Abstract
Interaction between metal surface waves and the periodic geometry of subwavelength structures is at the core of the recent renewal of interest in plasmonics. One of the most intriguing observations is Extraordinary Optical Transmission (EOT)—abnormal strong transmission through periodic structures that can exceed the classical limits defined by the filling factor of the plate by orders of magnitude. The current paradigm suggests that this transmission arises strictly from periodicity, implying that such effects should vanish in random geometries. This paper challenges that assumption, demonstrating that extra energy can be coupled through subwavelength structures by adding a controlled quantity of noise to the position of the apertures. This phenomenon is modeled within the statistical framework of stochastic resonance, revealing that specific levels of disorder can actually enhance energy coupling at the surface of the array.
Introduction
In the field of plasmonics, the discovery of Extraordinary Optical Transmission (EOT) revolutionized our understanding of how light interacts with subwavelength apertures. The standard model for EOT relies on the constructive interference provided by a perfectly periodic lattice, which allows incident light to couple into surface plasmon polaritons (SPPs). These surface waves then re-radiate through the holes, resulting in transmission levels far higher than those predicted by standard aperture diffraction theory.
For decades, the consensus has been that any deviation from perfect periodicity—any introduction of spatial disorder—would inevitably degrade the resonance and reduce transmission. However, by applying the principles of stochastic resonance (SR), this research demonstrates a counterintuitive reality: a controlled amount of "noise" or disorder in the hole positions can actually increase the total energy coupled through the array, outperforming the perfectly ordered grid.
Experimental Methodology
To investigate the relationship between disorder and transmission, the researchers utilized the Terahertz (THz) domain. This spectral range is particularly advantageous because metals behave nearly as perfect conductors at these frequencies, and the physical dimensions of the arrays (hundreds of micrometers) allow for high-precision fabrication and characterization.
Sample Fabrication and Controlled Disorder
The researchers fabricated free-standing nickel plates featuring square hole arrays with periods ($L$) of 300 $\mu$m and 400 $\mu$m. To test the impact of disorder, they introduced "positional noise" by shifting the center of each hole from its ideal lattice position. This shift followed a Gaussian distribution with a standard deviation ($\sigma$).
The study involved a comprehensive dataset of experimental samples:
- 20 distinct plates for the $L = 300$ $\mu$m period.
- 30 distinct plates for the $L = 400$ $\mu$m period.
The noise levels ranged from $\sigma = 0$ (perfectly periodic) to $\sigma = 60$ $\mu$m (significant disorder), allowing for a granular observation of how transmission evolves as a function of randomness.
Measurement and Modeling
Transmission spectra were measured using Terahertz Time-Domain Spectroscopy (THz-TDS). This technique provides a high-resolution look at the frequency-dependent transmission characteristics of each plate. To interpret the results, the authors utilized the framework of Array Enhanced Stochastic Resonance (AESR). This model, often used in signal processing and biological systems, describes how a system's response to a weak coherent signal can be optimized by the presence of a specific intensity of noise.
Key Insights: The "Sweet Spot" of Disorder
The experimental results revealed a non-monotonic relationship between disorder and transmission. Rather than a steady decline, the transmission energy followed a curve that peaked at an "optimal noise level" ($\sigma_M$).
Stochastic Resonance in Optics
By mapping the Fano model—which describes the interference between a continuum of states and a discrete resonant state—onto the AESR framework, the authors identified that noise helps "bridge" the energy coupling between incident light and surface waves. In a perfectly rigid periodic structure, the coupling is restricted to specific Bragg conditions. Introducing slight disorder appears to unlock a hidden reservoir of energy coupling that is otherwise inaccessible, effectively broadening the resonance conditions in a way that captures more incident energy.
Quantitative Results
The findings provide clear evidence that periodicity is not the absolute ceiling for optical transmission:
- Transmission Boost: For the 300 $\mu$m array, the addition of optimal noise resulted in an 18.2% increase in transmitted energy compared to the perfectly periodic sample. The 400 $\mu$m array showed a similar, though slightly lower, boost of 11%.
- Optimal Noise Threshold: The peak transmission occurred when the disorder was approximately 3.4% to 3.8% of the array period. This suggests that while a little chaos is beneficial, the system eventually degrades as it approaches a fully random state.
- Loss Reduction: At the specific resonance frequency for the 400 $\mu$m array, the energy loss was reduced by a factor of 2.7 through the optimization of positional noise.
- Model Accuracy: The experimental data showed an excellent fit with a mathematical model combining a Lorentzian function (representing the stochastic resonance) and a sigmoid function (representing the transition to a disordered state).
Discussion and Future Implications
These results have significant implications for both fundamental physics and practical nanotechnology.
Manufacturing and Design
The discovery that sub-5% disorder can significantly alter (and even improve) transmission is a double-edged sword for nanofabrication. It implies that even minor manufacturing tolerances can have a disproportionate impact on device performance. Conversely, it suggests a new design paradigm: instead of striving for impossible perfection, engineers can intentionally design "noisy" or quasi-periodic structures to maximize energy throughput or enhance the sensitivity of plasmonic sensors.
Scaling and Generalization
While this study was conducted in the THz range, the underlying physics of stochastic resonance should theoretically apply to the visible and near-infrared spectra. However, at higher frequencies, the inherent plasma resonance of metals and material absorption may introduce additional variables. Beyond optics, the authors suggest that this logic could extend to 3D photonic crystals and even phonon propagation, potentially allowing for the control of heat and sound through engineered disorder.
Conclusion
This work fundamentally shifts the understanding of Extraordinary Optical Transmission. By demonstrating that stochastic resonance can be harnessed to boost energy coupling in subwavelength arrays, the researchers have shown that periodicity is not the only path to high-efficiency plasmonic devices. In the delicate balance between order and chaos, a controlled amount of noise provides a powerful tool for enhancing light-matter interaction.
Summary updated