Optical, photonic and optoelectronic properties of graphene, h-BN and their hybrid materials

1.3.1 Linear optical absorption

Graphene absorbs 2.3% of the incident light, even at only one atomic thickness, and the frequency of the incident light has nothing to do with the absorption intensity (Figure 3) [26]. These unique optical properties are generated by the electrons’ conical band and holes at the Dirac point.

Figure 3:

(A) Schematic diagram of optical micrograph for different layers [24]. (B) Transmittance spectrum of single layer graphene [26].

Graphene on top of a Si/SiO₂ substrate can be identified by using optical image contrast [24]. Geim and co-workers studied the image where SiO₂ was used as a substrate representing the relationship between multilayer graphene interference [26]. They converted the incident light and the substrate’s thickness to maximize the contrast. According to the Fresnel equation [35], the photoconductivity of the suspended monolayer grapheme G₀=e²/4ℏ gives the transmissivity T_opt=(1+2πG/c)⁻²≈ 1–πα≈0.977. This is because the valence band and the conduction band of graphene intersect near the Dirac point and the photoconductivity of graphene in the photonic band-gap over a wide G₁ (w)=e² is independent of the frequency, but depends on the fine structure constant α, which is unrelated to the frequency. In the visible light region, the reflectance of each layer A=1–T≈πα≈2.3% as is shown in Figure 3A. At the atomic scale, graphene exhibits a strong broadband absorption (wavelength range of 300–2500 nm), which is about 50 times that of GaAs with the same thickness.

Electrons in graphene, as nonmass 2D particles, result in a very important nonwavenumber absorption (πα≈2.3%) for normal incident light below 3 eV [35]. In addition, when the light energy is less than twice the Fermi level, the monolayer and multilayer graphene become completely transparent due to the Pauli barrier effect. These properties are mainly attributed to the density of graphene in electron transport of the van der Hove singular point and the electron transition between the bands, which are suitable for use in many controllable photonic devices.

1.3.2 Photoluminescence

Graphene has special characteristics and potential applications in photoluminescence and electromagnetic transport, which have created a wide range of research interests [32], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46]. Graphene fragments and graphene quantum dots (GQD) show unique photoluminescence properties in the preparation process; the physical and chemical treatment of graphene to reduce the coherence between π electrons is another way to make graphene photoluminescent. After a slight oxidation plasma treatment, a single graphene sheet can have bright light [31], and at this time photoluminescence has a large uniform area (Figure 4A). Likewise, bulk graphene and dispersions also exhibit extensive photoluminescence [32], [48], [49]. Etching the top layer, and keeping the lower layer intact makes preparation of the mixture possible. Based on graphene materials’ photoluminescence being routinely prepared, the combination of this conductive layer and photoluminescence can be applied to sandwich-type light-emitting diodes (LEDs), with a wavelength range from infrared to blue spectral [31], [49].

Figure 4:

(A) Photoluminescence and elastic scattering image [31]. (B) Comparison of optical absorption of graphene quantum dots with NH₂-modified GQD [47].

A graphene quantum dot that can emit blue light by hydrothermal method has been prepared [40]. Li et al. [41] prepared green-emitting quantum dots by the electrochemical method, which can be made as the electron acceptor material for photovoltaic devices. Sun et al. used photoluminescence of oxidized graphene to image living cells in the near-infrared region [33]; Wang et al. studied the photoluminescence mechanism of GQD [50], which revealed the excited state transfer between electrons and holes and the influence of edge effect light luminescence (Figure 4B), made in visible photoluminescence devices within the scope of the preparation of a theoretical basis.

As the size of the quantum dots decreases, graphene exhibits quantum confinement effects and unique edge effects [36], [47], [50], [51]. Some teams interpret the photoluminescence of graphene oxide as radiative recombination of e-h pairs of localization states of the sp² cluster [32]. The energy gap between the π* and π states is defined by the size or the conjugate length of the sp² cluster [52], [53], which is more likely to be explained by the graphene edge effect and the defects associated with oxidation. Some researchers have explained that the photoluminescence of GQD is derived from radiative recombination of electrons in the free zigzag sites of the graphene edges [35], rather than the general transitions between π* and π. No matter which luminescence mechanism is used, the photoluminescence spectra of GQD have the same characteristics; the wavelength of the emission spectrum changes with the excitation wavelength.

Therefore, the fluorescent organic has a significant impact on the development of cheap optoelectronic devices [54]. On the other hand, their toxicity and potential environmental hazards are also limited to in vivo applications and other applications [55], [56].

1.3.3 Saturated absorption

The excitation of the ultra-fast optical pulse causes a nonequilibrium carrier cluster between the conduction and the valence band (Figure 5) [28]. In the time-resolved experiments [57], two relaxation time scales can be observed: one is the fast relaxation corresponding to the mutual collision and phonon scattering between the electrons between the conduction and the valence band, which is about 100 fs, and the other is the slower relaxation corresponding to the electron band relaxation and thermophonon cooling [58], [59].

Figure 5:

(A) Schematic diagram of optical excitation in graphene. (B) Low band relaxation [28].

On the picosecond scale, the linear dispersion relation of Dirac indicates that there is always resonance between electrons and holes for any excitation. The electrons and holes are solved quantitatively using the kinetic equation and the distribution function of holes [fe(p⇀) and fh(p⇀)], where p⇀ is the momentum of the Dirac point [28]. When the pulse time is greater than the relaxation time, the electrons reach a quiescent state during the pulse time. At the effective temperature, the electron and the hole reach the thermal equilibrium by collision. On the other hand, the number of electrons and holes determines their density of states and the total energy density [28]. Due to Paul blocking, by the factor 1+Δα1/α2=[1−fe(p⇀)][1−fh(p⇀)], the decrease of photon absorption for each layer given by laser energy transfer from electrons to phonons.

