semiconductor van der

3 downloads 0 Views 5MB Size Report
Jul 22, 2017 - b State Key Laboratory of Modern Optical Instrumentation, Zhejiang University, Hangzhou .... evidences [89] and molecular dynamics simulations [90]. ... ported graphene sheet/n-Si heterojunction solar cells have average ef- ..... sorption peak can be tuned by changing the shape, size and material of.
Nano Energy 40 (2017) xxx–xxx

Contents lists available at ScienceDirect

Nano Energy journal homepage: www.elsevier.com/locate/nanoen

Review

High performance graphene/semiconductor van der Waals heterostructure optoelectronic devices

MARK



Shisheng Lina,b, , Yanghua Lua, Juan Xuc, Sirui Fenga, Jianfeng Lic a

College of Microelectronics, Department of Information Science & Electronic Engineering, Zhejiang University, Hangzhou 310027, PR China State Key Laboratory of Modern Optical Instrumentation, Zhejiang University, Hangzhou 310027, PR China c MOE Key Laboratory of Spectrochemical Analysis and Instrumentation, State Key Laboratory of Physical Chemistry of Solid Surfaces, iChEM, and College of Chemistry and Chemical Engineering, Xiamen University, Xiamen 361005, PR China b

A R T I C L E I N F O

A B S T R A C T

Keywords: Graphene Heterojunction Solar cells Photodetector Surface plasmon

As a typical two-dimensional (2D) atomic thin material, the massless Dirac Fermions in graphene promise many unique physical properties, such as high carrier mobility and high light transmission. However, after a decade of research and a huge number of papers focusing on graphene, high performance graphene based optoelectronic devices is still lacked, which should be achieved for realizing competitive industrial graphene product. In this review, we point out high performance optoelectronic devices such as solar cells, photodetector and light emitting diodes can be demonstrated by properly marrying graphene with semiconductor. The fundamental physics of graphene/semiconductor heterostructure as well as its optoelectronic properties are comprehensively addressed. We suggest six strategies of improving the performance of graphene/semiconductor based optoelectronic devices. The outstanding light harvesting enhancement characteristic of surface plasmon is detailed represented to highlight its importance and potential. Especially, we indicate that graphene/semiconductor heterostructure solar cell can reach a power conversion efficiency of 30%.

1. Introduction Graphene has attracted intensive interests as its unique physical properties [1,2] were revealed experimentally since 2004 [3], which indicate some potentials in electronic and optoelectronic device applications [4–6]. Graphene can have a perfect body with carbon atom close-packed into two dimensional (2D) honeycomb crystal. The ‘body’ of graphene can be produced in mass production through liquid-phase exfoliation [7,8], solution chemical reduction [9,10] or chemical vapor deposition (CVD) [11–14] process. Besides the perfect body, graphene also has an outstanding ‘soul’ (fascinating physical properties), such as high carrier mobility [15,16], good light transmittance within wide wavelength range [17–19], great mechanical strength [20–22] and flexibility [23,24], excellent electrical [8,25,26] and thermal conductivity [27,28]. The carrier mobility of CVD-grown graphene achieves ~ 40,000 cm2 v−1 s−1 at room temperature [16,29–31]. The 2D plane with huge specific surface area and high thermal stability makes graphene used for sensors [32–35], electrochemical devices [36–38] and energy storage [39–42]. Besides, the high light transmittance and good electrical conductivity of graphene make it promising for transparent conductive electrodes in light-emitting diode [43–45],



touch screen [12,46] and solar cell [47–49], including organic [50–53], dye-sensitized [54–56] and perovskite [57,58] solar cells. The cost of solar power is beginning to reach price parity with cheaper fossil fuel-based electricity in many parts of the world, yet the clean energy source still accounts for just slightly more than 1 percent of the world's electricity mix. Solar, or photovoltaic (PV), cells, which convert sunlight into electrical energy, have a large role to play in boosting solar power generation globally, such as the graphene-based solar cell. However, the ‘soul’ of graphene is not perfect enough to convert the incoming light into electricity, since electron-hole pairs generated in zero band gap graphene will cool down in the time scale of femtoseconds through electron-electron or electron-phonon interaction. There are two ways to overcome this shortcoming of graphene: one is breaking the lattice symmetry of honeycomb carbon structure, which leads to a limited band gap and the other is forming rectified diode with semiconductor substrate with a suitable band gap for absorbing the incident light. Graphene/semiconductor heterojunction has exhibited good electrical properties [59–61]. Many efforts have been devoted to its photovoltaic applications including Si [62–64], III-V semiconductors [65,66] or nano arrays [67–70] substrate, and great achievements have been obtained. It is worthy noted that, the linear band gap structure and

Corresponding author at: College of Microelectronics, Department of Information Science & Electronic Engineering, Zhejiang University, Hangzhou 310027, PR China. E-mail address: [email protected] (S. Lin).

http://dx.doi.org/10.1016/j.nanoen.2017.07.036 Received 19 June 2017; Accepted 21 July 2017 Available online 22 July 2017 2211-2855/ © 2017 Elsevier Ltd. All rights reserved.

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 1. (a) A single layer of carbon atoms arranged in hexagonal crystal lattice for graphene. (b) The tunable Fermi level of the graphene.

low density states near the Dirac point make the Fermi level of graphene tunable by external electrical field [71–73] or chemical modification [74,75], benefiting for developing high performance graphene-based devices. In this review, we reach an important message that the power conversion efficiency of graphene/semiconductor heterostructure solar cell can exceed 30% in the future.

1.1. Electrical and mechanical properties of graphene Graphene is a monolayer of carbon atoms with close-packed hexagonal structure (Fig. 1a), which is considered as a building block for bulk graphitic materials when the planar layers are stacked and bonded

by a weak van der Walls force [76]. In 2004, Andre Geim and Kostya Novoselov revealed few-layer graphene is stable in air and the 2D carrier gas in graphene can be tuned by external gate voltage [77]. The massless Dirac fermions in graphene promise new physical phenomenon such as the anomalous quantum Hall effect [78]. In addition, the linear electronic structure of graphene makes the work function of graphene highly tunable, by external electrical field [72], chemical molecule modification [79–81], metal configuration [82] or thermal anneal [83]. Raman spectrum is a good technique to monitor the Fermi level of graphene [84]. It is reasonable that the Fermi level of the graphene can be tuned by as high as 3.0 eV [85], schematically shown in Fig. 1(b). The carrier mobility of mechanically exfoliated graphene can be as high as 200,000 cm2 V−1 s−1 [15]. The ‘soul’ of graphene mainly means graphene has a combination of high carrier mobility and tunable Fermi level. As a sequence of unique crystal lattice structure, graphene with a single layer of carbon atoms theoretically absorbs only 2.3% of incident white light, and the opacity increases with the increasing layer number w20hich adds another 2.3% [17]. As a result of interference, the mechanical exfoliation single or bilayer graphene can be seen by naked eyes through the optical microscope, as demonstrated in Fig. 2a. The Raman spectrum of the mechanical exfoliation graphene has mainly two peaks, one is G peak around 1580 cm−1 and the other is 2D peak around 2690 cm−1 (Fig. 2b). As for the transmittance of CVD-grown monolayer graphene at 550 nm is around 97% [12,18,25]. Fig. 2c shows long and continuous graphene transferred onto flexible and transparent substrate. Fig. 2d shows the Raman spectrum of monolayer graphene of the CVD-grown graphene, where a weak D peak around 1350 cm−1 can be found, indicating there are some crystal defects in

Fig. 2. (a) Optical microscope image of graphene fabricated by a “scotch-tape” method, the red curve denotes an area of bilayer graphene. (b) The Raman spectrum of bilayer graphene with the inset shows the Raman spectrum of monolayer graphene in (a), where the negligible D peak indicates the high crystal quality of graphene. (c)Photograph of graphene/epoxy/PET roll before doping and the graphene layer was produced by CVD method. The widths of the graphene/epoxy and the base PET film are 210 mm and 230 mm, respectively. The length of the graphene layer can be as long as 100 m. (d) Raman spectrum of graphene on SiO2/Si substrate fabricated by CVD method, where a weak D peak indicates some defects in the CVD grown large area graphene. Reproduced with permission [120] for (a), Copyright 2014, AIP;

Reproduced with permission [123] for (a), Copyright 2013, The Royal Society of Chemistry. 2

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 3. (a) The 3D and schematic structure of the graphene/semiconductor heterostructure. (b) Typical I-V curve of graphene/GaAs heterostructure.

is 42 Nm−1, and the corresponding stress is 130 GPa which are the highest among the values ever measured [20]. The fracture toughness of the CVD-graphene turns out to be extremely high [21,87]. Its strength hardly degrades despite the existence of grain boundary in the CVD-grown polycrystalline graphene [88]. Whereas, defects have significant effects on the fracture strength of graphene by experimental evidences [89] and molecular dynamics simulations [90]. Graphene could behave as structural stable material for a variety of applications, such as flexible electronics/optoelectronics [12,91] and strengthening components [92,93]. However, the zero band gap of graphene severely limits its industrial widespread applications [3]. In addition to a tunable Fermi level, by breaking the lattice symmetry of graphene, the band gap of graphene can be tuned. Single layer graphene exhibits semimetal character with zero bandgap, which can be slightly open up to 0.25 eV in bi- or tir- layer graphene [94–96]. Recently, we also have demonstrated that silicon alloying has the ability to tune the band gap of graphene [97]. We have synthesized monolayer silicon-doped graphene (SiG) with a large surface area using a CVD method. The work function of SiG, deduced from ultraviolet photoelectron spectroscopy, was 0.13–0.25 eV lager than that of graphene. The band gap of silicondoped graphene has been open as 0.28 eV.

Fig. 4. Mechanism schematic diagram of plasmon enhancement in solar cell and photocatalysis.

1.2. Graphene/semiconductor heterojunction with rectifying behavior Graphene can be combined with 3D semiconductor to form heterojunction with significant photovoltaic responses, and the first reported graphene sheet/n-Si heterojunction solar cells have average efficiencies up to 1.5% [62]. Basically, putting graphene over semiconductor can induce an asymmetric current flowing channel, which results in a rectifying behavior similar with a PN or Schottky diode. As the fabrication process of graphene/semiconductor heterostructure is very simple, with the schematic illustration of device structure shown in Fig. 3a, the research on graphene/semiconductor heterostructure has been ignited since 2010. Other than graphene and semiconductor, there are metal electrodes contacting graphene and semiconductor, respectively. Also, there can be one insulating layer, such as silicon nitride, separating graphene from semiconductor. As a van der Waals heterostructure, the Fermi level of graphene and semiconductor can be adjusted independently. The layer number of graphene sheets can be chosen as monolayer, bilayer and few-layers. As the Fermi levels of graphene and semiconductor are usually different, the built-in electric field can be established between graphene and semiconductor, which rectifies the current flow, such as the asymmetric current-voltage relationship found in graphene/GaAs heterostructure [77], as shown in Fig. 3b.

Fig. 5. Schematic structure of the graphene and the n-type or p-type semiconductor.

the graphene lattice [86]. The sheet resistance of monolayer graphene is usually 200–2000 Ω/sq, and it significantly reduces as the layer number increase [12,25]. The roll-to-roll produced four-layer graphene is superior to commercial ITO film [12], which is indicated to be the promising candidate for transparent conductive electrode in optoelectronic device applications. On the other hand, Graphene is experimentally established as the strongest material. The intrinsic break strength of defect-free graphene

3

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 6. (a) Schematic illustration of graphene/n-Si heterojunction device. Bottom-left inset: cross-sectional view, photogenerated holes (h+) and electrons (e−) are driven into the graphene and n-Si, respectively, by the built-in electric field. Bottom-right inset: photograph of a graphene/n-Si cell with a 0.1 cm 2 junction area. (b) Energy diagram of the forwardbiased graphene/n-Si heterojunction upon illumination. (c) Semilogarhythmic-scale dark current–voltage curves of two graphene/n-Si cells of different junction areas. The insets show the ideality factor (n) and the series resistance (RS) of the 0.1 cm 2 cell extrapolated from the linear regimes are 1.57 and 10.5 Ω, respectively. (d) Dark and light J-V curves of the cells with different areas under AM 1.5 G illumination.

Reproduced with permission [62]. Copyright 2010, Wiley.

Fig. 7. (a) Schematic structure of the graphene/GaAs solar cell. (b) Schematic electronic band structure of the graphene/n-GaAs heterojunction solar cell.

