Novel nanophotonic devices and nanomaterials are being explored for their potential in next-generation on-chip optical processing [1,2]. Surface plasmon polaritons (SPPs) [3], electromagnetic excitations that propagate along the interface between a metal and a dielectric, are a natural candidate for both integrated subwavelength light guiding and pronounced light–matter coupling [4–6]. Silver (Ag) nanowires have been studied extensively in this regard [7–11], and coupling has been demonstrated between Ag nanowires and other nanostructures, such as quantum dots [12–14] and nitrogen-vacancy centers [15,16].

Although there has been some investigation into graphene nanowire hybrids for nanophotonic circuitry [17,18], the vast potential for two-dimensional atomically thin materials in this realm is largely unexplored. Single-layer molybdenum disulfide (${\mathrm{MoS}}_2$) [19], a semiconductor being explored for its photoluminescence [20] potential as both a transistor [21] and photodetector [22,23], is an ideal choice to couple with nanoplasmonic circuitry. In this Letter, we explore the nanophotonics of a ${\mathrm{MoS}}_2/\mathrm{Ag}$ nanowire hybrid structure. We demonstrate coupling between a single-layer ${\mathrm{MoS}}_2$ flake and a single Ag nanowire. We show that a plasmon excited at the uncovered end of the nanowire can propagate and excite ${\mathrm{MoS}}_2$ photoluminescence (PL), both by direct plasmon-to-exciton conversion along the wire and by absorbing photons rescattering from the end of the wire. We also demonstrate that ${\mathrm{MoS}}_2$ excitons can decay to generate Ag nanowire plasmons. Finally, we show it is possible for the Ag nanowire to serve a dual role as both a channel for ${\mathrm{MoS}}_2$ excitation and subsequent extraction of the decaying ${\mathrm{MoS}}_2$ excitons.

Figure 1(a) presents an illustration of the fabricated ${\mathrm{MoS}}_2/\mathrm{Ag}$ nanowire device. An incident photon is converted to a plasmon that propagates along the wire. When the plasmon arrives at the ${\mathrm{MoS}}_2$, the plasmon may either be converted to an exciton, resulting in frequency-shifted photon emission from the ${\mathrm{MoS}}_2$, or it can be converted back to a photon at the end of the wire. An optical micrograph of the hybrid device studied in this Letter is shown in Fig. 1(b). Figure 1(c) is a Raman spectrum acquired at the overlap region between the end of the nanowire and the ${\mathrm{MoS}}_2$. The measured Raman spectrum reveals that the flake is single-layer ${\mathrm{MoS}}_2$ [24]. To fabricate the device, we used a PMMA liftoff technique [25] to transfer the ${\mathrm{MoS}}_2$ from a silicon with thermal oxide substrate to the wire on a standard glass cover slip. See Supplement 1 for more details on device fabrication.

 

Fig. 1. (a) Schematic of ${\mathrm{MoS}}_2/\mathrm{nanowire}$ structure. (b) Single-layer ${\mathrm{MoS}}_2$ on a silver nanowire on glass after transfer of the flake. (c) Raman spectrum collected at the end of the wire in ${\mathrm{MoS}}_2$; $\lambda =532\mathrm{nm}$, $\text{power}=70\mathrm{\mu W}$. (d) Demonstration of plasmon propagation along the Ag nanowire; $\lambda =635\mathrm{nm}$, $\mathrm{power}=20\mathrm{\mu W}$. All scale bars are 2 μm.

Download Full Size | PDF

The charge-coupled device (CCD) image in Fig. 1(d) demonstrates plasmon propagation and photon re-emission. Laser radiation ($\lambda =635\mathrm{nm}$), polarized parallel to the wire axis, is coupled from the far-field into the nanowire at the end labeled “1” in Fig. 1(b) using a $100\times$ oil-immersion objective with numerical aperture (NA) of 1.3. To reduce scattering and eliminate leakage radiation, the sample was covered in index-matching ($n=1.515$) oil. In order to convert a photon into an SPP, the laser must be focused onto one of the ends of the wire; this accounts for the momentum mismatch between the incoming photon and the plasmon [9]. The SPP $1/e$ propagation length increases for larger-diameter wires [26] and longer excitation wavelengths [4]. The wires used in our study (average diameter 386 nm) support two lower-order modes, $m=0$ and $m=1$, which couple to incident light polarized parallel and perpendicular to the wire, respectively [11]. The in-coupling efficiency is always greater for the $m=0$ mode than for the $m=1$ mode, but the $1/e$ propagation length of the $m=1$ mode becomes longer for larger diameter wires. For a wire of this length, following parallel excitation, we calculate the efficiency of photon re-emission at the end of the wire after plasmon propagation to be around 0.008%–0.012%. See Supplement 1 for a discussion of photon re-emission efficiencies.

