1. Introduction

Graphene, a single layer of carbon atoms in a hexagonal lattice, has been the focus of much research in recent years due to its exceptional physical properties, such as high carrier mobility [1], exceptional optical transparency [2], and ultrahigh mechanical strength [3]. Moreover, the photonic properties of graphene are remarkable, e.g., broad wavelength absorption from the visible to the infrared [4] and very high-energy photon excitation [5]. The strength of the interband transition in graphene can be controlled by shifting the electrical Fermi level [6, 7]. These features imply that graphene is a great candidate plane material in optoelectronics and optoelectronic devices. There are many diverse optoelectronic devices, including solar cells [8], optical modulator [9], photodetector [10, 11]. In addition, ultrahigh speed photodetectors with radio frequencies up to terahertz regime have been reported [12, 13]. However, it has been demonstrated that the external photo responsivity of graphene photodetectors is quite low [11], which limits its further development and practical application.

To overcome the photo responsivity limitation of graphene and to achieve practical application for the graphene photodetector, various strategies to improve the photo responsivity of graphene photodetector have been proposed. Those strategies can be mainly divided into two categories. One category is to directly enhance the external quantum efficiency of graphene. A method that integrates metallic plasmonic nanostructures into a graphene photodetector can greatly enhance the photocurrent and external quantum efficiency with variable resonance frequencies [14, 15]. However, the photodetection wavelength range becomes narrower with this method. In addition, it has been demonstrated that a hybrid graphene photodetector with a thin film of colloidal quantum dots on a monolayer or bilayer graphene exhibits ultrahigh photodetection amplification and high quantum efficiency [16], but such improvement of photo responsivity is achieved at the cost of response speed. The other category is to improve the light absorption ratio of graphene to indirectly enhance the quantum efficiency of graphene. For example, a graphene photodetector integrated into a resonant Fabry-Perot microcavity has been stated and the light absorption ratio of graphene is largely enhanced with the increase of optical field inside the microcavity [17]. However, this method is limited by the fixed wavelength because such microcavity is designed for a specific wavelength. In another study, by embedding a perfect metamaterial absorber into the graphene photodetector, the light absorption ratio increases by up to 40% [18]. However, with this approach, the working light wavelength is limited to the infrared range.

In this study, we propose a new approach to increasing the photo responsivity of a graphene photodetector by integrating PSNs onto the graphene surface for the reason that PSNs can effectively enhance the light field. This approach is not only to increase the photo responsivity of the graphene photodetector under incident light over a broad wavelength range, but also has potential to realize multicolor photodetection. Some experiments and simulations are performed to demonstrate the effectiveness of the approach.

2. Results

2.1 Sample preparation and experimental setup

The process to integrate PSNs onto the surface of the graphene photodetector and the experimental method to measure the photocurrent of the graphene photodetector under vertical incident light are illustrated in Fig. 1. In general, an Au electrode was fabricated on a Si/SiO₂ substrate using a pair of Au blocks of 400 nm thickness and 10 um width by photo etching technology, with a 2 um interval between the blocks, as shown in Fig. 1(a). Large single-crystal graphene films were grown on single-crystal Pt substrates using ambient pressure chemical vapor deposition (AP-CVD). Then a graphene film was chosen and transferred onto the Au electrode by the bubbling transfer process [19], as shown in Fig. 1(b). The graphene film was approximately tens of microns in width. For this study, PSNs solution with original concentration of 2.5% w/v was diluted in deionized water solution for investigating the effect of PSNs on the light responsivity of graphene photodetector at different PSNs concentrations. The concentrations of diluted PSNs solution were set to be 0.025%, 0.05%, and 0.075% w/v, respectively. Then, the PSNs were deployed onto the graphene surface by dropping the PSNs solution at a certain concentration using a glass micropipette with an inner diameter of 1 um at the tip, fabricated by a micropipette puller (Sutter Instrument Co., Novato, CA, USA), as shown in Fig. 1(c). Finally, the light-response properties of the graphene photodetector were investigated under incident light of wavelength of 470 nm (OGK4, Thorlabs, USA) using a Semiconductor Parameter Analyzer (4155C, Agilent Technologies, Santa Clara, CA, USA), by which the light is directed to one contact of the graphene photodetector, and the maximum output power of the light is 10.8 mW, as shown in Fig. 1(d).

