🔬 optics

Refractive indices of Y $_2$ SiO $_5$ in the near-infrared

This paper presents independent interferometric measurements of the principal refractive indices of Y $_2$ SiO $_5$ in the near-infrared telecom range, providing a new Sellmeier model with explicit coefficients to fill the gap in existing data while also demonstrating precise thin-film thickness determination via white-light interference.

Original authors: Shijun Zhang, Jérôme Debray, Benoît Boulanger, Pierre Lemonde, Thierry Chanelière

Published 2026-06-26

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Shijun Zhang, Jérôme Debray, Benoît Boulanger, Pierre Lemonde, Thierry Chanelière

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine Y2SiO5 (Yttrium Orthosilicate) as a very special, high-tech window glass used by scientists to build the computers of the future—specifically, quantum computers that use light instead of electricity.

For a long time, scientists knew exactly how this "glass" bends light when the light is visible (like the colors of a rainbow). They had a perfect map for that. But when it comes to near-infrared light (the kind used in fiber-optic internet cables and invisible to the human eye), the map was missing. Scientists were trying to guess the map by stretching the visible one, but as this paper explains, that guesswork was leading them astray.

Here is the story of how the authors fixed the map, told in simple terms:

1. The Problem: A Broken Compass

The crystal is shaped like a slightly squashed box (scientists call this "monoclinic"). Because of this shape, light bends differently depending on which direction it travels through the crystal.

The old map (from 1990) worked well for visible light, but when the authors tried to use it to predict how the crystal behaves in the infrared (like at 1536 nm, a common internet wavelength), the map broke. It predicted that light would bend in impossible ways, almost as if the crystal had two different "optical centers" that swapped places. This was a red flag. The old map was like a GPS that works great in your city but gives you a route through the ocean when you drive to the next town.

2. The Solution: The Crystal as a Musical Instrument

Instead of building a giant, complex machine to measure the light, the authors used the crystal itself as a musical instrument.

The Setup: They took a small, clear block of the crystal (about the size of a small Lego brick) and shone a bright, broadband light (like a super-bright flashlight covering a wide range of colors) through it.
The Magic: Because the crystal faces are slightly reflective, the light bounces back and forth inside the block, like sound echoing in a hallway. This creates a pattern of "ripples" or interference fringes in the light spectrum.
The Measurement: Think of these ripples like the strings on a guitar. The distance between the ripples tells you exactly how long the path the light took inside the crystal. By measuring these ripples very precisely, the authors could calculate the group index—a specific way of measuring how fast a pulse of light travels through the material.

They did this for three different directions through the crystal, effectively measuring the "speed limit" for light in every direction.

3. The New Map: Drawing the Real Route

With these new, precise measurements in the infrared, the authors went back to the drawing board. They combined their new infrared data with the old visible-light data to create a brand new mathematical formula (called a Sellmeier equation).

The Fix: They corrected the old formula so that it makes physical sense. The old formula had a "ghost" term that made the light speed up weirdly in the infrared. The authors removed this ghost, ensuring the new map stays smooth and logical all the way from visible light to infrared.
The Result: They now have a reliable "GPS" for Y2SiO5 that works for the entire range of light used in modern quantum technology and telecommunications.

4. The Bonus Trick: Measuring a Hair's Width

To prove their new map was accurate, they turned the method around. Usually, you need to know the thickness of an object to measure its optical properties. But what if you know the optical properties (thanks to their new map) and want to find the thickness?

They took a thin film of this crystal (about as thick as a human hair, roughly 14 micrometers) bonded to a piece of glass. They shone white light through it and watched the interference ripples. Because they knew exactly how the light bends in this material, they could calculate the thickness of the film just by looking at the pattern of the ripples.

They measured the film to be 14.13 micrometers thick. This is like being able to measure the thickness of a single strand of hair just by listening to the echo of a sound wave bouncing off it.

Summary

The Issue: The old data for how this crystal bends infrared light was wrong and physically impossible.
The Method: They used the crystal as a mirror-box to create light ripples, allowing them to measure the speed of light inside it with high precision.
The Outcome: They created a new, accurate mathematical model that works for both visible and infrared light.
The Proof: They used this new model to measure the thickness of a microscopic film with extreme precision, proving their method works both ways.

This work ensures that scientists building future quantum technologies have a reliable map for navigating light through this critical material.

