagrams of a), b), and c) [d), e), and f)] show the results from GOQDs [rGOQDs] fabricated with 10 mL/70 mL, 30 mL/50 mL, and 50 mL/30 mL of H2SO4/HNO3 [50, 5, and 0 mg of NaBH4], respectively.
6 www.small-journal.com
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201500206
www.MaterialsViews.com
Figure 6. a,b) Structural models of GOQDs and rGOQDs for chemical oxidation and reduction. Note that the white and yellow regions represent sp2 bonding and sp3 bonding formed by oxygen-functional groups, respectively, and the regions of v (red color) describe vacancies.
2.5. Luminescent Model from GOQDs to rGOQDs It is worth noting that since the optical properties of the GOQDs in the oxidation process are quite distinct from that of the rGOQDs in the reduction process (Figures 3–5), the structural model of the GOQDs under the oxidation process of pristine GQDs would be very dissimilar to that of the rGOQDs under the reduction process of GOQDs. For the latter, it has been sucessfully explained by the formation of numerous isolated sp2 nanodomains within the sp3 bonding of the oxygen-functional groups by removing the oxygenbonding randomly under the chemical reduction process (Figure 6d,e).[9,10] Based on this model, we can attribute the appearance of the blue light emission to the intrinsic states associated with isolated sp2 nanodomains in the rGOQDs under the reduction process, which have higher photon energy than the green light emission due to the extrinsic states formed by the oxygen-functional groups. For the former, on the other hand, we propose a model of oxygen-islands surrounded by a large area of sp2 carbon bonding during the oxidation process of the pristine GQD (Figure 6a,b). When the oxidation process proceeds, oxygen-islands are gradually formed on the graphene surface and then the size of oxygenislands increases.[27] As a result, the area of sp2 carbon bonding decreases, unlike the formation picture of sp2 nanodomains in rGOQDs during the reduction process. Therefore, for the GOQDs under the oxidation process of the pristine GQD, it would be difficult to observe a blue light emission originated from sp2 nanodomains, as opposed to the case of rGOQDs. In contrast to a simple speculation of a reversible process through chemical oxidation and reduction, we conclude that the chain of the oxidation and reduction process is irreversible in the luminescent GQDs, since graphene surface forms different structures during oxidation and reduction processes; the GOQDs are composed of oxygen-islands (with difference sizes depending on different oxidation conditions) surrounded by a large area of sp2 carbon bonding, whereas the rGOQDs have numerous isolated sp2 nanodomains surrounded by the sp3 bonding of the oxygen-functional groups.
3. Conclusion We successfully synthesized a series of GOQDs with different amounts of oxygen by controlling the oxidation conditions for GNPs and observed tunable PL of them. We also confirmed the irreversible spectral emission in the graphenesmall 2015, DOI: 10.1002/smll.201500206
based QDs modulated by a chemical route from oxidation to reduction, although both routes showed a similar tendency with regard to the changed amount of oxygen-functional components. Their light-emission phenomena were closely associated with their atomic structure on the graphene-based QDs. In particular, GOQDs controlled by the oxidation conditions have several advantages over rGOQDs, such as rapid fabrication process, clear tunable PL, and stable spectral PL for carrier relaxation. We believe that our studies provide important insights for understanding the optical properties of the graphene-based QDs and how they are affected by oxygen-functional groups. Our results can lead to future development of novel strategies for guiding the synthesis of the graphene-based QDs with tunable optical properties toward various applications such as bioimaging, sensor, biological labeling, and functional nanoscale devices.
4. Experimental Section Chemicals and Materials: Graphite nanoparticles (average diameter: 3–4 nm, 93%) were purchased by Skyspring Nanomaterials (USA). Ethanol (reagent grade) and sodium borohydride (NaBH4) were obtained from Sigma-Aldrich (USA). Extra-pure sulfuric acid (H2SO4, 95%), extra-pure nitric acid (HNO3, 60%–62%), and sodium hydroxide (NaOH, guaranteed reagent) were purchased from JUNSEI (Japan). Milli-Q water (Millipore, Milford, MA, USA) was used for all aqueous solutions. Preparation of the Samples: (i) The pristine GQD solution was prepared by shaking 5 mg of GNPs in a 10 mL ethanol/H2O (1:1 v/v ratio) mixture on a vortex mixer (Scientific Industries, USA) at 2700 rpm for 1 h. After centrifugation at 2000 rpm for 30 min, the pure GQD solution was collected from the supernatant. (ii) Graphene oxide QDs (GOQDs) were fabricated in which 1 g of GNPs was pre-oxidized in a concentrated H2SO4 and HNO3 solution (80 mL) for 3 h under mild ultrasonication (100 W). The ratio of H2SO4 and HNO3 ranged from 0 mL/80 mL to 60 mL/20 mL, and the oxidizing power increased as the H2SO4 portion was higher. After pre-oxidation, the mixture was stirred at 100 °C under a refluxing condition for 12 h. The mixture was cooled and diluted with 150 mL of deionized (DI) water in an ice bath. The suspension was centrifuged at 20000 rpm and the supernatant was carefully separated. The supernatant was pre-dialyzed to remove acidic ions for 1 d and the pH was adjusted to 8 with a diluted NaOH solution. The final product solution was further dialyzed in a dialysis bag (Molecular weight cut-off: 1000 Da) to remove the salt ions completely for 3 d. All GOQDs were synthesized under the identical
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
7
full papers www.MaterialsViews.com
conditions except the H2SO4/HNO3 volume ratio. (iii) rGOQDs were prepared by adding 1, 5, 10, and 50 mg of NaBH4 in 10 mL of the GOQD solution which was synthesized by a 60/20 H2SO4/HNO3 solution. Each sample was stirred at room temperature for 3 h. The products were dialyzed in a dialysis bag (Molecular weight cut-off: 1 kDa) to remove sodium ions or residues for 1 d. Optical Measurements and Characterization: Atomic force microscope (AFM, Veeco D3100, USA) images were obtained under the conditions of relative humidity of 43%, a tapping mode at a 1–3 Hz scan rate and a 512 × 512 pixel resolution. High resolution transmission electron microscope images were characterized with an ultra-thin carbon film on a holey carbon support film (400 mesh) at an acceleration voltage of 300 kV. X-ray photoelectron spectroscopy was conducted by a Thermo VG scientific Sigma Probe system, and FT-IR was conducted by IR Prestige-21 (DongIl SHIMADZU Corp.) over the range from 600 to 4000 cm−1. All of the luminescent data were obtained by a hand-made setup using precision cells made of Suprasil quartz. The PL and the EWD-PL behaviors were assessed using a cw 325 nm He-Cd laser, monochromatic light from a 300 W-xenon lamp, and UV spectrometers (Maya2000, Ocean Optics, USA) as a PL detector at room temperature. The PLE was measured by monochromatic light from a 300 W Xenon lamp and a high-sensitive photomultiplier tube as a PL detector. The absorption spectra were recorded on a UV–vis spectrophotometer (UV-2450, SHIMADZU, Japan). In order to elucidate the recombination dynamics, we carried out time-resolved PL experiments. A mode-locked femtosecond pulsed Ti:Sapphire laser (Coherent, Chameleon Ultra II) system was used as an excitation source [266 nm (third-harmonic-generation)], and a streak camera (Hamamatsu, C7700-01) was utilized to measure the decay profile of the PL spectra at room temperature. DFT Simulation: We employed projector-augmented wave (PAW)[28] pseudopotentials and the Perdew–Burke–Ernzerhof (PBE) exchange-correlation functional,[29] as implemented in the Vienna Ab Initio Simulation Package (VASP).[30] The kinetic energy cutoff of 400 eV are used for expansion of plane wave basis set and all atomic forces are fully relaxed until the Hellman–Feynman forces are less than 0.025 eV Å−1. The hexagon-like GQD models are used with vacuum separation of about 10 Å for calculating the electronic structure and the formation energy (Eform = Eoxgenous_grphene – ∑niμi, where i is atomic species, and μ is the chemical potential using graphene, oxygen molecule, and hydrogen molecule as a reference state of each species). All edges of GQDs are terminated with sp2 hydrogen atoms to remove the dangling electrons.
