Telecom-Band Optical Transitions of Erbium Atoms in Rare-Gas Solids

AI summary

Overview

Research area: Quantum photonics and solid-state quantum materials — specifically, matrix-isolation spectroscopy of neutral rare-earth atoms in cryogenic rare-gas solids for telecom-band spin–photon interfaces.

Technical level: Intermediate. The spectroscopy concepts (photoluminescence, phonon sidebands, zero-phonon lines, crystal-field splitting) are explained in the paper, but familiarity with atomic energy-level notation and rare-earth physics helps.

Scope (one sentence): The paper isolates neutral erbium atoms inside solid neon and solid argon at about 6 K and characterizes their near-infrared emission — including two telecom-band transitions near 1299 nm and 1437 nm — using steady-state and time-resolved photoluminescence spectroscopy, reporting line positions, linewidths, and excited-state lifetimes before and after thermal annealing.

What This Paper Is About

Rare-earth ions such as Er³⁺ are leading candidates for linking stationary qubits to optical photons, but the host crystal limits coherence through charge noise, magnetic impurities, and nuclear spins. The authors ask whether a chemically and magnetically "quiet" host — a van der Waals crystal of solid neon or solid argon — can preserve the narrow, shielded inner-shell transitions of a neutral erbium atom. Their goal is to establish the basic spectroscopy (which transitions appear, at what wavelengths, how broad, and how long-lived) as a foundation for future coherence and spin studies.

Key Contributions

1. First spectroscopic characterization of neutral Er atoms in rare-gas matrices. The authors state that, to their knowledge, Er atoms in rare-gas matrices had not previously been studied, and they report emission from the visible through the near-infrared (1000–1500 nm) in both solid Ne and solid Ar.

2. Identification of two contrasting telecom-band transitions. A predominantly 4f→5d transition near 1299 nm (labeled L13, telecom O band) and a shielded 4f→4f transition near 1437 nm (L14, telecom E band) are assigned and compared, showing sharply different host sensitivity and decay dynamics.

3. Quantitative line positions, matrix shifts, and linewidth bounds. The L13 multiplet is tabulated with ten components in Ar and eight in Ne, and the L14 line is shown to have a matrix shift below 1 cm⁻¹ in both hosts, with an inhomogeneous linewidth in Ne bounded by the spectrometer to <22 GHz (≈0.15 nm).

4. Lifetime measurements and annealing response. Decay times from microseconds (L13) to tens of milliseconds (L14) are measured, with L14 in Ne reaching up to 78.9 ms after annealing — close to the estimated radiative limit.

Main Findings

• Both hosts show narrow NIR emission between 1000 and 1500 nm. Several inner-shell transitions appear, including emission in the telecom O and E bands, following excitation at 532.2 nm.

• Ne is a gentler host than Ar. Features in Ar are typically broader and red-shifted relative to free-atom positions, while those in Ne are sharper and mostly lie close to the free-atom wavelengths. The authors relate this to the static dipole polarizability, α_Ar = 1.64 ų versus α_Ne = 0.40 ų — a factor of about four.

• The L13 multiplet near 1299 nm is host-dependent. In Ne, eight components span about 23 cm⁻¹, with the strongest component (Ne5) within 0.4 cm⁻¹ of the free-atom transition. In Ar, ten components spread over about 31 cm⁻¹ and are red-shifted by about 10 cm⁻¹. Because L14, which shares the same ground level, shows no comparably extended multiplet, the structure is attributed mainly to the 5d-derived upper level.

• L13 decays on a microsecond timescale. The Ar fit gives a mean decay time ⟨τ⟩ = 4.2 µs with stretching exponent β = 0.76; the Ne fit gives τ = 15.9 µs. These are about four orders of magnitude shorter than the free-atom value of 178(19) ms reported by Patscheider et al., indicating strong nonradiative quenching consistent with the 520 cm⁻¹ gap below the L13 upper level.

• L14 at 1437 nm is nearly host-independent. The emission maximum is 1436.74 nm in Ne and 1436.70 nm in Ar, corresponding to shifts of about −0.03 cm⁻¹ and +0.16 cm⁻¹ from the free-atom air wavelength of 1436.734 nm — both within spectrometer calibration accuracy.

• L14 is long-lived. The Ar fit gives ⟨τ⟩ = 11.4 ms with β = 0.64; the Ne fit gives τ = 44.2 ms, rising to as much as 78.9 ms after annealing. Both transitions decay more slowly in Ne than in Ar by a factor of about 4, and L14 is longer-lived than L13 by roughly three orders of magnitude.

