Controlling Random Lasing with Three- Dimensional Plasmonic ...
Supporting Information for
Controlling Random Lasing with ThreeDimensional Plasmonic Nanorod Metamaterials Zhuoxian Wang,† Xiangeng Meng,*,† Seung Ho Choi,‡ Sebastian Knitter,§ Young L. Kim,‡ Hui Cao,§ Vladimir M. Shalaev,† Alexandra Boltasseva*,† †
School of Electrical & Computer Engineering and Birck Nanotechnology Center, Purdue University, 1205 West State Street, West Lafayette, Indiana 47907, United States
‡
Weldon School of Biomedical Engineering, Purdue University, 206 S. Martin Jischke Drive, West Lafayette, Indiana 47907, United States
§
Department of Applied Physics, Yale University, New Haven, Connecticut 06250, United States *Correspondence: [email protected] (Xiangeng Meng) [email protected] (Alexandra Boltasseva)
Methods Device fabrication. Slanted silver NRs were grown on a glass substrate using the Glancing Angle Deposition (GLAD) technique. In order to produce a well-separated silver NR array, the substrate was tilted so that its surface normal was 86° with respect to the incident evaporation beam. The deposition rate was set to 3 Å/s and the deposition was performed in a high vacuum chamber with the background pressure around 1×10-6 torr. The gain layer in RL devices was fabricated by putting a drop of aqueous solution containing PVA and R6G on top of silver NRs, followed by a spin-coating process at 1,000 rpm for 30 seconds and subsequent baking at 55 °C for 8 hours to remove the residual solvent. The concentration of R6G in the resultant film was around 10 mM relative to PVA. The gain coefficient provided by R6G under this concentration is estimated to be g ~ 1.2 × 103 cm–1 by using g = ρσem1. Here, ρ (6.02 × 1018 cm–3) and σem (σem = 2 × 10–16 cm2) represent the number density and emission cross section of R6G, respectively. The thickness of the nanorod film layer varies with the nanorod length (Fig. S9). The total thickness of the nanorod layer and dye-polymer atop is ~2 μm. The test mark embedded with ‘PURDUE’ characters in a gold film was fabricated on an ITO-coated glass substrate using electron beam lithography and electron beam deposition. The thickness of the gold film layer is 80 nm. Optical Measurements. The optical scattering spectra were measured in a commercial UV-Vis-NIR spectrometer (Lambda 950). In lasing experiments, a frequency-doubled Nd: YAG picosecond laser (532 nm, 400 ps pulse duration, and 1 Hz repetition rate) was used as the pump source. The pump laser was normally incident on the sample, which was focused by an objective lens (5× NA = 0.1). The diameter of the pump spot size is ~148.0
μm. The polarization of the pump laser with respect to the orientation of the NRs was controlled by a half-wave plate. The emission signal was collected through an optical fiber and recorded by a spectrometer (SP-2150i, Princeton Instruments). All the measurements were performed at room temperature. In spatial-spectral measurements, a frequency-doubled Nd: YAG picosecond laser (Continuum Leopard, 30 ps pulse duration, and 10 Hz repetition rate) was used as the pump source. The emitted light was collected with an aspheric lens (f = 10 mm, NA = 0.25) and reimaged onto the entrance slit of an imaging spectrometer (Princeton Instruments, SpectraPro 300i), by means of a tube lens (f = 200mm). By using a twodimensional CCD array-sensor (Andor Newton, DU920P-BEX2-DD), synchronized with the pump source, the spectral-spatial emission profile for each pump/emission event could be detected by a single sensor exposure. As shown in Fig. 3, we acquired onedimensional spatial image data in the vertical direction and spectral information in the horizontal direction. The spatial/spectral resolution was on the order of 5 μm and 0.2 nm respectively. Numerical Simulations. The simulation was based on FEM and performed using a commercial software package (COMSOL Multiphysics 4.4). In all calculations, the dielectric constants of silver were described using the Drude-Lorentz model with five Lorentz oscillators2. The silver NR was surrounded by a host with the refractive index set as 1.5.
Figure S1. Top-view SEM images of the silver NR arrays with different length scales ranging from 192 nm to 710 nm on average.
Figure S2. Scattering properties of silver NR arrays with various lengths. Stronger scattering occurs for longer silver NR arrays. In the measurements, the polarization of the incident light is set as β = 0° (β is illustrated in Figure 2a).
Figure S3. Contour map of calculated σsca/σabs as a function of the aspect ratio (L/D) of silver nanorods. In the calculation, the incident wave is set as s-polarized at λ = 565 nm. The diameter of the NR is fixed at D = 100 nm while its length varies from 200 to 1000 nm. The incident angle is described with θ. The calculation shows that σsca/σabs is larger than 30 in the majority of the region while there are some minor exceptional regions with σsca/σabs > 20.
