Active plasmon injection scheme for subdiffraction imaging with imperfect negative index flat lens

Abstract

We present an active physical implementation of the recently introduced plasmon injection loss compensation scheme for Pendry's non-ideal negative index flat lens in the presence of realistic material losses and signal-dependent noise. In this active implementation, we propose to use a physically convolved external auxiliary source for signal amplification and suppression of the noise in the imaging system. In comparison with the previous passive implementations of the plasmon injection scheme for sub-diffraction limited imaging, where an inverse filter post-processing is used, the active implementation proposed here allows for deeper subwavelength imaging far beyond the passive post-processing scheme by extending the loss compensation to even higher spatial frequencies.Although this form of passive inverse filter provides compensation for absorption losses, it is also prone to noise amplification [56]. This is illustrated in figure 1 which shows an object with three Gaussian features separated by λ o /4, where λ o is the free space wavelength. Noise is prominent in the Fourier spectra beyond ky ko = 2.5 as seen in figure 2. However, the compensated image is still reasonably well resolved. Consider now the object shown in figure 3, which has four Gaussians separated by λ o /4. The Fourier spectra of the raw image, shown in figure 4, demonstrates how the feature at

Enhanced superlens imaging with loss-compensating hyperbolic near-field spatial filter

“…It is evident that the silver lens is incapable of resolving the object shown by the black line in figure 2 Figure 2(c) shows how the losses and noise start to progressively degrade the image spectrum from k y ≈ 2k 0 , where k 0 is the free-space wavenumber. Eventually, the image spectrum is completely overwhelmed by the noise and cannot be recovered by passive deconvolution or in- creased illumination intensity since both processes proportionally amplify noise [18]. In conclusion, a selective amplification with the auxiliary must be initiated from k y ≈ 2k 0 and progressively moved to higher spatial frequencies.…”

“…Hence, the selective amplification process can be applied till k y = 7k 0 . One important reason for the failure of the numerically calculated transfer function is the finite spatial extent of the image and object planes which introduces errors in the Fourier transform calculations as discussed in [18].…”

Plasmonic Superlens Imaging Enhanced by Incoherent Active Convolved Illumination

“…To push the performance of our compensation method further, we developed an active version which relies on the coherent convolution of a high spatial frequency function with an object field focused by a lossy NIFL. 52 Selective amplification of a small band of spatial frequencies by spatially filtering the object under a strong illumination beam can favorably alter the transfer function so that spatial frequency components which would originally be lost to noise can be successfully transferred to the image plane. In this article, we present a more advanced and versatile method that alternatively employs a simple plasmonic superlens structure illuminated by incoherent UV light, avoiding the complexity and practical difficulties related to phase retrieval or phase detection of coherent fields.…”

Hyperbolic Metamaterial as a Tunable Near-Field Spatial Filter to Implement Active Plasmon-Injection Loss Compensation

“…In conclusion, we have proposed a near-field spatial filter for the active implementation [23] of the recently introduced Π loss compensation scheme [18]. The "tunability" and "selective amplification" characteristics of the auxiliary source in [23] can be realized with layered metal-dielectric systems with hyperbolic dispersion. We have demonstrated that the convolution, which is vital for the construction of the auxiliary source, can be achieved in the layered system.…”

“…In [23], the auxiliary source is defined as the convolution of the object field with a function whose Fourier transform, P (k y ) = F {P (y)}, is a Gaussian. In the reciprocal space, the product of the object spectrum with this Gaussian represents the amplification provided to a band of Fourier components.…”