In recent years, different kinds of two-dimensional materials^[1–3] have attracted much attention for their applications in optics and optoelectronics. Among them, MoS₂ becomes one of the most important materials due to its unique optical properties, and has found applications in many areas, such as direct bandgap photon source,^[4] strong spin-orbit coupling,^[5,6] nonlinear optical imaging,^[7] spin Hall effect,^[8,9] and two-dimensional heterostructures.^[10–13] Meanwhile, the development of chemical vapor deposition (CVD) method^[14–16] makes the synthesis of MoS₂ more simple and efficient. For monolayer MoS₂, its photoluminescence (PL) wavelength is in visible range, and thus it is widely used in PL sources and optoelectronic devices. However, due to the inherent monolayer thickness (about 1 nm), monolayer MoS₂ suffers low PL efficiency and low quantum yield.^[17]

To compensate for this, different kinds of approaches to enhancing the emission of monolayer MoS₂ are proposed, including chemical doping,^[18] polymeric nano-spacing,^[19] defect engineering,^[20] and using surface plasmon polaritons (SPPs).^[21–23] Among these methods, SPPs which have been widely utilized for strong light-matter interaction applications such as photoluminescence enhancement^[24–31] and surface enhanced Raman scattering,^[32,33] could induce hot electrons and cause the phase transition of MoS₂,^[34–36] resulting in the enhanced PL in a metal-MoS₂ coupled system.^[37–40] For example, gap plasmons have been used in different areas, such as enhancing local chemical reactions,^[41] dielectric gratings,^[42,43] plasmonic nanocavities,^[44–46] and PL of single layer WSe₂.^[31]

One way to generate gap plasmons is to deposit metal particles on metal substrates or films, forming strongly localized plasmonic fields in the gap between the metal substrate and metal particles. Additionally, energetic electrons and holes could also be generated in the junction region, which may have applications in plasmon-mediated photocatalysis.^[47–50]

In this work, we use MoS₂ sandwiched between Au film and gold nanoparticle (AuNP), i.e., Au–MoS₂–AuNP structure, to study the PL enhancement effect of gap plasmons on monolayer MoS₂. For comparison, the PL of MoS₂ at the position without metal structure is also characterized. The experimental results show that the gap plasmons enhance the PL of monolayer MoS₂ by 4.43 times compared with the bare MoS₂ sample.

The experimental setup is shown in Fig. 1. A continuous-wave laser at 532 nm passes through a polarizer and shines through an objective lens (OLYMPUS, NA = 0.9, 100×) onto our sample, which is mounted on a three-axis piezo translation stage. The PL of the MoS₂ is collected with the same objective lens and then detected by a single photon avalanche diode (SPAD). By moving the three-axis piezo translation stage, we are able to obtain the PL map of the sample. We also use a spectrograph (Princeton Instruments, Acton SP2500) to analyze the spectrum of the PL emission. A dichroic plate is used to block the excitation laser and let the PL of MoS₂ pass. Two focal lenses each with a focal length of 40-mm together with a 30- -diameter pin-hole are used as spatial filtering. We put a long pass filter behind the lenses to remove other stray light in the light path. To obtain the Au–MoS₂–AuNP sample, we first grow monolayer MoS₂ on a 300-nm-thick silica substrate, and transfer it to the gold film using wet transfer method (see “method” titled section attached below). Then the AuNPs are spin-coated at the top of it. To prevent the monolayer MoS₂ from being damaged, the excitation power is kept below 1 mW. All the measurements are performed at room temperature.

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. (color online) Experimental setup for exciting MoS2 and collecting its emission.

(i) CVD growth of monolayer MoS₂ on SiO₂/Si substrate

Monolayer MoS₂ was grown on SiO₂(300 nm)/Si substrate in a quartz tube furnace at atmospheric pressure by a chemical vapor deposition (CVD) method. Prior to growth, SiO₂/Si substrates were cleaned by sonication in acetone, absolute ethanol and distilled water for 20 min in sequence. The substrates were then placed face-down above the ceramic boat that was filled with 30 mg of MoO₃ and which was placed in the center of the quartz tube. Another ceramic boat filled with 10 mg of sulphur was in the upstream of the tube 12 cm away from the center, after purging the system with ultrahigh-purity argon for 20 min, the furnace was kept at 650 °C for 5 min at a heating rate of 15 °C/min.