Supporting Information Supporting Information is available from the Wiley Online Library or from the author.
Acknowledgements M.-H.J. and H.D.H. contributed equally to this work. This work was supported by the National Research Foundation (NRF-
8 www.small-journal.com
2013R1A2A1A01016914, NRF-2013R1A1A2011750) of the Ministry of Education, the Converging Technology Project funded by the Korean Ministry of Environment (M112-00061-0002-0), and the GRC project of KI for the NanoCentury. E.-S.L. and Y.-H.K. were supported by the Global Frontier R&D program by the Center for Multiscale Energy Systems (2011-0031566) and the Institute for Basic Science (IBS) (IBS-R004-D1-2014-a00).
[1] L. Li, G. Wu, G. Yang, J. Peng, J. Zhao, J.-J. Zhu, Nanoscale 2013, 5, 4015. [2] Z. Luo, P. M. Vora, E. J. Mele, A. T. C. Johnson, J. M. Kikkawaa, Appl. Phys. Lett. 2009, 94, 111909. [3] H. Sun, L. Wu, W. Wei, X. Qu, Mater. Today 2013, 16, 433. [4] J. Shen, Y. Zhu, X. Yanga, C. Li, Chem. Commun. 2012, 48, 3686. [5] J. Peng, W. Gao, B. K. Gupta, Z. Liu, R. R. Aburto, L. Ge, L. Song, L. B. Alemany, X. Zhan, G. Gao, S. A. Vithayathil, B. A. Kaipparettu, A. A. Marti, T. Hayashi, J. J. Zhu, P. M. Ajayan, Nano Lett. 2012, 12, 844. [6] S. Kim, S. W. Hwang, M. K. Kim, D. Y. Shin, D. H. Shin, C. O. Kim, S. B. Yang, J. H. Park, E. Hwang, S. H. Choi, G. Ko, S. H. Sim, C. Sone, H. J. Choi, S. Bae, B. H. Hong, ACS Nano 2012, 6, 8203. [7] L. Tang, R. Ji, X. Cao, J. Lin, H. Jiang, X. Li, K. S. Teng, C. M. Luk, S. Zeng, J. Hao, S. P. Lau, ACS Nano 2012, 6, 5102. [8] W. Kwon, Y.-H. Kim, C.-L. Lee, M. Lee, H. C. Choi, T.-W. Lee, S.-W. Rhee, Nano Lett. 2014, 14, 1306. [9] G. Eda, Y.-Y. Lin, C. Mattevi, H. Yamaguchi, H.-A Chen, I.-S. Chen, C.-W. Chen, M. Chhowalla, Adv. Mater. 2010, 22, 505. [10] C.-T. Chien, S.-S. Li, W.-J. Lai, Y.-C. Yeh, H.-A. Chen, I.-S. Chen, L.-C. Chen, K.-H. Chen, T. Nemoto, S. Isoda, M. Chen, T. Fujita, G. Eda, H. Yamaguchi, M. Chhowalla, C.-W. Chen, Angew. Chem. Int. Ed. 2012, 51, 1. [11] S. H. Jin, D. H. Kim, G. H. Jun, S. H. Hong, S. Jeon, ACS Nano 2013, 7, 1239. [12] F. Liu, M.-H. Jang, H. D. Ha, J.-H. Kim, Y.-H. Cho, T. S. Seo, Adv. Mater. 2013, 25, 3657. [13] Q. Xu, Q. Zhou, Z. Hua, Q. Xue, C. Zhang, X. Wang, D. Pan, M. Xiao, ACS Nano 2013, 7, 10654. [14] Y. Li, Y. Hu, Y. Zhao, G. Shi, L. Deng, Y. Hou, L. Qu, Adv. Mater. 2011, 23, 776. [15] D. Pan, J. Zhang, Z. Li, M. Wu, Adv. Mater. 2010, 22, 734. [16] J. Luo, L. J. Cote, V. C. Tung, A. T. L. Tan, P. E. Goins, J. Wu, J. Huang, J. Am. Chem. Soc. 2010, 132, 17667. [17] Y. Dong, C. Chen, X. Zheng, L. Gao, Z. Cui, H. Yang, C. Guo, Y. Chi, C. M. Li, J. Mater. Chem. 2012, 22, 8764. [18] a) Z. Gan, S. Xiong, X. Wu, T. Xu, X. Zhu, X. Gan, J. Guo, J. Shen, L. Sun, P. K. Chu, Adv. Optical Mater. 2013, 1, 926; b) S. Zhu, J. Zhang, X. Liu, B. Li, X. Wang, S. Tang, Q. Meng, Y. Li, C. Shi, R. Hu, B. Yang, RSC Adv. 2012, 2, 2717; c) S. Zhu, J. Zhang, S. Tang, C. Qiao, L. Wang, H. Wang, X. Liu, B. Li, Y. Li, W. Yu, X. Wang, H. Sun, B. Yang, Adv. Funct. Mater. 2012, 22, 4732. [19] K. P. Loh, Q. L. Bao, G. Eda, M. Chhowalla, Nat. Chem. 2010, 2, 1015. [20] T. V. Khai, H. G. Na, D. S. Kwak, Y. J. Kwon, H. Ham, K. B. Shim, H. W. Kim, J. Mater. Chem. 2012, 22, 17992. [21] H. D. Ha, M.-H. Jang, F. Liu, T.-H. Cho, T. S. Seo, Carbon 2015, 81, 367. [22] Y. Qi-zhao, Ind. Eng. Chem. Prod. Res. Dev. 1982, 21, 356. [23] G. Zhang, S. Sun, D. Yang, J.-P. Dodelet, E. Sacher, Carbon 2008, 46, 196. [24] H.-J. Shin, K. K. Kim, A. Benayad, S.-M. Yoon, H. K. Park, I.-S. Jung, M. H. Jin, H.-K. Jeong, J. M. Kim, J.-Y. Choi, Y. H. Lee, Adv. Mater. 2009, 19, 1987.
