Bridging Theory and Experiment in Virtually Imaged Phased Array (VIPA) Spectrometers

TL;DR

The paper addresses the gap between ideal VIPA resolving power predictions and real-world performance by analyzing the impact of fabrication tolerances, aberrations, alignment, and coatings. It combines analytical VIPA models with non-sequential ray-tracing simulations and experimental validation in the mid-infrared to identify limiting factors and optimize a cross-dispersed VIPA spectrometer operating near 4.6 μm. The study predicts an ideal VIPA RP of about 830000 based on theory, but realistic simulations and measurements yield RP around 440000 (about 80% of the simulated limit) and demonstrate comb-resolving capability with a 250 MHz repetition-rate comb. The results provide practical design pathways for high-resolution, compact spectrometers with applications in space optics, line-by-line pulse shaping, and broadband spectral sensing.

Abstract

Virtually imaged phased array (VIPA) spectrometers provide high resolution and fast acquisition in a compact design, but their performance as dispersive instruments is sensitive to fabrication tolerances, component dimensions, and alignment. Here, leveraging numerical simulations validated by experimental data, we present a framework to identify the parameters that limit VIPA spectrometer resolution. This framework is applied to the construction of a new mid infrared VIPA spectrometer, tested at wavelengths near 4.6 um with both continuous-wave and frequency-comb laser sources, with a resolving power predicted by analytical expressions to be as high as RP = 830 000 (corresponding to a resolution of 78 MHz). Validated numerical simulations, however, provided a more realistic estimate that captures limits set by all the optical components. By correcting aberrations and optimizing alignment, a resolving power of RP = 440 000 (150 MHz) was experimentally achieved, corresponding to 80% of the value predicted by numerical simulation of the entire spectrometer. These results bridge the gap between analytical design expressions and experimental results for compact, high-resolution VIPA spectrometers to enable more efficient fabrication and advanced design across critical areas like applied space optics, line-by-line pulse shaping, and broadband spectral sensors.

Figures (6)

• Figure 1: (a) Schematic of virtually imaged phased array (VIPA) spectral disperser and the underlying operation mechanism. (b) Comparison of VIPA outputs for the "idealized" analytical model and "realistic" numerical simulations. (c) A magnified view of (b), highlighting the ideal Lorentzian line shapes based on both analytical expressions and ray-tracing numerical simulations. (d) Numerical simulation results showing the achievable resolution as a function of the VIPA–detector distance and VIPA height, comparing an ideal Fourier-transform (FT) case with a 200 mm focal length imaging lens.
• Figure 2: (a) Experimental setup and (b) layout for numerical simulation. (c,d) Detector images from experiment and simulation. (e,g) Cross-sectional profiles of the 2D images highlighting Lorentzian linewidth. (f,h) Zoomed-in view of the central peak with the corresponding Lorentzian fit.
• Figure 3: The effect of imaging lens focal length at fixed lens diameter on the VIPA resolving power. Numerical simulations (blue lines) were performed with (a) $f_\text{im}$ = 200 mm and (b) $f_\text{im}$ = 1000 mm imaging lenses and are compared to experimental results (red dashed lines). Peaks from the VIPA output were compared for both the numerical simulations and experiments for (c) $f_\text{im}$ = 200 mm and (d) $f_\text{im}$ = 1000 mm. (e-g) Schematics of spherical aberration from short focal-length, long focal-length, and ideal aberration-free lenses. The experimental resolving powers achieved are listed in the table in the image, along with corresponding values from numerical simulations.
• Figure 4: (a) Schematic of the angular dispersion measurement, along with the detector image captured for wavelengths ranging from (i) 4601.0 nm to (iv) 4601.5 nm. (b, c) Experimental and numerical simulation results show the laser wavelength as a function of the VIPA output angle, overlaid with the corresponding fit based on Eq. (4).
• Figure 5: The mid-infrared VIPA spectrometer under frequency-comb illumination. (a) Experimental two-dimensional spectrogram showing broadband spectrometer output; the red dashed line marks an extracted region for analysis. (b–c) Experimental line-out along the boxed region: (b) raw intensity profile with individual frequency-comb modes (red dots), (c) normalized peak profiles fitted with Lorentzian functions, yielding an average FWHM of 7.1 pixels $\pm$ 1.4 pixels and a mean peak spacing of 14.5 pixels $\pm$ 0.9 pixels. (d) Numerically simulated spectrogram under equivalent conditions. (e–f) Numerical simulation line-out: (e) raw intensity profile with detected peaks, (f) normalized peak profiles fitted with Lorentzian line-shape functions, yielding an average FWHM of 5.7 pixels $\pm$ 0.8 pixels and a mean peak spacing of 11.6 pixels $\pm$ 1.2 pixels.