Similarly, because of the Dirac points’ linear dispersion relation, the collision of the paired carriers does not cause relaxation between the conduction and the valence band, which protects the total number of electrons and holes, respectively [28], [60]. When the energy of electrons and holes approaches the Dirac point, the phonon emission causes interband relaxation, and the spontaneous emission of hot electrons and hole groups also occurs [29], [30].

Theoretically, for a certain amount of material, a decoupling single layer graphene (SLG) is able to provide higher saturable absorption than others, because of the two-order dispersion of the graphene sheet and the collision of the sub carriers, the relaxation of the band gap is produced [28].

1.3.4 Graphene plasmon

Surface plasmon polaritons (SPPs) are the collective oscillations of free electrons induced by external electromagnetic fields (such as photons or electrons), and have surface electromagnetic field propagation properties: the electric field intensity has the maximum value at the interface between the metal and the medium. When the vertical distance from the metal surface increases, the field strength exponentially decreases [61]. Recently graphene has been confirmed to be a plasma-derived waveguide material in the infrared frequency range [62], also known as terahertz metamaterials. Graphene has similar properties to metals in conductivity and can transport SPPs. As it has only a thin layer of atomic level, there is little need to construct surface properties, so it is called graphene plasmon (plasmons polartions, PPs).

The graphene plasmonic properties originate from the linear optical properties of the doped graphene materials. The response of graphene to high-frequency electromagnetic waves mainly relates to its concentration of carrier and carrier mobility. The graphene carrier concentration can generally be modulated by classical doping or chemical doping, and the carrier mobility relates to the quality of graphene and dielectric environment. As graphene is a single atomic layer material, the electronic behavior is more susceptible to the surrounding environment than the bulk material, which has a larger specific surface area (Figure 6) [63], [64], [65], [66], [67].

Figure 6:

(A) Schematic diagram of graphene plasma excitation. (B) Schematic diagram and test results of plasma resonance test of graphene. (C, D) Test patterns of different widths and layers [63].

The chemical potential μ and the Fermi level E_F determine the performance of graphene. By adjusting chemical doping or photoelectric reflection, the transition from electrolyte to metal can be achieved. The chemical potential μ between the conduction and the valence band gap and the in-band conductivity and graphene are relevant to the frequency ω of the incident light. When the chemical potential is μ<ℏω/2, the band gap transition plays a dominant role and controls the dynamic high-frequency conductivity; when the chemical potential is μ>ℏω/2, the transition in the band plays a dominant role and controls the megahertz range of conductivity. Under these conditions, the momentum of PPs is enhanced, so PPs can propagate in graphene. The competitive relationship between the transition of light in the band and the transition of the band gap makes it possible to couple the tunable optical response and the polarization selective coupling.

1.4.1 Flexible electronic devices

Electronic products widely use conductive coatings, such as touch screens [68], electronic paper, organic photovoltaic cells [69] and organic LEDs [70], hence the need for low surface resistance and high transmittance of special applications.

Figure 7 in comparison to the transparent conductive layer based on graphene is different from other properties of photoelectric materials, and shows their photoelectric properties [71], [72].

Figure 7:

Graphene meets the needs of electronic and optical equipment, single-layer transmittance of up to 97.7%, but in the past people thought that performance of indium tin oxide (ITO) would be better [73]. However, considering the annual increase in the quality of graphene, the price of ITO will increase, and the cost of the deposition method preparation of ITO will increase as well. So, graphene will certainly get a larger market share. The excellent flexibility and corrosion resistance of graphene are the most important properties of flexible electronic materials equipment, but ITO cannot achieve in this respect for it’s lack of flexibility and corrosion resistance.

The electrical properties required for different applications of the different motors are not the same (such as surface resistance). Due to the different production methods, there will be a variety of conductive coatings. Therefore, the electrodes of the contact screen (products requiring the chemical vapor deposition (CVD) method) have a relatively high surface conductance based on 90% light transmittance [71]. Graphene electrodes used in contact panels have the advantage of graphene having greater stability. In addition, graphene has a 10-fold higher fracture strain than ITO, which means that graphene can be used in collapsible and bendable equipment.

Flexible electronic paper is very attractive. Its bending radius is in the range of 5–10 mm. This requirement is very easy to achieve for graphene. And graphene can absorb visible light, which for the color of electronic paper is very important. However, the graphene electrode contact resistance and metal loop are still a big problem.

Despite the relatively high resistance of graphene films, the advantages of the flexibility and high mechanical strength of graphene ensure that the graphene equipment can have more flexible applications.

1.4.2 Photodetector

One of the most widely researched optoelectronic devices, graphene photodetectors, are and can be used in the wide-band spectral region between ultraviolet and infrared. The ultra-wide working bandwidth is an advantage of the graphene photodetector, allowing them to be used in high-speed data communications. The high carrier activity of graphene provides a carrier for rapid extraction of the picture, as it allows for higher bandwidth operation. At the reported velocity of saturated carriers, the bandwidth of the graphene photodetector due to time constraints is expected to reach 1.5 THz [88], [89]. In fact, the graphene photodetector has the maximum bandwidth of 640 GHz due to the delay of the capacitor rather than the delay of the transfer time [88].

At present, the graphene photodetector uses a local potential change near the surface of the metal-graphene to extract the photodetector carrier [90], [91]. The optical response rate can reach 40 GHz; the detector-operating rate can reach 10 GHz. However, the maximum response rate is relatively low due to the limitation of the smaller effective detection area and the thin graphene to the absorption rate.

There are many solutions for increasing the graphene photodetector sensitivity, for example, using nanostructured plasma to enhance the local optical electric field or to increase the photo-graphene interaction length by combining with a waveguide [92], [93].