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201500206
www.MaterialsViews.com [25] J. Zhang, H. Yang, G. Shen, P. Cheng, J. Zhang, S. Guo, Chem. Commun. 2010, 46, 1112. [26] S. H. Song, M.-H. Jang, J. Chung, S. H. Jin, B. H. Kim, S. H. Hur, S. H. Yoo, Y.-H. Cho, S. Jeon, Adv. Opt. Mater. 2014, 2, 1016. [27] Ž. Šljivancanin, A. S. Miloševic´, Z. S. Popovic´, F. R. Vukajlovic´, Carbon 2013, 54, 482. [28] P. E. Blöchl, Phys. Rev. B 1994, 50, 17953.
small 2015, DOI: 10.1002/smll.201500206
[29] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865. [30] G. Kresse, D. Joubert, Phys. Rev. B 1999, 59, 1758.
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Received: January 22, 2015 Revised: March 13, 2015 Published online:
www.small-journal.com
9
Graphene Quantum Dots
Is the Chain of Oxidation and Reduction Process Reversible in Luminescent Graphene Quantum Dots? Min-Ho Jang, Hyun Dong Ha, Eui-Sup Lee, Fei Liu, Yong-Hyun Kim, Tae Seok Seo,* and Yong-Hoon Cho*
Graphene-based
quantum dots (QDs) have received a tremendous amount of attention as a new type of light-emitting materials. However, their luminescence origins remain controversial due to extrinsic states of the impurities and disorder structures. Especially, the function of oxygen-contents should be understood and controlled as a crucial element for tuning the optical properties of graphene-based QDs. Herein, a series of graphene oxide QDs (GOQDs) with different amounts of oxygencontents are first synthesized via a direct oxidation route of graphite nanoparticle and thoroughly compared with a series of reduced GOQDs (rGOQDs) prepared by the conventional chemical reduction. Irreversible emission and different carrier dynamics are observed between the GOQDs and rGOQDs, although both routes show a similar tendency with regard to the variation of oxygen-functional components. Their luminescence mechanisms are closely associated with different atomic structures. The mechanism for the rGOQDs can be associated with a formation of small sp2 nanodomains as luminescent centers, whereas those of GOQDs may be composed of oxygen-islands with difference sizes depending on oxidation conditions surrounded by a large area of sp2 bonding. Important insights for understanding the optical properties of graphene-based QDs and how they are affected by oxygen-functional groups are shown.
M.-H. Jang, Prof. Y.-H. Cho Department of Physics and Graphene Research Center of KI for the NanoCentury Korea Advanced Institute of Science and Technology 291 Daehak-ro, Yuseong-gu Daejeon 305-701, South Korea E-mail: [email protected] H. D. Ha, Dr. F. Liu, Prof. T. S. Seo Department of Chemical and Biomolecular Engineering and Institute for the BioCentury Korea Advanced Institute of Science and Technology 291 Daehak-ro, Yuseong-gu, Daejeon 305-701, South Korea E-mail: [email protected] E.-S. Lee, Prof. Y.-H. Kim Graduate School of Nanoscience and Technology Korea Advanced Institute of Science and Technology and Center for Nanomaterials and Chemical Reactions Institute for Basic Science (IBS) 291 Daehak-ro, Yuseong-gu, Daejeon 305-701, South Korea DOI: 10.1002/smll.201500206 small 2015, DOI: 10.1002/smll.201500206
1. Introduction Since the first report of luminescence from graphene oxide (GO),[1,2] the field of graphene-based quantum dots (QDs), consisting of disc-like graphene with a few nanometers in size, has received a tremendous amount of attention, not only as a new type of light-emitting materials but also in relation to a wide variety of technical applications, especially those specific to graphene-based QDs in biological labeling, imaging, and functional nanoscale devices.[1–4] Currently, the luminescence origins are explained in terms of their size,[5–8] shape,[6] fraction of the sp2 domains in the sp3 matrix,[9,10] and/or various functional groups.[10–12] However, since photoluminescence (PL) is simultaneously observed by a combination of these factors, it is difficult to identify the main origin of the PL. For the size effect, both size-dependent[5,8] and size-independent[7,13] PL behaviors have been reported
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
1
full papers www.MaterialsViews.com
in a size range of 2–10 nm, and the calculated bandgap in this size range has shown much lower energy than that of the visible light. Therefore, the origins of the visible PL change due to the size-effect remain controversial, and it is difficult to exclude the effects of extrinsic states, which are derived from oxygen-functional groups (i.e., epoxy, carboxyl, and hydroxyl groups), the edge effect, and the disorder structure of vacancies. Hence, a full understanding of the optical properties from the defect states and judicious control of their PL properties would be essential for a variety of applications. In particular, oxygen-functional groups, a common type of defects, render graphene-based QDs to be helpful for the dissolution in water and to interact with a wide range of organic and inorganic materials, greatly widening the range of applications. Furthermore, they are more controllable than other types of defects when using oxidation and reduction methods. Therefore, the oxygen-functional groups were considered as a crucial element for tuning the optical properties of graphene-based QDs.[5,14–17] Thus far, graphene-based QDs have been mainly prepared with oxidation strategies by the treatment of various carbon sources.[5,14–17] Therefore, most graphene-based QDs are composed of a high level of oxygen contents, which can be called as graphene-oxide QDs (GOQDs). Hence, the GOQDs with different amount of oxygen contents have been commonly prepared by partially reducing the oxygen content through various reduction methods, such as the use of chemical agents,[18] high-temperature annealing,[19,20] and/or photothermal methods.[9,10] For reduced GOQDs (rGOQDs) prepared by a reduction process, the changes in the optical properties have been generally explained by a model of the isolated sp2 nanodomains, resulting in a bluecolored emission. Even though there are several researches about the PL properties of GOQDs depending on reduction degree,[9,10,18–20] the studies on the PL of the graphenebased QDs starting from the pure graphene quantum dot (GQDs) depending on oxidation degree has rarely been done. The PL properties of the graphene-based QDs have recently attracted huge attention, but the clarification for the PL mechanism is not satisfied so far simply by investigating the GOQD reduction process. Therefore, a series of GOQDs and rGOQDs should be prepared experimentally via both oxidation and reduction to understand the detailed whole PL mechanisms and to confirm whether the optical properties are the same or different depending on the degree of oxygen contents. Particularly, the oxidation approach will elucidate the fundamental PL mechanism of GOQDs with controlled oxygen content, and will overcome several drawbacks of the reduction process, such as the requirements of a high-power lamp source, long fabrication times, the use of harmful chemical reagents, and the generation of disordered structures like carbon mono and multivacancies. The limitations of previously reported oxidative cutting methods[5,14–17] will be overcome by exfoliation methods of graphite nanoparticle (GNP),[12,21] which result in the synthesis of single-layered graphene-based QDs with regular shapes and sizes regardless of the oxygen contents. The exfoliation methods of GNPs can generate a series of GOQDs with different amounts of oxygen by controlling chemical
