REVIEW 1 major objections 7 minor 29 references
Direct observation of a highly forbidden optical transition in Sm:SrF$_2$
T0 review · 1 major / 7 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read This paper reports the first direct laser excitation of the forbidden $\mathrm{^7F_0} \to \mathrm{^5D_0}$ optical clock transition in Sm:SrF$_2$, with a 12.4 ms excited-state lifetime and a $1.9\times 10^{-18}$ cm$^2$ excitation cross…
desk verdict Solid direct observation and lifetime for Sm:SrF2's forbidden transition, but the cross-section extraction has a systematic rise-time bias that needs correcting. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The mechanism that makes the forbidden transition weakly allowed is hyperfine-induced mixing: the nuclear magnetic field ($B_n \sim 100$ G) of the odd samarium isotopes admixes a small fraction $\xi = \mu B_n/\Delta E \sim 4\times 10^{-5}$ of the $\mathrm{^7F_1}$ state into the $\mathrm{^7F_0}$ ground state, giving the $\mathrm{^7F_0} \to \mathrm{^5D_0}$ line a finite oscillator strength. The quantitative extraction of the cross section is carried by a three-level rate-equation model (Eqs. 2 and 3) relating the steady-state $\mathrm{^5D_0} \to \mathrm{^7F_1}$ fluorescence to the applied photon flux, with the $\mathrm{^7F_1} \to \mathrm{^7F_0}$ decay assumed to be much faster than the $\mathrm{^5D_0}$ decay.
What would settle it
Directly measure the $\mathrm{^7F_1} \to \mathrm{^7F_0}$ relaxation rate, for example by pumping $\mathrm{^5D_0}$ and probing the transient population of $\mathrm{^7F_1}$ with time-resolved fluorescence or a pump-probe sequence, and compare it with the $\mathrm{^5D_0}$ decay rate $\gamma_e \simeq 80$ s$^{-1}$. If $\gamma_f$ is not much larger than $\gamma_e$, the reported $\sigma = 1.9(1)\times 10^{-18}$ cm$^2$ is wrong. An equally decisive test is to search for the line in an isotopically enriched even-isotope sample, where hyperfine mixing is absent and the transition should be strongly suppressed.
Extended reading notes
Core claim
The central claim is that the $4f^6$ $\mathrm{^7F_0} \to 4f^6$ $\mathrm{^5D_0}$ intra-configuration transition in Sm:SrF$_2$, formally forbidden at the octahedral Sm$^{2+}$ substitution site, has been directly excited with a continuous-wave laser near 438065 GHz and observed through the $\mathrm{^5D_0} \to \mathrm{^7F_1}$ fluorescence at 697 nm. The excited state decays as a single exponential with lifetime $\tau = 12.4(3)$ ms at 4.2 K, averaged across spectral classes between 438060 and 438068 GHz, and the excitation cross section is $\sigma = 1.9(1)\times 10^{-18}$ cm$^2$. The paper attributes the residual line strength to hyperfine-induced mixing of $\mathrm{^7F_0}$ with $\mathrm{^7F_1}$ in the $^{147}$Sm and $^{149}$Sm isotopes, estimating a mixing amplitude $\xi \sim 4\times 10^{-5}$ and a homogeneous cross section $\xi^2 \lambda^2/(2\pi) \sim 10^{-18}$ cm$^2$ that matches the measured value. On this basis the paper proposes Sm:SrF$_2$ as a candidate material for an optical frequency reference.
Load-bearing premise
The reported cross section stands on the assumption, stated just before Eq. (3), that the $\mathrm{^7F_1}$ state relaxes to the $\mathrm{^7F_0}$ ground state much faster than the $\mathrm{^5D_0}$ state decays; if that fast relaxation fails, the fitted cross section would be biased by roughly a factor $1 + \gamma_e/(2\gamma_f)$.
Editorial extensions
If this is right
- The measured lifetime of $\tau = 12.4(3)$ ms puts a lower bound of about $2\pi \times 13$ Hz on the homogeneous linewidth, roughly a factor of six below the lifetime-limited linewidth of the Eu:YSO clock transition.
- The transition can be excited with a continuous-wave laser and detected cleanly at 697 nm, so Sm:SrF$_2$ is immediately usable for spectral-hole-burning studies aimed at laser frequency stabilization.
- For the zero-nuclear-spin samarium isotopes, the transition should be far more forbidden, but their reduced coupling to the lattice makes them the promising route to an absolute optical frequency reference if they can be excited.
- Because the probe beam experiences negligible resonant absorption, the material is compatible with long path-length or cavity-enhanced excitation geometries for optical memories.
Reading between the lines
- Beyond the paper: if hyperfine-induced mixing is the mechanism, an externally applied magnetic field should tune $\xi$ and hence the excitation cross section in a calculable way, giving both a direct test of the model and a control knob for the line strength.