Without the band gap of graphene, low response and low optical gain and limits the applications of graphene based photoelectric detector, even the superfast and broadband optical response of graphene. Figure 8 shows a photodetector based on graphene-Bi₂Te₃ heteostructure which overcomes these difficulties [94]. Owing to the topological insulators family with a small hexagonal symmetry structure similar to that of graphene with zero gap materials, the photocurrent device can be effectively enhanced and fails to detect a decline in the spectral width. The result shows that the graphene Bi₂Te₃ photodetector has a higher optical response and a higher sensitivity than the pure monolayer graphene apparatus, and the detection wavelength range of the device is extended to the near infrared region (980 nm), and the communication band (1550 nm).

Figure 8:

Photodetector based on graphene-Bi₂Te₃ heterostructure [94].

1.4.3 Optical modulator

An excellent optical modulator is able to achieve the performance through stripping the resulting graphene, which absorbs a small amount of incident light from a broad spectrum of light and is able to respond quickly. In order to achieve these properties, in a single graphene layer [95], the inter-band transition photoelectrons are modulated by a driving voltage across a wide band to obtain an optical modulator with a bandwidth in the near-infrared region of more than 1 GHz [96]. The use of mutually constrained double graphene can reduce the delay in the RC delay by providing some structural change, providing an area that can reach hundreds of gigabits. Theoretically, operating light modulators with bandwidth in excess of 50 GHz cannot be achieved [93]. Graphene is a latent material for megahertz wireless communications because the loss of light in graphene is much less than in precious metals.

Liu et al. assembled and fabricated an electroabsorption modulator based on graphene waveguide integration for the first time [96], in which actively tuning the Fermi level of a monolayer graphene sheet achieves modulation (Figure 9). The gigahertz graphene modulator demonstrates a strong electroabsorption modulation of 0.1 dB μm⁻¹ and operates over a broad range of wavelength, with a range of 1.35–1.6 μm.

Figure 9:

Schematic diagram of graphene optical modulator [96].

1.4.4 Mode-locked laser/THz generator

Ultrafast passive mode-locked lasers have been used in spectroscopy, micro-materials processing [97], biomedical [98] and safety applications. They are often used as a saturable absorber, by selecting high-intensity light transmission to cause the modulation of light intensity. Compared to the semiconductor saturable absorber [99], graphene monolayer absorbance is very high, which in the low light intensity of the wide band area can be saturated [28], [100]. Controllable modulation depth, high thermal conductivity, wide frequency tuning, relaxation time and high damage threshold of the ultrafast carrier are all the advantages of the graphene saturated absorber [26], [101], [102].

Scientists have devoted much of their research to fiber and solid-state lasers [103]. However, people can also use graphene-saturated absorbers in semiconductor laser technology. Wavelength multiplexing eliminates the need for optical interconnections with a series of lasers of different wavelengths. Using a single laser in different longitudinal modes can generate different wavelengths such as mode-locked lasers [104]. Active mode-locked silicon hybrid lasers have been studied to meet the needs of laser technology [105], but graphene-saturated absorbers can provide simple passive mode-locked semiconductor lasers for operation and processing.

Wan et al. [106] prepared high-quality graphene using the CVD method on copper. It was the first time the use of high-quality single-graphene as a saturable absorber (SA) was used to prepare the stability of the diode-pumped passive mode-locked Tm: YAP laser (Figure 10). The absorption pump power is 7.67 W, and the output power of the laser can reach 256 mW at 1988.5 nm, which is the center wavelength. When the pulse frequency is 62.38 MHz, the pulse energy can reach 4.1 nJ [106]. Studies show that using graphene to create low-cost and ultrafast laser is promising.

Figure 10:

(A) Assembly diagram of mode-locked laser based on graphene. (B) Test pattern of graphene mode-locked laser [106].

2.2.1 Infrared characteristics of h-BN

As is known, Fourier transform infrared (FTIR) spectroscopy, which has been used to examine newly prepared products, is the effective method to characteristically determine the structure of new experimental products, so some researchers used FTIR to research the structural nature of h-BN [126], [127], [128], [129].

Some researchers obtained data on BN vibrations in h-BN, where h-BN is in the state of a single crystal. The peak value is 1365 cm⁻¹ [127]. The phonon modes of turbostatic h-BN are 792 and 1384 cm⁻¹. From some of the reports, we know that the phonon modes have shifted to 800 and 1372 cm⁻¹ when h-BN is in a multilayer state. On the other hand, the phonon modes of h-BN tubes shift to 811 and 1377 cm⁻¹, when h-BN is in the polycrystalline state [128]. From Figure 13, we can see that there are two peaks at 1374 and 818 cm⁻¹ from the FTIR spectrum, which is due to the strong vibrations of BN in the h-BN nanosheets. With the aid of FTIR, the structural properties and phase composition of experimental h-BN products were revealed. We also know that the peak is at 1374 cm⁻¹, which belongs to the transverse optical modes of the sp² bonded h-BN when B-N is in the plane. When nitrogen-boron-nitrogen is out of plane, a peak appears at 818 cm⁻¹, which could belong to the bending vibration of B-N-B [126], [129]. However, there are no absorption peaks belonging to the raw materials or any impurities that have emerged in the visible region.

Figure 13:

Infrared spectra of h-BN [126].

2.2.2 UV characteristics of h-BN

With deeper research, the theoretical calculation and experimental verification have shown the ultraviolet absorption spectrum of the 2D h-BN materials [110], [111], [130], [131].

By the simulation of the idea on 2D h-BN nanomaterials, Wang et al. calculated the single layer h-BN’s absorption spectrum by using DFT and DT-DFT methods, and the spectrum of the 2D single layer h-BN materials revealed the optical absorption property of h-BN [110]. From Figure 14A, we can see three strong absorption peaks in the spectrum. On the other hand, there is a shoulder around 210 nm, which in the deep UV region. The three absorption peaks are around at 197, 203, and 208 nm, respectively. By using the charge different density method, the absorption peaks are due to strong charge transfer and the vibration of the BN.

Figure 14:

UV-vis optical absorption spectra of 2D h-BN nanometer material. (A) Simulation model [110]. (B) Experimental h-BN nanosheets at 293K [111].

