Print temp file6

General Discussion
1

Near-infrared spectroscopy of fluoride species in molten FLiBe

Dissolved rare earth, transition metal, and uranium fluorides in molten LiF-BeF₂ present a rich near-infrared spectral landscape spanning 700–2500 nm, with at least a dozen species exhibiting absorption bands and several offering emission bands suitable for saturable absorber pairing. Direct molten-salt NIR spectroscopic data remains sparse — the overwhelming majority of measurements were performed by ORNL researchers in the 1960s–70s using UV-Vis instrumentation, with only recent (2023–2025) studies extending systematic measurements into the NIR for select transition metals. The best available approximations for most species come from fluoride crystals (LaF₃, LiYF₄) and fluoride glasses (ZBLAN), which share the low-phonon-energy, ionic-bonding environment of FLiBe. At 800 K, all spectral features broaden dramatically (FWHM ~~50–200 cm⁻¹ per band envelope), peak cross-sections decrease proportionally, and excited-state lifetimes shorten by 30–60% for well-protected levels — but the low maximum phonon energy of fluoride ligands (~~400–500 cm⁻¹) preserves radiative emission from levels with energy gaps exceeding ~5000 cm⁻¹ even at this temperature.

The foundational literature and where the best data lives

Three sources form the backbone of this compilation. The Carnall–Crosswhite ANL-78-XX-95 report (available from OSTI at osti.gov/servlets/purl/6417825) provides crystal-field energy levels, Judd-Ofelt parameters, matrix elements, and calculated radiative lifetimes for all trivalent lanthanides in LaF₃ — the single most important reference for f-f transition energies in a fluoride host. The Derdeyn et al. (2023) review in J. Mol. Liq. 385, 121936 is the definitive modern compilation of optical spectroscopic data for molten fluorides, containing 71 digitized spectra from historical studies and covering transition metals (Ni²⁺, Co²⁺, Cr²⁺, Cr³⁺), lanthanides (Nd³⁺, Pr³⁺), and actinides (U³⁺, U⁴⁺) in FLiBe and FLiNaK. The companion Derdeyn PhD thesis (UW-Madison, 2022) contains quantitative molar absorptivity spectra for CrF₃ and NiF₂ in FLiNaK at 600°C. The Runowski, Stopikowska, and Lis (2020) paper in Dalton Trans. 49, 2129–2137 provides UV-Vis-NIR absorption spectra of all twelve optically active LnF₃ compounds from 200–2500 nm with downloadable Excel datasets (rsc.org/suppdata/c9/dt/c9dt04921e/c9dt04921e1.xlsx) — an invaluable resource for transition assignments across the entire lanthanide series.

Additional key primary sources include J.P. Young’s 1967 Inorg. Chem. paper reporting U⁴⁺ and U³⁺ spectra in both FLiNaK and FLiBe, Toth’s 1971 J. Phys. Chem. paper analyzing coordination effects on U⁴⁺ spectra across fluoride melt compositions, and the LeCroy et al. (2025) JACS paper from INL providing temperature-dependent (500–600°C) molar absorptivity data for Co²⁺, Ni²⁺, and Cr³⁺ in FLiNaK. For cross-sections in fluoride laser crystals, the Payne, Chase, Smith, Kway, and Krupke (1992) paper in IEEE J. Quantum Electron. 28, 2619–2630 provides absorption and emission cross-sections for Er³⁺, Tm³⁺, and Ho³⁺ in LiYF₄, BaY₂F₈, and LaF₃. The Peijzel et al. (2005) J. Solid State Chem. 178, 448–453 extends the Carnall parameters to compute a complete energy level diagram for all Ln³⁺ ions.

Lanthanide ions with strong NIR transitions

The following table summarizes the NIR-active lanthanide ions with well-characterized absorption and emission data in fluoride hosts. All data are from solid-state fluoride crystals unless marked (melt) or (glass). Cross-sections are order-of-magnitude values reflecting the range across different fluoride hosts.