(ii) Wet transfer

The CVD MoS₂ was transferred from SiO₂/Si substrate using the wet transfer KOH method. Samples were first spin-coated at 2000 rpm with PMMA A4, resulting in a thick polymer film. These were detached in a 30% KOH solution, washed several times in DI water, and transferred onto our prepared samples sequentially, and then the samples were heated up to 180 °C slowly in order to obtain the dry samples. We use acetone to clean the PMMA, and blow-dry the sample by clean N₂ from IPA solution.

In Figs. 2(a) and 2(b), the optical microscope (OM) and scanning electron microscope (SEM) images of synthesized monolayer MoS₂ showtheir high crystallinity and uniformity in large area ( ). Figure 2(c) displays the Raman spectrum of the monolayer MoS₂. Two typical peaks for monolayer MoS₂, i.e., (in-plane) and (out-of-plane) separated by ∼20 cm⁻¹, are clearly seen. And from the PL spectrum of the MoS₂ shown in Fig. 2(d), we can also see two peaks around 635 nm and 670 nm, which are consistent with previous results.^[17]

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. (color online) (a) and (b)) Optical microscope (OM) and scanning electron microscope (SEM) images of the CVD-grown MoS2 deposited on SiO2/Si substrate. (c) Raman spectrum of CVD-MoS2. (d) Typical PL spectrum of MoS2 showing a strong emission peak at ∼670 nm.

For SPPs-assisted PL enhancement, it comes from two parts: enhanced excitation rate due to SPP resonance and enhanced emission rate via Purcell effect.^[51] To maximize the excitation rate, the pump wavelength should match with the plasmonic resonance of AuNP. Similarly, the Purcell factor should also be engineered by reducing the mode volume of the plasmonic mode. Therefore, it is better to utilize the extremely confined field of gap plasmons for PL enhancement experiments, rather than an “open” plasmonic cavity like single AuNP. In our experiment, the gap plasmons are formed between spin-coated AuNPs and Au film. The diameter of AuNPs is chosen to be 200 nm, to ensure enough confined electric field but not to block the pump laser.

In SPPs-assisted PL enhancement experiment, special attention should be paid to the quenching effect, which will increase the non-radiative decay of the MoS₂ emission and suppress the PL enhancement.^[52]

To better characterize the PL enhancement of gap plasmons, we prepared two kinds of samples: bare monolayer MoS₂ deposited on SiO₂/Si substrate (denoted as sample A) and the hybrid Au–MoS₂–AuNP structure (denoted as sample B). The preparation processes of samples A and B are schematically illustrated in Fig. 3(a). We scan the excitation position of the pump laser, and plot the PL maps of samples A and B in Figs. 3(b) and 3(c), respectively. Sample A has an average PL intensity of 2800, while the maximum PL intensity for sample B is about 6800. By removing a background noise count of ∼1500, an enhancement of 4.08 times is obtained for gap plasmons. This enhancement refers to all the fluorescent signals being collected.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. (color online) (a) Preparation process of samples A and B. (b) PL map and the corresponding SEM image for sample A. (c) PL map and the corresponding SEM image for sample B.

Furthermore, we analyze the spectra of the PL emissions of MoS₂ for different samples. As shown in Fig. 4, we extract the spectra from sample A (black curve), sample B where there is no AuNP (red curve), and sample B where there is an AuNP (blue curve). We can see that when only gold films exist, the count of MoS₂ decreases slightly, which is due to the quenching effect as we mentioned above. We define the enhancement factor as the ratio of the peak PL for Au–MoS₂–AuNP structure to the peak PL for bare MoS₂ (excluding the background noise). It is clearly shown that the PL emission of monolayer MoS₂ is enhanced by about 4.43 times due to gap plasmons.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. (color online) (a) Spectra of MoS2 in different circumstances. (b) PL enhancements of spectral peak of MoS2 in different circumstances.

This enhancement factor can be even greater, if we could reduce the quenching effect. One way is to add an insulator layer, like Al₂O₃. However, it could also weaken the Purcell effect, and thus reducing the PL enhancement of the hybrid structure. Therefore, an optimization of the insulator thickness is necessary. On the other hand, the enhancement factor of gap plasmon is actually higher than the number we give. Considering the objective lens we used, the diameter of the pump beam is about 720 nm. While the diameter of AuNP is just 200 nm, the contribution of gap plasmons to the collected spectrum is small. If we can further reduce the diameter of the incident light by using an oil objective lens with higher NA, we may obtain a better enhancement and real enhancement factor.

In this work, we have used a hybrid metal nanostructure to enhance the PL of monolayer MoS₂. We demonstrate that the gap plasmons between the gold particle and gold film enhances the PL by about 4.43 times. These findings inspire us to further apply the optical properties of two-dimensional materials to a series of new fields.