 

Fig. 1 Schematic illustration of integration process of PSNs onto the surface of a graphene photodetector and measurement of the photocurrent. (a) Fabrication of Au electrode on a Si/SiO2 substrate using a pair of Au blocks of 400 nm thickness and 10 um width with a 2 um interval between the both blocks. (b) Transfer graphene onto the Au electrode using the bobbling transfer process. (c) Deploy PSNs onto the graphene surface using a glass micropipette with an inner diameter of about 1 um at the tip. (d) Measurement of the photo responsivity of the graphene photodetector under the incident light of wavelength 470 nm.

Download Full Size | PDF

2.2 Experiments

A graphene photodetector with a monolayer graphene fabricated through AP-CVD method was used in the experiments and it was characterized by Raman spectroscope and optical microscope, as shown in Fig. 2(a), demonstrating the good quality of the graphene layer. Then, the PSNs (BaseLine ChromTech Research Centre, China) with diameter of 500 nm were deployed onto the graphene surface, as shown in Fig. 2(b), which was imaged by SEM. To characterize the effective enhancement of light responsivity of the graphene photodetector by PSNs under the incident light, the photocurrents were measured with and without PSNs deployed onto the graphene surface as well as with PSNs directly deployed onto the surface of Au electrode that has no graphene. As shown in Fig. 2(c), with the incident light of wavelength 470 nm, there was no photocurrent (blue curve) with PSNs on the surface of Au electrode if the graphene was not integrated onto the photodetector, inferring that the photocurrent was essentially generated by the graphene but not by the PSNs. It is also shown that, with PSNs onto the graphene surface, the photocurrent (black curve) was distinctly increased and higher than that (red curve) without PSNs onto the graphene, inferring that the photo responsivity of the graphene photodetector can be enhanced by PSNs.

 

Fig. 2 (a) The optical image (illustration) and Raman spectroscopy of the graphene photodetector. The yellow is Au electrode and the graphene cover the entire substrate. (b) SEM image of PSNs. (c) The I-V curves with PSNs (black square), without PSNs (red circle) deployed onto the graphene surface and with PSNs (blue triangle) deployed onto the substrate of electrode without graphene under the incident light of wavelength 470 nm. (d) Photocurrents of the graphene photodetector without (blue curve) and with (red curve) PSNs deployed onto the graphene surface under the incident light of wavelength 470 nm at the frequencies of 20 Hz (upper subfigure) and 10 Hz (lower subfigure) to switch the light on and off, respectively. (e) Photocurrents of the graphene photodetector at different PSNs concentrations under incident light of wavelength 470 nm with switching frequency of 10 Hz.

Download Full Size | PDF

To further investigate the enhancement effect of photo responsivity of the graphene photodetector, the photocurrent of the graphene photodetector was measured with various switching frequencies of incident light on and off at different concentrations of PSNs solution, as shown in Fig. 2(d) and 2(e). As shown in Fig. 2(d), compared with the photocurrent without PSNs on the graphene, the photocurrent of the graphene photodetector distinctly increases with PSNs deployed onto the graphene under the incident light at different light-switching frequencies. On average, the photocurrent of the graphene photodetector was enhanced by 23% with PSN concentration of 0.025%. It is obvious that the light switching frequency has no significant effect on the enhancement of photo responsivity of graphene photodetector. However, the photocurrent of the graphene photodetector is closely related to the PSNs concentrations, and the photocurrent increases as the PSNs concentration increases, as shown in Fig. 2(e).