2 www.small-journal.com
oxidation condition. Moreover, it is possible to comparatively analyze the optical properties of GOQDs and rGOQDs at similar levels of oxygen-functional groups. Herein, we suggest an easy route to obtain a series of GOQDs with tunable PL emissions controlled by the chemical oxidation of GNPs. To the best of our knowledge, this is the first report on the synthesis of oxygen-dependent GOQDs without a reduction process. Thus, we prepared all kinds of graphene-based QDs from the GQDs to GOQDs by oxidation as well as from the GOQDs to the rGOQDs by chemical reduction, and performed the in-depth comparative studies on the PL properties of the GQDs, rGOQDs, and GOQDs. The GOQDs have shown clear multicolor emissions depending on the amount of oxygen. We also found that the spectral emission of rGOQDs does not match that of GOQDs at all, in spite of similar oxygen contents. The effects of oxidation and reduction on the optical properties are discussed in details by means of PL, PL excitation (PLE), UV–vis absorbance, and time-resolved PL (TRPL) measurements as well as first-principles density-functional theory (DFT) calculations. This will provide important insights for understanding of how oxygen-functional groups affect the optical properties in graphene-based QDs and whether the successive process of oxidation and reduction is reversible or irreversible to the optical properties of them.
2. Results and Discussion 2.1. Preparation Figure 1 shows a schematic illustration of the proposed route in which we controlled the oxygen-functional groups in graphene-based QDs. Our studies focused on understanding how the optical properties of graphene-based QDs are affected when controlling the number of oxygen defects by means of chemical oxidation and reduction. In the gradual oxidation process, we employed the modified Hummer’s method using a sulfuric/nitric concentrated acid mixture (H2SO4/HNO3), and the number of the oxygen-functional groups on the surface could be controlled depending on the ratio between sulfuric acid and nitric acid.[22,23] The GNPs were chemically exfoliated to produce GOQDs, and the volume ratio of the H2SO4/HNO3 mixture was changed from 0/80 to 60/20. Note that 60/20 means the used volume ratio of the H2SO4 and HNO3 was 60 and 20 mL, respectively. Different oxidation degree resulted in the tunable number of the oxygen-functional groups in the GOQDs, which were monitored by the C1s core-level peaks in a high-resolution X-ray photoelectron spectroscopy (XPS) (Figure S1, Supporting Information). The C1s XPS data show a sp2 carbon peak at ≈284.5 eV, and sp3 carbon peaks including a carbonyl peak (C–O) at ≈286.6 eV, an epoxy (C–O–C) peak at ≈288.0 eV, and a carboxyl (–COOH) peak at ≈289.0 eV. Figure 2a shows a summarized proportion of the oxygen content for GOQDs fabricated with different oxidation, revealing a gradual increase under higher volume ratio of sulfuric acid to nitric acid. Then, we performed the partial reduction process using sodium borohydride (NaBH4), a well-known reducing
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201500206
www.MaterialsViews.com
2.2. Luminescent Characteristics of GOQDs and rGOQDs We examined the luminescent characteristics for all samples under the same UV excitation with a 325 nm at room temperature. Figure 3a–c shows the PL spectra and photographic images of various types of GOQDs and rGOQDs. For the GOQDs, the emission colors of the PL spectra were gradually red-shifted from sky-blue (0.21, 0.24) of the GOQD(00/80) to greenishyellow (0.33, 0.43) of the GOQD(50/30) in the international commission on illumination map (Figure S6, Supporting Information). The highest PL peaks also move toward a lower energy from 2.80 eV of the GOQD(00/80) to 2.35 eV of the GOQD(50/30) as shown in Figure 3d. Hence, GOQDs controlled by the oxiFigure 1. Schematic illustration of the tunable luminescent graphene-based QDs through the dizing conditions are easily able to offer reversible oxidation and reduction processes. tunable PL. In contrast to the oxidation approach, there were nearly unchangeable agent.[24] Starting from the GOQDs which were produced by peak positions of rGOQDs with different amounts of oxygen the treatment of H2SO4/HNO3 (60/20), the GOQD(60/20) (Figure 3d). However, an interesting finding is that a lumineswas gradually reduced by adding a different amount of cence peak centered at 2.8 eV (440 nm) appears after chemNaBH4 into the 10 mL GOQD(60/20) solution. The reduc- ical reduction, and its relative intensity gradually increases tion process was investigated by XPS (Figure S2, Supporting as the amount of reducing agent increases from 1 to 50 mg Information), and summarized in Figure 2a. The XPS results (Figure S7, Supporting Information). As shown in Figure 3d, display that the sp2-carbon structure could be clearly recov- although the amount of oxygen-functional groups of both ered with an increased reducing agents from 0 to 50 mg, by GOQDs and rGOQDs shows similar variation as a function showing the decreased intensity of the oxygenous carbon of chemical oxidation and reduction, the spectral emission peaks. Also, the Fourier transform infrared spectroscopy shows a wholly irreversible emission tendency from oxida(FT-IR) spectra support the existence of oxygenous groups tion to reduction. Therefore, their luminescent characteristics by showing C=O (1725 cm−1), and O–H peak (3400 cm−1) imply that there exist different recombination mechanisms (Figure S3, Supporting Information). Similar to the XPS data, between GOQDs and rGOQDs with different structures due the peaks of oxygenous functional groups tended to increase to the chemical reactions. Furthermore, since it is well known according to the oxidation ratio, whereas those gradually that the oxygen-functional groups are very sensitive to the decreased in proportion to the reduction ratio. pH condition than sp2 