Reproduced with permission [77]. Copyright 2015,

Elsevier. A modified model of heterojunction contacts between graphene and three-dimensional semiconductors has been developed. The model takes better account of the effective ‘zero mass’ and zero gap (semimetal) conduction/valence band structure of graphene with linear energy-wavevector relations. Metals and semiconductors are usually modeled with quadratic energy-wavevector relations, giving non-zero effective mass. The graphene/semiconductor heterojunction is expected to pass current in the forward bias (when the semiconductor is negatively biased) while becoming highly resistive in the reverse bias (when the semiconductor is positively biased). As seen in Fig. 3b, a typical J-V 4

data taken on graphene/semiconductor heterojunction displays strong rectification behavior. This rectification behavior is a consequence of heterojunction-barrier formation at the interface when electrons or holes flow from the semiconductor to the graphene as the Fermi energies equilibrate. In principle, any semiconductor with electron affinity (Χe) smaller than the work function of the mental (Φmetal) can create rectification with Schottky-barrier height, ФSBH = Φmetal-Χe, given by the SchottkyMott model. Similarly, the junction barrier height between graphene and semiconductor (Φbarrier) greatly decided the J-V characteristics of

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 8. (a) Schematic diagram of the MLG/h-BN/ZnO structure. (b) I-V curves of MLG/ZnO diode sandwiched with different layers of h-BN in dark. (c) I-V characteristics of the graphene/ three layers BN/ZnO UV detector under illumination of 365 nm UV light and the power density of UV ranging from 0 to 160 μW cm−2. (d) Time dependence of photocurrent when switching UV light on and off at a cycle of 30 s. The reverse bias is 1 V. Reproduced with permission [115]. Copyright 2015,

OSA Publishing.

Fig. 9. (a) Schematic illustration of the graphene/h-BN/GaN-heterostructure UV photodetector. (b) Dark J-V curves of the photodetector with and without h-BN. (c) J-V characteristics of the graphene/h-BN/GaN photodetector with and without ZnO QDs doping under illumination of 245 nm UV light. (d) Time-dependent photocurrent when switching UV light off and on at a cycle of 20 s on ZnO QD-doped graphene/h-BN/GaN heterostructure under a reverse-bias voltage of 1 V.

Reproduced with permission [101]. Copyright 2016, IOP Publishing. 5

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 10. (a) Band diagrams of the graphene/p-InP Schottky junction interface showing the tunable graphene Fermi level EF-gr. (b) J–V curves of the graphene/p-InP solar cells before and after tuned by chemical doping and applied gate voltage. (c) Performance stability of the graphene/p-InP solar cell devices under one sun illumination.

Reproduced with permission [66]. Copyright 2015, Elsevier.

Fig. 11. (a) Band diagrams of the graphene/p-GaN Schottky junction interface. (b) I-V curves of the devices under dark.

Reproduced with permission [114]. Copyright 2016, Elsevier.

Fig. 12. (a) J-V curves of solar cells based on graphene with 1–6 layers under illumination of 730 nm LED (2 mW/cm 2). (b) Theoretical analysis of the work function modulation on the PCE of graphene/n-Si solar cells as a function of graphene layer number. Reproduced with permission [120] for (a), Copyright 2014, AIP;

Reproduced with permission [123] for (a), Copyright 2013, The Royal Society of Chemistry. 6

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 13. Five sets of typical PCE (a), Voc (b), Jsc (c) and FF (d) of the graphene/GaAs heterojunction solar cells with different layer numbers of graphene.

Reproduced with permission [77]. Copyright 2015, Elsevier.

Fig. 14. (a) Schematic structure of the graphene/h-BN/GaAs sandwich heterostructure. (b) Electronic band structure of the graphene/GaAs heterostructure under thermal equilibrium. (c) Dark J-V curves of the solar cells based on graphene/h-BN/GaAs sandwich heterostructure with different layers of h-BN. (d) J-V curves of solar cells based on graphene/GaAs and graphene/h-BN/GaAs heterostructure. Reproduced with permission [100]. Copyright 2016,

OSA Publishing. semiconductor junction:

the graphene/semiconductor heterojunction. As graphene is atomic thin and the Dos near the Dirac point is quite low, the charge transfer from the semiconductor to the graphene shifts the Fermi level of the graphene and reduce the Φbarrier of the heterojunction. Without considering the accurate interfacial states between graphene and the semiconductor, the Φbarrier under thermal equilibrium condition can be considered as a simplified equation compared with the mental/

Φ barrier = Φgraphene − X e − Δg

(1)

where Φgraphene is the work function of graphene, Χe is the electron affinity of the semiconductor, Δg represents the Fermi level shift of graphene [98]. Electron transport over the Schottky barrier at the M/S interface or 7

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 15. (a) Graphene/n-Si solar cell geometry without (top) and with (bottom) TFSA doping for graphene. (b) Band diagram at the graphene/n-Si interface before (left) and after (right) the doping. (c) Light J-V characteristics of graphene/n-Si diodes before and after the doping. (d) External quantum efficiencies as a function of wavelength for pristine and TFSA-doped graphene/n-Si solar cells. Reproduced with permission [63]. Copyright 2012,

American Chemical Society. effective mass of the tunneling carrier, d is the thickness of interface hBN layer, A* is the effective Richardson constant of the semiconductor, T is the temperature, NIF is the ideality factor, h and K represent Plank constant and the Boltzmann constant, respectively. The effective Richardson constant of the heterojunction is influenced by the interface conditions. The effective Richardson constant and the Φbarrier are obtained based on the Arrhenius plotting of Eq. (4) at temperatures from 300 K to 350 K [100]. PCE, Voc, Isc and FF are the main parameters for a solar cell device. Light characteristics of the graphene/semiconductor heterojunction solar cells are usually tested with a solar simulator under AM 1.5 conditions. The photocurrent action spectra of the graphene/semiconductor heterojunction solar cells are measured with Agilent B1500A system, which can be calculated with the Eqs. (3) and (4). When the graphene/semiconductor heterojunction solar cell is short-circuited, the extracted photogenerated carriers can transit through the external circuit, generating a short-circuit current (Isc). When the solar cell is opencircuited, the separation of photogenerated electrons and holes will produce an open-circuit voltage Voc. The photocurrent is opposite to the forward-biased current of the solar cell. At V = Voc, these two currents will cancel each other and result in a zero net current. FF is the fill factor of the solar cell, representing the characteristic of the maximum power under optimal load, which is always less than 1.

Table 1 Characteristic parameters of graphene/n-silicon solar cells after graphene doped by different chemicals. Dopant

JSC (mA/cm 2)

VOC (V)

FF

PCE (%)

Improved factor

TFSA HNO3 HNO3 SOCl2 H2O2 HCl

25.3 23.9 15.1 17.9 16.9 16.8

0.54 0.55 0.54 0.55 0.55 0.55

0.63 0.68 0.67 0.61 0.55 0.54

8.6 8.9 5.47 5.95 5.12 4.93

4.5 2.4 2.1 2.3 2.0 1.9

heterojunction barrier at the graphene/semiconductor interface can be well described by thermionic-emission theory with the expression:

J = J0 [e

(

qV ) NIF KT −1]

(2)

where J is the current density across the graphene/semiconductor interface, V is the applied voltage, T is the temperature, and NIF is the ideality factor. The prefactor, J0 is the saturation current density and is expected as J0 = A*T 2exp(-qΦbarrier/KT), where qΦbarrier is the zerobias heterojunction barrier height and A* is the Richardson constant [99]. Based on Eq. (1), by decreasing the Δg with suppressing the charge transfer, higher Φbarrier can be obtained. In our recent work, we propose a sandwich structure device by inserting 2D h-BN into graphene/ semiconductor heterojunction to obtain high performance solar cell and photodetector devices. Considering the tunneling probability, J-V curve of the graphene/semiconductor can be expressed as below:

J = J0 e(−4π

2m*Φarrier d) h

J0 = A*T 2e(−

qΦ barrier ) KT

FF =

(5)

(3)

where Pmax is the maximum power of the graphene/semiconductor heterojunction solar cell. PCE is the power conversion efficiency of the solar cell, representing the maximum light energy conversion efficiency of the device under optical load.

(4)

PCE = Voc × Jsc × FF

qV

[e ( NIFKT )−1]

Pmax Voc × Isc

where J is the current density across the graphene/h-BN/semiconductor sandwich structure, J0 is the saturation current density, m* is the

(6)

Similarly, responsivity and detectivity are the main parameters of a photodetector device. Responsivity represents the photocurrent 8

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 16. (a) I(VB) dc characteristic of the device at different VG showing a current suppression for positive VG. The inset shows the electrical measurement configuration. (b) Photovoltaic response of the cell under AM 1.5 illumination showing the photocurrent as a function of VB for different gate voltages. (c) Band structure of graphene and Zn3P2 when VG increases to positive values, it raises the Fermi level in graphene inducing a higher junction barrier and depletion width in the Zn3P2. (D) For negative VG, the band bending and barrier height are reduced, easing the transport across the junction. Reproduced with permission [130]. Copyright 2014,

American Chemical Society. same period, Maxwell Garnett explained why the gold-ruby glass presenting different colors using Drude theory [103]. In 1908, Mie theory about light scattering of spherical particles was proposed by Gustav Mie [104]. These two characters laid the theoretical foundation of surface plasmon. However, surface plamon did not get enough development until 1950s, when surface plasmons was well studied theoretically and experimentally. In 1956, David Pines described the energy loss when fast electrons were scattered by thin metallic solid films theoretically, attributing this phenomenon to the collective oscillation of the free electrons in metal and nominating this behavior as “plasmon” [105]. Meanwhile, Ugo Fano characterized the coupling oscillation between photons and electrons in transparent medium, especially the property of polarization [102]. After which, Rufus Ritchie stated that the plasmonic mode might exist near the metal surface, firstly giving the theoretical description on “surface plasmon” [106]. In 1968, Andreas Otto and coworkers proposed several methods to inspire the surface plasmon on the metallic film, which made it easy to do the experimental research on surface plamon [107]. Until 1970, about seventy years later than the theoretical study on gold ruby glass, Kreibig and Peter Zacharias investigated the electrical and optical properties of gold and silver nanoparticles, and the optical properties of the metallic nanoparticles was firstly related to the concept of surface plasmons [108]. When surface plasmon is confined to the surface of metal nanostructures with subwavelength size, the free electrons in conduction band of metal participate in a coherence oscillation, ascribing to a localized surface plasmon resonance (SPR) [109]. According to the development history of surface plasmon, it can be known that the incident light or energy can be limited to the localized field of the plasmonic nanostructure. From the extinction spectra of the metallic nanoparticle

generated per unit of the incident light power by the device, while detectivity stands for the ability for sensing weak light of the photodetector. Responsivity can be determined with Eq. (7) shown below:

Responsivity =

(IL −ID) Plaser

(7)

where IL is the current under laser illumination, ID is the dark current, Plaser is the power of the incident laser. On the other hand, detectivity can be determined with Eq. (8):

Detectivity =

A •R 2qID

(8)

where A is the active area of the photodetector, R is the responsivity, q is the charge constant, ID is the dark current of the device [101]. 1.3. Surface plasmon resonance While graphene/semiconductor heterostructures are widely used in the field of solar energy conversion, the light harvesting capability of the devices turns to be of great importance to obtain a high performance. According to our study on surface plasmon resonance (SPR) effect, we believe SPR effect a promising strategy to improve the light field of the graphene/semiconductor heterostructures based optoelectronic devices. The science of surface plasmons was firstly observed at the beginning of the twentieth century, when artists utilized the special optical property of the metal nanostructures to decorate the windows of the church. In 1902, Rober W. Wood observed an unexplained phenomenon when he measured the reflectivity of the metal grating [102]. About the 9

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 17. (a) The schematic diagram of the gate tuning structure on graphene/p-InP solar cells. (b) The Fermi level of graphene EF-gr (the negative values represent the EF-gr below the Dirac point) and VOC of solar cells as a function of the applied gate voltage. (c) Measured transient photoluminescence in the InP substrate and graphene/InP heterostructures. (d) The schematic diagrams of carrier separation and recombination processes in graphene/InP and doped graphene/InP heterojunction after excited by the light source.

Reproduced with permission [66]. Copyright 2015, Elsevier.

Fig. 18. (a) Schematic cross-section view and top view structure of the field-effect graphene/GaAs solar cell. (b) Schematic electronic band structure of the field-effect graphene/GaAs solar cell at different Vgate. (c) Dark J-V curves of the graphene/GaAs solar cells with different Vgate. (d) The dependence of Raman G peak on Vgate.

10

Reproduced with permission [77]. Copyright 2015, Elsevier.

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 19. (a) Bright-field TEM image of SiC nanoflakes supported by an ultrathin graphene layer; the scale bar is 100 nm. (b) High resolution image of the bare area in (a), indicating the support layer is graphene, with a lattice parameter of 0.21 nm; scale bar is 5 nm. (c)Typical HRTEM image of small two-dimensional SiC nanoflake; the scale bar is 5 nm. (d) HRTEM image of large 2D SiC nanosheets, where a lattice parameter of 0.28 nm is resolved; the scale bar is 5 nm. (e) Fast Fourier transformation of (d), showing a hexagonal lattice configuration. (f) Schematic illustration of the atomic structure of single-layer graphitic SiC projected along the [0001] direction.