Nd³⁺ (4f³, ground state ⁴I₉/₂)

Nd³⁺ is the strongest NIR emitter among the lanthanides. Its key absorption bands in fluoride hosts are: 730–740 nm (⁴I₉/₂ → ⁴S₃/₂ + ⁴F₇/₂), 792–808 nm (⁴I₉/₂ → ⁴F₅/₂ + ²H₉/₂, σ_abs ≈ 1.2 × 10⁻¹⁹ cm² in YLF π-polarization), 863–880 nm (⁴I₉/₂ → ⁴F₃/₂, σ_abs ≈ 2 × 10⁻²⁰ cm²), and weak bands near 1680 nm (→ ⁴I₁₅/₂) and 2400 nm (→ ⁴I₁₃/₂). Emission bands occur at 1047–1065 nm (⁴F₃/₂ → ⁴I₁₁/₂, the dominant laser transition, σ_em ≈ 1.8 × 10⁻¹⁹ cm² in YLF π-pol at 300 K) and 1313–1374 nm (⁴F₃/₂ → ⁴I₁₃/₂, σ_em ≈ 5 × 10⁻²⁰ cm²). The ⁴F₃/₂ radiative lifetime is ~520 µs in YLF. The energy gap to the next-lower level is ~~5500 cm⁻¹ (~~11 fluoride phonons), ensuring predominantly radiative decay. At 800 K, the lifetime is estimated at 200–300 µs and peak emission cross-section decreases ~50% from the room-temperature value due to thermal broadening.

Er³⁺ (4f¹¹, ground state ⁴I₁₅/₂)

Key absorption bands: ~800 nm (⁴I₁₅/₂ → ⁴I₉/₂, σ_abs ≈ 2 × 10⁻²¹ cm²), 970–980 nm (→ ⁴I₁₁/₂, σ_abs ≈ 3 × 10⁻²¹ cm²), and 1480–1535 nm (→ ⁴I₁₃/₂, σ_abs ≈ 5–7 × 10⁻²¹ cm²). Emission: ~980 nm (⁴I₁₁/₂ → ⁴I₁₅/₂) and the telecom band at 1530–1560 nm (⁴I₁₃/₂ → ⁴I₁₅/₂, σ_em ≈ 5–7 × 10⁻²¹ cm²). The ⁴I₁₃/₂ lifetime is very long: ~10–15 ms radiative in LaF₃, with the 6500 cm⁻¹ energy gap providing excellent protection against multiphonon quenching even at 800 K (estimated 5–7 ms). The ⁴I₁₁/₂ level is more vulnerable, with only a 3500 cm⁻¹ gap to ⁴I₁₃/₂ (~7 phonons) and significant nonradiative decay at elevated temperature.

Ho³⁺ (4f¹⁰, ground state ⁵I₈)

Absorption: ~750 nm (⁵I₈ → ⁵S₂ + ⁵F₄, σ_abs ≈ 5 × 10⁻²¹ cm²), ~890 nm (→ ⁵I₅), ~1150 nm (→ ⁵I₆, σ_abs ≈ 3 × 10⁻²¹ cm²), and ~1950 nm (→ ⁵I₇, σ_abs ≈ 5 × 10⁻²¹ cm²). Emission at ~2000–2100 nm (⁵I₇ → ⁵I₈, σ_em ≈ 1 × 10⁻²⁰ cm² in YLF/BYF). The ⁵I₇ lifetime is ~~8–12 ms radiative in fluoride crystals, with a 5100 cm⁻¹ gap (~~10 phonons). At 800 K, lifetime estimated at 4–8 ms. Ho:CaF₂ has been experimentally demonstrated as a saturable absorber at 2.017 µm for Q-switching Tm,Cr:YAG lasers (51 mJ, 60 ns pulses), confirming the viability of Ho³⁺ ground-state absorption as a passive Q-switch mechanism in fluoride hosts.