carbon bonding,[11,26] we prepared graTo determine the size distribution, we conducted a high- phene-based QDs in the pH-dependent solution (Figures S8 resolution transmission electron microscopy (HRTEM) and and S9, Supporting Information). At pH of 1, the dominated an atomic force microscopy (AFM) analysis for the pristine- PL spectra clearly emit more bluish light than those of pH 7 graphene QDs (GQDs),[12] GOQD(40/40), GOQD(60/20),[21] due to the blue-shifted PL peak position or the increased and rGOQD (5 mg) (Figures S4 and S5, Supporting Infor- shoulder peak intensity of blue emission. On the other hand, mation). Due to the uniform size of the GNPs with average at pH of 14, the PL peak positions are similar to those of pH 7, diameters of 4 nm, all the produced graphene-based QD and but the full-width at half-maximum (FWHM) of the PL specits derivatives show a narrow size distribution between 2 and trum is broader than that of pH 7 due to the shoulder peak of 4 nm with an average diameter of ca. 3.37 nm (Figure 2b). the greenish emission. When the oxygen-functional groups are Thus, the optical properties of our graphene-based QDs were protonated under acidic condition, the greenish emission from dominantly depended on the number of oxygenous defects the oxygen-functional groups becomes inactive in the PL specrather than the size effect. In addition, the heights of the trum. Under alkaline condition, however, the oxygen-funcgraphene-based QDs display a slightly different distributions tional groups are deprotonated and the luminescence from among the categories of samples in shown Figure 2c. Inter- oxygen-functional groups can be maintained.[29] Especially, estingly, the individual sheets of GOQD (40/40) and GOQD in the case of rGOQDs, the degree of the blue-shift from pH (60/20) have an average thickness of 0.9–1 nm, significantly 7 to pH 1 is proportional to the strength of chemical reduchigher than that of pristine-GQDs and rGOQDs (about tion, 50 mg (143 meV) > 10 mg (129 meV) > 5 mg (81 meV). 0.7 nm). These findings could be rationally explained in terms This suggests that as reduction process is applied, more bluish of the oxygen-functional groups (e.g., epoxy) anchored on luminescence dominated than that of GOQDs. More detailed investigation of luminescence mechanisms is as follows. the basal plane of graphene.[25] small 2015, DOI: 10.1002/smll.201500206
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
3
full papers www.MaterialsViews.com
emitting states by losing energy through a relaxation process. Also, the red-shifted region in the 2D matrices may be formed due to the inhomogeneous distributions of graphene-based QDs.[14,17] Thus, we could extract the excitation wavelength (λex) and the invariant emission wavelength (λem) at the point of the starting spectral red-shift in the 2D matrices (Figure S11, Supporting Information). These points are organized according to the composition ratio of oxygenous-/sp2-carbon in the XPS data in Figure 4c,d. The values at both points of excitation and emission wavelength gradually decreased by increasing the fraction of oxygenous carbon (versus sp2 carbon). To further investigate the effect of oxygen-functional groups in a view of band-gaps in graphene-based QDs, we performed the first-principles DFT calculations for oxidized GQD models. The theoretical band-gap, defined as the energy difference between the highest occupied and lowest unoccupied molecular orbitals, was obtained by atomic structures of oxygen-functional groups (i.e., hydroxyl) that are serially conjugated on the nanodisk graphene (Figure S12, Supporting Information). The oxygen-functional groups split the degenerated energy levels in pure GQDs, causing that the band-gap of GQDs with oxygenous defects becomes smaller than that of pure GQDs. Accordingly, this is in good agreement with the optical properties of GOQDs, i.e., the color-tunable emission from blue to greenish-yellow and the increase of the absorption in the visible region (Figure 4e). Figure 2. a) The ratio of oxygenous-/sp2-carbons versus oxidation (red) or reduction Regarding the emission wavelength (blue) conditions. b) Diameter distribution analysis of the pristine-GQD, GOQD(40/40), of the invariant PL under similar oxygen GOQD(60/20), and rGOQD (5 mg). c) Height distribution analysis of the pristine-GQD (black), contents for GOQD and rGOQD in GOQD(40/40) (blue), GOQD(60/20) (green), and rGOQD (5 mg) (red). Figure 4c,d, the PL emission wavelength of GOQDs is shorter than that of the 2.3. Band-Gap Investigation of GOQDs and rGOQDs rGOQDs. This is because the reduction leads to other disordered structures (i.e., vacancies) on basal plane through the To gain insight into the luminescence features in details, removal of oxygen. Our DFT simulation indicates that the we measured the PL spectra of the samples in water sol- formation energy which represents the stability of atomic vent under different excitation wavelength (Figure S10, defects depends on the position of an oxygen atom on the Supporting Information). As shown in Figure 4a,b, the 2D nanodisk graphene, and an oxygen atom on the basal plane matrices were plotted by normalized PL peak intensity of has a higher formation energy than that of an oxygen atom excitation-wavelength-dependent PL (EWD-PL) spectrum. at the edge (Figure S13, Supporting Information). Therefore, Each 2D matrix could be divided into two regions based the reduction process preferentially removes the oxygen on their emission properties. One is mostly invariant with defects anchored on the basal plane of the GOQD. This increasing excitation wavelengths, whereas the other is red- strongly correlates with the increased sp2 fractions after shifted. The invariant PL with various excitation wavelengths chemical reduction and the decreased thickness by removing is closely related to the bandgap of the extrinsic states origi- oxygenous defects on the basal plane. For these reasons, nated from the oxygen-functional groups.[10,11] It reveals that rGOQDs show not only a small change of the luminescent the carriers excited by high energy photons arrive at the properties for extrinsic states, but also the generation of a
4 www.small-journal.com
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201500206
www.MaterialsViews.com
Figure 3. The photoluminescence spectra of a) GOQDs and b) rGOQDs irradiated by a cw 325 nm He–Cd laser at room temperature. c) The photographic images indicate tunable color emission by chemical oxidation (GOQDs) and reduction (rGOQDs), respectively. d) The peak positions of GOQDs (red circles) and rGOQDs (blue circles) were obtained from the photoluminescence spectra; they are plotted according to the oxygenous-/ sp2-carbon ratio.