Reproduced with permission [134]. Copyright 2012, American Chemical Society.

light scattering and near-field enhancement, or non-radiative effect such as hot-carrier transfer and plasmonic induced resonance energy transfer (Fig. 4) [112]. In detail, the unabsorbed photons can be reflected by the plasmonic nanoparticles in the solar cell until they are reabsorbed due to the far field light scattering; the incident photons can be concentrated in the near-field of the nanostructure to form a secondary light source with extremely high intensity, which might be orders of magnitude higher compared to the incident light. In the other hand, the surface plasmons can transfer the incident energy directly to the absorber through dipole-dipole coupling or transfer the hot carriers generated from plasmon decay to the nearby semiconductor. Consequently, the light trapping in the optoelectronic devices can be evidently enhanced to achieve high performance.

colloid, a maximum peak attributing to the surface plasmon resonance can be observed at resonant frequency, which is the first characteristic of surface plasmon. The extinction spectra are the combination of the absorption and scattering spectra of the nanoparticles. The SPR absorption peak can be tuned by changing the shape, size and material of the nanostructure, meanwhile, the light being absorbed or scattered at the SPR peak differs with the size of the structure, which allows us to regulate the light harvesting ability to the desired range of the light spectrum. Furthermore, the SPR absorption peak of the nanoparticles is sensitive to the refractive index of the surrounding environment, which is worth noticed during the application [110]. The second important characteristic of the plasmonic nanostructures is that the especially strong electromagnetic field on the surface of the nanoparticles, causing by the interaction between the incident light and the free electrons. The intensity of the localized electromagnetic field decays exponentially to the distance to the surface of the nanoparticle [111]. Owing to the two special properties of the plasmonic nanostructures, surface plasmon can break through the limitation of light absorption in optoelectronic devices, via radiative effects like far field

2. Graphene/semiconductor heterojunction: minority carrier device or majority carrier device? As the Fermi level of graphene is adjustable and the semiconductor can be N-doped type or P-doped type, such as GaAs, GaN, InP, CdTe, 11

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 20. TEM images of ultrathin SiC2 single crystal. (a) TEM image of a suspended SiC2 membrane with area larger than 50 µm2. ED patterns recorded from different areas of the membrane show that it is a single crystal. The single crystal ultrathin quasi-2D SiC2 is stable over three months. (b) Low magnification TEM image of ultrathin quasi-2D SiC2. (c) Electron diffraction image of the area in (b), where the distance between (101̅0) planes is 0.44 nm. (d) HRTEM image of the area labeled C in (b); the scale bar is 10 nm. (e) HRTEM image of the region folded by ultrathin SiC2 membrane marked D in (b); the scale bar is 10 nm. (f) The distribution of carbon atoms and (g) the distribution of silicon atoms in an area of 200 nm × 200 nm of quasi-2D SiC2. Reproduced with permission [135]. Copyright

2012, American Chemical Society. exfoliated or CVD-grown graphene on the top of Si substrate exhibits rectifying electrical behaviors. Under laser illumination, photocurrent and open-circuit voltage (VOC) generate [59]. The junction barrier heights are similar for n-Si and p-Si substrate (0.41–0.46 eV), while the ideality factors are obviously different, both of them affected by the anneal temperature [59]. The ideality factors of graphene/Si diode are usually larger than unity, which could be reduced close to ideal unity by pre-etched copper for avoiding the effect of impurities on CVDgrowth graphene [60]. The novel Landauer transport formalism is well explained for characteristics of the atomically thin graphene/Si diode, indicating that the carrier injection from graphene determines the transport properties of the heterojunction [60]. So graphene can be combined with Si to form PN junction like heterojunction with significant photovoltaic responses, and the first reported graphene sheet/n-Si heterojunction solar cells have average efficiencies up to 1.5% (Fig. 6) [62]. The schematic diagram and optical image of the solar cell are showed in Fig. 6a, where the graphene sheets are composed of monolayer, bi-layer and few-layer graphene. A spacecharge region is formed in n-Si near the graphene/n-Si interface (Fig. 6b), which indicates that graphene film serves as not only a transparent electrode, but also an active layer for electron-hole separation and hole transport. The dark I-V curves exhibit rectifying characteristics, with ideality factor of 1.57 (Fig. 6c). The power conversion efficiency (PCE) of 0.1 cm2 and 0.5 cm2 devices are 1.65% and 1.34% respectively (Fig. 6d), with a large room for further improvement. There are three generations of semiconductor, other than Si, graphene could form heterojunctions with the second and third generation of semiconductors, such as gallium arsenide (GaAs) [61], silicon carbide (SiC) [61,116–118], Znic Oxide (ZnO) [115] and gallium nitride (GaN) [119]. Compared to widely used N-Si, GaAs has advantages of

ZnO and Si, one question arises that what is the difference between graphene/semiconductor heterostructure and traditional PN junction or Schottky junction? Traditional PN junction is a minority device, which is widely used in logic circuit. In contrast, metal-semiconductor Schottky diode is a majority device, which can function as high-speed electronic devices. Graphene is a semimetal, which is different from metal and semiconductors. While graphene usually forms heterojunction with n-type silicon, it is still unclear whether we can form heterostructure by contacting graphene with p-type semiconductor. During the past studies, we found that graphene can form good heterostructure with p-type InP [66], p-type CdTe [113] and p-type GaN [114]. Those studies show that we should give a clear definition of the junction formed by graphene and semiconductors. Judging from the point that what kind of carrier transport through semiconductor, we can define it is a Schottky diode like or PN junction like device. We assume that the Fermi level of graphene lies near the dirac point and the as-grown graphene is weakly p-type doped [77]. For example, by contacting graphene with N-type ZnO [115], N-type GaN [101], N-type GaAs [100] and N-type silicon, we can find that holes inject into graphene and electrons transport through thick semiconductor layer, then we define this device as minority device similar with PN junction like devices. On the other hand, when graphene forms junction with p-type semiconductor and a forward bias is exerting at the graphene side, holes will transport through the p-type layer and electrons will move into graphene layer, similar with that case of Schottky junction. Thus, graphene/p-type semiconductor heterostructure can be defined as majority carrier device, similar with traditional Schottky diode (Fig. 5). 2.1. Minority carrier device: PN junction like device Primitively,

graphene/Si

diode

by

depositing

mechanically 12

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 21. (a) The optical image of SiG transferred to a SiO2/Si wafer. The inset is the schematic of SiG, where several carbon atoms (blue balls) in the lattice are replaced by silicon atoms (yellow balls). (b) Raman spectra of SiG and graphene on a SiO2/Si substrate. (c) and (d) XPS spectra of the SiG and graphene sheets. As shown in (c), there was no dramatic difference in the peak of graphite-like sp2 C between SiG and graphene. From (d), the clear peaks of Si 2p at 101.1 eV and 103.3 eV can be detected in SiG, providing reliable evidence for silicon doping. (e) J–V curves for the SiG/GaAs solar cell and graphene/GaAs solar cell. (f) EQE of the SiG and graphene/GaAs solar cell.

Reproduced with permission [97]. Copyright 2016, The Royal Society of Chemistry. promising for light-emitting-diodes, laser diodes, and ultraviolet (UV) photodetectors. Herein, we designed a novel photodetector based on graphene/ZnO single crystal heterostructure sandwiched with different layers of h-BN and the results show that h-BN layer can largely improve the rectification capability of graphene/ZnO interface (Fig. 8b), which leads to an enhanced photoresponse of the heterostructure. The schematic diagram and cross-section of the photodetector are shown in Fig. 8a and graphene/ZnO PN junction like heterojunction exhibits rectifying behaviors, which can be used to work as the high-performance photodetector (Fig. 8d). Fig. 8c shows the I-V characteristics of the graphene/three layers BN/ZnO UV detector under illumination of 365 nm UV light and the power density of UV ranging from 0 to 160 μW cm−2. The GaN is a typical III–V compound insulator with a band gap of

direct band gap and high electron mobility. Graphene/n-GaAs heterostructure can form PN junction like heterojunction [61,65] and exhibit excellent photovoltaic behaviors [65]. The schematic of the graphene/ GaAs junction solar cell are shown in Fig. 7a. Considering the electron affinity of GaAs and work function of graphene, the energy bands bend at the graphene/n-GaAs interface (Fig. 7b), as we mentioned before. The light generated carriers are separated by the built-in electric field and collected by graphene and GaAs respectively, yielding photovoltaic effects [77]. In addition, ZnO and GaN are the representative materials of the third generation semiconductor. Compared with the former semiconductor, they have the wider band gap and better stability. ZnO is a semiconductor with direct wide band-gap (3.3 eV at 300 K), high chemical stability and large exciton binding energy (60 meV) which is 13

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

reported graphene-based UV photodetectors. The various graphenecompound semiconductor contacts with rectifying behaviors are useful in solar cell, field effect transistor or photodetector applications. 2.2. Majority carrier device: Schottky diode like device Graphene based van der Waals heterostructure has attracted wide attention recently, especially for graphene/p-type semiconductor heterojunction. We have finished some excellent work of the Schottky junction like heterojunction of graphene with the p-type semiconductor. Through delicately designing and engineering the van der Waals heterostructure between graphene and p-type indium phosphide (p-InP), which has a suitable bandgap of 1.34 eV for solar energy conversion (Fig. 10a), we have achieved graphene/p-InP solar cells with power conversion efficiency (PCE) of 3.3% under AM 1.5G illumination [66], as shown in Fig. 10. The chemical doping or electrical field modulation has been used to tune the Fermi level of graphene, which leads to a PCE of 5.6% for the device under gating effect (Fig. 10b). Furthermore, the interface recombination rate could be reduced while graphene is doped or gated, as evidenced by transient photoluminescence measurements. Considering the stability of cell performance under illumination and the high resistance to space irradiation damage of InP, graphene/InP heterojunction may be promising for special applications such as space solar cells, which achieve a stable performance (Fig. 10c). GaN can also be a typical P type semiconductor, which can form Schottky junction like heterojunction with graphene. We have reported a type of LED based on graphene/p-GaN heterojunction [114], which is a PN junction like majority carrier device, as shown in Fig. 11a. But the J-V curve of the devices under dark shown in Fig. 11b indicates bad rectifying behaviors. However, we introduce high performance surface plasmon enhanced graphene/p-GaN LEDs by inserting Ag nano-

Fig. 22. Calculated band gaps of the ground state SixC1-x sheets as the function of x. Blue solid triangles, black hollow hexagons, and red solid hexagons designate semimetallic, indirect, and direct band gap structures, respectively.

Reproduced with permission [136]. Copyright 2015, American Chemical Society. 3.4 eV, which has outstanding chemical stability and can be used to fabricate the blue LEDs or UV photodetectors. And excellent rectifying behaviors and thermal stability up to 550 K have been observed in graphene/n-GaN PN junction like heterojunction [119]. In our recent work, the graphene/GaN heterojunction is designed as a UV photodetector, as shown in Fig. 9a. By inserting h-BN into the interface of the graphene/GaN heterostructure to decrease the dark current and by employing photo-induced doping to increase the photocurrent (Fig. 9b.c), the responsivity and the detectivity reach as high as 1915 A W−1 and 1.02 × 1013 Jones (Fig. 9d), respectively. The responsivity and detectivity obtained are the highest among all the

Fig. 23. (a) Schematic structure of the CdSe QDs covered graphene/CdTe solar cell. (b) J-V curves of the devices under AM1.5 G illumination with and without CdSe QDs. (c) Mechanism of the CdSe QDs introduced photo-induced doping in graphene. (d) Schematic electronic band structure alignment of CdSe QDs covered graphene/CdTe solar cell.

14

Reproduced with permission [113]. Copyright 2015, AIP.

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 24. (a) Electronic band structure of graphene/h-BN/GaN heterojunction with ZnO QDs covering graphene under illumination. (b) J-V characteristics of the graphene/h-BN/GaN photodetector with and without ZnO QDs doping under illumination of 245 nm UV light. (c) High-magnification TEM image of the ZnO QDs. (d) Responsivity values of the graphene/hBN/GaN photodetector with different power intensity of UV lights.