The theoretical calculation is consistent with the experimental verification. Jin et al. got the deep UV absorption spectrum of the h-BN experiment materials. There is a strong absorption peak at 234.8 nm, and the optical energy band gap corresponds to 5.28 eV (Figure 14B) [111]. Their data are basically the same as the previous researchers, just as 5.2 and 5.14 eV, which were made by Hoffman et al. and Stenzel et al. [130], [131], respectively. On the other hand, an absorption peak around 263.1 nm appeared in the optical absorption spectrum, which corresponds to the optical energy band gap with the value of 4.71 eV. They made sure that the defect level of the boron nitride nanosheets resulted in this absorption peak, and the absorption peak energy corresponds to the difference between 5.28 (corresponding to the optical energy band gap) and 0.63 eV (corresponding to the activation energy of the donor level). By using the TSC method, the observed result can prove the influence of the donor level, which leads to the emergence of the absorption peak at 263 nm.

Both the results of the experiment and simulation have proved once again that 2D h-BN has a strong optical response in the deep UV area.

2.2.3 Luminescence property of h-BN

As one of the semiconductor materials, the bandgap energy properties and the band structure of h-BN materials have been studied for some years. With the accelerated development of research, both direct bandgap and indirect bandgap have been revealed slowly. Luminescence is one unique optical feature of h-BN [111], [116], [132], [133], [134], [135].

2.2.3.1 Cathodoluminescence property

Watanabe et al. studied the optical properties and electrical properties of the h-BN crystal [116]. There is a cathodoluminescence (CL) spectrum of experimental h-BN materials at room temperature shown in Figure 15A. From Figure 15A, there is a strong peak in the spectrum of the h-BN single layer crystal, which had been successfully made by them. The peak appeared at 215 nm with the bandgap of 5.76 eV at the room temperature. On the other hand, there is a shoulder structure appearing at 300 nm, which corresponds to the bandgap of 4 eV. By researching this shoulder, either the structure of h-BN defects or the impurities in h-BN might cause this result, just as in oxygen and carbon [132]. The ratio of two peaks of about 100 times are all in the deep UV region.

Figure 15:

(A) Cathodoluminescence spectrum of h-BN [116]. (B) PSL excitation spectrum and PLE spectrum of polycrystalline h-BN [133]. (C) Phosphorescence decay curves of different photon excitation energies [133]. (D) Laser emission spectrum of h-BN [116].

2.2.4 Phonon polaritons on h-BN

Optical phonons coupling of the photons in polar crystals, in the collective modes are called phonon polaritons. These materials show optical phonons, which are similar to some polar Van der Waals solids, and the property might be usually called polaritonic effects. Because of the strong phonon resonances, h-BN becomes a good polaritonic effect material.

Dai et al. studied this character of h-BN by using s-SNOM [137]. They adjusted the intensity of incident light and changed the number of layers to achieve the realization of the phonon polariton. Figure 16A is the dispersion relation of phonon polaritons of h-BN. From the image we can know that AFM tip alters the range of momenta, which supports phonon palaritons of h-BN. The polariton propagates rapidly on the surface of boron nitride. With the change of the incident wavelength, the dispersion relation of the phonon polariton is expressed in Figure 16A. Figure 16B shows the image of different layers of h-BN and different lengths of the edge whereas Figure 16C and D shows experimental measurement of polariton wavelength and the data are predicted according to the principle, which are in good agreement.

Figure 16:

(A) Sketch map of phonon polaritons of h-BN [137]. The test spectrum line of h-BN with different layers, width in (B)–(D).

From their study, we can know that the phonon polaritons can confine and tune electromagnetic energy on the nanometer scale. By altering the layers number, the wavelength and the confinement of phonon polaritons can be designed.

2.3.1 MSM photodetector

The h-BN epilayers have some features, such as optical absorption and dielectric strength, and as a result, they have potential for application as deep UV detector materials.

Li et al. synthesized h-BN epilayers by using the CVD method [138]. Analogous to the optical absorption property of graphene, they tested and found the absorption coefficient as about 7×10⁵/cm of band edge in h-BN. Both the dielectric strength and the absorption coefficient are better than those of a wurtzite AlN. Figure 17 shows the typical I–V performance of MSM detectors, which is based on the h-BN epilayer. When the bias voltage is 100 V, the low dark current is 200 pA and the current density is 10⁻¹⁰ A/cm², respectively. There are other data of spectral responses corresponding to the different bias voltages in the figure. From Figure 17B, there is a sharp peak of optical responsivity at 220 nm and the cut-off wavelength at 230 nm, respectively. It is in good agreement with the fluorescence emission peak of the band-edge of h-BN, which is at 227 nm (5.48 eV). On the other hand, there are no detectable responses in low frequency region up to 800 nm. However, the visible rejection ratio of h-BN MSM detectors in the UV region is two to three orders of magnitude shorter than that of the AlN detectors [139], [140].

Figure 17:

(A) I–V characteristics of photodetector. (B) Relative spectral response of detector [138].

Based on the above study, they summed up several advantages such as high deep ultraviolet to visible rejection ratio, due to the high absorption of the thin active layers; it is suitable for oxidation resistance and chemical inertness and high dielectric strength.

2.3.2 Deep UV emitters

In the P-type conductivity demonstration progress of the epitaxial growth the h-BN method of the epitaxial layer, which represents the p-layer revolutionary and overcomes the Al-rich AlGaN deep purple problems inherent in low p-type conductivity using a special deep UV device application. However, the epitaxial growth of h-BN on AlGaN is a prerequisite for the formation of a P-type AlGaN deep UB device structure [141], [142], [143], [144].

