arXiv subjects

Shravan Raghunathan

Publications and source records attributed to Shravan Raghunathan.

2 recordsLinked to original sources

arXiv2026

Optically dense nanowire metamaterials are transparent to polarization

We study the transport of light through dense opaque anisotropic metamaterials consisting of oriented nanowires. The nanowires consist of polymer photoresist that is structured by direct laser writing (DLW) with two-photon induced polymerization, with radii between $a = 0.5$ and $1~\mu \text{m}$. Our flat samples have a thickness up to 9 layers, from $L = 3~\mu \text{m}$ to $20~\mu \text{m}$. Within each layer, the nanowires are parallel and spaced with random nearest-neighbor distances; nanowires in adjacent layers are perpendicular. The diffuse optical transmission at $\lambda = 633~$nm is as low as $T = 12 \%$, typical of optically dense, multiple scattering metamaterials, with a mean free path down to $\ell = 1.1~\mu \text{m}$, much less than the sample thickness. It is striking that the linear polarization of the input light is maintained at the output of the dense nanowire samples, and not scrambled as in dense nanosphere arrays. Moreover, the linear output polarization faithfully tracks the input polarization. We propose that the polarization is maintained in our optically thick samples, since light is predominantly transported perpendicularly to the nanowire layers. The polarization vector then lies in the nanowire plane, consisting of a linear combination of parallel and perpendicular vectors that are both conserved upon subsequent scattering. Hence, the polarization remains independent of nanowire orientation, even after multiple scattering events. We propose that anisotropic scattering samples may find practical uses in white LEDs and its applications in lighting luminaires, optical communication, and encryption systems.

physics.optics↗

arXiv2026

Denoising and Baseline Correction of Low-Scan FTIR Spectra: A Benchmark of Deep Learning Models Against Traditional Signal Processing

High-quality Fourier Transform Infrared (FTIR) imaging usually needs extensive signal averaging to reduce noise and drift which severely limits clinical speed. Deep learning can accelerate imaging by reconstructing spectra from rapid, single-scan inputs. However, separating noise and baseline drift simultaneously without ground truth is an ill-posed inverse problem. Standard black-box architectures often rely on statistical approximations that introduce spectral hallucinations or fail to generalize to unstable atmospheric conditions. To solve these issues we propose a physics-informed cascade Unet that separates denoising and baseline correction tasks using a new, deterministic Physics Bridge. This architecture forces the network to separate random noise from chemical signals using an embedded SNIP layer to enforce spectroscopic constraints instead of learning statistical approximations. We benchmarked this approach against a standard single Unet and a traditional Savitzky-Golay/SNIP workflow. We used a dataset of human hypopharyngeal carcinoma cells (FaDu). The cascade model outperformed all other methods, achieving a 51.3% reduction in RMSE compared to raw single-scan inputs, surpassing both the single Unet (40.2%) and the traditional workflow (33.7%). Peak-aware metrics show that the cascade architecture eliminates spectral hallucinations found in standard deep learning. It also preserves peak intensity with much higher fidelity than traditional smoothing. These results show that the cascade Unet is a robust solution for diagnostic-grade FTIR imaging. It enables imaging speeds 32 times faster than current methods.

eess.IV↗