Reproduced with permission [101]. Copyright 2016, IOP Publishing.

increases with the graphene layer number, then decreases when it is over four-layer as a result of the reduced transmittance. As the work function of graphene increases, less graphene layers are needed for achieving the maximum PCE of the graphene/n-Si solar cells through theoretical analysis (Fig. 12b) [123]. Similarly, the bi-layer graphene/n-GaAs solar cell also shows higher conversion efficiency compared to that of single-layer one [65]. So, we should also use the appropriate graphene layer number in the graphene/compound semiconductor heterojunction to achieve the maximum PCE of solar cells. Using multi-layer graphene can not only modify the work function of graphene, but also decrease the sheet resistance. Five sets of typical PCE, Voc, Jsc and FF of the solar cells with different layer numbers of graphene are shown in Fig. 13a-d, respectively. For the graphene/GaAs solar cell, Voc is mainly influenced by the recombination rate in the graphene/GaAs interface region because of the recombination rate in GaAs substrate is similar for all the devices. During the multi transfer process for the devices with different layer numbers of graphene, unexpected recombination centers can be introduced into the interfaces between graphene layers, which lead to higher recombination rates in the interface region. Thus, devices with more layer numbers of graphene have lower values of Voc. Based on the zero band gap of graphene, sunlight absorbed in graphene cannot effectively converted into electricity. Monolayer graphene absorbs 2.3% of sunlight, thus, one more layer of graphene will cause 2.3% decrease of Jsc. During the graphene transfer, supporting PMMA layer cannot be cleaned completely, leading to larger decrease ratio of Jsc for the device

particles (Ag NPs) into the graphene/p-GaN interface. Bidirectional LEDs have been realized with a broad band emission from 550 nm to 650 nm at a forward bias of graphene side and a sharp emission of ~ 400 nm at a reversed bias of graphene. The emission intensity of graphene/Ag NPs/p-GaN is largely enhanced in both forward and reverse bias situations when compared with the bare graphene/p-GaN heterostructure, which is attributed to the surface plasmon resonance of Ag NPs. These results indicate that graphene/Ag NPs/p-GaN heterojunction is a promising candidate for high brightness LEDs. 3. Strategies of improving graphene-based heterojunction optoelectronic devices 3.1. Controlling the number of graphene's layers Graphene-based heterojunction has distinct advantages than ITO/ semiconductor junction due to the tunability of graphene's work function. The work function of graphene is affected by the number of layers. The J-V characteristics of graphene/n-Si heterojunction solar cells under illumination are dependent on the number of graphene layer (Fig. 12a) [120]. The VOC shows a linear increase with the graphene layer number, which is the opposite trend for graphene/p-Si junction device [121]. These are accordance with the increased work function of graphene as the number of graphene layers increases [122], which improves the barrier height of graphene/n-Si, while reduces that of graphene/p-Si junction. The short-circuit current density (Jsc) firstly 15

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 25. (a) Illustration of the solar cell model. (b) SEM image of the deposited Ag NPs on the 10 nm SiO2 layer of the solar cell. (c) The light absorptions for the solar cells for different designs. (d) External quantum efficiency (EQE) for the solar cell using a-Si:H absorber layer with a thickness of 150 nm. A for ITO contact only, B for front only NPs, C for both sides NPs, and D for back only NPs. Reproduced with permission [149]. Copyright 2015,

OSA Publishing. cells (Fig. 14d).

with more layers of graphene. FF of the graphene/GaAs solar cell is mainly influenced by series resistance for the devices with different layer numbers of graphene. The more layers of graphene in the device, the lower the series resistance, and thus higher FF is expected. The best PCE obtained with as-grown graphene is 9.20% with trilayer graphene on GaAs.

3.3. Static electrical gating/chemical doping of graphene for Fermi level tuning The performance of graphene/semiconductor heterojunction solar cells has been greatly improved by comprehensively chemical doping [63,124,125]. A single layer graphene/n-Si exhibits PCE of 8.6% under AM 1.5 illumination upon bis(trifluoromethanesulfonyl)-amide [((CF3SO2)2NH)] (TFSA) chemical doping, exceeding the native device performance by a factor of 4.5 (Fig. 15) [63]. TFSA was coating on the graphene/n-Si device surface (Fig. 15a), which increases the work function of graphene via charge transfer and improves the barrier height of graphene/n-Si junction (Fig. 15b). The VOC thus increases and the fill factor (FF) is improved due to the reduced series resistance, as well as for the JSC (Fig. 15c). The increase of carrier collection leads to the improved quantum efficiency of the doped graphene/n-Si solar cells (Fig. 15d). Kinds of p-type dopants of graphene on performances of graphene/ n-Si heterojunction solar cells are listed in Table 1. The highest improved factor of performance is achieved by TFSA [63], and similar values (~ 2) are got for other dopants (HNO3, SOCl2, H2O2, HCl) [64,124]. Graphene/n-Si nanowire-arrays heterojunction solar cells also obtain significant improvements by chemical doping [67,126]. The p-doping of graphene by such oxidants is due to ions impacting into graphene surface and causing electrons transferring from graphene [81,125]. The work function of graphene thus increases, and the junction barrier height is improved. In addition to chemical doping, electrical field effect can also tune the work function of graphene [72] and modulate electrical behaviors of graphene/semiconductor heterojunction [127]. Yu et al. [128] have applied external electrical field or a ferroelectric material on the graphene/n-Si heterojunction solar cells, resulting in the performance of

3.2. Interface engineering of graphene/semiconductor heterostructure Great achievements have been obtained on graphene/semiconductor devices by lots of efforts, however, many devices often exhibit distinctive s-shaped kink in the measured I-V curves under illumination. For conventional semiconductor photovoltaic devices, reducing recombination at the surface by depositing a passivation dielectric film is of great importance to achieve high performance. The quality of the interface at atomically thin graphene/semiconductor heterostructure would affect carrier transport and collection behaviors significantly. As shown in Eq. (1), Φbarrier = Φgraphene – Χe –Δg, by decreasing Δg with suppressed charge transfer, high Φbarrier can be obtained. Here, our recent work demonstrates that the interface of the graphene/semiconductor heterojunction play an important role in the high performance of the photoelectronic devices [100]. We recently found that 2D hexagonal boron nitride (h-BN) used as an interlayer in graphene/GaAs heterojunction can significantly improve the device performance by suppressing the charge transfer between the graphene and semiconductor (Fig. 14). A tunneling diode structure is designed (Fig. 14a) and the interface charge transfer could be suppressed to decrease the Δg (Fig. 14b). In the I-V tests of graphene/GaAs rectified diodes, the current reduces as the introducing of h-BN layer but the threshold voltage of the device increases with the increase of the number of h-BN layers (Fig. 14c). A PCE of 10.18% has been achieved with five-layer BN insertion in graphene/GaAs heterostructure compared with 8.63% without BN in graphene/GaAs heterojunction solar 16

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 26. (a) Illustration of the synthesis process for Al nanoparticles. (b) SEM image of the Al nanoparticles on the front side of the device. (c) JSC enhancement of the solar cell integrated with 100 nm Ag NPs and Al NPs with different diameter under different surface coverage. (d) External quantum efficiency enhancement (EQE) of the solar cell when using 100 nm Al NPs and 100 nm Ag NPs. (e) Absorption cross-sections (Qabs) for 100 nm Al NPs and 100 nm Ag NPs. Reproduced with permission [150]. Copyright 2013,

Nature Publishing Group. increases and the PCE improves correspondingly. The PCE comes to 18.5% when the gate voltage equals to −15 V, which is the record high for graphene based heterojunction solar cells. Furthermore, it could be theoretically improved up to 23.8% using a higher gate voltage of −60 V. The simulated electronic band structure of gated graphene/ GaAs solar cells (Fig. 18b) well illustrates that the Fermi level of graphene touching GaAs moves downwards as the applied gate voltage increases, thus the barrier height of heterojunction increases and the VOC of solar cells improves (Fig. 18c), so as the G peak of the graphene's Raman (Fig. 18d). Although the performance of graphene-based heterojunction solar cells has been improved significantly by chemical doping or electrical field modulation, the increase of Voc or junction barrier height is smaller than that of its direct effect on the work function of graphene film. For example, the work function of graphene increases by ~ 0.5 eV with AuCl3 doping [81], and it reduces by 1.1 eV after reduced viologen doping [131]. The improvement of VOC is 0.1–0.2 eV when using the dopants on graphene based solar cells [63,66]. More work is needed to explore the origin and optimized effect of graphene doping on the graphene/semiconductor solar cell performance.

graphene/n-Si solar cells shows an improvement more than twice of the control samples. Similar improvements with PCE from ~ 0.5% to ~ 1.8% for graphene/n-Si junction solar cells by negative gate voltages have been achieved [129]. The electric field-effect is feasible for other graphene/semiconductor heterojunction solar cells, such as graphene/ p-znic phosphide (Zn3P2) heterojunction (Fig. 16) [130]. The current across the graphene/p-Zn3P2 junction can be modulated by the gate voltage (Fig. 16a). The illuminated I-V curves exhibit an Ohmic contact at negative gate voltage and an improved rectifying behavior at positive gate voltage, with the PCE improved by 2-fold at +2 V (Fig. 16b). The energy band structure illustrates that the Fermi level goes higher and increases the band bending as the positive gate voltage increases (Fig. 16c). In the opposite way, the band bending and barrier height decrease with negative gate voltage (Fig. 16d). Recently, our work demonstrates that Fermi energy of graphene increases with applied positive gate voltage, evidenced by the Raman G band shift of graphene, and the VOC of graphene/p-indium phosphide (InP) heterojunction solar cells is improved (Fig. 17a, b) [66]. On the other hand, the electron transport across graphene/InP interface becomes faster through transient photoluminescence measurements when applied with chemical doping or gate voltage, thus the interface recombination is suppressed (Fig. 17c, d). With applied negative gate voltage on the top electrode (Fig. 18a), high efficiency field-effect graphene/n-GaAs solar cells also have been demonstrated [77]. As the voltage increases, VOC of the solar cells

3.4. Band gap engineering of graphene by two dimensional silicon-carbon system The band gap engineering of graphene has been proved as difficult 17

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 27. Illustration of OPVs with Au@Ag CNs embedded in (a) PEDOT:PSS layer and (b) BHJ active layer. (c) TEM images of Au@Ag core-shell CNs with different Ag shell thickness. Current density-voltage (d), EQE and EQE enhancement curves (e) of OPVs based on PTB7:PC71BM when Au@Ag CNs were integrated in different layers.

ability to tune the band gap of graphene [97]. We have synthesized monolayer silicon-doped graphene (SiG) with a large surface area using a CVD method, as shown in Fig. 21(a). The silicon doped graphene has been realized by reaction between SiH4 and CH4 on copper substrate using H2 as the carrying gas. The Raman spectrum of SiG shown in Fig. 21b reveals that the D peak enhances after silicon doping. The silicon atoms are doped into the graphene lattice at a doping level of 2.7–4.5 at%, as evidenced by XPS scan of Si 2p electron shown in Fig. 21c. The work function of SiG, deduced from ultraviolet photoelectron spectroscopy, was 0.13–0.25 eV lager than that of graphene. The band gap of silicon-doped graphene has been open as high as 0.28 eV, as shown in Fig. 21(d). Moreover, compared with graphene/ GaAs heterostructure, SiG/GaAs exhibits the enhanced performance and the performance of 3.4% silicon doped SiG/GaAs solar cell has been improved for 33.7% in average shown in Fig. 21e, which is attributed to the increased barrier height and improved interface quality as evidenced by EQE of the device shown in Fig. 21f. It could be thought that 2D silicon-carbon crystal can be existed with a variation of silicon concentration. Indeed, the calculations show

and several proposals have been explored, such as fabrication of graphene nanoribbons, boron nitride doping of graphene, bilayer or trilayer graphene, fluorination of graphene et al. However, it has been demonstrated as difficult especially for controlling the shape of edge lattice of graphene nanoribon, the homogenous distribution of boron and nitride atoms in graphene lattice. The band gap of bilayer or trilayer graphene can be open up to 0.3 eV, which is still low for many optoelectronic applications. The fluorination [132] or hydrogenation [133] of graphene will usually result in an insulating state of graphene. Recently, two dimensional silicon-carbon system, such as the 2D SiC, SiC2 and silicon doped graphene has been demonstrated experimentally [134,135]. We have fabricated 2D SiC by sonication method and chemical vapor deposition method using graphene as the templates. Also, 2D SiC2 has been discovered when the reaction between graphene and silicon takes place near 1500 °C. As shown in Figs. 19 and 20, the TEM image of the 2D SiC and SiC2 was demonstrated clearly, which indicated that high quality and ultrathin 2D SiC and SiC2 single crystal was obtained. Recently, we also have demonstrated that silicon alloying has the

18

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 28. (a) Illustration of the photocatalytic water splitting system. (b) SEM image of the Au nanoparticles deposited on the surface of TiO2 electrode. (c) Photocurrent of the TiO2 electrode deposited with and without Au nanoparticles under visible light irradiation. (d) Electric field distribution at the interface of TiO2 and Au nanoparticles (simulated by FDTD method).

