1. Introduction

The ultrafast optical excitation of nonlinear optical materials and semiconductors is widely used as a source for electromagnetic radiation in the terahertz (THz) range [1–3 ]. Recent research focuses on the development of suitable materials or structures that can achieve high conversion efficiency from the exciting laser’s energy to THz energy. Plasmonics offers the prospect of the subwavelength confinement of light and the associated enhancement of the electromagnetic field strength [4–6 ]. Several research groups have used the femtosecond laser irradiation of metal surfaces to generate THz radiation; in these cases, thin metal films are typically deposited onto dielectric substrates with periodic structures [7–12 ]. Moreover, it has been reported that the thermal emission of light from metals is significantly improved by femtosecond laser-induced surface structures [13]. Our previous research demonstrated high-power terahertz-to-infrared (THz-to-IR) thermal radiation through the femtosecond laser excitation of metal films with randomly arranged nanoscale pore arrays [14].

In this letter, we focus on THz-to-IR emission from different metal surface structures by modifying the thickness of porous metal films. The thermal emission in both directions of the laser beam axis (i.e., forward and backward) was observed. The intensity ratio between the backward and forward radiation is exponentially dependent on the metal film thickness. The bi-directional thermal emission is in agreement with the modeled behavior of surface plasmon polaritons propagating simultaneously on both metal layer interfaces (air/metal and metal/substrate).

2. Sample description

A commercially available 60 μm thick anodic aluminum oxide (AAO) membrane (Whatman, Germany) serves as the substrate of our sample, which consists of randomly arranged pore arrays. The average pore diameter is approximately 200 nm and the pore density is 50%. Ruthenium (Ru), a noble metal, is deposited on the substrate through magnetron sputtering, thereby creating a metal film with a nominal thickness. As a result of the through-pore arrays of the AAO membranes, the deposited metal film acquires a nanoscale porous roughness.

In order to modify the metal surface structures, the metal films are prepared with different nominal thicknesses (e.g., 10 nm, 30 nm, 50 nm, 70 nm, 100 nm, 150 nm, 200 nm, and 300 nm). The morphology of the metal surface is observed by scanning electron microscope (SEM). Figure 1(a) shows representative images of the metal films with thicknesses of 30 nm, 100 nm, and 200 nm. The images show that the metal surfaces have a variety of nanoscale structures, including nanoscale voids and nanoprotrusions. Moreover, with increasing thickness, the average diameter of the nanoscale voids in the metal film decreases.

 

Fig. 1 (a) Scanning electron microscope (SEM) images of porous metal films with thicknesses of 30 nm, 100 nm, and 200 nm. The darkest regions indicate the absence of metal. The scale bar is 500 nm. (b) Optical absorptivity as a function of the thickness of the porous metal films.

Download Full Size | PDF

The surface roughness of the metal film, including the size and shape of the metal nanoparticles and the dielectric environment – governs the plasmon resonance excitation [15]. The optical properties of the nanostructured metal films are therefore influenced by the excitation of surface plasmons. The optical absorptivity ($A$) of the metal surface consists of two components:

A Ru thin film with a thickness of 150 nm was also deposited on a polished sapphire crystal substrate with magnetron sputtering. This film was very flat, and the Ru evenly distributed itself on the substrate. The optical absorptivity of the flat metal film maintained a value of approximately 44.5% across the surface. Figure 1(b) depicts the measured optical absorptivity for a wavelength of 800 nm as a function of the thickness of the porous metal film. It is obvious that the absorptivity of the nanostructured metal films is dramatically enhanced because of plasmonic effects. Moreover, the surface roughness that corresponds to the 100 nm thick metal film results in the strongest absorptivity.

3. Experimental results

A fraction of the absorbed pulse energy retains in the heat-affected zone, dissipates into the bulk sample because of heat conduction, and remains in the sample as residual thermal energy [18,19 ]. The remaining energy causes the bulk sample temperature to rise. The heated sample acts as a thermal radiation source when the bulk metal reaches thermal equilibrium.

In THz-to-IR radiation measurements, an amplified Ti:Sapphire laser provided pulses centered at a wavelength of 800 nm. The pulses had a duration of 100 fs with 1 mJ of pulse energy and a repetition rate of 1 kHz. The samples were placed in the beam’s path with the metal surface facing the incident beam. The optical beam was focused on the sample with a spot diameter of 6 mm. The THz-to-IR radiation was detected by a calibrated Golay cell equipped with a 6 mm diameter diamond input window (Microtech SN:220712-D). It had an approximately flat response over a broad spectral range (0–150 THz).

Figure 2(a) illustrates the setup used for measuring the incidence angle ($\theta$) dependence of THz-to-IR thermal radiation. In this configuration, the sample and detector were spatially fixed at an optimized position. Figure 2(b) shows the modified setup to measure the in-plane angular distribution of the radiation by rotating the Golay cell detector in the horizontal plane around the center of the sample. The quantity $\phi$ measures the angle between the sample-detector direction and the incident laser’s direction. The distance between the sample and detector was fixed. The power of the incident laser was 700 mW for all measurements.

