Photonic Lantern-Fed Spectrometer

• Photonic Lantern-Fed Spectrometer is a device that converts multimode astronomical light into diffraction-limited single-mode outputs using an adiabatic taper, enabling precise spectral measurements.
• It employs advanced reformatting and integrated photonic techniques to enhance throughput, reduce modal noise, and improve coupling efficiency under turbulent conditions.
• Applications include high-resolution spectroscopy, precise wavefront sensing, and astrometry, particularly in systems utilizing adaptive optics and extremely large telescopes.

A photonic lantern (PL)–fed spectrometer is a device architecture that utilizes a photonic lantern to efficiently interface multimode astronomical light to one or more single-mode spectrographs. The PL enables the reformatting of seeing- or turbulence-limited, multimode input—delivered via direct fiber injection or a focal-plane coupling—into a set of diffraction-limited single-mode waveguides. This transition is critical for leveraging compact, high-stability integrated photonic spectrographs and for executing advanced spatial and spectral measurements. PL-fed spectrometers are studied as solutions for overcoming the scaling, throughput, and modal noise limitations associated with classical fiber-fed or bulk-optic astronomical spectrographs, especially in the context of adaptive optics–corrected or Extremely Large Telescopes (ELTs).

1. Photonic Lantern Operation and Optical Principles

A photonic lantern is an adiabatic waveguide transition that enables efficient conversion between a multimode fiber or waveguide (MMF) input and a bundle of single-mode fibers or waveguides (SMF) outputs. In astronomical PL-fed spectrometers, the MMF input is matched to the étendue and modal content of the seeing- or AO-corrected point spread function. The number of supported spatial modes $N_\text{MM}$ at the input is approximately preserved along the taper and mapped to the same number of SMF outputs to ensure minimal loss:

$N_\text{MM} \approx N_\text{SM}$

The success of the transition is determined by both the geometry of the taper and the matching of the number of modes.

The single-mode outputs can be arranged into a linear pseudo-slit, remapped into an array for multiplexed processing, or directly coupled to integrated photonic dispersers such as arrayed waveguide gratings (AWGs) (Harris et al., 2012, Harris et al., 2014, 1311.0578). The adiabatic nature of the transition minimizes coupling and mode-mismatch losses, routinely achieving transmission of $75$–$93$\% under optimal conditions.

Relevant performance metrics include:

• Coupling Efficiency: Fraction of incident flux coupled from the telescope to the lantern, frequently measured both in the laboratory (e.g., $51\% \pm 10\%$ mean, up to $80\%$ peak (Vievard et al., 11 Sep 2024, Vievard et al., 22 Jul 2024)) and on-sky ($14$–$43$\% under 1" seeing with large tip/tilt residuals).
• Slit Reformatting Efficiency: Determines how well the output pseudo-slit matches the spectrograph input (can significantly reduce modal noise (Pike et al., 2020)).
• Transition Losses: Typically $<5\%$ per transition in optimized laser-written devices (1311.0578).

2. Integration Strategies and Instrument Architectures

PL-fed spectrometers are typically realized in two main forms:

A. Conventional Spectrograph Feeding (Semi-Photonic):

The SMF outputs are reformatted into a diffraction-limited slit and coupled into a conventional (bulk optics) dispersive spectrograph. This method acts as an optical image-slicer at the modal level, preserving spatial resolution and optimizing detector use (Harris et al., 2012).

B. Fully Integrated Photonic Spectrographs:

Each SMF output is delivered to an individual or shared integrated disperser (such as an AWG). This enables an ultra-compact, fully photonic system in which the spectral resolution is given by

$R = m \cdot N_{\text{wg}} \cdot C$

where $m$ is diffraction order, $N_\text{wg}$ is the number of waveguides, and $C$ is a fabrication quality factor. This approach is scalable and allows for detector reduction schemes by combining outputs onto a single linear detector, adapted to the AWG's focal surface curvature (Harris et al., 2012).

A key challenge is the redundancy introduced in modal decomposition: the number of SMF channels per spaxel scales as $(\chi D_T)^2$ (field angle × telescope diameter), impacting both component count and detector usage.