• The L14 linewidth in Ne is resolution-limited. With 200 µm slits the measured width is about 0.24 nm (1.2 cm⁻¹); at higher resolution it narrows to about 0.15 nm (22 GHz), reported as an upper bound on the ensemble linewidth. At least two partially overlapping components (L14a and L14b) are resolved in both matrices.

• Emission intensity follows energy-gap structure, not free-atom line strength. L1 and L2 are about three orders of magnitude weaker than L6, because their upper levels sit only 108 cm⁻¹ and 50 cm⁻¹ above the next lower level, whereas the L6 upper level has a 662 cm⁻¹ gap. With maximum phonon energies of only about 50–65 cm⁻¹ in these hosts, multiphonon relaxation controls which levels emit.

• ³H₅ acts as the lower-manifold bottleneck. It lies 1923 cm⁻¹ above ³F₄ — a gap requiring about 30–40 phonons — so nonradiative decay is negligible and population accumulates there, giving rise to the long-lived L14 emission.

• The 890 nm region shows vibronic structure. In Ne, two overlapping series appear: one starting at 885.6 nm with spacing decreasing from about 24 to 15 cm⁻¹, and a second of five components between 891.4 and 898.0 nm with nearly constant 20.6 cm⁻¹ (2.6 meV) spacing, peaking at 896.5 nm. Different excitation spectra point to two distinct trapping sites, each coupled to a pseudolocalized vibrational mode. In Ar, a broad maximum at 893.8 nm with a weaker feature at 897.2 nm and a shoulder at 901.5 nm shows spacings of about 40–55 cm⁻¹ within the Ar phonon band.

• Annealing affects the two hosts differently. In Ar, annealing suppresses the phonon sidebands relative to the L11/L12 zero-phonon lines and increases L13 peak counts without shifting features. In Ne, annealing reduces NIR emission by about an order of magnitude, shifts the L11 band maximum from about 1027 nm to 1032 nm, and decreases the ZPL/PSB peak-height ratio by about 40%. After annealing, decay times vary by up to about 30% between positions on the same sample.

• A two-step excitation channel is identified. A feature near 1187–1190 nm appears under strong 532.2 nm excitation but not under weaker broadband excitation, suggesting absorption of a second 532.2 nm photon from the long-lived ³H₅ level, reaching about 25,740 cm⁻¹.

• No Er ion emission was found. Within the investigated spectral range and excitation conditions, the authors found no PL features that could be confidently assigned to Er ions.

Methodology in Plain English

The team grew their samples inside a closed-cycle optical cryostat with a base temperature of 4 K; the sample itself sits at 5–6 K because of heat load from the positioner stack, which allows three-axis translation and 360° rotation. They mounted a single-crystal sapphire substrate and simultaneously delivered two things onto it: a rare gas (neon or argon), which condenses into a solid, and erbium atoms, which are produced by firing a focused frequency-doubled Q-switched Nd:YAG laser (532 nm, 5 ns pulses, about 25 mJ) at a 99.99% pure Er foil. The erbium atoms get trapped individually in the growing solid — "matrix isolation" — which prevents them from clumping together.

For measurements, they used a continuous-wave 532.2 nm Nd:YVO₄ laser, or a supercontinuum source filtered by an acousto-optic tunable filter with about a 2 nm passband for wavelength-selective excitation between 450 and 700 nm. The same AOTF gates the excitation for lifetime measurements, with a fall time of about 1.2 µs. Light is focused to a roughly 0.1 mm spot, and the collected photoluminescence is dispersed by a 500 mm spectrograph with a 1200 grooves/mm grating and detected by a Si detector (visible) or an InGaAs detector (near-infrared), giving about 0.1–0.2 nm spectral resolution.

Decay curves were fitted with the simplest model that worked for each host: a single exponential for Ne and a stretched (Kohlrausch–Williams–Watts) exponential for Ar, the latter reported as a mean decay time computed from the fitted parameters. To test how the Matrix environment affects the emitters, they briefly heated samples to nominally 15 K (Ne) or 40 K (Ar) — near the sublimation regime — then rapidly cooled back to 6 K, comparing all spectra at the same 6 K measurement temperature.

Why This Matters

Impact on research: The work opens a new materials direction for telecom-band quantum emitters. Rare-gas solids combine high chemical purity, a nearly nuclear-spin-free lattice (only ²¹Ne carries nuclear spin, at 0.27% natural abundance, and all three stable Ar isotopes have zero nuclear spin), and weak polarizability. Demonstrating that a neutral Er atom's shielded 4f→4f transition survives nearly unshifted and with millisecond-scale lifetimes establishes the spectroscopic baseline needed before anyone can attempt optical coherence or spin measurements in this system. The 1437 nm transition sits in a fiber window with loss of about 0.3 dB·km⁻¹, making it directly relevant for fiber-based quantum networks.

