1. Introduction

Recently, high-efficient random lasers based on plasmonic nanostructures have attracted significant attention due to their appealing physical pictures and potential applications [1–7]. Plasmonic random lasers in both solutions and thin films have been investigated extensively by a number of research groups over the last decade [8–14]. By comparison, the solid-state plasmonic random laser is considered as a more promising solution for practical applications. The miniaturization and flexibility of plasmonic random lasers are essential for the continued development of laser technologies. In our previous work, a flexible plasmonic random laser has been proposed to achieve the tunability of laser output [15]. Note that all solid-state random lasers mentioned above are constructed on the substrates. In fact, the substrate is unnecessary for most configurations, which is not involved in the laser oscillation process. Thus, removing the substrate of the laser may be an alternative to significantly reduce the thickness of the device, which miniaturizes the size of the random laser device efficiently. Furthermore, the flexibility and transplantation of the free-standing membrane are much better than its counterpart with a substrate. It can be attached to some special surfaces to achieve plasmonic random lasing, such as an optical fiber facet. The random laser on the optical fiber facet can be equalized to a kind of random fiber lasers [16–20], which is potentially used in speckle free imaging [21] or adaptive pumping [22].

In this work we demonstrate an ultrathin plasmonic random laser, which consists of a free-standing polymer membrane embedded with silver nanoparticles (Ag NPs). The laser device is fabricated using a simple lift-off technique. Low threshold random lasing can be observed when the polymer membrane device is optically pumped due to the plasmon-assist waveguide resonance. The membrane device is very flexible, which can be transplanted to the end facet of an optical fiber to obtain a plasmonic random fiber laser.

2. Fabrication of ultrathin plasmonic random lasers

The fabrication process is illustrated in Fig. 1. A solution of the water-soluble polyvinyl alcohol (PVA, 107, Celanese Chemicals, Germany) with a concentration of 0.04 g/mL is spin-coated onto a glass substrate (10 mm × 10 mm × 1mm) at a speed of 4000 rpm, forming a 400-nm-thick film, which acts as a removable layer. Then, the solution of a typical light-emitting polymer, poly [(9,9- dioctylfluorenyl-2,7-diyl)-alt-co-(1,4-benzo-(2,1’,3)-thiadiazole)] (F8BT) doped with Ag NPs is spin-coated on the 400 nm PVA film with a speed of 2000 rpm, forming a 200-nm-thick film. The concentration of F8BT in xylene is 24 mg/mL and that of the Ag NPs is 4 mg/mL, which are optimized after a series of experiments and employed in the following context. It will be discussed in detail later. An ultrasonic dispersion process is applied to guarantee the dispersity of distribution of Ag NPs in the solution.

 

Fig. 1 Fabrication procedures and simplified optical layout of ultrathin plasmonic random lasers. (a) A sandwich structure is fabricated by spin-coating the solution of the PVA and F8BT dopped with Ag NPs onto the glass substrate. (b) The sample is immersed in deionized water at room temperature for 30 mins, so that the PVA layer is sufficiently dissolved. (c) The F8BT layer peels off from the substrate, forming a free-standing membrane device with a thickness of 200 nm.

Download Full Size | PDF

So, a sandwich structure consisting of F8BT/PVA/glass substrate is fabricated by spin coating, as shown in Fig. 1(a). The F8BT layer can be peeled off from the substrate by a simple lift-off technique. The whole sample is immersed into deionized water at room temperature for 30 mins. The F8BT film will get removed automatically from the substrate due to that the PVA layer is etched off through dissolving into water as shown in Fig. 1(b), forming a free-standing membrane device in Fig. 1(c). The thickness of the device is about 200 nm, which is measured using a Veeco dektak 150 surface profilometer.

When the polymer membrane is optically pumped, random lasing based on the waveguide plasmonic feedback can be observed as illustrated in Fig. 1(c). The radiation of the polymer molecules is scattered strongly multiple times by the Ag NPs due to the localized surface plasmon resonance, which is guaranteed by the overlap of the localized surface plasmon resonance of Ag NPs and the photoluminescence (PL) spectra of the polymer. It will be discussed later in detail.