Tm³⁺ (4f¹², ground state ³H₆)

Absorption: ~790 nm (³H₆ → ³H₄, σ_abs ≈ 5 × 10⁻²¹ cm²), ~1200 nm (→ ³H₅), ~1700 nm (→ ³F₄, σ_abs ≈ 3 × 10⁻²¹ cm²). Emission at ~800 nm (³H₄ → ³H₆), ~1470 nm (³H₄ → ³F₄), and 1800–2000 nm (³F₄ → ³H₆, σ_em ≈ 4 × 10⁻²¹ cm²). Both the ³F₄ (~5700 cm⁻¹ gap, ~~12 phonons) and ³H₄ (~~4300 cm⁻¹ gap, ~9 phonons) levels are reasonably well-protected in fluorides. Lifetimes: ³F₄ ~12–16 ms, ³H₄ ~14–16 ms in fluoride crystals. The well-known “two-for-one” cross-relaxation (³H₄ + ³H₆ → 2 × ³F₄) doubles ~1.8 µm emission efficiency.

Yb³⁺ (4f¹³, ground state ²F₇/₂)

The simplest lanthanide system with only two manifolds. Absorption at 960–980 nm (²F₇/₂ → ²F₅/₂, σ_abs ≈ 1 × 10⁻²⁰ cm²). Emission at 980–1050 nm (²F₅/₂ → ²F₇/₂, σ_em ≈ 1 × 10⁻²⁰ cm² peak). The massive ~10,000 cm⁻¹ energy gap (~20 phonons) renders Yb³⁺ completely immune to multiphonon quenching at any temperature. Fluorescence lifetime is ~2.0 ms in YLF and is essentially unchanged at 800 K. Spectral broadening will reduce peak cross-sections but the integrated oscillator strength is conserved.

Pr³⁺, Sm³⁺, and Dy³⁺ — weaker NIR species

Pr³⁺ (4f²) has NIR absorption at ~1440–1530 nm (³H₄ → ³F₃ + ³F₄) and ~2000–2200 nm (→ ³F₂), but closely spaced energy levels make it susceptible to multiphonon quenching even in fluorides. Significant thermal quenching is expected at 800 K. Sm³⁺ (4f⁵) absorbs in the NIR at ~900 nm (⁶H₅/₂ → ⁶F₁₁/₂), ~1080 nm (→ ⁶F₉/₂), ~1230 nm (→ ⁶F₇/₂), ~1380 nm (→ ⁶F₅/₂), and ~1480 nm (→ ⁶F₃/₂), but is not a strong NIR emitter. Dy³⁺ (4f⁹) shows NIR absorption at ~810 nm (⁶H₁₅/₂ → ⁶F₅/₂), ~900 nm (→ ⁶F₇/₂), ~1100 nm (→ ⁶F₉/₂), and a hypersensitive band at ~1300 nm (→ ⁶H₉/₂ + ⁶F₁₁/₂), but NIR emission is generally too weak for practical use.

Lanthanide ions with no useful NIR transitions

Six lanthanide ions are effectively transparent in the 700–2500 nm range. La³⁺ (4f⁰) and Lu³⁺ (4f¹⁴) have empty and full 4f shells respectively — zero f-f transitions at any wavelength. Ce³⁺ (4f¹) has only the ²F₅/₂ → ²F₇/₂ transition at ~4 µm (mid-IR) and 4f→5d transitions in the UV (250–350 nm in fluorides). Gd³⁺ (4f⁷, half-filled shell) has a colossal 32,000 cm⁻¹ gap between the ⁸S₇/₂ ground state and the ⁶P₇/₂ first excited state — there are literally no energy levels anywhere in the 300–2500+ nm range. Eu³⁺ (4f⁶) has an empty spectral desert from ~5000 cm⁻¹ (⁷F₆) to ~17,200 cm⁻¹ (⁵D₀), making it NIR-inactive except for a marginal ⁵D₀ → ⁷F₄ emission tail near 700 nm. Eu²⁺ emits in the UV-blue (350–450 nm) in fluoride hosts via 4f→5d transitions. Tb³⁺ (4f⁸) has only very weak ground-manifold inter-multiplet transitions (⁷F₆ → ⁷F₅,₄,₃) in the 1.8–2.5 µm range; all useful optical activity is in the visible.

