In this section, we arrange the SiNPs having a specific dimension and an
internal magnetic dipole resnance upon the CVD-as-grown graphene/Cu
foil and cleverly give rise to a dramatically intense electromagnetic field
around the graphene. Through the strong electromagnetic field
(E
2/E
02~123.2), the graphene can be employed to enhance its Raman
signals (up to 206 times).
5.1 Introduction
Due to graphene emerging extraordinary electronic [98], thermal [99], optical [100] and mechanical properties [101], this unique two-dimensional material, graphene, has been attracting considerable interests in both scientific and technologic communities. In the past two decades, following the successful producing single layer graphene by micromechanical exfoliation [102], many methods have been developed to produce large-area and high-quality graphene. For example, to grow uniformly ordered graphene on an insulating substrate, the techniques of preparing large-area graphene by thermal decomposition of silicon carbide (SiC) has been developed [103, 104]; to produce graphene using cost-effective graphite as raw material, the chemical method of reduction graphene oxide has been proposed [105, 106]; to obtain graphene with smooth edges and controllable widths, the approach of plasma cutting carbon nanotubes has been presented [107, 108]. In addition to the above described methods, to stably obtain graphene, chemical vapor deposition (CVD) large-area crystal carbon atom on metal surface is widely accepted as an effective method [109-116]. To improve the quality of CVD-grown graphene, scientists launched a series of investigations and started discussing differences of various metal substrates, such as Ni [109-111], Ru [112, 113], and Cu. Because of carbon displaying extremely low solid-solubility in Cu, CVD growth of graphene on polycrystalline Cu foil gradually
becomes a mainstream [114-116]. More recently, comparison to CVD growth of graphene on polycrystalline Cu foil, CVD-grown single layer crystal carbon atom on single crystal Cu substrate has been reported to produce higher quality graphene by removing the influences of Cu grain boundaries and reducing the concentration of nucleation sites of carbon atom [117-119]. While the gradual maturity of CVD-grown graphene, in-situ and non-destructive characterizing the properties of the graphene on a Cu substrate became a considerably important issue.
Raman scattering spectroscopy, a powerful, efficient and non-destructive method, provides abundant information relating to the bonding, quality, defect, and vibration of molecules. Raman scattering is also a practical method for characterizing the structural and physical properties of the graphene, such as the atomic structures of disorder, defects, edge and strain. Although Raman spectroscopy is very useful for identifying the properties of graphene, unfortunately, the as-grown graphene on Cu foil provides a very weak Raman signal. Previous investigations generally considered the weak Raman signal of graphene on Cu foil surface because of following reasons:
(1) the monolayer carbon atoms of graphene absorbs only a quite small portion of incident light to generate the Raman scattering signal [120]. (2) Due to the interference between a Raman incident light and its reflection light, the Cu foil surface presents a low localized electric field which is a disadvantageous environment
for the Raman scattering of the graphene [121]. Besides, the background noises of Cu foil are also attributed to an adverse influence for characterizing Raman signals of CVD-grown graphene [122, 123]. To improve the weak Raman signal of CVD-grown graphene on Cu foil, transferring the graphene from Cu foil onto particularly designed substrates is a fairly common practice. These substrates, such as oxide/Si substrates [124, 125], one-dimensional photonic crystals [126, 127], and nanocavity structure [121] etc., provide a high localized electric field at the interface between graphene and substrates to increase the light absorption and the Raman scattering of the graphene. However, the transferring involving chemical processes could cause unavoidable changes of graphene like defects, impurities, or stresses. In these manners, the transferred graphene could not faithfully display the original characteristics of the CVD-as-grown graphene on Cu foil. Therefore, the graphene transferring process substantially reduces the advantages of the in-situ and non-destructive Raman characterization.