The Ag NPs are synthesized by a one-step synthesis approach [23], and are then dispersed in a xylene solution of F8BT to obtain the Ag NPs ink. The typical diameter of the Ag NPs is about 200 nm as shown in Fig. 2(a), which is characterized by a scanning electron microscopy (SEM, Hitachi S-4800). The shape of the Ag NPs is approximately ellipsoidal. The Ag NPs are scattered randomly in the polymer solution after doping and ultrasonic dispersion. Thus, a random distribution of the Ag NPs in the polymer film can be obtained, as illustrated in Fig. 2(b). The random distribution of Ag NPs provides strong plasmonic feedback with the assistance of the high-quality waveguide confinement of the polymer membrane, which is essential to ensure the low-threshold random lasing.

 

Fig. 2 SEM images of (a) the Ag nanoparticle on an ITO glass substrate and (b) the F8BT film doped with Ag NPs. (c) Photograph of the ultrathin plasmonic random laser based on the free-standing polymer membrane. The PET plate with a round hole in the center acts as a frame. The diameter of the hole is about 5 mm.

Download Full Size | PDF

The polymer membrane will float on the surface of the water after peeling off the glass substrate. To facilitate the spectroscopic measurement, the free-standing polymer membrane is attached to a polyethylene terephthalate (PET) frame as shown in Fig. 2(c). There is a round hole with a diameter of 5 mm in the center of the PET plate (15 mm × 15 mm × 0.5 mm). Thus, most of the membrane device is free standing, which provides stronger waveguide confinement than its counterpart with a substrate. The radiation of the polymer is scattered by the disordered plasmonic Ag NPs at their interfaces in the polymer membrane, which is multiplied by the reflection or total reflection at the two F8BT/air interfaces. The minimum mean free path length of such a waveguide configuration can be defined as [8]

Low-threshold laser emission can be generated through strong waveguide confinement and plasmonic feedback when the 200-nm-thick polymer membrane is pumped optically. A femtosecond laser beam (200 fs) with a wavelength of 400 nm and a repetition frequency of 1 kHz is employed as a pump source. The diameter of the laser spot is about 3 mm. The pump beam impinges on the random laser with an incident angle (~30 degree) to facilitate the actual test. The intensity of the laser beam can be tuned continuously using a variable optical attenuator. The laser emission is collected by a fiber lens onto an optical fiber that is coupled to a spectrometer (Maya 2000 Pro, Ocean Optics). Figure 3(a) shows the output intensity of the ultrathin plasmonic random laser at different pump powers. A random laser emission peak can be observed clearly at 567 nm (λ₀) when the pump power intensity is around 5.1 μJ/cm². The full width at half maximum (FWHM) of the emission peaks is less than 10 nm at 46.8 μJ/cm². It can be seen that the minimum mean free path length (l_min) equals about 1.48λ₀. The real l_min may be smaller after taking into account the multiple scattering of Ag NPs, which means the proposed configuration is a strongly scattering system [24].

 

Fig. 3 (a) Measured emission spectra of the ultrathin plasmonic random laser. The inset denotes the enlarged view of the laser mode. (b) The output intensity and linewidth of the laser device as a function of the pump power density. The threshold of the laser emission (4 mg/mL) is around 2 μJ/cm² as indicated by the black arrow. The brown- green-, and cyan-dot curves indicate different concentrations of Ag NPs, respectively.

Download Full Size | PDF

Figure 3(b) presents the intensity and FWHM of the output laser as a function of the pump power density. The threshold of the ultrathin plasmonic random laser can be estimated as 2 μJ/cm² as indicated by the black arrow in Fig. 3(b), which is much lower than the plasmonic random laser with substrate [8]. It can be attributed to the strong confinement provided by the waveguide gain channels without a substrate and the intrinsic scattering at the polymer-metal interfaces.

A control experiment was performed to check the influence of the concentration of the Ag NPs on the performance of the random laser, as shown in Fig. 3(b). It can be seen that the laser performance will be lowered if the concentration of Ag NPs is too high or too low. The optimized concentration reflects the balance between the photoluminescence enhancement and quenching by the Ag NPs. For a F8BT membrane without Ag NPs, no random lasing can be observed.