Figure 4. Tunable luminescent mechanism. 2D matrices via the normalized excitation-wavelength-dependent PL behavior of various types of GOQDs and rGOQDs. The white-circles in the 2D matrices express the peaks of the PL spectra. a) The diagrams show the GOQDs fabricated by 10 mL/70 mL, 30 mL/50 mL, and 50 mL/30 mL (H2SO4/NHO3). b) The diagrams exhibit the cases of 50 mg, 5 mg, and GOQD (0 mg) according to the reducing agent of NaBH4. c,d) The points of the excitation (red-circle) and emission (blue-circle) energy of GOQDs and rGOODs, respectively. The error bar of the x-axis relate to the XPS analysis by several measurement, whereas that of y-axis (blue-circle) is standard deviation of PL peak from 2D matrices. e,f) Normalized UV–vis absorbance spectra of GOQDs and rGOQDs, respectively. The inset in f) shows that a shoulder peak at 285 nm in UV–vis absorbance corresponds to the peak of the PLE spectrum. g) The intensity ratio compared with 285 and 325 nm in PLE spectrum of the detection wavelength at 450 nm. The ratio is calculated by intensity of PLE at 280 nm/intensity of PLE at 320 nm. small 2015, DOI: 10.1002/smll.201500206
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
5
full papers www.MaterialsViews.com
luminescence centered at 2.8 eV (440 nm) due to isolated sp2 domains. The PLE measurement of the rGOQDs shows the distinct peak at 285 nm, corresponding to the shoulder peak at 285 nm as shown in the UV–vis absorbance, as well as broad PLE spectra with a center at approximately 325 nm (Figure 4f). The distinct PLE peak of the rGOQDs is similar to the previously reported optical properties of the intrinsic state in the pristine-GQDs, which is responsible for the blue-colored emission.[12] In addition, the fraction of the intensities of the two parts in the PLE spectrum, a peak at 285 nm and shoulder at 325 nm in Figure S14, Supporting Information, was summarized in Figure 4g, indicating the increase in the strength of the intrinsic states of the rGOQDs as a function of reduction degree. On the other hand, the GOQD(60/20), a fully oxidized sample, shows only broad PLE spectra with a center of about 325 nm (Figure S15, Supporting Information). Therefore, the broad PLE spectra can be attributed to optical transitions from oxygen-functional groups. Interestingly, all GOQD samples have broad PLE spectra regardless of the ratio of oxygen contents (Figure S14, Supporting Information). This means that the emission of GOQDs is mainly associated with oxygen-functional groups instead of the intrinsic states originated from the sp2 carbon.
2.4. In-Depth Study of Carrier Dynamics and Recombination An in-depth study of carrier dynamics and recombination is necessary for accurately understanding the PL phenomenon. The TRPL was carried out using a femtosecond Ti:Sapphire laser and a streak camera detector at room temperature. Figure 5 exhibits the temporal profile of time-integral PL
(TIPL) spectra obtained from streak images. Interestingly, the TIPL of GOQDs and rGOQDs showed very different spectral evolution characteristics as time passed after the pump event. As shown in Figure 5a–c, upon a time delay, the spectral signals of the respective GOQDs retain an invariable emission wavelength due to the presence of only single luminescence origin derived from the extrinsic states of the oxygen-functional groups. All of the rGOQDs, in contrast, show red-shifted spectral migrations as time proceeds, as shown in the Figure 5d–f. These spectra show a fixed peak position at 440 nm in a range of 0–1 ns, and a slight blueshift of the peak position from 515 to 480 nm as a function of the reducing agent in a range of 15–31 ns. This difference between GOQDs and rGOQDs is attributed to different recombination dynamics. According to TIPL spectra in a time range of 15–31 ns, both results of the red-shift from GOQD(10/70) to GOQD(50/30) and the blue-shift from GOQD(60/20) to rGOQD(50 mg) are associated with the change of the bandgap with the extrinsic states originated from oxygen-functional groups.[12] The emission with 440 nm of rGOQDs has significantly shorter PL lifetimes than that of the extrinsic state (Figure S17, Supporting Information). These properties are in good agreement with previously reported results,[12] corresponding to the fast decay of the blue emission in the pristine-GQDs and the slow decay of green emission in the GOQDs (Figure S16, Supporting Information).The blue emission with fast recombination in the rGOQDs originates from the intrinsic states of the sp2 nanodomains. Their dynamical spectral migration clearly supports that the luminescent origin of GOQDs is only the extrinsic states derived from oxygen-functional groups, whereas that of the rGOQDs include the extrinsic states as well as the intrinsic states formed by chemical reduction.
Figure 5. The time-dependent integral PL spectra from the GOQDs (blue) and rGOQDs (red) groups under excitation at 266 nm by a thirdharmonic-generation femtosecond pulse laser. The diagrams of a), b), and c) [d), e), and f)] show the results from GOQDs [rGOQDs] fabricated with 10 mL/70 mL, 30 mL/50 mL, and 50 mL/30 mL of H2SO4/HNO3 [50, 5, and 0 mg of NaBH4], respectively.
6 www.small-journal.com
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201500206
www.MaterialsViews.com
Figure 6. a,b) Structural models of GOQDs and rGOQDs for chemical oxidation and reduction. Note that the white and yellow regions represent sp2 bonding and sp3 bonding formed by oxygen-functional groups, respectively, and the regions of v (red color) describe vacancies.
2.5. Luminescent Model from GOQDs to rGOQDs It is worth noting that since the optical properties of the GOQDs in the oxidation process are quite distinct from that of the rGOQDs in the reduction process (Figures 3–5), the structural model of the GOQDs under the oxidation process of pristine GQDs would be very dissimilar to that of the rGOQDs under the reduction process of GOQDs. For the latter, it has been sucessfully explained by the formation of numerous isolated sp2 nanodomains within the sp3 bonding of the oxygen-functional groups by removing the oxygenbonding randomly under the chemical reduction process (Figure 6d,e).[9,10] Based on this model, we can attribute the appearance of the blue light emission to the intrinsic states associated with isolated sp2 nanodomains in the rGOQDs under the reduction process, which have higher photon energy than the green light emission due to the extrinsic states formed by the oxygen-functional groups. For the former, on the other hand, we propose a model of oxygen-islands surrounded by a large area of sp2 carbon bonding during the oxidation process of the pristine GQD (Figure 6a,b). When the oxidation process proceeds, oxygen-islands are gradually formed on the graphene surface and then the size of oxygenislands increases.[27] As a result, the area of sp2 carbon bonding decreases, unlike the formation picture of sp2 nanodomains in rGOQDs during the reduction process. Therefore, for the GOQDs under the oxidation process of the pristine GQD, it would be difficult to observe a blue light emission originated from sp2 nanodomains, as opposed to the case of rGOQDs. In contrast to a simple speculation of a reversible process through chemical oxidation and reduction, we conclude that the chain of the oxidation and reduction process is irreversible in the luminescent GQDs, since graphene surface forms different structures during oxidation and reduction processes; the GOQDs are composed of oxygen-islands (with difference sizes depending on different oxidation conditions) surrounded by a large area of sp2 carbon bonding, whereas the rGOQDs have numerous isolated sp2 nanodomains surrounded by the sp3 bonding of the oxygen-functional groups.