To overcome the drawbacks of transferring, the use of metal nanostructures is one of major approaches toward enhancing the Raman scattering of graphene [128-131]. Through integrating graphene with plasmonic nanostructures, the localized electric field between the graphene and metal nanostructures can be used to significantly increase the light-graphene interaction. To enhance the weak Raman
signals of CVD-as-grown graphene on Cu foil, some methods of integrating metal nanostructures and the graphene/Cu foil have been proposed [22, 131]. For instances, Zhao et al. deposited Au nanoislands onto the surface of graphene/Cu foil for enhancing the Raman signals of as-grown graphene [22]; Xiang et al. observed higher Raman signals of as-grown graphene by thermal annealing Au film to form Au nanoparticles on a graphene/Cu foil [131]. These studies successfully enhanced Raman signals of the CVD-grown graphene on Cu foils by metal nanostructures.
However, the approach of using metal nanostructure to enhance Raman signals of the graphene could result some serious disadvantages. For example, integrating a graphene with metal nanostructures would affect the carrier mobility and surface properties of the graphene [132]. Moreover, metal nanostructures might difficultly be removed from the surfaces of the CVD-grown graphene on Cu foils. These changes of graphene, similarly, are contrary to the original intention of characterizing the as-grown graphene by in-situ and non-destructive Raman scattering method.
The dielectric nanostructure having high refractive index, such as silicon nanoparticle (SiNP), was experimentally demonstrated to generate excellent magnetic response in the visible spectral range [133-136]. The magnetic resonance which is accompanied with displacement current would induce a strong electromagnetic field inside and around the resonant SiNP [136]. In this study, we arrange the SiNPs having
a specific dimension upon the CVD-as-grown graphene/Cu foil and cleverly give rise to a dramatically intense electromagnetic field around the graphene. Through numerical simulations, we found the strong electromagnetic field around the graphene originates from the coupling between the concentrated electromagnetic field of SiNP of magnetic resonance and the free carriers of the underlying Cu foil. We also demonstrated the electromagnetic field around the graphene can be employed to enhance the Raman signals of the graphene. We observed the enhanced Raman signal of as-growth graphene on a Cu foil is even higher than the Raman signal of the graphene transferred on a 300 nm oxide/Si substrate. It is worth mentioning that we experimentally observed this approach can enhance the Raman signal of graphene under causing extremely low influence of original characteristics of the CVD-as-grown graphene. Moreover, the arranged SiNPs on the CVD-grown graphene can also be easily removed without destroying the graphene. We simultaneously present that the proposed Raman signal enhancement approach can be used to easily characterize the quality difference of the graphene growing on the Cu foils having various sizes of grain boundaries. The enhanced approach could assist to practice the in-situ and non-destructive Raman characterization of CVD-as-grown graphene on a Cu foil. It’s an interesting and valuable Raman signal enhancement approach which may open a door of new vision for optically measuring the
characteristics of the CVD-grown graphene on a Cu foil.
5.2 Experimental Section
Preparations of CVD-Grown Graphene:
CVD was used for the growth of graphene on various Cu foils [137]. The polycrystalline Cu foils which were purchased from Nilaco Inc. were placed on a hot wall furnace comprising a 1 inch fused silica tube. For growing the single layer graphene on the Cu foils which have various grain sizes, we introduced two different procedures. First, the furnace was heated to 1000° C. The polycrystalline Cu foils in the furnace were then annealed during 3 hours for changing the grain sizes. After annealing, a reduction process was conducted in H2 flow prior to the introduction of CH4. Then, the single layer graphene was synthesized by Cu-catalyzed using CH4 as the carbon source. After a growth time of 30 min, CH4 was then turn off and the system was cooled in Ar flow to reach room temperature. For comparison, another batch of polycrystalline Cu foils was used to directly grow graphene without annealing. Both procedures were conducted at low pressure (~500 mTorr).