3. Transplantation of a plasmonic random laser onto the optical fiber facet

The most intriguing feature of the ultrathin plasmonic random laser is that it can be attached on an optical fiber end face to achieve random lasing. The wet polymer membrane can be pasted on the fiber facet easily without using adhesives. After drying at room temperature, the polymer membrane will stick tightly to the fiber end face due to the surface tension.

Figure 4(a) shows the spectroscopic characterization and the microscopy image of the plasmonic random laser on the optical fiber end face. A multimode optical fiber is used in the experiment, which has a core diameter of 600 μm, a cladding thickness of 30 μm, and a coating thickness of 120 μm. The polymer membrane is indicated by a red arrow in the inset in Fig. 4(a). In the experiment, a nonpolarized white light from a tungsten halogen lamp (HL-2000) is employed to characterize the extinction and PL spectra. The PL spectrum (red curve) is centered at about 564 nm. The bandwidth of the extinction spectrum of Ag NPs is very broad, which overlaps the PL spectrum of F8BT. There exists an arrow peak at about 450 nm in the extinction spectrum of Ag NPs. The localized surface plasmon resonance of the Ag NPs has induced enhancement of the PL spectrum of F8BT, resulting to the stimulated emission or even the random lasing as indicated by the black curve in Fig. 4(a). When the pump beam impinges on the polymer membrane through the optical fiber, a laser emission peak located at about 562 nm emerges as shown in Fig. 4(a). A small shift of the emission peak can be observed in Figs. 3(a) and 4(a), which is decided by both the scattering mean free path and the refractive index of the substrate. The direction of the pump beam is denoted by a blue arrow in the inset in Fig. 4(a). Several narrower peaks can be observed around 562 nm in the emission spectrum, which implies that the broadband random laser emission actually contains multiple oscillation modes. Note that the emission wavelength of the membrane device changes after being pasted on the fiber tip. The blue-shift of the emission wavelength is derived from the increase of the refractive index of the substrate.

 

Fig. 4 (a) Measured spectra of plasmonic random laser on the fiber end face. The red curve presents the laser emission above the pump threshold. The black curve denotes the PL spectrum of F8BT. The blue curve indicates the extinction spectrum of Ag NPs in the F8BT film. The inset is a microscopy image of the plasmonic random laser on the fiber end face. The pump beam propagates in the optical fiber as indicated by the blue arrow. (b) PL lifetime of the F8BT membrane with and without Ag NPs.

Download Full Size | PDF

The influence of the localized surface plasmon resonance of Ag NPs on the radiation of polymer molecules is confirmed further by the comparison between the PL decay dynamics of F8BT with (red curve) and without (blue curve) Ag NPs, as shown in Fig. 4(b). The interaction between the localized surface plasmon resonance of Ag NPs and the polymer molecules reduced the PL lifetime of F8BT from about 1.2 to 0.7 ns, which indicates possible amplification of the fluorescence through stimulated emission.

To further understand the plasmonic enhancement in such an Ag NP/polymer configuration, the electric field distribution near the Ag NP is calculated at an excitation wavelength of 560-570 nm with a step of 1 nm. Strong localized field enhancement can be observed at all excitation wavelengths as shown in Fig. 5. The simulations were done using the finite element method with the commercial software COMSOL. The permittivity of Ag is chosen from Johnson and Christy's experimental data [25].

 

Fig. 5 The typical electric field distribution near the Ag NP embedded in the F8BT membrane. The black arrow indicates the direction of the incident light. The red arrow denotes the polarization direction of the excitation light. The refractive index of polymer (F8BT) is about 1.6. The scale bar is 100 nm.

Download Full Size | PDF

4. Conclusions

In conclusion, we demonstrate an ultrathin plasmonic random laser fabricated by a simple lift-off technique, which consists of a polymer membrane embedded with Ag NPs. A low threshold can be observed when the 200-nm thick laser device is optically pumped, which is about 2 μJ/cm². It attributes to the strong plasmonic scattering by the Ag NPs and the high-quality confinement by the polymer membrane. Furthermore, the plasmonic random laser device can be attached to a fiber end face to achieve random lasing. These results can be utilized in the applications of plasmonic random lasers.