In addition, not only the concentration but also the size of PSNs significantly affects the photocurrent amplification. To investigate such effect, we repeated the experiment to measure the photocurrent amplification by choosing different diameters of PSNs, 0.1 um, 0.2 um, 0.5 um and 1.0 um, and varying the concentrations of PSN solution, 0.025%, 0.050%, and 0.075% respectively. As shown in Fig. 3(a), the photocurrent is almost linear with the concentration of PSNs at the same diameter, and varies with the diameters of PSNs at the same concentration. However, the photocurrent amplification is not linearly correlated with the PSN diameter and, under the incident light of 470 nm, the maximum photocurrent amplification for a particular concentration of PSNs solution is always achieved for the PSNs with the diameter of 0.5 um. Additionally, the maximum enhancement ratio is about 71% with the PSNs of 0.5 um diameter at 0.075% solution concentration. It is noted that the maximum photocurrent amplification is achieved when the light wavelength in this study is 470 nm, which is approximate to the diameter of PSNs. Therefore we infer from these results that the maximum photocurrent amplification occurs when the PSNs diameter is equal to the wavelength of the incident light. To validate this hypothesis, the experiments were performed with different wavelength of incident light for each of the PSNs at the concentration of 0.025% of PSNs. In this study,the wavelengths for the available incident lights are 380 nm, 470 nm and 850 nm respectively. As shown in Fig. 3(b), for a fixed light wavelength, the maximum enhancement of photo responsivity was always achieved with the PSNs that have a diameter closest to the wavelength of light incident.

 

Fig. 3 (a) Enhancement ratio of photocurrents of the graphene photodetector with different diameters and concentrations of PSNs. (b) Enhancement ratio of photocurrents of the graphene photodetector under incident light with different wavelengths.

Download Full Size | PDF

2.3 Discussion

To further verify the above hypothesis, the amplification of the light intensity by PSNs of different diameters with various light wavelength were simulated based on near-field optics by using the Finite Difference Time Domain (FDTD) software. As shown in Fig. 4(a), in the simulation, a single layer of PSNs is evenly deployed onto the graphene surface and the incident light is vertically applied onto the PSNs layer and then transmitted through the PSNs. In addition, the requisite parameters of PSNs for the simulation were set, and the Permittivity and Dielectric constants of polystyrene are 23 × 10⁻¹² C²N⁻¹m⁻² and 2.6, respectively [20]. The total light intensity $P_{NG}$ impacting on the surface of the graphene photodetector through PSNs is equal to the sum of the light intensity at all individual points on the entire surface of the graphene, i.e., $P_{NG}={\displaystyle \int p_{Nk}dk}$, where $p_{Nk}$ is the light intensity at any point on the graphene surface through the PSNs layer. Therefore, the PSNs amplification in light field is calculated by the equation, $G_A=P_{NG}/{\displaystyle \int p_{i}di}$, where $p_i$ is the light intensity at any point on the graphene surface without PSNs. It is shown in Fig. 4(b) that the maximum amplification of light intensity occurs when the diameter of the PSNs is 450 nm for the incident light with a wavelength of 470 nm. It is obvious in the simulation that, among the diameter set of PSNs, 450 nm is mostly approximate to the wavelength of 470 nm. Therefore, the result is consistent with the assumption and agrees with the experimental results as shown in Fig. 3. Furthermore, the set of wavelengths of incident light, 300 nm, 400 nm, 470 nm, 600 nm, 700 nm, and 800 nm, were selected for the simulation and it is clear that the maximum amplification of the light intensity by PSNs always occurs when the wavelength of incident light is equal to the diameter of PSNs, as shown in Fig. 4(c), which validates the hypothesis that the maximum photocurrent amplification occurs when the PSNs diameter is equal to the wavelength of the incident light.