Figure S4. Volume fraction of metal in RLs based on silver NRs of different length scales.
Figure S5. Calculated σsca/σabs for a 12 and 55 nm-diameter silver sphere used as light scatterers in references 13 and 19, respectively. In the calculation, the incident wave is set as p-polarized at λ = 565 nm which is close to the lasing wavelength when using R6G as the gain medium in random lasers. The magnitude of σsca/σabs is much lower than that of the silver nanorod used in this work. The results calculated under s-polarization are similar to those obtained under p-polarization due to the geometric symmetry of the sphere.
Figure S6. Strong anisotropic scattering properties of the silver NR metamaterial (sample of L = 917 nm). The measurements were conducted in a commercial UV-Vis-NIR spectrometer (Lambda 950). A polarizer was inserted in the spectrometer to control the polarization of the pump beam so as to obtain the polarization-dependent scattering spectra. The measurements were conducted at an interval of 10 degree of β.
Figure S7. Lasing from the RL device based on TiO2 nanorod array. TiO2 nanorod array was prepared by the GLAD approach and then the RL device was fabricated and characterized in the same way as the device based on silver nanorod array. (a) SEM image of the TiO2 nanorod array, the length of the nanorods is ~700 nm on average. (b) Emission spectra recorded at various pump energies of a picosecond pulsed pump laser (λ = 532 nm). The pump spot size is ~148.0 μm in diameter. (c) Integrated intensity versus the pump energy. The arrow indicates the lasing threshold. (d) Normalized peak intensity versus the polarization of the pump laser. No obvious polarization dependence is observed. The pump energy is ~0.88 μJ. The decrease of the emission signal after β = 40° is ascribed to the photo-bleaching effects of the dye molecules.
Figure S8. (a) SEM image of the test mark embedded with “PURDUE” characters in an 80 nm-thick gold film. The enlarged figure shows that the smallest feature on the test mark is ~3 μm. The scale bar is 100 μm. (b) Images of ‘PURDUE’ test mark under the illumination of the 532 nm laser. Speckles are clearly observed.
Figure S9. The thickness of the nanorod film layer versus the nanorod length.
References 1. 2.
Meng, X. G., Fujita, K., Murai, S., Motaba, T., Tanaka, K. Nano Lett. 2011, 11, 1374-1378. Chen, W. Q., Thoreson, M. D., Ishii, S., Kildishev, A. V., Shalaev, V. M. Opt. Express 2010, 18, 5124-5134.
Controlling Random Lasing with ThreeDimensional Plasmonic Nanorod Metamaterials Zhuoxian Wang,† Xiangeng Meng,*,† Seung Ho Choi,‡ Sebastian Knitter,§ Young L. Kim,‡ Hui Cao,§ Vladimir M. Shalaev,† Alexandra Boltasseva*,† †
School of Electrical & Computer Engineering and Birck Nanotechnology Center, Purdue University, 1205 West State Street, West Lafayette, Indiana 47907, United States
‡
Weldon School of Biomedical Engineering, Purdue University, 206 S. Martin Jischke Drive, West Lafayette, Indiana 47907, United States
§
Department of Applied Physics, Yale University, New Haven, Connecticut 06250, United States *Correspondence: [email protected] (Xiangeng Meng) [email protected] (Alexandra Boltasseva)
Methods Device fabrication. Slanted silver NRs were grown on a glass substrate using the Glancing Angle Deposition (GLAD) technique. In order to produce a well-separated silver NR array, the substrate was tilted so that its surface normal was 86° with respect to the incident evaporation beam. The deposition rate was set to 3 Å/s and the deposition was performed in a high vacuum chamber with the background pressure around 1×10-6 torr. The gain layer in RL devices was fabricated by putting a drop of aqueous solution containing PVA and R6G on top of silver NRs, followed by a spin-coating process at 1,000 rpm for 30 seconds and subsequent baking at 55 °C for 8 hours to remove the residual solvent. The concentration of R6G in the resultant film was around 10 mM relative to PVA. The gain coefficient provided by R6G under this concentration is estimated to be g ~ 1.2 × 103 cm–1 by using g = ρσem1. Here, ρ (6.02 × 1018 cm–3) and σem (σem = 2 × 10–16 cm2) represent the number density and emission cross section of R6G, respectively. The thickness of the nanorod film layer varies with the nanorod length (Fig. S9). The total thickness of the nanorod layer and dye-polymer atop is ~2 μm. The test mark embedded with ‘PURDUE’ characters in a gold film was fabricated on an ITO-coated glass substrate using electron beam lithography and electron beam deposition. The thickness of the gold film layer is 80 nm. Optical Measurements. The optical scattering spectra were measured in a commercial UV-Vis-NIR spectrometer (Lambda 950). In lasing experiments, a frequency-doubled Nd: YAG picosecond laser (532 nm, 400 ps pulse duration, and 1 Hz repetition rate) was used as the pump source. The pump laser was normally incident on the sample, which was focused by an objective lens (5× NA = 0.1). The diameter of the pump spot size is ~148.0