Preparations of CVD-Grown Graphene on 300 nm Oxide/Si Substrate:
The commercial 300 nm oxide/Si substrates were cleaned sequentially with acetone, isopropyl alcohol (IPA), and deionized water and then dried under a flow of N2. The single layer CVD-grown graphene was transferred onto the 300 nm oxide/Si
substrate over an area of approximately 1 cm2 through polymer-mediated transfer [127]. First, a small amount of PMMA (MicroChem Corp.) as the transfer medium was drop-coated upon the CVD-grown graphene on the Cu foil. After the PMMA was slowly cured at room temperature, the Cu foil was then etched away by using an aqueous solution of iron nitrate (1 M) for 1 h. When the Cu foil was completely removed, the PMMA/graphene stack was cleaned by deionized water, and then placed on the 300 nm oxide/Si substrate and dried under a flow of N2. The precoated PMMA was dissolved by dropping a second liquid PMMA solution onto the PMMA/graphene stack. Finally, the transferred graphene on the 300 nm oxide substrate were soaked in acetone for 1 h to remove the residual PMMA.
Preparations of SiNPs on Graphene/Cu Foil:
SiNPs (particle size: ca. 120 nm) were purchased from Emaxwin Technology.
The SiNPs were dispersed (at concentrations 5 10–3 wt%) in ethanol under ultrasonication for 30 min. The SiNP suspensions were then spun onto graphene/Cu foils at a spinning rate of 3000 rpm. The SiNPs can be removed from the graphene/Cu foil by placing the SiNPs/graphene/Cu foil in ethanol under ultrasonication for 1 hour.
Preparations of AuNPs on Graphene/Cu Foil:
The graphene/Cu foil was immersed in 1 mM (3-aminopropyl)trimethoxy-silane (APTMS, Sigma–Aldrich) in ethanol and incubated for 6 h. After a self-assembled
monolayer (SAM) of APTMS had formed on the graphene/Cu foil surface, the graphene/Cu foil was rinsed sequentially with ethanol and deionized water to remove any non-bound APTMS. An aliquot (300 μL) of an AuNP suspension (Ted Pella) was placed on the graphene/Cu foil for 1 h. After the AuNPs were immobilized on the graphene/Cu foil surface, the samples were rinsed with deionized water and dried under a flow of N2.
Measurements and Characterizations:
The SiNPs/graphene/Cu foils were observed by using scanning electron microscopy (SEM, NOVA NANO 450). To observe the grain boundaries of Cu foils,
the optical images were obtained from an optical microscope (Olympus) focusing by a 20 objective with the numerical aperture of 0.4. The Raman spectra of the graphene
were recorded using a commercial micro-Raman microscope (UniRAM, UniNanoTech) equipped with a monochromator having a focal length of 75 cm. The
wavelength of the excitation laser line was fixed at 532 nm (diode laser). The laser beam was focused by a 100 objective with the numerical aperture of 0.95. The spot
size of the excitation laser was approximately 0.4 μm2; the integration time was 25 s.
5.3 Results and Discussion
Mie theory illuminates that the first and second lowest frequency resonances of dielectric particles correspond to the terms from magnetic and electric dipoles. The
magnetic dipole resonance occurs at a value of /n of approximately dp, where is the
wavelength of incident light, n is the refractive index of the particle, and dp is the particle diameter [133]. In the case of SiNPs, the magnetic dipole of a SiNP having a diameter of 120 nm would resonate at a wavelength of approximately 530 nm. The resonant SiNPs would induce a strong electromagnetic field inside and around the SiNPs [136]. This electromagnetic field from magnetic dipole resonance of the SiNPs could be coupled to the CVD-grown graphene/Cu foil to generate a concentrated electromagnetic field around the graphene. In other words, the Raman signal of the CVD-grown graphene on Cu foil could be enhanced through the locally concentrated electromagnetic field. Here, we extend a practical study to investigate a Raman scattering enhancement technique of CVD-grown graphene on Cu foils.