Majety et al. used the method mentioned earlier to assemble the heterostructure, which was prepared for the p-type h-BN and AlGaN deep UV device structure [144]. They used the CVD method to prepare h-BN and the AlGaN/AlN/Al₂O₃ templates. The epitaxial layers of h-BN were also highly insulated in the AlN and N-type AlGaN templates, trying to prove that h-BN/AlGaN p-n junction is grown by Mg doping. By testing the diode performance of the heterostructure, they demonstrated the feasibility of the p-type h-BN as electron blocking of its high conduction, which can be made as a p-contact layer of deep ultraviolet emitters (Figure 18A,B). The characteristics of I–V correspond to the p-n structure, which is doped Mg in a buffer layer annealed at a temperature of 1020°C. The test values under different temperatures are shown in Figure 18B. The test results showed that these p-n heterostructures have wide application prospects in the application of high efficiency deep UV photovoltaic devices.

Figure 18:

(A) I–V characteristics of a p-BN: Mg/n-Al_0.62Ga_0.38 N/AlN structure. (B) I–V characteristics of emitters at different times and temperatures [144].

2.3.3 Far UV plane-emission device (LED)

Some researchers have been hoping to replace the traditional far UV lamps with the high efficiency of a solid-state device with a long service life. But, due to the low efficiency of UV lamps, research still continues for developing far UV devices with high performance [145], [146], [147].

Watanabe et al. studied the potential application of h-BN as far UV fluorescent materials [147]. The height of the luminescence of h-BN far UV emission planar compact device is equipped with a field emission array as the excitation source, which proves that it is a stable operation and has an output power of 0.2 MW at 225 nm. Due to its low current consumption during operation, the device can be driven by a dry battery. This convenient far UV device may prove useful in chemical and biotechnological applications such as chemical modification, photo-catalysis and sterilization. The operating state of a FUV plane-emitting device consisting of a field emitter as a cathode, the power of h-BN on the screen, a vacuum chamber, and a wire electrode around the screen is shown in Figure 19A, which is driven by a dry battery. Figure 19B shows a FUV plane-emission device. The output power of the device is shown in Figure 19C, from it, we know that, with the increase of the acceleration voltage from 3 to 8 kV, the output power increases nonlinearly from 0.02 to 0.25 mV. The output spectra of this device are shown in Figure 19D. From Figure 19D, there are three spectral lines in the figure corresponding to different anode voltage, which the emission intensity increases with the increase of anode voltage. The emission wavelength ranges from 215 to 400 nm and there is a strong peak around 227 nm.

Figure 19:

(A) Photograph of different far-ultraviolet (FUV) plane-emission device in operation. (B) Photograph of a battery-driven FUV plane-emission device in operation. (C, D) Test spectrum line of FUV [147].

3.3.1 Optical advantages of graphene/h-BN in visible and SW-NIR regions

Wang et al. calculated the optical properties of this structure in the entire optical region and the electron transfer in the excited state by DFT and TD-DFT [133]. The results show that the graphene has a strong optical absorption peak in the visible and near UV regions, whereas the single layer hexagonal boron nitride has optical absorption in the deep UV region. When they compose heterostructures, boron nitride has little absorption effect on graphene; from the situation of SLG and heterostructures in the excited state electron-hole transfer, it can also be seen in the visible region, between graphene and boron nitride there is no charge transfer, but inside graphene itself there is charge and hole transfer (Figure 24A). The graph shown in Figure 24B also confirms that the influence of boron nitride on the graphene as a graphene substrate is negligible, which further theoretically explains that for the production of graphene-based optoelectronic devices, boron nitride is a very good substrate.

Figure 24:

(A) Charge-transfer densities for the strong electronic transitions in single-layer on single-layer h-BN substrate with top and side views, where the green and the red stand for the hole and electrons, respectively. (B) The calculated absorption spectra of SLG and graphene/h-BN [133].

3.3.2 UV features of graphene/h-BN

Recently, Liu Zhongfan Research Group of Peking University successfully completed the preparation of patterned graphene/h-BN single crystal heterojunctions [169]. Prior to CVD growth, PMMA was coated on the copper foil surface by electron beam etching. These PMMA particles, like “seed”, act as centers of nucleation, allowing patterned growth of graphene during preparation of CVD. Meanwhile, PMMA particles also controlled the nucleation density and grain size, and provided a possible direction for the high quality graphene/h-BN heterojunction growth. In the experiment, they also concluded that the optical properties of the heterostructure, UV-visible absorption spectra showed that at 270 and 200 nm there are graphene and h-BN absorption peaks; at 550 nm graphene/h-BN van der Waals heterojunction has high visible light transparency, which provides a possibility for heterojunction to be used in optoelectronic devices (Figure 25).

Figure 25:

Electrical and optical properties of patterned G/h-BN. (A) Schematic depiction of FET devices based on G/h-BN stacks. (B) Source-drain current (Ids) and channel resistance (R) as a function of back gate voltage (V_gate) measured at room temperature. (C) Optical image of patterned FET devices, fabricated using transferred G/h-BN patterns on 300 nm SiO₂/Si. (D) Close-view of (C) showing 3×2 devices. (E) UV-vis transmittance of G/h-BN (blue) and h-BN (red). (F) Photograph of the transparent graphene FET arrays fabricated using G/h-BN stacks on the quartz glass [169].

3.3.3 Negative refraction effect

Recently, the theoretical and experimental research on the hybridization between the graphene plasmon polariton and hexagonal boron nitride phonon polaritons have been extensively studied. In this regard, the study of the stacked structure of graphene and h-BN as the multilayer structure of the crystal is less [175], [176], [177], [178].