Coupling between far-field photons, SPPs, and the single-layer ${\mathrm{MoS}}_2$ was studied using an inverted confocal microscope utilizing an oil-immersion objective ($\mathrm{NA}=1.4$). A nanopositioning stage (Mad City Labs, Inc.) was used to scan and position the sample. Excitation polarization was controlled by a half-wave plate. The signal from the sample was sent to either an APD or a spectrometer. Longpass filters to block the laser line were used in front of both detectors. A fluorescence image of the single-layer flake on the wire from Fig. 1(b) using a laser ($\lambda =633\mathrm{nm}$) is shown in Fig. 2(a). For photon-counting images, we adopt the convention of using a solid red circle to indicate the excitation and a white star to represent the approximate center of the collection focal volume, respectively. For this data set, the excitation and collection focal volumes are coincident so the dot and star overlap, indicating no displacement between excitation and collection. In Fig. 2(a), we observe strong PL from the ${\mathrm{MoS}}_2$ flake, characteristic of single-layer, as well as a large increase in counts in the area where the flake overlaps with the wire.

 

Fig. 2. (a) Confocal fluorescence image of the ${\mathrm{MoS}}_2/\mathrm{wire}$ device from the dotted white box in Fig. 1(b). The red dot and white star represent no displacement between the excitation and collection focal volume, respectively. Scale bar is 1 μm. (b) Normalized photoluminescence spectra for the end of the wire covered by ${\mathrm{MoS}}_2$ excited with light polarized parallel (blue decorated with squares) and perpendicular (red decorated with triangles) to the wire. Spectra of the ${\mathrm{MoS}}_2$ on substrate were also acquired using parallel (green with star above) and perpendicular (black) excitation polarizations. The spectra for the ${\mathrm{MoS}}_2$ on substrate are rescaled for clarity. Inset: absolute photoluminescence counts for the end of the wire in ${\mathrm{MoS}}_2$ with light polarized parallel to the wire. (c) Spectra taken in 150 nm steps along the line in (a). For all images, $\lambda =633\mathrm{nm}$, $\mathrm{power}=5\mathrm{\mu W}$.

Download Full Size | PDF

To investigate the origin of the increased counts, Fig. 2(b) presents spectra acquired with the excitation polarization parallel and perpendicular to the wire on the bare flake [location indicated by the white circle in Fig. 2(a)] and at the end of the wire covered by ${\mathrm{MoS}}_2$ [location indicated by the break in the solid vertical line in Fig. 2(a)]. Spectra taken at the end of the wire covered by ${\mathrm{MoS}}_2$ (bare flake) with excitation polarized parallel to the wire are shown in blue (green), and the red (black) curve shows the perpendicular polarization case. By summing the recorded photo counts for all detected wavelengths when the excitation polarization is parallel ($I_{\text{par}}$) and perpendicular ($I_{\text{perp}}$) to the wire axis, we calculate an on-wire contrast $(I_{\text{par}}-I_{\text{perp}})/(I_{\text{par}}+I_{\text{perp}})$ of 9.5% and an off-wire contrast of 2.6%. We attribute this difference to two effects. First, by removing direct contact with the substrate, ${\mathrm{MoS}}_2$ fluorescence is known to increase [19,27]. Second, the enhancement of the ${\mathrm{MoS}}_2$ fluorescence when the excitation is polarized parallel to the nanowire is a manifestation of an antenna-like enhancement of the excitation field (the wire influences emission the same way for the two different excitation conditions). Comparing the counts from the flake/wire region to the bare flake, we observed an enhancement of about $6.4\times$ in fluorescence. We would expect an enhancement of about $3–5\times$ due to removal of direct contact with the substrate [27], so the enhancement due to antenna effects is approximately $1.3–2.1\times$. The enhancement in ${\mathrm{MoS}}_2$ fluorescence in the vicinity of the region where the Ag nanowire end overlaps with the ${\mathrm{MoS}}_2$ flake is consistent on all devices we have fabricated, and we have observed a device exhibiting a 40-fold enhancement in fluorescence. See Supplement 1 for data on this device.