To study the complete picture of magnetic dipole resonance occurring in SiNPs placed upon graphene/Cu foil system, we started to investigate the electric and magnetic field distributions of a Raman excitation light inside a SiNP/graphene/Cu foil system. We used a three-dimensional finite-difference time-domain (3D-FDTD) approach to simulate the behavior of excitation light interacting with a SiNP/graphene/Cu foil system. The 3D-FDTD simulations revealed the electromagnetic fields over the entire computational domain as they evolved over time, providing analysis of the electric and magnetic field distributions on the various
visual angles and profiles. Figure 5-1a displays a schematic representation of the simulation model: an incident plane wave having a wavelength of 532 nm is incident onto a SiNP/graphene/Cu foil system. In the system, the SiNP having a diameter of 120 nm is placed upon a single layer graphene (thickness of 0.3 nm) and a connected Cu substrate (thickness of 200 nm). According to the incident light setting an x polarization direction of the electric field, we arranged flat monitors at y-z and x-z planes to observe the magnetic and electric field distributions of cross-section, respectively. Figure 5-1b displays the simulated magnetic field distribution of cross-section of SiNP/graphene/Cu system; the corresponding electric field distributions in the x-z plane were displayed in Figure 5-1c. In these two figures, the black round lines represent the SiNP; the black dotted lines indicate the positions of the single layer graphene; the black arrows show the directions of the electric fields;
and the colors map the intensities of magnetic/electric fields. Due to the wavelength of incident light corresponding to magnetic dipole resonance condition of the SiNP, Figure 5-1b displays an obvious and intense magnetic hot spot near the center of the SiNP. The directions of the electric fields (black arrows) inside the SiNP reveal the magnetic dipole resonance occurring accompanies a displacement current loop. To observe the electric field distribution of the SiNP/graphene/Cu system when the magnetic dipole resonance occurring on the SiNP, we found two remarkably
concentrated electric field hot spots between the SiNP and Cu substrate, as shown in Figure 5-1c. These two electric field hot spots which are separated at the both sides by the SiNP exactly exist at the position of graphene (red dotted line). Simultaneously, we observed the circular electric field distribution inside the SiNP has a notch faces downward Cu substrate. The electric field distribution apparently suggests the electromagnetic field in the SiNP has an interplay with the underneath graphene/Cu substrate. Similar to the case of a plasmonic nanoparticle/spacer/film system [138, 139], the SiNP positioned upon a Cu substrate (metallic film) could induce image charges of high density within the Cu substrate; the interplay between the SiNP and its image charges would induce a large enhancement in the electromagnetic field in the graphene (spacer). Therefore, we infer the intense electric field hot spots are generated from the charge coupling between the displacement current loop inside the SiNP and the concentrated free carriers inside the Cu substrate. We also describe a schematic representation of the displacement current loop, as shown in Figure 5-1a (red portion).
(a)
(b)
(c)
(e) (d)
Figure 5-1 (a) Schematic representation of our simulation model: an incident plane wave having a wavelength of 532 nm is incident onto a SiNP/graphene/Cu substrate system. The simulated magnetic field intensity (b) and electric field intensity (c) distributions of cross-section of the SiNP/graphene/Cu system. The electric field intensity distributions of a graphene coating with a SiNP (d) and a bare graphene (e), respectively, on a Cu substrate.
To further understand the enhancement in the electric field within graphene, we used a flat monitor along x-y plane located at the position of the graphene to observe the electric field distribution. For a comparison, Figure 5-1d and 5-1e display electric field distributions of a graphene coating with a SiNP both on Cu substrates and a bare graphene, respectively. Obviously, the electric field distribution of the graphene coating a SiNP appears two significant hot spots located at both sides of the SiNP corresponding to the simulation of Figure 5-1c, as shown in Figure 5-1d. Remarkably, an intense enhancement in electric field intensity arises at the locations of the hot spots (E2/E02~123.2). In addition, the range of the enhanced electric field intensity apparently exceeds the physically projective area of the SiNP. This result implies the resonant SiNP can efficiently enhance a large-scale electric field hot zone within the graphene in the SiNP/graphene/Cu foil system. We will further discuss the range of the enhanced electric field intensity in following section. Differentially, the electric field intensities within the bare graphene on the Cu substrate uniformly distribute including the entire range of simulated domain, as shown in Figure 5-1e. Due to the interference between the incident light and its reflective light occurring at the metal surface [121], the electric field intensity within the bare graphene on the Cu substrate is quite low (E2/E02~0.42).