Fig. 29. (a) Illustration of the excitation of the hot carriers from the decay of the surface plasmon resonance. (b) Transfer of the plasmonic hot electrons from gold to a n-type gallium arsenide through a Schottky barrier. (c) Plasmonic hot holes transfer from gold to p-type gallium arsenide without barrier. (Ef is the Fermi energy, Ec and Ev are the bottom of conduction band and the top of valence band, respectively.)

performance of the photoelectronic devices [100,101,113]. The graphene forms heterojunction with p-type or n-type semiconductor and the photo-induced electrons or holes of the quantum dots will transport into the graphene layer, which will lead to p-type or n-type doping of the graphene and then change the Fermi level. In our research, graphene/CdTe heterojunction exhibits photovoltaic responses and its performance could be improved by coating CdSe quantum dots, from 2.1% to 3.1% (Fig. 23a, b) [113]. Photo-induced doping is mainly accounted for this improvement (Fig. 23c, d), which is feasible to other graphene heterostructure solar cells, such as the graphene/GaAs heterostructure solar cell. In our recent research, we also use the ZnO quantum dots (Fig. 24c) to dope the graphene/h-BN/GaN heterostructure photodetector (Fig. 24a) [101]. The graphene /GaN heterojunction is designed as a UV photodetector. By inserting h-BN into the interface of the graphene/ GaN heterostructure to decrease the dark current and by employing photo-induced doping to increase the photocurrent, the responsivity

that the variation of silicon doping concentration can promisingly introduce the band gap engineering of graphene [136] as shown in Fig. 22. Forming heterostructure between 2D silicon-carbon systems is attractive for optoelectronic applications. For example, we have improved the performance of graphene/GaAs solar cell by silicon doping of graphene. Our results suggest silicon doping can effectively engineer the band gap of monolayer graphene and SiG has great potential in optoelectronic device applications. Besides, a series of g-SiC2/GaN bilayer and g-SiC2 nanotube/ZnO monolayer excitonic solar cells (XSCs) have been proposed, which exhibit considerably high PCEs in the range of 12–20% [137]. 3.5. Dynamic photo-induced doping of graphene by semiconductor quantum dots We also found that the photo induced doping of the quantum dots such as Si/CdSe/ZnO quantum dots was a feasible way to enhance the 19

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 30. (a) Schematic electronic structure and hot electron transfer process in a CdSe-Au nanorod. (b) Plasmon-induced hot-electron transfer (PHET) mechanism. (c) Direct metal-to-semiconductor interfacial charge-transfer transition (DICTT) mechanism. (d) Plasmon-induced metal-to-semiconductor interfacial charge-transfer transition (PICTT) mechanism.

and the detectivity reach as high as 1915 A W−1 and 1.02 × 1013 Jones, respectively, as shown in Fig. 24b. The responsivity and detectivity obtained are the highest among all the reported graphenebased UV photodetectors. The photo-doping effect increases the barrier height and carrier concentration at the graphene/h-BN/GaN heterojunction. The photo-induced doping by the ZnO QDs provides a feasible and stable way to develop high-performance graphene/semiconductor heterostructure photodetectors (Fig. 24d).

plasmon nanoparticles, as these parameters of the particles dictate the spectral features of their plasmon resonance, typically for gold nanosphere and nanorod [148]. Consequently, we can utilize the metal nanoparticles with specially appointed spectral absorption to improve the light harvesting or charge carriers transfer of the graphene/semiconductor heterostructure. We believe that SPR effect can open up a new situation in the field of solar energy conversion. Then, the specific mechanism and the application of the four correlated effects and the application in optoelectronic devices and photocatalysts are explained in the following section.

4. New strategy of surface plasmon enhancement

4.1. Far-field light scattering effect

Graphene is compatible to be integrated with semiconductor, however, the device suffers from light loss due to the high light reflection of semiconductor. Several strategies have been advanced to reduce the light loss of the photocatalyst and optoelectronic device. For example, nanowire or nanohole arrays on or beneath the surface of semiconductor are believed to achieve high light absorption compared to planar semiconductor [126]. According to our judgment, introducing metallic nanoparticles which support surface plasmon resonance effect seems to be a noteworthy and promising way to improve the light absorption and the energy conversion efficiency of the device [138]. Surface plasmon resonance (SPR), an optical phenomena of coherence oscillation of conduction band electrons in nanoparticles [139], has been extensively used in fields of surface enhanced spectroscopy [140,141], sensing [109,142], photocatalysis [143,144], photodetectors [145], and photovoltaics [146,147]. The plasmonic metal nanoparticles exhibit specific optical and electromagnetic properties such as far-field light scattering, near-field enhancement, hot charge carrier transfer, and plasmon-induced resonanse energy transfer, which lead to the wide application of SPR [138]. It is possible to regulate the SPR effect through changing the metal material, shape, or size of the

As mentioned above, plasmonic metal nanoparticles exhibits farfield light scattering signature, which mainly depends on the geometry of the nanoparticles and is mostly observed from the nanoparticles relatively larger than 90 nm [112]. The incident photons can be scattered repeatedly by the plasmonic nanoparticles in the structure, causing a reabsorption of the lost light by the light absorbers in the device. The plasmonic nanoparticles can be placed to different locations in the device to acquire different effect. For example, when the plasmonic nanoparticles are placed to the frontward or backward position of the semiconductor film, the incident light will be effectively trapped in the film as the photons will be reflected several times by the metal nanoparticles and the optical length can be increased. Joshua D. Winans and coworkers added Ag nanoparticles into an ultrathin amorphous silicon solar cell at the place of front, back and both (Fig. 25) [149]. The cell with Ag NPs located at both front and back side showed the highest enhancement, showing improvement in power conversion efficiency of 50%, while the front modified cell exhibited better performance than the backward modified one. The nanoparticles located to the frontward 20

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 31. (a) TEM image of the Au/Ag/SiO2 nanostructures. (b) Illustration of the plasmonic embedded dye-sensitized solar cell. (c) Current density-potential curves of the solar cells with different particle density of Au/Ag/SiO2 nanostructures in the devices. (d) The maximum and average power conversion efficiencies achieved from five devices with different nanoparticle concentration.

electromagnetic intensity decays exponentially with the distance to the particle surface, so it is important to place the light absorbers around the nanoparticles, approximately within 50 nm from the surface. As a result, the metallic nanoparticles are always located in the absorption layer of the solar cell. Feng Yan and co-workers tuned the localized SPR effect of the Au@Ag core-shell nanocuboids (NCs) through regulating the thickness of Ag shell to match the spectral absorption of the organic photovoltaics (OPVs) (Fig. 27) [151]. The Au@Ag NCs embedded into the active layer of OPVs showed higher efficiency than the ones located in other place. While coated with SiO2 shell to prevent the charge carrier recombination on the particle surface, the integrated OPVs exhibited even higher PCE attributed to the synergistic effect of scattering and nearfield enhancement. The average PCE of the OPVs was enhanced to 10.42%, firstly realizing a higher efficiency than 10% for palsmonic OPV cell. Another work using gold nanocubes to improve the performance of dye-sensitized solar cells was conducted by Rizia Bardhan [152]. Nanocubes coated with SiO2 shell were incorporated into the TiO2 layer. The strong electromagnetic field at the sharp corners derived from surface plasmon resonance could effectively facilitate the generation of the charge carriers in the devices. A PCE of 7.8% was obtained with 34% enhancement, attributing to the proper concentration of nanocubes. The near-field enhancement effect is also applied to augment the activity of the photocatalysts. By introducing Ag nanoparticles with SiO2 protective shell into TiO2 layer, the photodegradation activity was enhanced by 7 times, attributing to the electronic near-field

should extend the optical path of the incident photons and the forward placed nanoparticles should reflect the light not being absorbed by the front film, so that the light harvesting can be augmented. Moreover, the light wavelength being trapped by the nanoparticles can be tuned through altering the material and size, for instance, Al for ultraviolet region, Ag for visible region, and Au for visible and near infrared region. Min Gu and coworkers introduced Al nanoparticles on the front side of the silicon solar cell, a QCE of 19.54% was achieved by integrating graphene sheets simultaneously (Fig. 26) [150]. The JSC and EQE enhancement changed with the material, size and the surface coverage of the nanoparticles. While Al nanoparticles attributed to a higher capability of trapping the photons in the ultraviolet region, the solar cells showed higher efficiency for Al nanoparticles rather than Ag nanoparticles. The size and the surface coverage of the particles could affect the scattering angle and the amount of the incident photons, so these two parameters should be optimized to get a higher efficiency. 4.2. Near-field enhancement effect Near-field enhancement by localized SPR is also beneficial to increase the solar cell's efficiency. For plasmonic nanoparticles, incident light will be confined at the local field of the particle surface, bringing about a localized intense electromagnetic field at the surface, and the nanoparticles can also be regarded as a secondary light source. The nonspherical shaped particles such as nanocubes and nanostars always show higher intensity of electromagnetic field than spherical particles, especially at the corners and edges [139]. The near-field 21

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 32. (a) HRTEM image of Au@SiO2@Cu2O sandwich nanoparticle. (b) Schematic representation of the different energy transfer mechanisms during the photocatalytic process for Au@Cu2O nanostructure. (c) Time course of photodegradation process for methyl orange under visible light irradiation. (d) PIRET (energy upconversion) and FRET (energy downconversion) distinguished spectrally.

Nordlander and Govorov stated that Au or Ag nanoparticles smaller than 20 nm could generate more efficient hot electrons with over barrier energy according to the calculation result, and the importance of the polarization of the incident light was also stated [155,157]. While the hot electrons are able to transfer to the n-type semiconductor, the hot holes can also transfer to p-type semiconductor according the theoretical predictions (Fig. 29) [158]. The electronic structure and hotcarrier distribution from the decay process of surface plasmon resonance were simulated on gold, silver, copper, and aluminium. It was predicted that gold and copper could produce hot holes with much more energy than hot electrons, while the energy of the hot carriers generated by silver and aluminium were much more equitable. In consequence, hot carrier transfer effect can be applied into different kind of solar cells with different composition. The hot charge carrier transfer effect is widely applied in photocatalysis field as well because it can extend the spectral response range of the photocatalysts. To study the transfer mechanism between metallic nanoparticle and the semiconductor, Lian et al. synthesized CdSe/ Au nanorod heterostructure and investigated the charge transfer process through transient absorption spectroscopy. It was proposed that the excited hot electrons transferred to CdSe directly via plasmon-induced interfacial charge-transfer transition (PICTT) mechanism, which was different to the plasmon-induced hot-electron transfer (PHET) and the direct metal-to-semiconductor interfacial charge-transfer transition (DICTT) mechanism (Fig. 30). A high quantum efficiency of 24% was achieved in this model, indicating that hot electron transfer might be another effective way to transfer solar energy to chemical energy through photocatalytic reaction [159]. Wei David Wei and coworkers

enhancement in neighbouring places of Ag nanoparticles. While the SiO2 shell became thinner, the enhancement factor of the photocatalytic activity became higher as TiO2 layer was closer to the electromagnetic field of Ag nanoparticles [153]. Furthermore, the enhancement effect was also demonstrated in photocatalytic water splitting system. By depositing Au nanoparticles on TiO2 electrode, a great number of “hot spots” with high intensity electromagnetic field were formed, increasing the generation rate of the electron-hole pairs in TiO2, and photocatalytic performance was improved by 5 times under visible light irradiation consequently (Fig. 28) [154].

4.3. Hot charge carrier transfer effect Plasmonic nanoparticles with relatively small size can generate hot electron-hole pairs not in thermal equilibrium via plasmon decay and the hot electrons or holes can be injected into the conduction band or valance band of the semiconductor that is directly contacted with the nanoparticles, respectively [155]. This provides a new way to transfer energy from the plasmonic nanoparticle to semiconductor. The efficient excitation and capture of the hot charge carriers is dominated by the size, shape, and metal composition of the nanoparticles. The size dependent effect of the plasmonic hot electron transfer mechanism applied in solar cells was studied by Udo Bach and coworkers [156]. The gold nanoparticles with the size ranging from 5 nm to 40 nm were located at the interface of TiO2/spiroOMe TAD (hole conductor). The results showed that nanoparticles with smaller size (5 nm) showed higher photon-to-electron conversion efficiency (APCE, 13.3%), while a APCE of 3.33% was obtained from 40 nm nanoparticles. Besides, both 22

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 33. (a) Illustration of the surface plasmon-enhanced graphene/GaAs solar cell. (b) AFM image of graphene coved with 80 nm Au NPs. (c) J–V curves of the original surface plasmonenhanced graphene/GaAs solar cell and the device with chemical doping and ARC layer. (d) Stability of the surface plasmon-enhanced graphene/GaAs solar cell under AM1.5 G illumination.

enhancement in the DSSC performance by introducing silica-coated bimetallic nanostructures (Au/Ag/SiO2 NSs) into its active layer (Fig. 31) [163]. The power conversion efficiency reached 7.51% relative to 5.91% for the reference, which means 26% enhancement is achieved by just adding 0.44 wt% plasmonic nanoparticles into the matrix. They attributed the enhancement to the intense near-field interactions between nanostructures and dye, far field light scattering, and plasmon resonant energy transfer from metal to TiO2. The PIRET effect was confirmed by the findings of a decrease in amplitude weighted lifetimes with the increasing particle density in transient absorption spectroscopy (TAS). Actually, the PIRET effect was denoted by Wu et al. and proved in photocatalytic process. A series of pure Cu2O, Au@Cu2O core shell, and Au@SiO2@Cu2O sandwich nanostructures were prepared and the photodegradation activity was measured (Fig. 32). It was found that, the sandwich nanostructure, with SiO2 shell preventing the electron transfer and the recombination at the interface, showed the highest activity due to the resonant energy transfer from Au to Cu2O [162,164]. The mechanism was investigated using transient absorption and timeresolved fluorescence spectroscopy. According to the investigation, the spectral overlap between metal (donor) and semiconductor (acceptor) was found to be indispensable while the direct contact was not necessary for the resonant energy transfer. This was also demonstrated in the Au@TiO2 and Ag@TiO2 core shell structure, while SiO2 shell is used to separate the two materials and Au or Ag core was chosen to change the spectral overlap [165]. The above studies also stated that the PIRET effect can extend the absorption of the metal-semiconductor heterostructure by taking advantage of sun light near and below the band edge of the semiconductor, enhancing the photocatalytic activity of the photocatalysts consequently.

decorated Au nanoparticles with different size on the surface of titanium dioxide and demonstrated hydrogen evolution at long wavelength region which cannot be utilized by titanium dioxide directly. Hot electrons could be excited on the relatively bigger size Au nanoparticles and transferred to TiO2 to complete the reduction reaction. It was also found that no hydrogen was detected when Au nanoparticles employed were too small and did not show any plasmonic property [160]. Besides, a plasmonic enhanced photocatalytic oxygen production reaction was also realized benefitting from hot hole transfer from Ag nanoparticles to BiOCl(001). Xiong and coworkers loaded Pd and Ag nanoparticles on the different facet of BiOCl single crystal, owing to the synergetic effect of Schottky junction and plasmonic hot hole injection, the photocatalytic activity of BiOCl was the significantly improved [161]. 4.4. Plasmon-induced resonance energy transfer effect Plasmon-induced resonance energy transfer (PIRET) is another way to integrate energy into the solar cell. Just like Förster resonance energy transfer (FRET), energy transfers from emitter to absorber by non-radiative dipole–dipole coupling when they have an overlapping spectra [162]. In the case of plasmonic-semiconductor solar cell, the plasmonic metal acts as the donor and absorbs sunlight, then transfers the absorbed energy to the semiconductor, which generates electron-hole pairs in the nearby semiconductor. The energy transfer rate depends on the spectral overlap integral between plasmonic metal and semiconductor as well as the distance between them. Different from hot electron transfer, direct contact or band alignment between plasmonic nanoparticle and semiconductor is unnecessary for PRET to take place, thus offering a new strategy in enhancing the device performance. Rizia Bardhan and her co-workers demonstrated a remarkable 23

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Fig. 34. (a) Schematic structure of the graphene/Ag NPs/p-GaN. (b) EL spectra of graphene/p-GaN LED at a reverse bias. (c) EL spectra of graphene/Ag NPs/p-GaN LED at reverse bias. (d) The variation of EL peak when increases the reverse current. Fig. 35. Illustration of the graphene/semiconductor solar cell combining gating effect and surface plasmon enhancement effect and the proposed performance.