Integration with advanced AO systems or phase-induced amplitude apodization (PIAA) optics further boosts effective throughput—PIAA optics can improve SMF/PL coupling, particularly in systems supporting only a few modes (Lin et al., 2021).

3. Performance, Modal Noise, and Throughput Considerations

PL-fed spectrometers offer advantages over traditional MMF or direct SMF feeds:

• Throughput Enhancement: PLs can achieve significantly higher coupling efficiencies in turbulent or low-Strehl environments compared to SMF feeds. In practical on-sky conditions where residual tip/tilt is substantial, SMF injection is very inefficient (e.g., 18 mas FoV at 700 nm), whereas a 19-port PL maintains substantial coupling (up to 43\% at 680 nm in 1″ seeing (Vievard et al., 11 Sep 2024)).
• Reduced Modal Noise: The conversion to multiple SM outputs and pseudo-slit reformatting dramatically suppresses spectrograph modal noise, with reductions by factors of up to six compared to MMF, approaching the performance of a SMF under broadband coupling (Pike et al., 2020).
• PSF Stabilization: Certain architectures exploit the spatial signatures in the output ports for real-time wavefront sensing and feedback or tip/tilt stabilization (Corrigan et al., 2018, Lin et al., 2023).
• Detector Optimization: Compact pseudo-slits and efficient remapping of modal outputs can reduce the physical extent of the entrance slit (e.g., 2400 μm → 240 μm (1311.0578)) and thus the required number of detector pixels.
• Scattered Light and Spectral Purity: PL-fed instruments achieve very low levels of scattered light (≪1\%) thanks to the clean Gaussian PSFs of SMF outputs, supporting high-dynamic-range and high-resolution spectroscopy (Betters et al., 2013).

However, PL-fed systems can be affected by modal noise arising from modal mismatch at MMF-to-lantern interfaces. Wavelength-dependent coupling fluctuations up to 20% (peak-to-peak) have been observed and studied; these can be mitigated by direct lantern injection or MMF agitation to average speckle effects (Cvetojevic et al., 2017).

4. Advanced Capabilities: Wavefront Sensing and Spectroastrometry

PL-fed spectrometers enable advanced measurement modalities not accessible to classical designs:

• Focal-Plane Wavefront Sensing (PLWFS):

By exploiting the sensitivity of each output channel to low-order aberrations (e.g., tip, tilt, astigmatism, petaling modes), the PL serves as a focal-plane WFS. The response can be described using transfer matrices capturing mode intensity variations:

$p_{\text{out}} = |A u_{\text{in}}|^2$

where $A$ is the system transfer matrix and $u_\text{in}$ encodes the input phase. Linear and quadratic models are used for reconstruction (Lin et al., 2022, Lin et al., 2023). On-sky, real-time demonstration of petaling mode and Zernike aberration correction has reached $\sim$95\% WFE reduction and $10\times$ dynamic error rejection at 1-Hz (Lin et al., 2023). Spectral dispersion of outputs further increases the number of orthogonal modes that can be probed simultaneously (Lin et al., 1 May 2025).

• Spectroastrometric Precision and Subdiffraction Imaging:

By measuring the wavelength-dependent distribution of intensities in the output SMFs, PL-fed spectrometers can recover two-dimensional spectroastrometric signals, offering centroid precision well below the classical diffraction limit. New calibration strategies using simultaneous PSF imaging and spectral-differential self-calibration have delivered on-sky Hα photocenter precision of 50 μas in 10 min with a single telescope—a factor of several beyond the Rayleigh limit (Kim et al., 22 Oct 2025). The underlying measurement equations tie the output intensity vector to the centroid shift:

$$I_n(\lambda) \approx I_{n,0}(\lambda) + B_n(\lambda) \cdot \boldsymbol{\alpha}_{\text{centroid}}(\lambda)$$

with $B_n$ a wavelength-dependent response matrix. Sinusoidal variation in tip–tilt sensitivity with wavelength has been experimentally validated (Kim et al., 4 Nov 2024), and coupling maps are employed for calibration and recovery of pointing errors.