As tunable materials, graphene and h-BN stack of hyperbolic crystal can be adjusted by adjusting the chemical potential of graphene to give more freedom. Sayem et al. studied on the properties of the crystal structure stacked by graphene and h-BN [178]. There are two schematic diagrams of negative refraction in Figure 26B, h-BN and graphene/h-BN hyper-crystals, respectively. Which the hyper-crystals stacked by SLG and h-BN. By matching the wave-vector (θ_rk,hc) and pointing vector (θ_rs,hc), a function corresponding to the incidence angle at wavenumber 810 cm⁻¹, which corresponds to the difference from the graphene chemical potential value. The 2D map of refraction angles is shown in Figure 26C, for μ_c=0.25 eV. Figure 26D shows us the reflection for different μ_c. From 725 to 850 nm, the reflection increases slowly with the increase in wavenumber. When the wavenumber increases to 800 cm⁻¹ this reflex is stopped. In the range of wavenumbers from 800 to 820 cm⁻¹, the reflection decreases sharply. After 820 cm⁻¹, the reflection increases sharply, reaching a maximum value near 830 cm⁻¹, and then the reflection decreases slowly, but the small μ_c refraction but high.

Figure 26:

(A) Schematic representation of negative refraction in bare h-BN and GhHC. (B) Refraction angles for the pointing vector (symbols “ο”) and wave-vector (solid lines) for GhHC as a function of incidence angle at wavenumber 800 cm⁻¹ for different values of the chemical potential of graphene. (C) 2D map of the refraction angles for the pointing vector as a function of wavenumber and incidence angle for μ_c=0.25 eV. (D) Reflection of h-BN and GhHC as a function of wavenumber with an incidence angle 30° [178].

The all-angle negative refraction of the hyper crystal is much higher than that of bare h-BN. On the other hand, graphene can completely control the negative refraction of the optical properties of boron nitride and the transmission characteristics of the boron nitride structure without hindering the negative refraction.

3.3.4 Raman characteristics of graphene/h-BN superlattices

As a fast and nondestructive detection technique, Raman spectroscopy is widely used in the qualitative study of some materials. In particular, it is much sensitive to the alignment between graphene and h-BN [179], [180], [181].

Eckmann et al. studied the Raman spectra of the heterosturcture stacked by graphene on h-BN, where the lattice of two kinds of materials is perfectly aligned [182]. Figure 27A is the schematic diagram for graphene/h-BN. Due to the difference of the length of the B-N bond and the length of the C-C bond, graphene and h-BN, which is the same as the honeycomb lattice structure of the six party, have been shown to form a superlattice after being stacked into a heterostructure. The wavelength λ_M and the low superlattice Dirac point (SPD) energy E_SDP corresponds to the stacking angle θ between graphene and h-BN. The linear dependence of FWHM (2D peak) and moiré wavelength for twist angles is below 2°. They make the function FWHM (2D) Δ5+2.6λ_M. The comparing spectral lines between graphene/h-BN and tBLG (twisted bilayer graphene) of peck R′ as a function of the mismatch angle are shown in Figure 27C. From it, we can know that, the position of peck R′ increases with the increase of the lattice alignment angle of graphene and h-BN. The theoretical calculation of the two structures is basically the same. Figure 27D is the Raman spectrum, which has been calculated corresponding to Lorentz, Gaussian and Voigt, respectively.

Figure 27:

(A) Superlattice potential of graphene/h-BN. (B) The test spectrum of FWHM and λ_M as the function. (C) The test spectral line of different heterosturcture by θ and R′ as the function. (D) The Raman shift of h-BN/graphene superlattice with different calculated method [182].

The Raman spectra of these superlattices reveal the relationship between the energy and the angle of the superlattice, which provides a reliable theoretical basis for the structure of the optoelectronic devices.

3.3.5 Plasmon-phonon coupling and plasmon delocalization, high pressure and low loss of plasmons in graphene/h-BN heterostructure

High quality graphene/h-BN heterostructures including other two-dimensional materials are considered to have broad application prospects [183], [184], [185], [186].

Figure 28A and B shows the SINS (synchrotron infrared nanospectroscopy) spectra of the graphene/SiO₂ and graphene/h-BN. From Figure 28A, the wavenumber at 765 cm⁻¹ belongs to a small additional band of G//SiO₂, which is the enhancement corresponding to the band at 1120 cm⁻¹ (red point line). In the Figure 28B, the wavenumber at 817 cm⁻¹ belongs to the longitudinal optical (LO) phonon of h-BN (green point line); the wavenumber at 1365 cm⁻¹ belongs to the transverse optical (TO) phonon of h-BN (green point line). On the other hand, there is a strength enhancement at 817 cm⁻¹, the same strength at 1365 cm⁻¹, respectively. The enhancement belongs to the transverse surface plasmons of graphene with the LO phonon band of h-BN. But, the coupling of plasmons-phonon polaritons has no effect on TO modes [186].

Figure 28:

SINS spectra of graphene/SiO₂ in (A) and graphene/h-BN in (B) [186], (C) and (D) are the spectral lines of plasmons damping in graphene/h-BN [187].

Compared with other metal materials, graphene has the unique property of strong field confinement and relatively low transport decay in the plasmons transport capacity [188], [189], [190]. Woessner et al. studied the propagating plasmons of high quality in graphene with h-BN by near-field microscopy and also do a quantitative study on the plasmon damping, which corresponds to the spectrum in Figure 28C and D [187]. From the figure, with the increase of the distance from the boundary of the heterostructure, the plasma damping decreases nonlinearly. The complex parameter ξ_opt, is made ordinate as shown in Figure 28C and D. The oscillating signal is fitted with equation 1:

The attenuation of fringes away from the edge is caused by the combination of the damping and the geometric diffusion of the circular wave. The results show that the main damping is due to the intrinsic thermal phonons scattering in graphene and the dielectric loss in h-BN.

3.4.1 Light-harvesting devices (the potential applications in solar cell and photodetector)

Under the limitation of the current conversion efficiency standards, such as the Shockley-Queisser limit, the absorption of a single photon can only excite an electron in a conventional light harvesting device [172].

A device, using h-BN/graphene/h-BN heterostructures was prepared by Wu et al. [172], the schematics are shown in Figure 29A. Through experiments, the collection of multiple hot carriers in the six molar boron nitride semiconductor superlattice was investigated.

Figure 29:

Anomalous photo-Nernst effect in graphene/BN superlattices [172].

