Phase-Labeled Spider in Photonic Interferometry

• Phase-labeled Spider is a dual-domain technique that directly measures both amplitude and phase for ultrafast optical pulse characterization and interferometric imaging.
• It employs SPIDER in integrated photonic circuits and degenerate four-wave mixing to reconstruct complex fields without mechanical delay scanning.
• Advanced sparse reconstruction and calibration methods enable high-fidelity imaging, bridging optical and radio-interferometric techniques for diverse applications.

A phase-labeled spider encompasses two major domains: ultrafast optical pulse characterization using Spectral Phase Interferometry for Direct Electric-Field Reconstruction (SPIDER) in integrated photonic platforms, and direct phase-labeled optical interferometric imaging realized by the Segmented Planar Imaging Detector for Electro-optical Reconnaissance (SPIDER) telescope. In both contexts, the defining feature is the capability to record and process both amplitude and phase information—integrally and directly—enabling precise field reconstruction and high-fidelity imaging beyond the limits of traditional amplitude-only modalities.

1. Principle of Phase Labeling in SPIDER Modalities

Phase labeling refers to direct measurement and association of the spectral or complex phase with each amplitude datum in an interferometric or spectroscopic context. In optical pulse characterization, the SPIDER methodology exploits sheared-spectral interferometry by mixing time-delayed replicas of the pulse-under-test (PUT) with a highly chirped pump, generating idler spectra that are frequency-shifted by a well-defined shear $\Omega$. The phase difference $\phi(\omega+\Omega) - \phi(\omega)$ is then extracted from the resulting interferogram, providing direct spectral phase labeling (Pasquazi et al., 2014).

In the SPIDER telescope architecture, phase labeling is achieved via integrated photonic circuits that maintain coherence across multiple Mach–Zehnder interferometers fed by single-mode waveguides. Each lenslet pair generates a complex cross-correlation, recording both amplitude and phase for each baseline. Accordingly, observables are measured in the form $V(u,v) = |V|\,e^{i\phi}$, rendering each datum phase-labeled (Pratley et al., 2019).

2. Integrated SPIDER for Ultrafast Optical Pulse Characterization

The CMOS-compatible SPIDER device utilizes degenerate four-wave mixing (FWM) in a 45 cm spiral Hydex waveguide. The process unfolds as follows:

• The PUT is split into two replicas with delay $\Delta t$.
• These replicas interact with a stretched pump $p(t)\approx \exp[-i\,t^2/(2\phi_P)]$ inside the $\chi^{(3)}$ waveguide, producing an idler $s(t)\propto e(t+\Delta t)p(t)+e(t)p(t)$.
• In the spectral domain, the output is $$S(-\omega)\propto E(\omega+\Omega)\exp[-i(\omega+\Omega)\Delta t]+E(\omega)$$ with $\Omega=2\Delta t/\phi_P$.
• The measured spectrum $|S(-\omega)|^2$ yields an interferogram $I(\omega)$ whose phase fringes encode the phase difference $\phi(\omega+\Omega)-\phi(\omega)$.

Phase recovery proceeds algebraically, concatenating these differences to reconstruct $\phi(\omega)$. Complete amplitude and phase characterization ($E(\omega)=A(\omega)\exp[i\phi(\omega)]$) is realized without mechanical delay scanning, enabling single-shot, ultrafast, phase-sensitive measurements at peak powers below 100 mW and bandwidths exceeding 1 THz (Pasquazi et al., 2014).

3. SPIDER Optical Interferometric Telescope: Architecture and Phase-Labeling

The SPIDER telescope design integrates a dense two-dimensional lenslet array (typical diameter $\sim$8.75 mm) atop a planar wafer. Each lenslet focuses light into a single-mode waveguide leading to an on-chip Mach–Zehnder interferometric network, which balances path lengths to preserve coherence. Unlike classical Fizeau or pupil-remapping instruments, which record intensity, SPIDER directly measures complex visibility (amplitude and phase) per baseline and per spectral channel.