Transition metals produce broadband NIR absorption but no emission at 800 K

Transition metal d-d transitions in octahedral fluoride coordination create broad absorption features spanning hundreds of nanometers — ideal for saturable absorber applications if ground-state absorption bleaching can be achieved. All five transition metal ions adopt octahedral coordination in fluoride melts (confirmed by LeCroy et al. 2025 for Co²⁺, Ni²⁺, Cr³⁺ in FLiNaK at 500–600°C). The fluoride nephelauxetic ratio β is high (0.79–0.96), reflecting ionic bonding.

Ni²⁺ (d⁸): The ν₁ band (³A₂g → ³T₂g = 10Dq) falls at ~7000–8500 cm⁻¹ (1175–1430 nm), directly measured in FLiNaK. The ν₂ band (³A₂g → ³T₁g(³F)) at ~12,000–14,000 cm⁻¹ (714–833 nm) straddles the visible/NIR boundary. Molar absorptivity is low (ε ≈ 2–10 M⁻¹cm⁻¹). Derdeyn (2022) calibrated NiF₂ detection in FLiNaK down to 114 wppm.

Co²⁺ (d⁷): In fluoride melts, Co²⁺ is octahedral (unlike tetrahedral in chloride melts). The ν₁ band (⁴T₁g → ⁴T₂g) is predicted at ~7000–9000 cm⁻¹ (1100–1430 nm) but falls outside the UV-Vis window of most molten salt studies and has not been directly measured. The ν₂ band (→ ⁴A₂g) at ~14,000–16,000 cm⁻¹ (625–714 nm) was observed by LeCroy et al. Octahedral Co²⁺ has ε ≈ 5–20 M⁻¹cm⁻¹.

Fe²⁺ (d⁶, high spin): Has a single spin-allowed transition (⁵T₂g → ⁵Eg = 10Dq) predicted at ~7000–8500 cm⁻¹ (1175–1430 nm) in octahedral fluoride. This has not been measured in molten fluorides; the estimate is based on crystal-field scaling from oxide hosts. Molar absorptivity is very low (ε ≈ 1–5 M⁻¹cm⁻¹).

Cr²⁺ (d⁴): Exhibits Jahn-Teller-distorted octahedral coordination. In fluoride environments, the ⁵Eg → ⁵T₂g transition falls at ~8000–10,000 cm⁻¹ (1000–1250 nm), significantly blue-shifted from the well-known Cr²⁺:ZnSe laser absorption at 1775 nm (tetrahedral coordination). Emission in fluorides would also be blue-shifted to ~1000–1300 nm but is likely quenched at 800 K.

Cr³⁺ (d³): Absorbs primarily in the visible — the ⁴A₂g → ⁴T₂g band at ~14,000–16,000 cm⁻¹ (625–714 nm) with a tail extending to ~750 nm. Measured directly in both FLiNaK and FLiBe. Broadband emission (650–900 nm) occurs in fluoride crystals like K₃AlF₆:Cr³⁺ but is expected to be quenched in the melt at 800 K.

Mn²⁺ (d⁵): No NIR absorption whatsoever. All d-d transitions are both spin-forbidden and Laporte-forbidden, falling in the visible/UV (400–560 nm) with extremely weak extinction (ε ≈ 0.01–0.1 M⁻¹cm⁻¹). Mn²⁺ is effectively invisible in the NIR.

Critically, no NIR emission is expected from any transition metal ion in molten FLiBe at 800 K. The d-d excited states relax nonradiatively via rapid multiphonon processes at these temperatures. The transition metals therefore serve exclusively as broadband ground-state absorbers in saturable absorber schemes.