(A) Schematics of device and photocurrent measurement. (B) The superlattices is made of graphene and h-BN. (C) Optical image of the device (A). (D) and (E) are the test results of device (A). (F) Typical spatially resolved scanning photocurrent map of the device (A).

Figure 29A and C shows the results and optical images of the photonic collection device. Figure 29B shows a moiré pattern formed by graphene stack on the h-BN. Figure 29D shows the change of the resistance of the Dirac point and secondary Dirac point in the excitation field with the base graphite as the back gate and the longitudinal resistance as a function of 50 mT. When the back gate voltage is −4.7 and 4.4 volts, they correspond to two peaks, corresponding to the sPDs. A strong peak at the back gate voltage of −0.1 volts is the main Dirac point. Figure 29E is an image of the photocurrent variation corresponding to Figure 29D. Near the back gate voltage −4 and 4 volts, there is a phenomenon of photocurrent enhancement, corresponding to the electron and hole of secondary Dirac point position. The excitation laser is 660 nm. Figure 29F shows the device’s scanning photocurrent image.

The use of graphene has been demonstrated by the photo-Nernst effect (which has been shown to absorb at least one of the five carriers of a photon), so that the zero-bias optical response is recorded at a 0.3A/W. This phenomenon is due to the enhancement of the energy coefficient of the Lifshtiz transition at the low energy Van Hov singularities. This electronic collection device provides an effective and flexible means for the fabrication of optoelectronic resonators and LED, based on the flexible optoelectronic devices of Van Der Waals heterostructure.

3.4.2 Light-emitting diodes

Withers et al., contrary to the conventional epitaxial heterostructures, stacked in the vertical direction by a mechanical transfer method [191], [192], [193], which interacts weakly with each other by the van der Waals force. Figure 30A is a simulation diagram of LED heterostructure; the heterogeneous structure is different from the traditional one as it contains only the semiconductor; they used as three different 2D materials: h-BN insulator, graphene metal, and transition metal sulfides with direct band gap (MoS₂). Figure 30B is a schematic diagram of the LED, which corresponds to Figure 30A TMDC, in which the electrons and holes pass through from the graphene layer tunnel through the h-BN dielectric layer to TMDC at the external bias. The inset is a physical image of LED corresponding to Figure 30A. Figure 30C is the comparison of PL MoS₂ monolayer, and EL spectra, it can be seen from the figure, the PL frequency is close to the frequency of EL around V_b=2.4 V. After analysis, this phenomenon is the radiative recombination of X⁻ corresponding to the EL spectral line overlap.

Figure 30:

(A) The schematic of LED [192]. (B) The working principle diagram of LED [192] (the inset is the physical optical image). (C) The comparison diagram of PL and EL in single layer MoS₂ [191].

In addition, the authors also show that the external emission efficiency of multilayer TMDCs LED is close to 10%, which is comparable to that of modern LED based on organic materials. Although the wavelength of the LED heterostructure in the work of the single semiconductor from TMDCs range from 600 to 700 nm, it has the potential to extend the communication range of emission wavelength, and could reach the midinfrared band.

In fact, the authors attempted to make thin, transparent, and flexible polyester film LED. The main factors restricting the development of this technology are the lack of a scalable synthesis strategy for heterostructures. However, it is encouraging because simple Van Der Waals heterostructures have been proven to grow and the mechanics of two-dimensional layers of mechanical transfer have been made, suggesting that this barrier may eventually be overcome.

3.4.3 Solar cell

As the application of a low-cost photovoltaic, the structure of heterostructure based on graphene has become an excellent candidates for it [194], [195], [196].

Meng et al. [196] prepared the heterostructure using the CVD method. Figure 31A is the graphic illustration of graphene/h-BN/Si. Figure 31B shows that h-BN is employed in an effective electron blocking layer, which is inset in the graphene and n-Si. Owning to the wide band gap, h-BN can reduce recombination of the unfavorable carriers. With a negative electron affinity, ΔE_C of h-BN/Si is determined to be larger than that of Si (4.05 eV). The valence band offset ΔEV is estimated to be less than 0.63 eV from the formula ΔΔEV=EghBN−EgSi−ΔΔEC by taking the room temperature band gaps of h-BN (5.80 eV) and Si (1.12 eV). The existence of small ΔΔE_v is also an important problem in avoiding the adverse effects of the increase of the series resistance of the cell in the effective transmission of the hole from the silicon layer through the h-BN layer to graphene. On the other hand, they used two different methods for preparing the heterostructures, which are shown in Figure 31C and D. Figure 31E shows the J-V characteristics of different solar cells, which is a comparison of the spectral lines of the different structures prepared by the different methods corresponding to Figure 31C and D. From the comparison, we know that the heterostructure of the one-step method is better than that of the two-step method. The power conversion efficiency (PCE) of solar cell reaches 10.93%.

Figure 31:

(A) Schematic diagram of solar cells. (B) Schematic diagram of working principle of solar cells. Schematic diagram of solar cells prepared by one step (C) and two step method (D). (E) Comparison of two methods for solar cell performance test J–V characteristics [196].

The layered h-BN is also used to improve the performance of the device. Because of its unique physical and chemical properties and the appropriate band structure, h-BN acts as an effective electronic barrier, hole transport, thus inhibiting the interface complex, so that the open circuit voltage significantly improved. On the other hand, based on h-BN, the carrier conductivity increases significantly and the resistance of the solar cell is reduced, which leads to the increase in the short-circuit current density. By improving the preparation method of the heterogeneous, the graphene/h-BN heterostructure can be directly grown, and the interface defect and contamination can be avoided, and the performance of the device can be greatly improved.

3.4.4 Nanoresonators (based on graphene/h-BN/SiO₂ heterostructures)

The optical properties of graphene and h-BN have been changed due to graphene/h-BN heterostructures. In particular, the coupling between the surface plasmons of the graphene and the phonons of h-BN leads to a new frequency-wavevector dispersion relation [95], [183], [184], [185], [186].