In addition to the enhancement, there is a clear spectral shift in the peak of the PL for the ${\mathrm{MoS}}_2$ over the wire (656 nm) compared to on the substrate (668 nm). We attribute this shift to the ${\mathrm{MoS}}_2$ flake not being in direct contact with the Ag nanowire as a result of the transfer process. The main PL peak of single-layer ${\mathrm{MoS}}_2$ consists of two peaks: the A peak centered at 655 nm attributed to uncharged excitons and the ${\mathrm{A}}^{-}$ peak centered around 670 nm due to negatively charged trions [28]. It has been reported that interaction of the ${\mathrm{MoS}}_2$ with the substrate suppresses exciton emission due to doping [27]. When removed from the substrate, the A peak becomes dominant. This shift is not due to strain, as strain would redshift the spectra [29]. For comparison, Fig. 2(c) presents ${\mathrm{MoS}}_2$ spectra as we measure along the line in Fig. 2(a), starting from the bottom. Each spectrum is independently normalized. The spectral position of the peak is consistent along the wire covered by the ${\mathrm{MoS}}_2$, and as the collection region moves off the wire, the peak redshifts.

To explore plasmon excitation of ${\mathrm{MoS}}_2$ PL, the collection and excitation focal volumes are displaced vertically by the length of the wire. Figure 3(a) shows the resulting fluorescence image (with a CCD image of the structure overlaid) when the sample is scanned in this configuration. The prominent feature in the fluorescence image results when the laser excites plasmons at the end of the wire opposite the ${\mathrm{MoS}}_2$, which propagate along the wire and excite ${\mathrm{MoS}}_2$ PL. Figures 3(b) and 3(c) display a scan of this feature when light is polarized parallel and perpendicular to the wire, respectively. There is a reduction in intensity when the light is polarized perpendicular to the wire. Figure 3(d) shows the normalized polarization dependence of the signal with a contrast of 21%. This observed modulation indicates that the coupling is stronger when the excitation is parallel to the wire axis. This contrast is also greater than when the excitation was not displaced, suggesting that this feature is the result of plasmon propagation and ${\mathrm{MoS}}_2$ excitation. The largest contrast that we observed of all of the devices tested was 80%. See Supplement 1 for data on this device. Simulations using the finite-difference time-domain (FDTD) method in Lumerical to further investigate the plasmon excitation are shown in Supplement 1.

 

Fig. 3. (a) Fluorescence image resulting from displacing the collection and excitation focal volumes by the length of the wire, overlaid with a CCD image of the device. Scale bar is 2 μm. (b) and (c) Feature in (a) when light is polarized parallel and perpendicular to the wire, respectively. Scale bar in each is 1 μm. (d) Normalized polarization contrast of the ${\mathrm{MoS}}_2$ fluorescence as a function of excitation polarization angle with respect to the nanowire axis; 0° corresponds to polarization parallel to the wire. (e) Spectra taken at the positions marked in the inset when the laser excitation is at the uncovered end of the wire. Scale bar is 2 μm. For all images, $\lambda =633\mathrm{nm}$, $\mathrm{power}=5\mathrm{\mu W}$.

Download Full Size | PDF

We anticipate that plasmon-excited ${\mathrm{MoS}}_2$ PL is not limited to the end of the wire. To investigate this, the displacement between the laser excitation and collection is adjusted to be a fraction of the wire length. The sample is then translated so that the laser excitation is at the uncovered end of the wire. Figure 3(e) presents the spectra corresponding to several points starting from the top circle, labeled “1,” and walking downward in the fluorescence scan of the full sample in the inset. We observe that the PL is strongest near the end of the wire. However, we also obtain significant signal over the entire length that the wire is covered by the ${\mathrm{MoS}}_2$. This is the result of two mechanisms at play. First, plasmons that propagate to the end of the wire are rescattered as photons and reabsorbed by the ${\mathrm{MoS}}_2$, exciting an exciton. The electron-hole recombination then produces the PL signal. Second, plasmons in the wire are directly converted to excitons in the ${\mathrm{MoS}}_2$, which then fluoresces.