JSC, the final device achieves PCE as high as 16.2% (Fig. 33c). At the same time, we have tested the stability of the device, which behaved an excellent stability (Fig. 33d). We also introduce high performance surface plasmon enhanced graphene/p-GaN LEDs by inserting Ag nano-particles (Ag NPs) into the graphene/p-GaN interface, as shown in Fig. 34a [114]. Bidirectional LEDs have been realized with a broad band emission from 550 nm to 650 nm at a forward bias of graphene side and a sharp emission of ~ 400 nm at a reversed bias of graphene. The emission intensity of graphene/Ag NPs/p-GaN is largely enhanced in both forward and reverse bias situations when compared with the bare graphene/p-GaN heterostructure (Fig. 34b.c), which is attributed to the surface plasmon resonance of Ag NPs. These results indicate that graphene/Ag NPs/p-GaN heterojunction is a promising candidate for high brightness LEDs (Fig. 34d). According to the above discussion, the light absorption and the performance of the solar cell or the graphene/semiconductor optoelectronic devices are effectively enhanced through regulating the shape, size, material, composition or the location of the plasmonic

4.5. Plasmon enhanced graphene/semiconductor optoelectronic devices According to previous discussion, Au nanoparticles formed on graphene/semiconductor device surface through deposition of Au films with different thicknesses followed by anneal or spin of the Au nanoparticle solution in ethyl alcohol might be useful to improve the graphene/semiconductor device performance, which can improve the work function and carrier concentration of graphene due to the enhanced light field and built-in electric field effect. Our work recently demonstrates that the Au nanoparticle can be used to enhance the performance of graphene/GaAs solar cell by simply spinning the Au nanoparticles on the surface of graphene. After coating the surface plasma on to the surface of graphene/semiconductor device, its performance could be improved significantly. The illustration of monolayer graphene/n-GaAs device with surface plasma on the surface is shown in Fig. 33a. The distribution of the Au plasma is relatively uniform, which significantly reduce the light decay and enhance the light field (Fig. 33b). With the main function of TFSA doping to improve the VOC and FF, and surface plasmon enhanced light field to increase the 24

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

and 21522508), Special Foundation of Young Professor of Zhejiang University (No. 2013QNA5007), Plan of Outstanding Young Teacher of Zhejiang University, Postdoctoral Science Foundation of China (2014M561759), and Thousand Youth Talents Plan of China.

nanostructures in the devices. With better understanding on the specific enhancement mechanism, the absorption, scattering, or the localized electromagnetic field can be tuned to match the demand with the different mechanism. It is worth mentioning that these different effects always arise at the same time but not independently. It is obvious that surface plasmon can be applied to gain high performance of the optoelectronic devices, but how to make full use of these effects is still a problem remains to be solved.

References [1] A.H. Castro Neto, F. Guinea, N.M.R. Peres, K.S. Novoselov, A.K. Geim, Rev. Mod. Phys. 81 (2009) 109–162. [2] A.K. Geim, K.S. Novoselov, Nat. Mater. 6 (2007) 183–191. [3] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos, I.V. Grigorieva, A.A. Firsov, Science 306 (2004) 666. [4] F. Bonaccorso, Z. Sun, T. Hasan, A.C. Ferrari, Nat. Photonics 4 (2010) 611–622. [5] G. Fiori, F. Bonaccorso, G. Iannaccone, T. Palacios, D. Neumaier, A. Seabaugh, S.K. Banerjee, L. Colombo, Nat. Nanotechnol. 9 (2014) 768–779. [6] Y. Sun, Q. Wu, G. Shi, Energy Environ. Sci. 4 (2011) 1113–1132. [7] Y. Hernandez, V. Nicolosi, M. Lotya, F.M. Blighe, Z. Sun, S. De, I.T. Mcgovern, B. Holland, M. Byrne, Y.K. GunKo, J.J. Boland, P. Niraj, G. Duesberg, S. Krishnamurthy, R. Goodhue, J. Hutchison, V. Scardaci, A.C. Ferrari, J.N. Coleman, Nat. Nanotechnol. 3 (2008) 563–568. [8] X. Li, G. Zhang, X. Bai, X. Sun, X. Wang, E. Wang, H. Dai, Nat. Nanotechnol. 3 (2008) 538–542. [9] V.C. Tung, M.J. Allen, Y. Yang, R.B. Kaner, Nat. Nanotechnol. 4 (2008) 25–29. [10] S. Stankovich, D.A. Dikin, R.D. Piner, K.A. Kohlhaas, A. Kleinhammes, Y. Jia, Y. Wu, S.T. Nguyen, R.S. Ruoff, Carbon 45 (2007) 1558–1565. [11] X. Li, W. Cai, J. An, S. Kim, J. Nah, D. Yang, R. Piner, A. Velamakanni, I. Jung, E. Tutuc, S.K. Banerjee, L. Colombo, R.S. Ruoff, Science 324 (2009) 1312–1314. [12] S. Bae, H. Kim, Y. Lee, X. Xu, J.S. Park, Y. Zheng, J. Balakrishnan, T. Lei, H.R. Kim, Y.I. Song, Y.J. Kim, K.S. Kim, B. Ozyilmaz, J.H. Ahn, B.H. Hong, S. Iijima, Nat. Nanotechnol. 5 (2010) 574–578. [13] A. Reina, X. Jia, J. Ho, D. Nezich, H. Son, V. Bulovic, M.S. Dresselhaus, J. Kong, Nano Lett. 9 (2009) 30–35. [14] Z. Sun, Z. Yan, J. Yao, E. Beitler, Y. Zhu, J.M. Tour, Nature 468 (2010) 549–552. [15] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos, I.V. Grigorieva, A.A. Firsov, Science 306 (2004) 666–669. [16] H. Lv, H. Wu, J. Liu, J. Yu, J. Niu, J. Li, Q. Xu, X. Wu, H. Qian, Appl. Phys. Lett. 103 (2013) 193102. [17] R.R. Nair, P. Blake, A.N. Grigorenko, K.S. Novoselov, T.J. Booth, T. Stauber, N.M.R. Peres, A.K. Geim, Science 320 (2008) 1308. [18] A. Kasry, M.A. Kuroda, G.J. Martyna, G.S. Tulevski, A.A. Bol, ACS Nano 4 (2010) 3839–3844. [19] Y. Lee, S. Bae, H. Jang, S. Jang, S.E. Zhu, S.H. Sim, Y. Il. Song, B.H. Hong, J.H. Ahn, Nano Lett. 10 (2010) 490–493. [20] C. Lee, X. Wei, J.W. Kysar, J. Hone, Science 321 (2008) 385–388. [21] P. Zhang, L. Ma, F. Fan, Z. Zeng, C. Peng, P.E. Loya, Z. Liu, Y. Gong, J. Zhang, X. Zhang, P.M. Ajayan, T. Zhu, J. Lou, Nat. Commun. 5 (2014) 3782. [22] Y. Hwangbo, C.K. Lee, S.M. Kim, J.H. Kim, K.S. Kim, B. Jang, H.J. Lee, S.K. Lee, S.S. Kim, J.H. Ahn, S.M. Lee, Sci. Rep. 4 (2014) 4439. [23] L.G.P. Martins, Y. Song, T. Zeng, M.S. Dresselhaus, J. Kong, P.T. Araujo, P. Natl. Acad. Sci. USA 110 (2013) 17762–17767. [24] Z. Chen, W. Ren, L. Gao, B. Liu, S. Pei, H.M. Cheng, Nat. Mater. 10 (2011) 424–428. [25] X. Li, Y. Zhu, W. Cai, M. Borysiak, B. Han, D. Chen, R.D. Piner, L. Colombo, R.S. Ruoff, Nano Lett. 9 (2009) 4359–4363. [26] M.A. Worsley, P.J. Pauzauskie, T.Y. Olson, J. Biener, J.H. Satcher Jr., T.F. Baumann, J. Am. Chem. Soc. 132 (2010) 14067–14069. [27] S. Ghosh, I. Calizo, D. Teweldebrhan, E.P. Pokatilov, D.L. Nika, A.A. Balandin, W. Bao, F. Miao, C.N. Lau, Appl. Phys. Lett. 92 (2008) 151911. [28] S. Chen, Q. Wu, C. Mishra, J. Kang, H. Zhang, K. Cho, W. Cai, A.A. Balandin, R.S. Ruoff, Nat. Mater. 11 (2012) 203–207. [29] Y.M. Lin, A.V. Garcia, S.J. Han, D.B. Farmer, I. Meric, Y. Sun, Y. Wu, C. Dimitrakopoulos, A. Grill, P. Avouris, K.A. Jenkins, Science 332 (2011) 1294–1297. [30] F. Schwierz, Nat. Nanotechnol. 5 (2010) 487–496. [31] F. Xia, D.B. Farmer, Y. Lin, P. Avouris, Nano Lett. 10 (2010) 715–718. [32] H. Bi, K. Yin, X. Xie, J. Ji, S. Wan, L. Sun, M. Terrones, M.S. Dresselhaus, Sci. Rep. 3 (2013) 2714. [33] S. Borini, R. White, D. Wei, M. Astley, S. Haque, E. Spigone, N. Harris, J. Kivioja, T. Ryhanen, ACS Nano 7 (2013) 11166–11173. [34] C. Hou, H. Wang, Q. Zhang, Y. Li, M. Zhu, Adv. Mater. 26 (2014) 5018–5024. [35] S. Chun, Y. Kim, H. Jung, W. Park, Appl. Phys. Lett. 105 (2014) 041907. [36] R. Raccichini, A. Varzi, S. Passerini, B. Scrosati, Nat. Mater. 14 (2015) 271–279. [37] C. Punckt, M.A. Pope, J. Liu, Y. Lin, I.A. Aksay, Electroanalysis 22 (2010) 2834–2841. [38] Y. Wang, Y. Shao, D.W. Matson, J. Li, Y. Lin, ACS Nano 4 (2010) 1790–1798. [39] J. Hassoun, F. Bonaccorso, M. Agostini, M. Angelucci, M.G. Betti, R. Cingolani, M. Gemmi, C. Mariani, S. Panero, V. Pellegrini, B. Scrosati, Nano Lett. 14 (2014) 4901 (1906). [40] H.F. Xiang, Z.D. Li, K. Xie, J.Z. Jiang, J.J. Chen, P.C. Lian, J.S. Wu, Y. Yu, H.H. Wang, RSC Adv. 2 (2012) 6792–6799. [41] C. Liu, Z. Yu, D. Neff, A. Zhamu, B.Z. Jang, Nano Lett. 10 (2010) 4863–4868. [42] Z.S. Wu, K. Parvez, X. Feng, K. Mullen, Nat. Commun. 4 (2013) 2487. [43] N. Li, S. Oida, G.S. Tulevski, S.-J. Han, J.B. Hannon, D.K. Sadana, T.C. Chen, Nat. Commun. 4 (2013) 2294.