Uranium ions provide uniquely broad NIR absorption

U⁴⁺ (5f², isoelectronic with Pr³⁺, ground state ³H₄) shows multiple f-f absorption bands extending across the entire NIR. In molten fluorides, the key bands include: ~2000–2500 nm (³H₄ → ³H₅, weak), ~1220–1540 nm (→ ³H₆ + ³F₂, moderate), ~950–1110 nm (→ ³F₃ + ³F₄, moderate), and ~645–667 nm (→ ¹G₄, the strongest f-f band used for analytical monitoring). Molar absorptivities for 5f-5f transitions are ε ≈ 5–40 M⁻¹cm⁻¹, somewhat higher than lanthanide f-f bands due to greater 5f radial extension. The coordination number differs between solvents: ~8–9 in FLiBe (due to the polymeric BeF₂ network) versus ~6–8 in FLiNaK, producing measurably different spectral profiles. Carnall et al. (1991) J. Chem. Phys. 95, 7194 provides the definitive crystal-field analysis of solid UF₄.

U³⁺ (5f³, isoelectronic with Nd³⁺, ground state ⁴I₉/₂) has NIR f-f absorption bands at approximately ~1210 nm (→ ⁴I₁₁/₂), ~1020–1060 nm (→ ⁴F₇/₂ + ⁴I₁₅/₂), ~870–893 nm (→ ⁴S₃/₂, ²H₉/₂), and ~752 nm (→ ⁴F₅/₂). Additionally, U³⁺ possesses intense parity-allowed 5f³ → 5f²6d transitions in the visible (400–600 nm) with ε > 100 M⁻¹cm⁻¹, producing the characteristic deep red/brown color of U³⁺ solutions. The U³⁺/U⁴⁺ ratio in MSRE fuel salt was monitored spectrophotometrically using these distinctive color differences.

No emission is expected from either U³⁺ or U⁴⁺ in molten salts at operating temperatures. In solid LiYF₄ at cryogenic temperatures, U³⁺ shows emission from the ⁴G₇/₂ level with lifetimes of 2.5–7.6 µs, and laser transitions at 2.50, 2.83, and 2.97 µm have been observed. These are completely quenched by multiphonon relaxation at melt temperatures.

Spectral overlap map for saturable absorber pairing

The central application requires identifying pairs where one species’ emission band overlaps another’s absorption band. Three tiers of overlaps emerge from the compiled data.

Tier 1 — well-established, high-confidence overlaps:

  • Yb³⁺ emission (980–1050 nm) → Er³⁺ absorption (⁴I₁₅/₂ → ⁴I₁₁/₂, 950–1000 nm): The classic sensitization pair. Energy mismatch is only ~39 cm⁻¹ between lowest crystal-field components. Extremely efficient in all fluoride hosts. At 800 K, thermal broadening enhances the spectral overlap further.
  • Tm³⁺ emission (³F₄ → ³H₆, 1700–2000 nm) → Ho³⁺ absorption (⁵I₈ → ⁵I₇, 1850–2150 nm): Overlaps in the ~1850–1950 nm region. This is the basis of Tm-sensitized Ho lasers, widely demonstrated in fluoride crystals. Ho³⁺ ground-state absorption at ~2 µm has been experimentally demonstrated as a saturable absorber in Ho,Er:CaF₂ at 2.017 µm.
  • Er³⁺ emission (⁴I₁₃/₂ → ⁴I₁₅/₂, 1450–1650 nm) → Tm³⁺ absorption (³H₆ → ³F₄, 1600–1900 nm): Overlaps near ~1600–1650 nm. Cross-relaxation Er³⁺(⁴I₁₃/₂) + Tm³⁺(³F₄) → Er³⁺(⁴I₁₅/₂) + Tm³⁺(³H₄) is well-documented in fluoride hosts with up to 88% energy transfer efficiency.