Real-world applications (potential):

• Telecom-band single-photon sources for fiber-optic quantum communication
• Spin–photon interfaces and quantum repeater nodes operating at standard telecom wavelengths
• Hybrid quantum architectures linking photonic links to superconducting microwave circuits
• A testbed material for studying decoherence mechanisms in solid-state quantum emitters, since the host can be made almost free of nuclear spins and magnetic impurities

Industry relevance: Telecom-wavelength compatibility with existing fiber infrastructure lowers the engineering barrier for deployed quantum communication hardware. The ability to grow these matrices in a tabletop cryostat with a 4 K base temperature, rather than requiring dilution refrigeration, is relevant for systems engineering — though the 5–6 K sample temperature and the matrix-growth process remain practical constraints the paper does not attempt to solve.

Future Directions

• Optical coherence measurements. The reported inhomogeneous linewidth of <22 GHz in Ne is limited by the spectrometer, not the material. Higher-resolution techniques (such as spectral hole burning or photon echoes) are needed to determine the homogeneous linewidth and whether it approaches the free-atom natural linewidth of 0.9(1) Hz reported for ultracold Er.

• Spin coherence and spin manipulation. The paper motivates this host precisely because of its low nuclear-spin background, but spin coherence times for Er in these matrices are not measured here and remain an open question.

• Identifying the L14a/L14b components and the two 890 nm trapping sites. The present spectra cannot distinguish crystal-field splitting from inequivalent trapping sites; site-selective spectroscopy or better-controlled growth could resolve this.

• Controlling and exploiting annealing. Annealing extends the L14 lifetime in Ne toward the radiative limit and alters ZPL/PSB ratios, but it also introduces spatial inhomogeneity of up to about 30% and, in Ne, reduces NIR emission by about an order of magnitude. Understanding the microscopic origin of these changes would allow deliberate optimization of the matrix.

• Integration with photonic and superconducting devices. The authors position these results as a foundation for hybrid quantum architectures, but no device integration is demonstrated in this work.

Target Audience

Researchers in quantum photonics, solid-state quantum emitters, and rare-earth spectroscopy — particularly groups working on spin–photon interfaces, quantum memories, and telecom-band single-photon sources. It is also useful for matrix-isolation spectroscopists and for materials scientists interested in cryogenic van der Waals solids as host matrices. Graduate students with a background in atomic or optical physics will find the paper accessible; the experimental section is detailed enough to support replication attempts. Readers looking for device demonstrations, coherence measurements, or spin data will not find them here — those are explicitly framed as future work.

Note: The paper content provided is truncated partway through Fig. 6, so the full temperature-dependent photoluminescence results are not available for summary. Values such as the 78.9 ms lifetime and the <22 GHz linewidth are taken from the abstract and the text sections included.

Authors’ abstract

Rare-earth (RE) atoms, such as erbium (Er), have inner-shell optical transitions in the near-infrared, some of which are in the telecom bands. These transitions are largely insensitive to the host environment due to effective shielding by the filled outer shells. Cryogenic rare-gas (RG) solids, such as solid neon (Ne) and argon (Ar), combine high chemical purity with a nearly nuclear-spin-free environment, making them attractive solid hosts for preserving optical and spin coherence. In this work, we isolate neutral Er atoms in solid Ne and Ar at 6 K and characterize their emission by steady-state and time-resolved photoluminescence (PL) spectroscopy. In both hosts, we observe several near-infrared transitions between 1000--1500 nm, including inner-shell emission in the telecom O and E bands. The $4f \rightarrow 5d$ transition near 1299 nm forms a multiplet of narrow components, consistent with crystal-field splitting, and decays on a microsecond timescale. The $4f \rightarrow 4f$ transition near 1437 nm shows a matrix shift below 1 cm$^{-1}$ in both hosts. In solid Ne, its inhomogeneous linewidth is bounded by the spectrometer resolution to &lt; 22 GHz, and its PL lifetime reaches up to 78.9 ms, close to the estimated radiative limit. Thermal annealing extends the 1437 nm lifetime in Ne and suppresses the phonon sidebands relative to the zero-phonon lines in Ar. These results identify RG solids as promising hosts for long-lived telecom-band Er emission and provide a foundation for future spin-photon interfaces and hybrid quantum architectures.