5. Conclusions and outlook Graphene/semiconductor heterostructures have exhibited great potentials for optoelectronic applications. High performance optoelectronic devices such as solar cells, photodetector and light emitting diodes can be demonstrated by properly marring graphene with kinds of semiconductor. And graphene/semiconductor heterostructure has now been successfully applied in a wide variety of fields, such as solar cells, photodetectors, the light emitting diodes, electrochemistry and sensors. Heterojunction is formed between graphene and various n-type semiconductors (Si, GaAs, n-GaN, CdTe) and p-type semiconductors (ZnO, InP, p-GaN). In our past work, we explored many kinds of graphene/semiconductor heterojunction based optoelectronic devices (solar cells, photodetectors, LEDs), such as GaAs, InP, GaN, ZnO and so on. Graphene acts as active layer for carrier separation and transparent electrode for carrier collection. The Fermi level of graphene can be largely tuned through static methods (chemical doping or electric fieldeffect) or dynamic method (photo-induced doping), thus the carrier concentration and barrier height of the devices are improved. The work function of graphene can be regulated by increasing the electrical conductivity of graphene (control the layer number of graphene), respectively. The light absorption and the carrier concentration of the devices can be improved by introducing plasmonic nanostructures (changing the shape, size, material, composition or the location of the plasmonic nanostructures). The quality of graphene/semiconductor interface also affects the performance of optoelectronic devices significantly. Especially, there have been great and fast breakthroughs in the performance of graphene/semiconductor heterostructure solar cells. The PCE of first graphene/Si heterojunction solar cell is only 1.5% in 2010, and recently comes to 15.6%. Graphene/other semiconductors junctions also present remarkable photovoltaic responses. The highest PCE of 18.5% has been achieved on graphene/GaAs heterostructure solar cells. Although the PCE of graphene heterojunction solar cell is still lower than that of currently commercial silicon based photovoltaic cells, it comes close. Through summarizing our and other group's pervious work, we reach an important message that the power conversion efficiency of graphene/semiconductor heterostructure solar cell can exceed 30% in the future (Fig. 35). Fig. 35 shows the graphene solar cell combing gating effect and surface plasmon enhancement effect, where the gate effect can effectively increase the Voc up to 1.2 eV and the surface plasmon effect can effectively increase the Jsc up to 36 mA/cm2. It is emphasized that reducing the resistance of the ohmic contact between the metal and graphene can also effectively increase the FF value of the solar cell. It is highly expected that high performance graphene/ semiconductor heterostructure solar cell with a PCE over 30% can be achieved under sustainable efforts [166]. Considering of the simple process of junction formation by graphene coating and the further improvement of the solar cell performance, the graphene/semiconductor heterojunction solar cells are very promising for practical applications in the near future. Acknowledgements This project is supported by National Natural Science Foundation of China (No. 51202216, 61431014, 61171037, 61171038, 61376118, 25

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

[90] H. Zhao, N.R. Aluru, J. Appl. Phys. 108 (2010) 064321. [91] J.H. Ahn, B.H. Hong, Nat. Nanotechnol. 9 (2014) 737–738. [92] Y. Kim, J. Lee, M.S. Yeom, J.W. Shin, H. Kim, Y. Cui, J.W. Kysar, J. Hone, Y. Jung, S. Jeon, S.M. Han, Nat. Commun. 4 (2013) 2114. [93] S.E. Shina, H.J. Choib, J.H. Shina, D.H. Bae, Carbon 82 (2015) 143–151. [94] E.V. Castro, K.S. Novoselov, S.V. Morozov, N.M.R. Peres, J.M.B. Lopes dos Santos, J. Nilsson, F. Guinea, A.K. Geim, A.H.C. Neto, Phys. Rev. Lett. 99 (2007) 21680. [95] E. McCann, Phys. Rev. B 74 (2006) 161403. [96] C.H. Lui, Z. Li, K.F. Mak, E. Cappelluti, T.F. Heinz, Nat. Phys. 7 (2011) 944–947. [97] S. Zhang, S. Lin, X. Li, X. Liu, H. Wu, W. Xu, P. Wang, Z. Wu, H. Zhong, Z. Xu, Nanoscale 8 (2016) 226–232. [98] H. Zhong, K. Xu, Z. Liu, G. Xu, L. Shi, Y. Fan, J. Wang, G. Ren, H. Yang, J. Appl. Phys. 115 (2014) 013701. [99] S. Tongay, M. Lemaitre, X. Miao, B. Gila, B.R. Appleton, A.F. Hebard, Phys. Rev. X 2 (2012) 011002. [100] X. Li, S. Lin, X. Lin, Z. Xu, P. Wang, S. Zhang, H. Zhong, W. Xu, Z. Wu, W. Fang, Opt. Express 24 (2016) 134–145. [101] Y. Lu, Z. Wu, W. Xu, S. Lin, Nanotechnology 27 (2016) 48LT03. [102] U. Fano, Phys. Rev. 103 (1956) 1202. [103] J.M. Garnett, Philos. Trans. R. Soc. A (1906) 237–288. [104] G. Mie, Ann. Phys. 330 (1908) 377–445. [105] D. Pines, Rev. Mod. Phys. 28 (1956) 184. [106] R. Ritchie, Phys. Rev. 106 (1957) 874. [107] A. Otto, Z. Phys. 216 (1968) 398–410. [108] U. Kreibig, P. Zacharias, Z. Phys. A Hadrons Nucl. 231 (1970) 128–143. [109] K.A. Willets, R.P. Van Duyne, Annu. Rev. Phys. Chem. 58 (2007) 267–297. [110] L.J. Sherry, R. Jin, C.A. Mirkin, G.C. Schatz, R.P. Van Duyne, Nano Lett. 6 (2006) 2060–2065. [111] S.A. Maier, Plasmonics: Fundamentals and Applications, Springer Science & Business Media, 2007. [112] W.R. Erwin, H.F. Zarick, E.M. Talbert, R. Bardhan, Energy Environ. Sci. 9 (2016) 1577–1601. [113] S. Lin, X. Li, S. Zhang, P. Wang, Z. Xu, H. Zhong, Z. Wu, H. Chen, Appl. Phys. Lett. 107 (2015) 191106. [114] Z. Wu, Y. Lu, W. Xu, Y. Zhang, J. Li, S. Lin, Nano Energy 30 (2016) 362–367. [115] Z. Wu, X. Li, H. Zhong, S. Zhang, P. Wang, T.H. Kim, S.S. Kwak, C. Liu, H. Chen, S.W. Kim, S. Lin, Opt. Express 23 (2015) 18864–18871. [116] Th Seyller, K.V. Emtsev, F. Speck, K.Y. Gao, L. Ley, Appl. Phys. Lett. 88 (2006) 242103. [117] S. Shivaraman, L.H. Herman, F. Rana, J. Park, M.G. Spencer, Appl. Phys. Lett. 100 (2012) 183112. [118] S. Sonde, F. Giannazzo, V. Raineri, R. Yakimova, J.R. Huntzinger, A. Tiberj, J. Camassel, Phys. Rev. B 80 (2009) 241406 (R). [119] S. Tongay, M. Lemaitre, T. Schumann, K. Berke, B.R. Appleton, B. Gila, A.F. Hebard, Appl. Phys. Lett. 99 (2011) 102102. [120] Y.F. Li, W. Yang, Z.Q. Tu, Z.C. Liu, F. Yang, L.Q. Zhang, R. Hatakeyama, Appl. Phys. Lett. 104 (2014) 043903. [121] K. Ihm, J.T. Lim, K.J. Lee, J.W. Kwon, T.H. Kang, S. Chung, S. Bae, J.H. Kim, B.H. Hong, G.Y. Yeom, Appl. Phys. Lett. 97 (2010) 032113. [122] H. Hibino, H. Kageshima, M. Kotsugi, F. Maeda, F.Z. Guo, Y. Watanabe, Phys. Rev. B 79 (2009) 125437. [123] Y. Lin, X. Li, D. Xie, T. Feng, Y. Chen, R. Song, H. Tian, T. Ren, M. Zhong, K. Wang, H. Zhu, Energ. Environ. Sci. 6 (2013) 108–115. [124] T. Cui, R. Lv, Z.H. Huang, S. Chen, Z. Zhang, X. Gan, Y. Jia, X. Li, K. Wang, D. Wu, F. Kang, J. Mater. Chem. A 1 (2013) 5736–5740. [125] X. Li, D. Xie, H. Park, M. Zhu, T.H. Zeng, K. Wang, J. Wei, D. Wu, J. Kong, H. Zhu, Nanoscale 5 (2013) 1945–1948. [126] X. Zhang, C. Xie, J. Jie, X. Zhang, Y. Wu, W. Zhang, J. Mater. Chem. A 1 (2013) 6593–6601. [127] C.C. Chen, C.C. Chang, Z. Li, A.F.J. Levi, S.B. Cronin, Appl. Phys. Lett. 101 (2012) 223113. [128] X. Yu, L. Yang, Q. Lv, M. Xu, H. Chen, D. Yang, Nanoscale 7 (2015) 7072–7077. [129] W. Regan, S. Byrnes, W. Gannett, O. Ergen, O.V.- Mena, F. Wang, A. Zettl, Nano Lett. 12 (2012) 4300–4304. [130] O.V. Mena, J.P. Bosco, O. Ergen, H.I. Rasool, A. Fathalizadeh, M. Tosun, M. Crommie, A. Javey, H.A. Atwater, A. Zettl, Nano Lett. 14 (2014) 4280–4285. [131] Hae Kyung Jeong, Ki-jeong Kim, Soo Min Kim, Y.H. Lee, Chem. Phys. Lett. 498 (2010) 168–171. [132] R.R. Nair, W.C. Ren, R. Jalil, I. Riaz, V.G. Kravets, L. Britnell, P. Blake, F. Schedin, A.S. Mayorov, S.J. Yuan, M.I. Katsnelson, H.M. Cheng, W. Strupinski, L.G. Bulusheva, A.V. Okotrub, I.V. Grigorieva, A.N. Grigorenko, K.S. Novoselov, A.K. Geim, Small 6 (2010) 2877–2884. [133] D.C. Elias, R.R. Nair, T.M.G. Mohiuddin, S.V. Morozov, P. Blake, M.P. Halsall, A.C. Ferrari, D.W. Boukhvalov, M.I. Katsnelson, A.K. Geim, K.S. Novoselov, Science 323 (2009) 610–613. [134] S.S. Lin, J. Phys. Chem. C 116 (2012) 3951–3955. [135] S.S. Lin, S.J. Zhang, X.Q. Li, W.L. Xu, X.D. Pi, X.Y. Liu, F.C. Wang, H.A. Wu, H.S. Chen, J. Phys. Chem. C 119 (2015) 19772–19779. [136] Z. Shi, Z. Zhang, A. Kutana, B.I. Yakobson, ACS Nano 9 (2015) 9802–9809. [137] L.J. Zhou, Y.F. Zhang, L.M. Wu, Nano Lett. 13 (2013) 5431–5436. [138] Y.H. Jang, Y.J. Jang, S. Kim, L.N. Quan, K. Chung, D.H. Kim, Chem. Rev. (2016). [139] O. Nicoletti, F. de La, Peña, R.K. Leary, D.J. Holland, C. Ducati, P.A. Midgley, Nature 502 (2013) 80–84. [140] J.F. Li, Y.F. Huang, Y. Ding, Z.L. Yang, S.B. Li, X.S. Zhou, F.R. Fan, W. Zhang, Z.Y. Zhou, B. Ren, Nature 464 (2010) 392–395. [141] J.F. Li, Y.J. Zhang, S.Y. Ding, R. Panneerselvam, Z.Q. Tian, Chem. Rev. 117 (2017)