In addition to plasmons exciting the ${\mathrm{MoS}}_2$, excitons in the ${\mathrm{MoS}}_2$ can be converted to plasmons that propagate along the wire and are rescattered as photons. The excitation is aligned with the overlap region of the ${\mathrm{MoS}}_2$ flake/nanowire end, and the collection focal volume is aligned to the uncovered nanowire end (the reverse configuration of Fig. 3). Figure 4(a) shows the resulting fluorescence image with a CCD image overlaid when the sample is scanned in this configuration. Compared to the localized feature in Fig. 3(a), the present image shows an attribute that extends beyond the end of the wire. This is suggestive of plasmonic excitation along the ${\mathrm{MoS}}_2/\mathrm{wire}$ interface. As the laser excitation scans over the ${\mathrm{MoS}}_2/\mathrm{wire}$ interface, photons that re-emerge from the uncovered end are still detected by the confocal volume of the APD. If plasmons could only be excited at the end of the wire, the attribute in this image would look similar to the feature in Fig. 3(a). The upper (lower) inset displays a scan of the feature when the excitation light is polarized parallel (perpendicular) to the wire. Evident from comparing the two insets is an increase in the emission from the uncovered end when the excitation is parallel to the nanowire. The enhancement again suggests the nanowire provides an antenna-like enhancement of the excitation. Because ${\mathrm{MoS}}_2$ absorption does not prefer a linear polarization, any mismatch results from excitation effects.

 

Fig. 4. (a) Image resulting from displacing the collection and excitation focal volumes by the length of the wire, overlaid with an image of the device. Opposed to Fig. 3, in this case the excitation is located at the ${\mathrm{MoS}}_2$ end, and the collection is at the uncovered wire end. Scale bar is 2 μm. Top (bottom) inset: feature when light is polarized parallel (perpendicular) to the wire. Scale bar in each is 1 μm; $\lambda =635\mathrm{nm}$, $\mathrm{power}=20\mathrm{\mu W}$. (b) Confocal fluorescence image of the sample with the uncovered end of the wire rescaled. Scale bar is 2 μm. (c) Spectra collected at the rescaled feature in (b) for light polarized parallel (blue) and perpendicular (red) to the wire. For (b) and (c), $\lambda =633\mathrm{nm}$, $\mathrm{power}=5\mathrm{\mu W}$.

Download Full Size | PDF

Finally, we demonstrate it is possible to use the Ag nanowire both as a channel for near-field excitation of the ${\mathrm{MoS}}_2$ flake and to recollect the resultant ${\mathrm{MoS}}_2$ fluorescence. The re-excited plasmons, at the ${\mathrm{MoS}}_2$ photon energy, can propagate back along the wire and rescatter to the far-field as photons. Figure 4(b) shows a fluorescence image of the entire sample with the excitation and collection aligned. For this image, we have rescaled the end of the wire not covered by the ${\mathrm{MoS}}_2$. There is a pronounced feature at the excitation end of the wire that is stronger than the background. To investigate this feature, spectra were collected for light polarized parallel and perpendicular to the wire [Fig. 4(c)]. The spectra reveal this is indeed PL from the ${\mathrm{MoS}}_2$ flake. The pronounced polarization contrast in the two different excitation directions suggests that the Ag nanowire plasmons mediate this excitation and collection process.

In summary, we have demonstrated photonic and plasmonic interactions between an individual Ag nanowire and single-layer ${\mathrm{MoS}}_2$. We found it is possible to excite ${\mathrm{MoS}}_2$ with Ag nanowire plasmons as well as convert decaying ${\mathrm{MoS}}_2$ excitons into Ag wire plasmons. This first step shows that there is pronounced nanoscale light–matter interaction between plasmons and atomically thin material that can be exploited for nanophotonic integrated circuits. A natural next step is the creation of a near-field detector based on ${\mathrm{MoS}}_2$, as well as ${\mathrm{MoS}}_2$ light-emitting diodes coupled to on-chip nanoplasmonic circuitry.