Each measured visibility is phase-labeled, formalized as: $$V(u,v) = \int\int I(l,m)A(l,m)\exp[-2\pi i(ul+vm)]\,dl\,dm$$ and, upon discretization,

$$\mathbf{y} = \boldsymbol{\Phi}\mathbf{x} + \mathbf{n}$$

where $\mathbf{y}\in\mathbb{C}^M$ is the vector of phase-labeled visibilities and $\boldsymbol{\Phi}$ is the non-uniform FFT-based sampling operator. This approach enables radio-synthesis-like image reconstruction strategies in the optical regime (Pratley et al., 2019).

4. Sparse Reconstruction Algorithms and Convex Optimization

The phase-labeled visibility data enable sparse, high-fidelity image recovery via convex optimization methods. For $M\ll N$ (number of visibilities $\ll$ number of pixels), the imaging inverse problem is ill-posed. By exploiting compressibility in wavelet or gradient bases: $$\hat{\mathbf{x}} = \arg\min_{\mathbf{x}}\;\frac{1}{2\sigma^2}\|\mathbf{y}-\boldsymbol{\Phi}\mathbf{x}\|_2^2 + \gamma\|\Psi^\dagger\mathbf{x}\|_1$$ where $\Psi^\dagger\mathbf{x}$ are the transform coefficients and $\gamma$ regularizes sparsity.

A more robust constrained formulation is: $$\hat{\mathbf{x}} = \arg\min_{\mathbf{x}\geq 0}\;\|\Psi^\dagger\mathbf{x}\|_1\quad \text{subject to}\; \|\mathbf{y}-\boldsymbol{\Phi}\mathbf{x}\|_2\leq\epsilon$$ solved effectively by Alternating-Direction Method of Multipliers (ADMM), separating the data-consistency and sparsity terms (Pratley et al., 2019). This method precisely reconstructs both amplitude and phase, preserving asymmetric features and suppressing noise artifacts.

5. Calibration, Phase Referencing, and Self-Consistency

Phase-labeling in SPIDER demands calibration to mitigate instrumental phase offsets (thermal drifts, path-length errors). Calibration employs phase-reference sources (e.g., bright stars) with well-known visibilities, solving for per-channel offsets $\theta_k$ by minimizing the deviation between observed and model visibilities: $$\sum_k \|V^{\text{obs}}_k e^{-i\theta_k} - V^{\text{model}}_k\|_2^2$$ Calibrated visibilities $y_k=\Phi_k x + n_k$ are then consistent with true sky phases. Iterative self-calibration between updating $\theta_k$ and image reconstruction refines phase accuracy, directly analogous to radio-interferometric self-cal procedures. Neglecting measured phases ($\phi_k=0$) leads to constrained, artifact-ridden images, reinforcing the essential role of phase labeling (Pratley et al., 2019).

6. Comparative Advantages and Applications

Both ultrafast, phase-labeled SPIDER methods confer significant advantages over historical techniques:

• Direct joint phase and amplitude measurement in a single (spectral or spatial) shot.
• Compatibility with CMOS photonics, enabling integrated, compact, real-time devices.
• Telecommunications regime operation with peak powers $<100$ mW and time-bandwidth products $>100$.
• Absence of scanning delay lines, variable-delay spectrometers, or large free-space interferometers (Pasquazi et al., 2014).

Applications include:

• On-chip pulse metrology for ultrafast coherent optical communications, phase-noise characterization.
• Space-based optical imaging with high resolution, low mass, and fidelity surpassing traditional intensity-only interferometers.
• Employing mature radio-interferometric reconstruction toolkits (e.g., PURIFY, multi-wavelet dictionaries) for optical imaging—demonstrated by recovery of spiral arms and H II regions in simulated astronomical data (Pratley et al., 2019).

A plausible implication is that the future SPIDER optical instruments, via phase labeling and integrated photonic circuits, may extend the precision and computational efficacy of astronomical imaging, bridging radio and optical domains through direct field synthesis techniques.