3. Conclusion We successfully synthesized a series of GOQDs with different amounts of oxygen by controlling the oxidation conditions for GNPs and observed tunable PL of them. We also confirmed the irreversible spectral emission in the graphenesmall 2015, DOI: 10.1002/smll.201500206
based QDs modulated by a chemical route from oxidation to reduction, although both routes showed a similar tendency with regard to the changed amount of oxygen-functional components. Their light-emission phenomena were closely associated with their atomic structure on the graphene-based QDs. In particular, GOQDs controlled by the oxidation conditions have several advantages over rGOQDs, such as rapid fabrication process, clear tunable PL, and stable spectral PL for carrier relaxation. We believe that our studies provide important insights for understanding the optical properties of the graphene-based QDs and how they are affected by oxygen-functional groups. Our results can lead to future development of novel strategies for guiding the synthesis of the graphene-based QDs with tunable optical properties toward various applications such as bioimaging, sensor, biological labeling, and functional nanoscale devices.
4. Experimental Section Chemicals and Materials: Graphite nanoparticles (average diameter: 3–4 nm, 93%) were purchased by Skyspring Nanomaterials (USA). Ethanol (reagent grade) and sodium borohydride (NaBH4) were obtained from Sigma-Aldrich (USA). Extra-pure sulfuric acid (H2SO4, 95%), extra-pure nitric acid (HNO3, 60%–62%), and sodium hydroxide (NaOH, guaranteed reagent) were purchased from JUNSEI (Japan). Milli-Q water (Millipore, Milford, MA, USA) was used for all aqueous solutions. Preparation of the Samples: (i) The pristine GQD solution was prepared by shaking 5 mg of GNPs in a 10 mL ethanol/H2O (1:1 v/v ratio) mixture on a vortex mixer (Scientific Industries, USA) at 2700 rpm for 1 h. After centrifugation at 2000 rpm for 30 min, the pure GQD solution was collected from the supernatant. (ii) Graphene oxide QDs (GOQDs) were fabricated in which 1 g of GNPs was pre-oxidized in a concentrated H2SO4 and HNO3 solution (80 mL) for 3 h under mild ultrasonication (100 W). The ratio of H2SO4 and HNO3 ranged from 0 mL/80 mL to 60 mL/20 mL, and the oxidizing power increased as the H2SO4 portion was higher. After pre-oxidation, the mixture was stirred at 100 °C under a refluxing condition for 12 h. The mixture was cooled and diluted with 150 mL of deionized (DI) water in an ice bath. The suspension was centrifuged at 20000 rpm and the supernatant was carefully separated. The supernatant was pre-dialyzed to remove acidic ions for 1 d and the pH was adjusted to 8 with a diluted NaOH solution. The final product solution was further dialyzed in a dialysis bag (Molecular weight cut-off: 1000 Da) to remove the salt ions completely for 3 d. All GOQDs were synthesized under the identical
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
7
full papers www.MaterialsViews.com
conditions except the H2SO4/HNO3 volume ratio. (iii) rGOQDs were prepared by adding 1, 5, 10, and 50 mg of NaBH4 in 10 mL of the GOQD solution which was synthesized by a 60/20 H2SO4/HNO3 solution. Each sample was stirred at room temperature for 3 h. The products were dialyzed in a dialysis bag (Molecular weight cut-off: 1 kDa) to remove sodium ions or residues for 1 d. Optical Measurements and Characterization: Atomic force microscope (AFM, Veeco D3100, USA) images were obtained under the conditions of relative humidity of 43%, a tapping mode at a 1–3 Hz scan rate and a 512 × 512 pixel resolution. High resolution transmission electron microscope images were characterized with an ultra-thin carbon film on a holey carbon support film (400 mesh) at an acceleration voltage of 300 kV. X-ray photoelectron spectroscopy was conducted by a Thermo VG scientific Sigma Probe system, and FT-IR was conducted by IR Prestige-21 (DongIl SHIMADZU Corp.) over the range from 600 to 4000 cm−1. All of the luminescent data were obtained by a hand-made setup using precision cells made of Suprasil quartz. The PL and the EWD-PL behaviors were assessed using a cw 325 nm He-Cd laser, monochromatic light from a 300 W-xenon lamp, and UV spectrometers (Maya2000, Ocean Optics, USA) as a PL detector at room temperature. The PLE was measured by monochromatic light from a 300 W Xenon lamp and a high-sensitive photomultiplier tube as a PL detector. The absorption spectra were recorded on a UV–vis spectrophotometer (UV-2450, SHIMADZU, Japan). In order to elucidate the recombination dynamics, we carried out time-resolved PL experiments. A mode-locked femtosecond pulsed Ti:Sapphire laser (Coherent, Chameleon Ultra II) system was used as an excitation source [266 nm (third-harmonic-generation)], and a streak camera (Hamamatsu, C7700-01) was utilized to measure the decay profile of the PL spectra at room temperature. DFT Simulation: We employed projector-augmented wave (PAW)[28] pseudopotentials and the Perdew–Burke–Ernzerhof (PBE) exchange-correlation functional,[29] as implemented in the Vienna Ab Initio Simulation Package (VASP).[30] The kinetic energy cutoff of 400 eV are used for expansion of plane wave basis set and all atomic forces are fully relaxed until the Hellman–Feynman forces are less than 0.025 eV Å−1. The hexagon-like GQD models are used with vacuum separation of about 10 Å for calculating the electronic structure and the formation energy (Eform = Eoxgenous_grphene – ∑niμi, where i is atomic species, and μ is the chemical potential using graphene, oxygen molecule, and hydrogen molecule as a reference state of each species). All edges of GQDs are terminated with sp2 hydrogen atoms to remove the dangling electrons.
Supporting Information Supporting Information is available from the Wiley Online Library or from the author.