[44] J. Wu, M. Agrawal, H.A. Becerril, Z. Bao, Z. Liu, Y. Chen, P. Peumans, ACS Nano 4 (2010) 43–48. [45] D.H. Youn, Y.J. Yu, H.K. Choi, S.H. Kim, S.Y. Choi, C.G. Choi, Nanotechnology 24 (2013) 075202. [46] J. Wang, M. Liang, Y. Fang, T. Qiu, J. Zhang, L. Zhi, Adv. Mater. 24 (2012) 2874–2878. [47] Z. Liu, J. Li, F. Yan, Adv. Mater. 25 (2013) 4296–4301. [48] H. Kim, S.H. Bae, T.H. Han, K.G. Lim, J.H. Ahn, T.W. Lee, Nanotechnology 25 (2014) 014012. [49] X. Wang, L. Zhi, K. Mullen, Nano Lett. 8 (2008) 323–327. [50] J. Wu, H.A. Becerril, Z. Bao, Z. Liu, Y. Chen, P. Peumans, Appl. Phys. Lett. 92 (2008) 263302. [51] H. Park, S. Chang, M. Smith, S. Gradecak, J. Kong, Sci. Rep. 3 (2013) 1581. [52] J. Liu, Y. Xue, L. Dai, J. Phys. Chem. Lett. 3 (2012) 1928–1933. [53] H. PilKim, A.R.M. Yusoff, J. Jang, Sol. Energy Mater. Sol. Cells 110 (2013) 87–93. [54] X. Xu, D. Huang, K. Cao, M. Wang, S.M. Zakeeruddin, M. Gratzel, Sci. Rep. 3 (2013) 1489. [55] D.W. Zhang, X.D. Li, H.B. Li, S. Chen, Z. Sun, X.J. Yin, S.M. Huang, Carbon 49 (2011) 5382–5388. [56] S. Sun, L. Gao, Y. Liu, Appl. Phys. Lett. 96 (2010) 083113. [57] J.T.W. Wang, J.M. Ball, E.M. Barea, A. Abate, J.A.A. Webber, J. Huang, M. Saliba, I.M.- Sero, J. Bisquert, H.J. Snaith, R.J. Nicholas, Nano Lett. 14 (2014) 724–730. [58] K. Yan, Z. Wei, J. Li, H. Chen, Y. Yi, X. Zheng, X. Long, Z. Wang, J. Wang, J. Xu, S. Yang, Small 11 (2015) 2269–2274. [59] C.C. Chen, M. Aykol, C.C. Chang, A.F.J. Levi, S.B. Cronin, Nano Lett. 11 (2011) 1863–1867. [60] D. Sinha, J.U. Lee, Nano Lett. 14 (2014) 4660–4664. [61] S. Tongay, T. Schumann, A.F. Hebard, Appl. Phys. Lett. 95 (2009) 222103. [62] X. Li, H. Zhu, K. Wang, A. Cao, J. Wei, C. Li, Y. Jia, Z. Li, X. Li, D. Wu, Adv. Mater. 22 (2010) 2743–2748. [63] X. Miao, S. Tongay, M.K. Petterson, K. Berke, A.G. Rinzler, B.R. Appleton, A.F. Hebard, Nano Lett. 12 (2012) 2745–2750. [64] E. Shi, H. Li, L. Yang, L. Zhang, Z. Li, P. Li, Y. Shang, S. Wu, X. Li, J. Wei, K. Wang, H. Zhu, D. Wu, Y. Fang, A. Cao, Nano Lett. 13 (2013) 1776–1781. [65] W. Jie, F. Zheng, J. Hao, Appl. Phys. Lett. 103 (2013) 233111. [66] P. Wang, X. Li, Z. Xu, Z. Wu, S. Zhang, W. Xu, H. Zhong, H. Chen, E. Li, J. Luo, Q. Yu, S. Lin, Nano Energy 13 (2015) 509–517. [67] G. Fan, H. Zhu, K. Wang, J. Wei, X. Li, Q. Shu, N. Guo, D. Wu, ACS Appl. Mater. Interfaces 3 (2011) 721–725. [68] L. Zhang, L. Fan, Z. Li, E. Shi, X. Li, H. Li, C. Ji, Y. Jia, J. Wei, K. Wang, H. Zhu, D. Wu, A. Cao, Nano Res. 4 (2011) 891–900. [69] Y. Ye, Y. Dai, L. Dai, Z. Shi, N. Liu, F. Wang, L. Fu, R. Peng, X. Wen, Z. Chen, Z. Liu, G. Qin, ACS Appl. Mater. Interfaces 2 (2010) 3406–3410. [70] H. Park, S. Chang, J. Jean, J.J. Cheng, P.T. Araujo, M. Wang, M.G. Bawendi, M.S. Dresselhaus, V. Bulovic, J. Kong, S. Gradecak, Nano Lett. 13 (2013) 233–239. [71] Jun Yan, Yuanbo Zhang, Philip Kim, A. Pinczuk, Phys. Rev. Lett. 98 (2007) 166802. [72] Y.-J. Yu, Y. Zhao, S. Ryu, L.E. Brus, K.S. Kim, P. Kim, Nano Lett. 9 (2009) 3430–3434. [73] Q. Hao, S.M. Morton, B. Wang, Y. Zhao, L. Jensen, T.J. Huang, Appl. Phys. Lett. 102 (2013) 011102. [74] H.-Y. Kim, K. Lee, N. McEvoy, C. Yim, G.S. Duesberg, Nano Lett. 13 (2013) 2182–2188. [75] S.J. Jeon, S.Y. Kwak, D.B. Yim, J.M. Ju, J.H. Kim, J. Am. Chem. Soc. 136 (2014) 10842–10845. [76] N.D. Mermint, Phys. Rev. 176 (1968) 250–254. [77] X. Li, W. Chen, S. Zhang, Z. Wu, P. Wang, Z. Xu, H. Chen, W. Yin, H. Zhong, S. Lin, Nano Energy 16 (2015) 310–319. [78] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, M.I. Katsnelson, I.V. Grigorieva, S.V. Dubonos, A.A. Firsov, Nature 438 (2005) 197–200. [79] S. Tongay, K. Berke, M. Lemaitre, Z. Nasrollahi, D.B. Tanner, A.F. Hebard, B.R. Appleton, Nanotechnology 22 (2011) 425701. [80] M.F. Khan, M.Z. Iqbal, M.W. Iqbal, J. Eom, Sci. Technol. Adv. Mater. 15 (2014) 055004. [81] Y. Shi, K.K. Kim, A. Reina, M. Hofmann, L.J. Li, J. Kong, ACS Nano 4 (2010) 2689–2694. [82] Y. Wu, W. Jiang, Y. Ren, W. Cai, W.H. Lee, H. Li, R.D. Piner, C.W. Pope, Y. Hao, H. Ji, J. Kang, R.S. Ruoff, Small 8 (2012) 3129–3136. [83] A. Nourbakhsh, M. Cantoro, A. Klekachev, F. Clemente, B. Soree, M.H. van der Veen, T. Vosch, A. Stesmans, B. Sels, S.D. Gendt, J. Phys. Chem. C 114 (2012) 6894–6900. [84] A. Das, S. Pisana, B. Chakraborty, S. Piscanec, S.K. Saha, U.V. Waghmare, K.S. Novoselov, H.R. Krishnamurthy, A.K. Geim, A.C. Ferrari, A.K. Sood, Nat. Nanotechnol. 3 (2008) 210–215. [85] C.F. Chen, C.H. Park, B.W. Boudouris, J. Horng, B. Geng, C. Girit, A. Zettl, M.F. Crommie, R.A. Segalman, S.G. Louie, F. Wang, Nature 471 (2011) 617–620. [86] S. Bae, H. Kim, Y. Lee, X. Xu, J.S. Park, Y. Zheng, J. Balakrishnan, T. Lei, H. Ri Kim, Y.I. Song, Y.J. Kim, K.S. Kim, B. Ozyilmaz, J.H. Ahn, B.H. Hong, S. Iijima, Nat. Nanotechnol. 5 (2010) 574–578. [87] Y. Hwangbo, C.K. Lee, S.M. Kim, J.H. Kim, K.S. Kim, B. Jang, H.J. Lee, S.K. Lee, S.S. Kim, J.H. Ahn, S.M. Lee, Sci. Rep. 4 (2014) 4439. [88] G.H. Lee, R.C. Cooper, S.J. An, S. Lee, A. Zande, N. Petrone, A.G. Hammerberg, C. Lee, B. Crawford, W. Oliver, J.W. Kysar, J. Hone, Science 340 (2013) 1073–1076. [89] A. Zandiatashbar, G.H. Lee, S.J. An, S. Lee, N. Mathew, M. Terrones, T. Hayashi, C.R. Picu, J. Hone, N. Koratkar, Nat. Commun. 5 (2014) 3186.

26

Nano Energy 40 (2017) xxx–xxx

S. Lin et al.

Yanghua Lu received his B.S. degree in Information and Communication Engineering from Zhejiang University, China in 2015. He is currently a master student in the Department of Electronic and Communication Engineering, Zhejiang University. His research focuses on application of 2D materials based electronic and optoelectronic devices, such as solar cells and photodetectors.

5002–5069. [142] Z. Zeng, S. Mizukami, K. Fujita, K. Kikuchi, Chem. Sci. 6 (2015) 4934–4939. [143] S. Linic, P. Christopher, D.B. Ingram, Nat. Mater. 10 (2011) 911–921. [144] H. Yang, Z.H. Wang, Y.Y. Zheng, L. q. He, C. Zhan, X. Lu, Z.Q. Tian, P.P. Fang, Y. Tong, J. Am. Chem. Soc. (2016). [145] B.Y. Zheng, Y. Wang, P. Nordlander, N.J. Halas, Adv. Mater. 26 (2014) 6318–6323. [146] L.Q. Xie, L. Chen, Z.A. Nan, H.X. Lin, T. Wang, D.P. Zhan, J.W. Yan, B.W. Mao, Z.Q. Tian, J. Am. Chem. Soc. 139 (2017) 3320–3323. [147] L. Chen, J.R. Wang, L.Q. Xie, C. Zhan, Z. Qiu, J.Z. Zhou, J.W. Yan, B.W. Mao, Z.Q. Tian, Electrochem. Commun. 68 (2016) 40–44. [148] S.Y. Ding, J. Yi, J.F. Li, B. Ren, D.Y. Wu, R. Panneerselvam, Z.Q. Tian, Nat. Rev. Mater. 1 (2016) 16021. [149] J.D. Winans, C. Hungerford, K. Shome, L.J. Rothberg, P.M. Fauchet, Opt. Express 23 (2015) A92–A105. [150] X. Chen, B. Jia, Y. Zhang, M. Gu, Light.: Sci. Appl. 2 (2013) e92. [151] S. Liu, R. Jiang, P. You, X. Zhu, J. Wang, F. Yan, Energy Environ. Sci. 9 (2016) 898–905. [152] H.F. Zarick, O. Hurd, J.A. Webb, C. Hungerford, W.R. Erwin, R. Bardhan, ACS Photonics 1 (2014) 806–811. [153] K. Awazu, M. Fujimaki, C. Rockstuhl, J. Tominaga, H. Murakami, Y. Ohki, N. Yoshida, T. Watanabe, J. Am. Chem. Soc. 130 (2008) 1676–1680. [154] Z. Liu, W. Hou, P. Pavaskar, M. Aykol, S.B. Cronin, Nano Lett. 11 (2011) 1111–1116. [155] A. Manjavacas, J.G. Liu, V. Kulkarni, P. Nordlander, ACS Nano 8 (2014) 7630–7638. [156] P. Reineck, D. Brick, P. Mulvaney, U. Bach, J. Phys. Chem. Lett. 7 (2016) 4137–4141. [157] A.O. Govorov, H. Zhang, Y.K. Gun’ko, J. Phys. Chem. C 117 (2013) 16616–16631. [158] R. Sundararaman, P. Narang, A.S. Jermyn, W.A. Goddard III, H.A. Atwater, Nat. Commun. (2014) 5. [159] K. Wu, J. Chen, J. McBride, T. Lian, Science 349 (2015) 632–635. [160] K. Qian, B.C. Sweeny, A.C. Johnston-Peck, W. Niu, J.O. Graham, J.S. DuChene, J. Qiu, Y.C. Wang, M.H. Engelhard, D. Su, J. Am. Chem. Soc. 136 (2014) 9842–9845. [161] S. Bai, X. Li, Q. Kong, R. Long, C. Wang, J. Jiang, Y. Xiong, Adv. Mater. 27 (2015) 3444–3452. [162] S.K. Cushing, J. Li, F. Meng, T.R. Senty, S. Suri, M. Zhi, M. Li, A.D. Bristow, N. Wu, J. Am. Chem. Soc. 134 (2012) 15033–15041. [163] H.F. Zarick, W.R. Erwin, A. Boulesbaa, O.K. Hurd, J.A. Webb, A.A. Puretzky, D.B. Geohegan, R. Bardhan, ACS Photonics 3 (2016) 385–394. [164] J. Li, S.K. Cushing, F. Meng, T.R. Senty, A.D. Bristow, N. Wu, Nat. Photonics 9 (2015) 601–607. [165] S.K. Cushing, J. Li, J. Bright, B.T. Yost, P. Zheng, A.D. Bristow, N. Wu, J. Phys. Chem. C 119 (2015) 16239–16244. [166] X. Li, S. Zhang, P. Wang, H. Zhong, Z. Wu, H. Chen, C. Liu, S. Lin, arXiv preprint arXiv:1409.3500, 2014.

Sirui Feng received his B.S. degree in Information and Communication Engineering from Zhejiang University, China in 2016. He is currently a master student in the Department of Electronic and Communication Engineering, Zhejiang University. His research focuses on application of 2D materials based electronic and optoelectronic devices, such as the solar cells and light emitting diodes.

Dr. Juan Xu received a Ph.D. in Biochemical Engineering from East China University of Science and Technology in 2015. She is currently a Postdoctoral Fellow in the group of Prof. Jian-Feng Li. She has experience in material characterization and photocatalytic hydrogen production from water splitting. Her current work focuses on surface plasmon enhanced phototcatalysis.

Prof. Jianfeng Li is a Full Professor of Chemistry at Xiamen University. He received a B.Sc. in chemistry from Zhejiang University, and a Ph.D. in chemistry from Xiamen University. Dr. Li is the principal inventor of shell-isolated nanoparticle-enhanced Raman spectroscopy (SHINERS) technique. His research interests include plasmonic materials, plasmon-enhanced Raman / fluorescence spectroscopy, electro-(photon-)catalysis.

Prof. Shisheng Lin leads the work and a pioneering group in the department of information science and electronic engineering in Zhejiang University. He implements the novel physics carried by novel materials into the traditional devices and create high performance optoelectronic and electronic devices. He has achieved high performance twodimensional materials based solar cells, photodetectors and light emitting diodes. Professor Lin has demonstrated the possibility of fabrication of two dimensional SiC, SiC2, and silicon doped graphene, which provides a solution for band gap engineering of graphene. Professor Lin leads the systematic research on 2D materials based heterostructure solar cells and has achieved 18.5% efficient graphene/ semiconductor heterostructure solar cells.

27