Due to the Raman signal enhancement involving the enhanced excitation effect
of an incident laser and the enhanced spontaneous emission of a Raman scattered light, it has generally been considered that the enhancement factor of Raman signals approximates proportionally to the fourth power of the electric field (E4) [140].
Therefore, to employ the concentrated electric field intensity within the graphene in a SiNP/graphene/Cu foil system for enhancement the Raman signals of the as-grown graphene, we arranged SiNPs upon the CVD-grown graphene on Cu foil to carry out the Raman signal measurement of the graphene. Figure 5-2a depicts a schematic representation of an experimental Raman spectra measurement of SiNPs/graphene/Cu foil. The SiNPs having a diameter of 120 nm are chosen to induce the magnetic dipole resonance for the wavelength of the Raman excitation laser (532 nm). Here, we deliberately arranged the low surface coverage of SiNPs to ensure that (1) the well-dispersed SiNPs would generate the magnetic dipole resonance without the interference from vicinal SiNPs; (2) the strong electromagnetic field within the graphene is only originated from the coupling effect between SiNPs and the Cu foil;
(3) the incident excitation laser and the spontaneous Raman scattering emission from the graphene would not be substantially absorbed by aggregated SiNPs. The SEM image of the well-dispersed SiNPs on a surface of a graphene/Cu foil will be presented later in this study. Figure 5-2b displays the measured Raman spectra of the graphene under different treatment. We compare the Raman spectra of the graphene
on a Cu foil, the graphene on a 300 nm oxide/Si substrate, and the graphene in the SiNPs/graphene/Cu foil system, respectively. First, we observed a weak Raman signal from a CVD-grown graphene on a Cu foil, as the black line shown in figure. Although 2D band of the graphene can be observed, G band is not quite obvious to be found.
After transferring the graphene to the 300 nm oxide/Si substrate, the Raman signal of the transferred graphene is significantly enhanced so that both 2D and G bands can be easily observed, as the red line shown in figure. Then, interestingly, after placing the SiNPs upon the CVD-grown graphene/Cu foil of the same batch, the Raman signals of both 2D and G bands of the sandwiched graphene is dramatically enhanced and even beyond the Raman signals of the graphene on the 300 nm oxide/Si substrate, as the blue line shown in figure. In addition, the graphene in these three cases are not observed any D band. Generally, due to an oxide/Si substrate is beneficial to have a higher surface electric field intensity than a Cu foil, transferring the CVD-grown graphene from the Cu foil onto the oxide/Si substrate is usually employed to enhance the graphene's visibility and Raman scattering signals. However, placing the SiNPs upon the CVD-grown graphene/Cu foil reveals a more efficient Raman enhancement
After transferring the graphene to the 300 nm oxide/Si substrate, the Raman signal of the transferred graphene is significantly enhanced so that both 2D and G bands can be easily observed, as the red line shown in figure. Then, interestingly, after placing the SiNPs upon the CVD-grown graphene/Cu foil of the same batch, the Raman signals of both 2D and G bands of the sandwiched graphene is dramatically enhanced and even beyond the Raman signals of the graphene on the 300 nm oxide/Si substrate, as the blue line shown in figure. In addition, the graphene in these three cases are not observed any D band. Generally, due to an oxide/Si substrate is beneficial to have a higher surface electric field intensity than a Cu foil, transferring the CVD-grown graphene from the Cu foil onto the oxide/Si substrate is usually employed to enhance the graphene's visibility and Raman scattering signals. However, placing the SiNPs upon the CVD-grown graphene/Cu foil reveals a more efficient Raman enhancement