Tier 2 — plausible overlaps requiring broadened-spectrum verification at 800 K:

  • Nd³⁺ emission (~1060 nm) → broad Ni²⁺/Co²⁺/Fe²⁺/Cr²⁺ absorption (1000–1430 nm): The transition metal ν₁ bands in octahedral fluoride are broad enough to overlap the Nd³⁺ 1.06 µm emission region. Ni²⁺ absorption starting near 1000 nm and extending past 1400 nm would encompass the full Nd³⁺ ⁴F₃/₂ → ⁴I₁₁/₂ emission band.
  • Nd³⁺ emission (~1340 nm) → Sm³⁺ absorption (⁶H₅/₂ → ⁶F₅/₂, ~1380 nm) or Dy³⁺ absorption (⁶H₁₅/₂ → ⁶H₉/₂, ~1300 nm): Near overlap; thermal broadening at 800 K may bring these into alignment.
  • Er³⁺ emission (~1530 nm) → U⁴⁺ absorption (³H₄ → ³H₆ + ³F₂, ~1220–1540 nm): The long-wavelength edge of U⁴⁺'s ³H₆/³F₂ absorption manifold may overlap with Er³⁺ 1.5 µm emission.
  • Ho³⁺ emission (~2050 nm) → U⁴⁺ absorption (³H₄ → ³H₅, ~2000–2500 nm): Direct overlap expected in the 2000–2100 nm region.

Tier 3 — broadband transition metal absorbers as universal saturable absorbers:

The Ni²⁺, Co²⁺, Fe²⁺, and Cr²⁺ ν₁ absorption bands (spanning roughly 1000–1500 nm collectively) are broad enough to potentially serve as saturable absorbers for multiple lanthanide emission wavelengths simultaneously (Nd³⁺ ~1060 nm, Er³⁺ ~980 nm, Ho³⁺ ~1150 nm). However, the low molar absorptivities (ε ≈ 1–20 M⁻¹cm⁻¹) and extremely fast nonradiative relaxation at 800 K make ground-state bleaching difficult to achieve with these species.

Temperature effects at 800 K and thermal stability ranking

At molten salt operating temperature (~800 K), the spectroscopic environment resembles a fluoride glass with additional thermal disorder. Three effects dominate. First, spectral broadening: individual Stark transitions (1–10 cm⁻¹ wide in crystals) merge into broad envelopes with FWHM ~~50–200 cm⁻¹, similar to glass spectra but with additional homogeneous broadening from the liquid-state dynamics. Band centers shift by ~~50–300 cm⁻¹ from crystalline LaF₃ values, generally red-shifting in the disordered melt. Second, lifetime quenching follows the multiphonon relaxation rate W_mp = W₀ · [n(T)+1]^p, where n(T) is the Bose-Einstein phonon occupation at 800 K (n ≈ 0.59 for a 500 cm⁻¹ phonon) and p = ΔE/ℏω_max is the phonon order. Third, Boltzmann thermal population of higher Stark levels within each manifold redistributes oscillator strength across broader spectral ranges, reducing peak cross-sections while conserving integrated intensity.

The thermal stability ranking for NIR emitters at 800 K, from best to worst:

  • Yb³⁺ (ΔE ~10,000 cm⁻¹, ~20 phonons): essentially immune to thermal quenching; lifetime unchanged
  • Er³⁺ ⁴I₁₃/₂ (ΔE ~6,500 cm⁻¹, ~13 phonons): excellent protection; ~5–7 ms lifetime
  • Tm³⁺ ³F₄ (ΔE ~5,700 cm⁻¹, ~12 phonons): very good; ~6–10 ms lifetime
  • Nd³⁺ ⁴F₃/₂ (ΔE ~5,500 cm⁻¹, ~11 phonons): very good; ~200–300 µs lifetime
  • Ho³⁺ ⁵I₇ (ΔE ~5,100 cm⁻¹, ~10 phonons): good; some quenching, ~4–8 ms lifetime
  • Tm³⁺ ³H₄ (ΔE ~4,300 cm⁻¹, ~9 phonons): moderate; measurable quenching
  • Er³⁺ ⁴I₁₁/₂ (ΔE ~3,500 cm⁻¹, ~7 phonons): significant quenching
  • Pr³⁺ excited states (closely spaced levels): poor; severe quenching expected