 

Fig. 4 Simulation of amplification of light intensity by PSNs under assumption that the PSNs were evenly deployed into a layer of $2.5\text{um}\times 2.5\text{um}$. (a) Profile of light intensity within a layer of 500 nm PSNs under 470 nm incident light along vertical direction. (b) Amplification of light intensity by PSNs of different diameters under 470 nm incident light. (c) Amplification of the light intensity of PSNs with different diameters under incident light of different wavelengths.

Download Full Size | PDF

Next, a simulation based on the time-harmonic wave equation for electric field [21] using Comsol Multiphysics 4.0 software was performed to demonstrate the effect of incident light intensity on the photo responsivity of the graphene photodetector. In the simulation, the graphene photodetector was 2 um long, which was equal to the width between the two Au electrodes, and 0.5 nm thick, which was approximately equal to the thickness of a single layer of graphene. The wavelength of incident light was 470 nm. The electric field on the graphene surface was simulated by continuously varying the incident light intensities. It is clearly confirmed that the photo responsivity of the graphene photodetector is approximately proportional to the incident light intensity, as shown in Fig. 5.

 

Fig. 5 Simulation of electric field on a graphene layer under 470 nm incident light by continuously varying the light power.

Download Full Size | PDF

In conclusion, by combining the results of two simulations, on one hand, it is proved that the PSN layer can amplify the light intensity illuminated onto the surface of graphene and indirectly increase the light absorbed by the graphene. The underlying reason for the light enhancement is that the micro-/nano-scale PSNs can generate the ‘photonic nanojet’ with high intensity and super-resolution foci at the shadow-side [22, 23]. The size parameter of such super-resolution foci, defined as $q=2\pi a/\lambda$ with $a$ denoting the radius of the nanoparticle and $\lambda$ the wavelength of the incident light, determines where the photonic nanojet is located and explains why the maximum photocurrent amplification occurs when the PSNs diameter is equal to the wavelength of the incident light. If $q$ is less than $\pi$, then the enhanced light field by the photonic nanojet is located within the nanoparticle, and otherwise, the enhanced light field is located outside of the nanoparticle [24]. In the case that $q=\pi$, the enhanced light field is located on the shadow-side surface, as shown in Fig. 4(a). In this study, PSNs were closely deployed onto the graphene surface and the maximum light enhancement was achieved with the PSNs that have diameters close to the wavelength of incident lights, in which the enhanced light field was located onto the graphene surface [24]. Moreover, the interference enhancement of light field between the neighboring PSNs also contributes to the light field enhancement of the PSNs. On the other hand, the enhancement of light field in turn strengthens the electric field within the graphene and results in photocurrent enhancement because, upon the light illumination, electron-hole pairs are separated and then transported to opposite directions due to the strong electric field near the metal-graphene contacts [25]. In addition, since the PSN layer does not generate photocurrent (Fig. 2(c)) under the incident light, we can conclude that the photo responsivity enhancement of graphene is due to the light intensity amplification by the PSN layer.

3. Conclusion

In summary, in this study we have presented a new approach to increasing the photo responsivity of the graphene photodetector by integrating the PSNs onto the graphene surface. The experimental results imply that the enhancement of photo responsivity is almost linearly correlated with the concentration of PSNs solution and indicate that the maximum enhancement ratio of photo responsivity occurs when the diameter of PSNs is equal to the wavelength of incident light. The assumption is validated by simulating the light intensity amplification by PSNs and the effect of light intensity on the photo responsivity of graphene layer. This approach to enhancing the photo responsivity of graphene photodetector is suitable for incident light of any wavelength because PSNs of a particular diameter can be chosen to match the incident light for the maximum enhancement ratio of photo responsivity. On the other hand, the wavelength of a given light can be determined by the size of PSNs that achieve the maximum enhancement of photo responsivity of graphene photodetector. Therefore, this approach has a potential application to multicolor photodetection by designing a structure of PSNs with various diameters for their selective enhancement of photo responsivity with respect to the light wavelength.