Acknowledgements M.-H.J. and H.D.H. contributed equally to this work. This work was supported by the National Research Foundation (NRF-
8 www.small-journal.com
2013R1A2A1A01016914, NRF-2013R1A1A2011750) of the Ministry of Education, the Converging Technology Project funded by the Korean Ministry of Environment (M112-00061-0002-0), and the GRC project of KI for the NanoCentury. E.-S.L. and Y.-H.K. were supported by the Global Frontier R&D program by the Center for Multiscale Energy Systems (2011-0031566) and the Institute for Basic Science (IBS) (IBS-R004-D1-2014-a00).
[1] L. Li, G. Wu, G. Yang, J. Peng, J. Zhao, J.-J. Zhu, Nanoscale 2013, 5, 4015. [2] Z. Luo, P. M. Vora, E. J. Mele, A. T. C. Johnson, J. M. Kikkawaa, Appl. Phys. Lett. 2009, 94, 111909. [3] H. Sun, L. Wu, W. Wei, X. Qu, Mater. Today 2013, 16, 433. [4] J. Shen, Y. Zhu, X. Yanga, C. Li, Chem. Commun. 2012, 48, 3686. [5] J. Peng, W. Gao, B. K. Gupta, Z. Liu, R. R. Aburto, L. Ge, L. Song, L. B. Alemany, X. Zhan, G. Gao, S. A. Vithayathil, B. A. Kaipparettu, A. A. Marti, T. Hayashi, J. J. Zhu, P. M. Ajayan, Nano Lett. 2012, 12, 844. [6] S. Kim, S. W. Hwang, M. K. Kim, D. Y. Shin, D. H. Shin, C. O. Kim, S. B. Yang, J. H. Park, E. Hwang, S. H. Choi, G. Ko, S. H. Sim, C. Sone, H. J. Choi, S. Bae, B. H. Hong, ACS Nano 2012, 6, 8203. [7] L. Tang, R. Ji, X. Cao, J. Lin, H. Jiang, X. Li, K. S. Teng, C. M. Luk, S. Zeng, J. Hao, S. P. Lau, ACS Nano 2012, 6, 5102. [8] W. Kwon, Y.-H. Kim, C.-L. Lee, M. Lee, H. C. Choi, T.-W. Lee, S.-W. Rhee, Nano Lett. 2014, 14, 1306. [9] G. Eda, Y.-Y. Lin, C. Mattevi, H. Yamaguchi, H.-A Chen, I.-S. Chen, C.-W. Chen, M. Chhowalla, Adv. Mater. 2010, 22, 505. [10] C.-T. Chien, S.-S. Li, W.-J. Lai, Y.-C. Yeh, H.-A. Chen, I.-S. Chen, L.-C. Chen, K.-H. Chen, T. Nemoto, S. Isoda, M. Chen, T. Fujita, G. Eda, H. Yamaguchi, M. Chhowalla, C.-W. Chen, Angew. Chem. Int. Ed. 2012, 51, 1. [11] S. H. Jin, D. H. Kim, G. H. Jun, S. H. Hong, S. Jeon, ACS Nano 2013, 7, 1239. [12] F. Liu, M.-H. Jang, H. D. Ha, J.-H. Kim, Y.-H. Cho, T. S. Seo, Adv. Mater. 2013, 25, 3657. [13] Q. Xu, Q. Zhou, Z. Hua, Q. Xue, C. Zhang, X. Wang, D. Pan, M. Xiao, ACS Nano 2013, 7, 10654. [14] Y. Li, Y. Hu, Y. Zhao, G. Shi, L. Deng, Y. Hou, L. Qu, Adv. Mater. 2011, 23, 776. [15] D. Pan, J. Zhang, Z. Li, M. Wu, Adv. Mater. 2010, 22, 734. [16] J. Luo, L. J. Cote, V. C. Tung, A. T. L. Tan, P. E. Goins, J. Wu, J. Huang, J. Am. Chem. Soc. 2010, 132, 17667. [17] Y. Dong, C. Chen, X. Zheng, L. Gao, Z. Cui, H. Yang, C. Guo, Y. Chi, C. M. Li, J. Mater. Chem. 2012, 22, 8764. [18] a) Z. Gan, S. Xiong, X. Wu, T. Xu, X. Zhu, X. Gan, J. Guo, J. Shen, L. Sun, P. K. Chu, Adv. Optical Mater. 2013, 1, 926; b) S. Zhu, J. Zhang, X. Liu, B. Li, X. Wang, S. Tang, Q. Meng, Y. Li, C. Shi, R. Hu, B. Yang, RSC Adv. 2012, 2, 2717; c) S. Zhu, J. Zhang, S. Tang, C. Qiao, L. Wang, H. Wang, X. Liu, B. Li, Y. Li, W. Yu, X. Wang, H. Sun, B. Yang, Adv. Funct. Mater. 2012, 22, 4732. [19] K. P. Loh, Q. L. Bao, G. Eda, M. Chhowalla, Nat. Chem. 2010, 2, 1015. [20] T. V. Khai, H. G. Na, D. S. Kwak, Y. J. Kwon, H. Ham, K. B. Shim, H. W. Kim, J. Mater. Chem. 2012, 22, 17992. [21] H. D. Ha, M.-H. Jang, F. Liu, T.-H. Cho, T. S. Seo, Carbon 2015, 81, 367. [22] Y. Qi-zhao, Ind. Eng. Chem. Prod. Res. Dev. 1982, 21, 356. [23] G. Zhang, S. Sun, D. Yang, J.-P. Dodelet, E. Sacher, Carbon 2008, 46, 196. [24] H.-J. Shin, K. K. Kim, A. Benayad, S.-M. Yoon, H. K. Park, I.-S. Jung, M. H. Jin, H.-K. Jeong, J. M. Kim, J.-Y. Choi, Y. H. Lee, Adv. Mater. 2009, 19, 1987.
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201500206
www.MaterialsViews.com [25] J. Zhang, H. Yang, G. Shen, P. Cheng, J. Zhang, S. Guo, Chem. Commun. 2010, 46, 1112. [26] S. H. Song, M.-H. Jang, J. Chung, S. H. Jin, B. H. Kim, S. H. Hur, S. H. Yoo, Y.-H. Cho, S. Jeon, Adv. Opt. Mater. 2014, 2, 1016. [27] Ž. Šljivancanin, A. S. Miloševic´, Z. S. Popovic´, F. R. Vukajlovic´, Carbon 2013, 54, 482. [28] P. E. Blöchl, Phys. Rev. B 1994, 50, 17953.
small 2015, DOI: 10.1002/smll.201500206
[29] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865. [30] G. Kresse, D. Joubert, Phys. Rev. B 1999, 59, 1758.
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Received: January 22, 2015 Revised: March 13, 2015 Published online:
www.small-journal.com
9