 

Fig. 2 (a) Experimental setup for measuring the incidence angle dependence of terahertz-to-infrared (THz-to-IR) radiation. (b) Modified setup for measuring the angular distribution of THz-to-IR radiation.

Download Full Size | PDF

To demonstrate the involvement of surface plasmons in THz-to-IR emission, the incidence angle dependence of the radiation intensity for a p-polarized laser irradiated on the metallic nanostructures was measured using the setup shown in Fig. 2(a). The angle $\theta$ ranged from −60° to 60° with a step of 5° and with a normal incidence corresponding to an angle of 0°. The range was limited by the available space of the setup. Figure 3(a) shows the radiation intensity as a function of the angle of incidence for the metal films with different thickness. Each curve is symmetric and reaches a minimum at 0°. Maximum radiation intensity is observed around a resonant angle, which is similar to the angle that corresponds to the phase-matching condition for a grating [8]. The distribution of resonant angles appears to be somewhat wider than the surface plasmon absorption resonance. It is attributed to the variety of nanoscale structures in our roughened metal surface. Furthermore, the emission spectrum of the femtosecond pulse was broad relative to the monochromatic light. One may expect the excitation of surface plasmons to widen the distribution.

 

Fig. 3 (a) Measured terahertz-to-infrared (THz-to-IR) thermal radiation intensity as a function of the incidence angle emitted from nanostructured metal films with different thicknesses under p-polarized femtosecond laser irradiation. (b) THz-to-IR emission intensity at a normal incidence angle (red data points) and resonant angles (blue data points) as a function of metal film thickness. The lines are merely guides to the eye.

Download Full Size | PDF

Figure 3(b) shows the radiation intensity at an angle of 0° as a function of the metal film thickness. It is clear that the 100 nm thick sample – which was shown in Fig. 1(b) to have the strongest optical absorptivity – emits the maximum radiation intensity. This implies that higher absorptivity contributes to more efficient thermal radiation. The resonant angle as a function of the thickness is also plotted in Fig. 3(b). It is clear that the resonant angle is minimized for the thickness of 100 nm. The results can be explained in terms of the excitation of surface plasmons.

The resonant angle ${\theta }_{SP}$can be derived using the modified phase-matching condition:

Angular distributions of the porous metal films were then measured using the setup illustrated in Fig. 2(b). The optical laser hits the metal film at the normal incidence. For this configuration, thermal THz-to-IR emission was observed both in the direction of laser propagation and the reverse direction. Figure 4(a) displays representative results obtained with the thicknesses of 30 nm, 100 nm, and 200 nm. The angle ranges from −120° to 140° in steps of 5°, and is limited by the available space of the setup. The patterns of both the forward and backward radiation are circles tangent to the metal surface, satisfying Lambert’s cosine law.

 

Fig. 4 (a) Representative angular distribution of the terahertz-to-infrared (THz-to-IR) emission from the nanostructured metal films with thicknesses of 30 nm (magenta), 100 nm (blue), and 200 nm (green). The red arrow indicates the propagation direction of the optical laser. The solid lines are experimental data. The dashed lines are guides to the eye. (b) Top: the intensity of the forward (red data points) and backward (blue data points) emission as a function of the metal film thickness. Bottom: the intensity ratio between backward and forward radiation as a function of the thickness. The blue dots are experimental data. The red curve represents an exponential fit.

Download Full Size | PDF

The graph at the top of Fig. 4(b) shows the intensity of the forward and backward radiation as a function of metal film thickness. The forward radiation is always stronger than that of the backward direction, except for the sample with a 10 nm thick metal film, which displays almost equal intensities for the two directions. The maximum intensity of the forward radiation corresponds to the thickness of 100 nm. However, the maximum intensity of the backward radiation occurs around a thickness of 50 nm. There are two main factors that explain the observed optimal thickness for the forward and backward radiation directions. First, the interplay between the absorption of the emitted THz-to-IR waves by the metal film itself and the “gain” obtained by exciting the film with 800 nm pulses gives rise to an optimal THz-to-IR emission efficiency that occurs at certain film thickness in the transmission geometry. Second, in metal films with a thickness around the skin depth of the electric field, a strong coupling of the surface plasmons in both interfaces is possible, resulting in simultaneous thermal emission from both interfaces of the sample [20,21 ].

The graph at the bottom of Fig. 4(b) plots the intensity ratio between backward and forward radiation (which decays exponentially) as a function of the metal film thickness. We performed experimental data fitting using the following formula:

4. Conclusion

In conclusion, we demonstrated the involvement of surface plasmons in the dramatic enhancement of optical absorptivity and efficient THz-to-IR emission from nanoscale porous metal films that are irradiated by a femtosecond laser. We characterized the emission along the forward and backward directions of laser propagation. With the evidence that the intensity ratio of the forward and backward emission is exponentially dependent on the metal film thickness, we surmise that the strongest localized surface plasmon electric field (and, therefore, the most efficient thermal radiation source) exists at the metal/AAO interface.