Brar et al. [197] assembled graphene nanoresonators on a single layer h-BN nanosheet, which was used to measure the coupling between the plasmons and the excitons near graphene. They found that small mode volume and high modulus graphene plasma oscillator strengths of h-BN and the formation of two phonon coupling allows for strong coupling, thus forming the two clearly separated hybridization pattern show anticrossing behavior.

Figure 32A is the schematic of the nanoresonator based on graphene/h-BN heterostructure. The figure shows the schematic diagram of the motion of the graphene plasmons and the h-BN phonon under the laser excitation. The resonator is etched into an electron rod array pattern with a width ranging from 30 to 300 nm and a spacing ratio of 1:2. FTIR spectroscopy was used to measure the transmission mode with the vertical irradiation of nanoresonators, the test result is shown in Figure 32B, which is 1.0×10¹³ cm⁻² carrier density. The wavenumber at 1370 cm⁻¹ belongs to h-BN optical phonon, which the spectra lines below the Figure 32B. With the increase in the width, the frequency has a significant red shift and the strength of the transmission modulation (−ΔT/T_CNP) becomes higher. Figure 32C is the carrier-induced change in the transmission under different wavenumbers and energy of the same carrier density, which corresponds to Figure 32B. These features indicate that the h-BN optical phonon energy, which appears above and below the experimentally measured characteristics, and shows a strong inverse cross behavior pattern with a strong correspondence.

Figure 32:

(A) Schematic of graphene nanoresonators. (B) Experimental device physical test line of nanoresonators. (C) Theoretical calculation spectrum corresponding to (B) [197].

This work is based on graphene/h-BN heterostructure nanoresonator to test the graphene plasmons and h-BN phonon coupling results. However, the fabrication of the nanoresonator shows that the heterostructure has a broad prospect in the application of optoelectronic devices.

3.4.5 Photodetector and autocorrelator

Recently, 2D materials have shown a number of new properties, such as optical emission, parametric nonlinearity, broadband ultrafast optoelectronic detection, saturable absorption, on-chip electro-optic modulation, etc. These unique properties make them an alternative to the optoelectronic platform for semiconductor devices [198], [199], [200], [201].

A heterostructure consisting of a single layer of graphene encapsulated in a six party boron nitride, has high efficiency of photoelectric detection and ultrafast metrology [200]. Figure 33A is the structure of the photodetector. The illustration of the red circle is the side simulation of the heterogeneous structure formed by the encapsulation of SLG with h-BN. Figure 33B is the result of an electrical spectrum analyzer obtained by a high speed RF probe, with the measured power Δf(V_DS=V_GS=0 V). The measurement results show that the cut-off frequency of 42 GHz is 3 dB, and the results match the maximum velocity of the graphene photodetector. At the same time, the optical response of the zero drain source bias is also observed, a typical distinction between graphene and semiconductor high speed photodiode. Figure 33C is the delay time and the intensity of the optical current as the coordinates of the different input power, the change of the photocurrent image. From the figure, at Δt=0, the spectrum of the relaxation time and the width corresponds obviously to the sink.

Figure 33:

(A) Schematic of the h-BN/SLG/h-BN photodetector on a buried silicon waveguide. The inset shows the electric field energy density of the fundamental TE waveguide mode, obtained by finite-element simulation. (B) High-speed response of the graphene photodetector. (C) Autocorrelation traces of the graphene-based autocorrelator with different input average power of the laser pulses [201].

The results of the measurement of photocurrent by using the gate voltage and drain source voltage as a function show that the thermal electrons mediate the photoelectric response. At the same time, the observed saturation photocurrent corresponds to the super collisional cooling mechanism of the electron phonon. This nonlinear optical response can be tuned to picosecond scale timing resolution on the optical on-chip autocorrelation measurements and extremely low peak power [201].

3.4.6 High-speed electro-optic modulator

Efficient, nanoscale, electro-optic modulator is an important part of optical interconnects. Due to the weak photoelectric effect of silicon, the silicon is a technical bottleneck in the modulator. As a substitute for silicon, graphene has special optical and electronic properties, which makes it an alternative to silicon-based materials in optoelectronic applications [202], [203], [204].

Based on the graphene/h-BN heterostructure and the silicon photonic crystal resonant cavity, Gao et al. integrated a high-speed graphene electro-optic modulator. There is a strong interaction between the light and the material in the submicron cavity, which makes the reflection of the resonator have high efficiency. After testing, the modulation depth is 3.2 dB, and the cut-off frequency is 1.2 GHz. Figure 34A is a schematic diagram of the structure of the electro-optic modulator, which the preparation based on the graphene/h-BN heterostructure. Double layer graphene capacitors on the quartz substrate coupled with a planar photonic crystal cavity. The gate voltage (V_G) is scanned slowly, and the optical response of the modulator is measured by observing the reflection spectrum of the cavity, as shown in Figure 34B. During the gate voltage increase in the range from 2.5 to 6.7 V, the carrier density of two graphene layers increased slowly. For the incident wavelength at 1551 and 1570 nm, there are two peaks corresponding to the carrier density, moreover, the intensity becomes suddenly stronger at V_G=±5 V. Figure 34D shows the high speed response characteristics of the low-pass filter with a 3 dB cut-off frequency of 1.2 GHz. The RC time constant of the bilayer graphene capacitors limits the cut-off frequency, which is consistent with the measurement of the impedance of the device.

Figure 34:

(A) Schematic of modulator based on graphene/h-BN. (B) Cavity reflection spectrum as a function of V_G and wavelength λ. (C, D) The test spectrum of modulator with different parameter [204].

This work shows that the electric adsorption of graphene can facilitate the powerful high performance electro-optic modulator in the photonic crystal cavity, with low energy consumption, told the wavelength scale mark, etc., wide broadband characteristics, which makes the equipment with high efficiency in photonic interconnection conversion, stability.