μm. The polarization of the pump laser with respect to the orientation of the NRs was controlled by a half-wave plate. The emission signal was collected through an optical fiber and recorded by a spectrometer (SP-2150i, Princeton Instruments). All the measurements were performed at room temperature. In spatial-spectral measurements, a frequency-doubled Nd: YAG picosecond laser (Continuum Leopard, 30 ps pulse duration, and 10 Hz repetition rate) was used as the pump source. The emitted light was collected with an aspheric lens (f = 10 mm, NA = 0.25) and reimaged onto the entrance slit of an imaging spectrometer (Princeton Instruments, SpectraPro 300i), by means of a tube lens (f = 200mm). By using a twodimensional CCD array-sensor (Andor Newton, DU920P-BEX2-DD), synchronized with the pump source, the spectral-spatial emission profile for each pump/emission event could be detected by a single sensor exposure. As shown in Fig. 3, we acquired onedimensional spatial image data in the vertical direction and spectral information in the horizontal direction. The spatial/spectral resolution was on the order of 5 μm and 0.2 nm respectively. Numerical Simulations. The simulation was based on FEM and performed using a commercial software package (COMSOL Multiphysics 4.4). In all calculations, the dielectric constants of silver were described using the Drude-Lorentz model with five Lorentz oscillators2. The silver NR was surrounded by a host with the refractive index set as 1.5.
Figure S1. Top-view SEM images of the silver NR arrays with different length scales ranging from 192 nm to 710 nm on average.
Figure S2. Scattering properties of silver NR arrays with various lengths. Stronger scattering occurs for longer silver NR arrays. In the measurements, the polarization of the incident light is set as β = 0° (β is illustrated in Figure 2a).
Figure S3. Contour map of calculated σsca/σabs as a function of the aspect ratio (L/D) of silver nanorods. In the calculation, the incident wave is set as s-polarized at λ = 565 nm. The diameter of the NR is fixed at D = 100 nm while its length varies from 200 to 1000 nm. The incident angle is described with θ. The calculation shows that σsca/σabs is larger than 30 in the majority of the region while there are some minor exceptional regions with σsca/σabs > 20.
Figure S4. Volume fraction of metal in RLs based on silver NRs of different length scales.
Figure S5. Calculated σsca/σabs for a 12 and 55 nm-diameter silver sphere used as light scatterers in references 13 and 19, respectively. In the calculation, the incident wave is set as p-polarized at λ = 565 nm which is close to the lasing wavelength when using R6G as the gain medium in random lasers. The magnitude of σsca/σabs is much lower than that of the silver nanorod used in this work. The results calculated under s-polarization are similar to those obtained under p-polarization due to the geometric symmetry of the sphere.
Figure S6. Strong anisotropic scattering properties of the silver NR metamaterial (sample of L = 917 nm). The measurements were conducted in a commercial UV-Vis-NIR spectrometer (Lambda 950). A polarizer was inserted in the spectrometer to control the polarization of the pump beam so as to obtain the polarization-dependent scattering spectra. The measurements were conducted at an interval of 10 degree of β.
Figure S7. Lasing from the RL device based on TiO2 nanorod array. TiO2 nanorod array was prepared by the GLAD approach and then the RL device was fabricated and characterized in the same way as the device based on silver nanorod array. (a) SEM image of the TiO2 nanorod array, the length of the nanorods is ~700 nm on average. (b) Emission spectra recorded at various pump energies of a picosecond pulsed pump laser (λ = 532 nm). The pump spot size is ~148.0 μm in diameter. (c) Integrated intensity versus the pump energy. The arrow indicates the lasing threshold. (d) Normalized peak intensity versus the polarization of the pump laser. No obvious polarization dependence is observed. The pump energy is ~0.88 μJ. The decrease of the emission signal after β = 40° is ascribed to the photo-bleaching effects of the dye molecules.
Figure S8. (a) SEM image of the test mark embedded with “PURDUE” characters in an 80 nm-thick gold film. The enlarged figure shows that the smallest feature on the test mark is ~3 μm. The scale bar is 100 μm. (b) Images of ‘PURDUE’ test mark under the illumination of the 532 nm laser. Speckles are clearly observed.
Figure S9. The thickness of the nanorod film layer versus the nanorod length.
References 1. 2.
Meng, X. G., Fujita, K., Murai, S., Motaba, T., Tanaka, K. Nano Lett. 2011, 11, 1374-1378. Chen, W. Q., Thoreson, M. D., Ishii, S., Kildishev, A. V., Shalaev, V. M. Opt. Express 2010, 18, 5124-5134.