For quantitative modeling at 800 K in FLiBe, the recommended approach is to use ZBLAN glass cross-section values as starting estimates (already inhomogeneously broadened), apply an additional thermal broadening factor of ~2–3× to linewidths, reduce peak cross-sections proportionally, and use Carnall/Crosswhite LaF₃ energy levels for transition assignments with Judd-Ofelt parameters from fluoride glass hosts.

Comprehensive master reference list flagged for ongoing use

The following resources are flagged as particularly valuable for ongoing reference and downloadable data:

  • Carnall, Crosswhite, Crosswhite (1978), ANL-78-XX-95: Complete energy levels, J-O parameters, radiative lifetimes for all Ln³⁺ in LaF₃. Downloadable PDF from OSTI (osti.gov/servlets/purl/6417825).
  • Runowski et al. (2020), Dalton Trans. 49, 2129: All LnF₃ absorption spectra 200–2500 nm with downloadable Excel datasets (rsc.org/suppdata/c9/dt/c9dt04921e/c9dt04921e1.xlsx).
  • Derdeyn et al. (2023), J. Mol. Liq. 385, 121936: Definitive review of all molten fluoride spectroscopic data, 71 digitized spectra, molar absorptivity tables.
  • Derdeyn PhD thesis (2022), UW-Madison: Quantitative CrF₃ and NiF₂ absorptivity in FLiNaK (asset.library.wisc.edu/1711.dl/GRIYUQT2I4VG38B/R/file-aac45.pdf).
  • Payne et al. (1992), IEEE J. Quantum Electron. 28, 2619: Emission/absorption cross-sections for Er³⁺, Tm³⁺, Ho³⁺ in fluoride crystals.
  • Peijzel et al. (2005), J. Solid State Chem. 178, 448: Complete calculated energy level diagram for all Ln³⁺ (dspace.library.uu.nl/bitstream/handle/1874/26006/1064.pdf).
  • Carnall et al. (1991), J. Chem. Phys. 95, 7194: Crystal-field analysis of UF₄ — definitive U⁴⁺ energy levels.
  • LeCroy et al. (2025), JACS 147, 24814: Latest Co²⁺, Ni²⁺, Cr³⁺ spectra in FLiNaK with temperature dependence.
  • Nielsen and Sørensen (2025), Nature Comm. 16, 10754: Updated Dieke diagram (open access).

Conclusion

The most promising saturable absorber pairs for the FLiBe NIR application are the Tm³⁺/Ho³⁺ pair near 1.9–2.0 µm (experimentally validated in fluoride hosts, both ions thermally robust at 800 K), the Yb³⁺/Er³⁺ pair near 980 nm (highest energy transfer efficiency known), and the Er³⁺/Tm³⁺ pair near 1.6 µm. Transition metal ions — particularly Ni²⁺ and Co²⁺ — offer broadband NIR absorption spanning the 1.0–1.4 µm region but cannot serve as emitters at melt temperatures, limiting them to the absorber role only. U⁴⁺ provides broad background absorption across most of the NIR that may interfere with or complement specific lanthanide saturable absorber schemes. The critical data gap remains the absence of quantitative NIR cross-section measurements in actual molten FLiBe for any species; all emission cross-sections derive from solid-state fluoride hosts and require thermal correction factors for 800 K melt conditions. The Derdeyn et al. (2023) review and Runowski et al. (2020) datasets together provide the most accessible starting point for detailed spectral overlap modeling.