5. Instrument Design Trade-offs and Optimization

Key design parameters and trade-offs include:

• Number of SM Outputs ($N_\text{SM}$): Higher $N_\text{SM}$ increases modal acceptance but requires more detector area and backend dispersers.
• Taper Profile and Length: Controls radiative mode losses and broadband performance. Longer lead-in sections can allow weakly guided modes to attenuate, but excess length may impose throughput penalties (Lin et al., 2021).
• Beam-shaping Optics: Integration with PIAA optics can boost coupling in low–mode-number PLs (up to $25$–$30$\% gain), but the effect diminishes for lanterns supporting more than three modes.
• Detector Technology: The tension between fast, low-noise detectors for wavefront sensing and high-dynamic range, slow detectors for spectroscopy remains an open challenge (Lin et al., 1 May 2025).
• Operating Wavelength: The modal response and tip–tilt sensitivity shows chromaticity—phase differences among PL supermodes introduce oscillatory astrometric sensitivity with wavelength, requiring careful PL geometry and calibration (Kim et al., 4 Nov 2024, Kim et al., 13 Sep 2024).

Redundancy due to decomposing seeing-limited focal spots into many modes increases system complexity, detector and component count. Optimized calibration and instrument design (e.g., stacking AWG outputs for 1D detector compatibility (Harris et al., 2012)) as well as the use of few-mode PLs for moderate throughput can alleviate some of this burden.

6. Applications, Scientific Impact, and Future Directions

PL-fed spectrometers support a diverse array of applications:

• Compact, Stable, High-Resolution Spectrographs: Especially for AO-corrected or diffraction-limited inputs, where the size and complexity of bulk-optic instruments become prohibitive (Harris et al., 2012).
• Extreme Precision Spectroscopy: Modal noise suppression and output stability make PL-fed systems suitable for radial velocity measurements of M-dwarfs and exoplanet searches (Pike et al., 2020).
• High-Angular Resolution and Astrometry: Subdiffraction photocenter precision ($\sim$50 μas) for disk kinematics, spectroscopic binaries, and inner disk imaging (Kim et al., 22 Oct 2025, Kim et al., 2023, Kim et al., 13 Sep 2024).
• Wavefront Sensing for AO and Nulling: Integrated PLWFS demonstration for petaling, Zernike, and non-common-path errors in high-contrast imaging, with near-nanometric performance anticipated (Lin et al., 2023).
• ELT Instrumentation: The scaling properties and phase-stable output of PLs enable efficient and robust coupling to backend photonic interferometers and spectrographs for ELTs (Kim et al., 2023).

Directions for further research include the development of mode-selective PLs for enhancing measurement orthogonality, improvement of spectral calibration and removal of instrumental fringe structure, extension to broadband and visible-wavelength operation, and the integration of on-chip photonic processing for even more compact instrument footprints.

7. Limitations, Calibration, and Technological Prospects

Key limitations and areas for further refinement include:

• Modal Mismatch and Chromaticity: Wavelength-dependent coupling and modal crosstalk can impose calibration requirements; spectral self-calibration and auxiliary PSF imaging are deployed to mitigate time-varying errors (Kim et al., 22 Oct 2025, Kim et al., 4 Nov 2024).
• Dynamic Range and Environmental Effects: The linearity of the response matrix is maintained only for moderate phase excursions (typically $\lesssim$0.25 rad); neural network or nonlinear inversion methods are under paper for extended range (Romer et al., 21 Jul 2025).
• Instrumental Stability: Long-term drift of PL response matrices due to mechanical or thermal changes is observed ($\sim$2\% day$^{-1}$); periodic recalibration mitigates this effect in practical systems (Lin et al., 2023).
• Hardware Constraints: Detector limitations (e.g., overlapping spectral traces with large $N_\text{SM}$, readout speed) may require trade-offs or hybrid solutions for integrated wavefront sensing and spectroscopy (Lin et al., 1 May 2025).

Continued advances in photonic device fabrication, calibration techniques, and photonic integration are anticipated to further extend the capabilities and adoption of PL-fed architectures in high-precision astronomical instrumentation and beyond.