Technical Summary: Refractive Indices of Y2SiO5 in the Near-Infrared

Problem Statement
Yttrium orthosilicate (Y2SiO5) is a critical birefringent material for optical quantum technologies, particularly when doped with rare-earth ions like ytterbium (979 nm) and erbium (1536 nm). However, the principal refractive indices of Y2SiO5 in the dielectric frame (axes b, D1, and D2) are only approximately known in the visible spectrum. The primary reference, a 1990 study by Beach et al., provides measurements only between 430 nm and 650 nm using a heavily doped sample. Extrapolating these historical data into the near-infrared (NIR) and mid-infrared is scientifically unsound; the original Sellmeier model predicts a non-physical crossover of refractive indices around 1100 nm and relies on a positive mid-infrared pole coefficient that lacks physical justification. Furthermore, the monoclinic nature of Y2SiO5 introduces intrinsic limitations: the dielectric reference frame rotates with wavelength, making high-precision refractometry on fixed samples challenging.

Methodology
To address the lack of reliable NIR data, the authors employed a spectral interferometric technique, utilizing the Y2SiO5 crystal itself as a Fabry-Perot etalon.

Experimental Setup: A millimeter-sized, all-side-polished Y2SiO5 crystal (slightly doped with 50 ppm Er3+, treated as pure for refractive index purposes) was used. The crystal faces provided sufficient Fresnel reflection (~8%) to generate interference fringes.
Measurement: A broadband Er-doped fiber amplifier (EDFA) source (centered at 1550 nm) was collimated and passed through the crystal. The transmission spectrum was recorded using a high-resolution optical spectrum analyzer (0.01 nm resolution).
Data Processing: By rotating a half-wave plate and a polarizer, the authors probed the birefringence along different propagation axes (b, D1, D2). The free spectral range (FSR) of the observed interference fringes was extracted via Fourier transform of the spectral interferograms. This FSR yields the group index ( $N$ ), not the phase refractive index ( $n$ ).
Modeling: The authors fitted the group index data to a Sellmeier equation. To derive the phase refractive index, they assumed the difference between the group and phase indices ( $N - n$ ) is constant across the principal axes at 1550 nm. They also imposed a physical constraint by setting the mid-infrared pole coefficient ( $D_i$ ) to zero, as the measurement range is far from the Si-O-Si bond resonance (~10,000 nm).

Key Contributions and Results

Independent NIR Measurements: The authors determined the principal group indices at 1550 nm with an uncertainty of approximately $10^{-2}$ (dominated by systematic errors in sample length measurement). The values are $N_b \approx 1.784$ , $N_{D1} \approx 1.788$ , and $N_{D2} \approx 1.807$ .
Revised Sellmeier Coefficients: By combining the new NIR group index data with the historical visible spectrum measurements, the authors generated an extended set of Sellmeier coefficients valid from the visible to the near-infrared.
- The revised model eliminates the non-physical index crossover observed in the historical model.
- The coefficients $C_i$ (ultraviolet pole) are negative, consistent with physical expectations.
- The coefficients $D_i$ (mid-infrared pole) are set to zero.
- The root mean square deviation (RMSD) of the fit is $3.2 \times 10^{-5}$ , comparable to the historical fit but physically consistent.
Demonstration of Thin Film Thickness Measurement: As a practical application, the authors used the newly established refractive index model to measure the thickness of a Y2SiO5 thin film (~10–20 $\mu$ m) bonded to a glass substrate. Using white-light interferometry and a standard UV-Vis-NIR spectrometer, they determined the film thickness to be 14.13 $\mu$ m with nanometer-scale precision, validating the utility of the index model for characterizing integrated photonic structures.

Significance and Claims
The paper claims that while Y2SiO5 poses extreme challenges for traditional high-precision refractometry due to its monoclinic symmetry and the rotation of its dielectric axes, the proposed interferometric method provides a sufficiently accurate model for quantum technology applications.

Accuracy Limit: The authors acknowledge that the intrinsic accuracy of the index estimation is limited to approximately $1 \times 10^{-4}$ due to the wavelength-dependent rotation of the dielectric reference frame. They argue that achieving precision beyond this order of magnitude without measuring the frame orientation at every wavelength is unrealistic.
Utility: The derived Sellmeier coefficients are presented as a reliable tool for modeling Y2SiO5 in both the visible and near-infrared, covering the operational wavelengths of major rare-earth dopants (Nd, Yb, Er).
Methodological Value: The work clarifies the distinction between group and phase indices in interferometric measurements and demonstrates how group index data can be integrated into a Sellmeier model to extend spectral coverage.

The authors conclude that while further improvements (such as Kramers-Kronig linked absorption measurements in the UV and MIR) are possible, the current model represents a significant step forward in providing a physically sound description of Y2SiO5's optical properties for active material integration in quantum devices.

1. The Problem: A Broken Compass

2. The Solution: The Crystal as a Musical Instrument

3. The New Map: Drawing the Real Route

4. The Bonus Trick: Measuring a Hair's Width

Summary

More like this