- Beyond the paper: the two-peak profile in Fig. 2 predicts that isotopically purified $^{147}$Sm and $^{149}$Sm crystals should each show a single inhomogeneous line, and measuring the isotope shift would separate the nuclear-moment contribution to the broadening from static strain.
- Beyond the paper: since $\mathrm{^5D_0}$ decays almost exclusively to $\mathrm{^7F_1}$, the system is a natural three-level lambda scheme; a second laser resonant with the $\mathrm{^7F_1} \to \mathrm{^7F_0}$ far-infrared transition could coherently drain or repump the population, opening possibilities for optical-to-THz transduction that the paper does not discuss.
- Beyond the paper: the observed absence of phonon sidebands up to 2 THz suggests very weak electron-phonon coupling; if this persists at higher doping, inhomogeneous broadening should be dominated by static strain, so strain-engineered or isotopically purified hosts could approach the homogeneous linewidth limit.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript reports the first direct continuous-wave laser excitation of the 4f^6 7F0 → 4f^6 5D0 intra-configuration transition in Sm:SrF2 at 4.2 K. The authors locate the transition near 438065 GHz, assign the observed inhomogeneously broadened structure to the 147Sm and 149Sm isotopes, measure the 5D0 excited-state lifetime τ = 12.4(3) ms, and extract an excitation cross-section σ = 1.9(1)×10^-18 cm^2 from the intensity dependence of the steady-state fluorescence. They attribute the finite transition strength to hyperfine-induced mixing with the 7F1 state and discuss the implications for optical frequency references and spectral hole-burning.
Significance. If the lifetime and cross-section claims withstand reanalysis, this is a notable experimental result: the 12 ms 5D0 lifetime is the longest-lived optically excited state reported in a solid, and the small cross-section quantitatively supports the proposed hyperfine-mixing mechanism. The lifetime measurement, the line-position determination, and the isotope assignment are plausible and well supported by the figures. The cross-section measurement, however, is the load-bearing quantitative claim, and its extraction contains an internal inconsistency with the authors' own rate-equation model. Since that issue is correctable by reanalysis, the central result is defensible but needs revision before the paper can be accepted.
major comments (1)
- [Section III, Eq. (3), and Fig. 6] The rising fluorescence was fit to S(t) = S0(1−e^(−t/τ)) + b with τ set to the measured 5D0 lifetime, but the rate-equation model in Eq. (2) predicts a different rise constant. In the γf ≫ γe limit used by the authors, adiabatic elimination of the 7F1 population gives dNe/dt = ΦσN − (2Φσ + γe)Ne, so the fluorescence rise time constant is 1/(γe + 2Φσ), not τ = 1/γe. At the highest intensity in Fig. 6 (Φ ≈ 1×10^23 m^-2 s^-1, σ = 1.9×10^-18 cm^2), 2Φσ ≈ 40 s^-1 while γe ≈ 80 s^-1, so the correct rise time is about 8.4 ms rather than 12.4 ms. Fitting with the longer time constant introduces an intensity-dependent overestimate of S0 that grows with laser intensity; a simple least-squares estimate for the 50 ms pulse length gives an overestimate of roughly 10% at the highest intensity. This bias weakens the apparent saturation and therefore biases σ low by an amount comparable to the claimed 5% uncertainty. Because the reduced χ² = 0.98 does not diagnose this problem, the data in Fig. 6 must be reanalyzed using the correct rise function, or S0 must be determined from the steady-state portion of the trace, before the reported σ and ηcN can be considered reliable.
minor comments (7)
- [Section II] The text contains a typo: "photomultipler tube" should be "photomultiplier tube".
- [Abstract and Section IV] The phrase "longest lived excited state ever observed in a solid" is too broad; please qualify it as the longest-lived optically excited electronic state in a solid, since nuclear or other long-lived excitations are outside the paper's scope.
- [Fig. 2 caption and axis label] The caption says the fluorescence is normalized to laser power, but the vertical axis in the figure is labeled "Fluorescence (arb units)"; please make the normalization explicit in the axis label.
- [Section III] Because the lifetime varies by about 10% across the inhomogeneous profile (Fig. 4 inset), the quoted value τ = 12.4(3) ms should report the spread of the measured spectral classes as well as the statistical error of the average, so that the uncertainty reflects the inhomogeneous distribution.
- [Section III] The statement that Eq. (3) fits with reduced χ² = 0.98 is presented as support for the γf ≫ γe assumption, but a finite γf modifies the denominator by a factor that is partially degenerate with σ and ηcN; the fit alone cannot validate that assumption. Please report the number of degrees of freedom and, if possible, an independent bound on γf.
- [Section III] The independent estimate of ηcN is mentioned but not described; a short account of the calculation, or a reference to a supplement, would allow the claimed consistency to be verified.
- [Section IV] The hyperfine-mixing estimate uses Bn ∼ 100 G without derivation or uncertainty; since σ_hyp scales as Bn^2, a factor-of-two uncertainty in Bn changes the predicted cross-section by a factor of four, so the agreement with the measured σ should be presented as order-of-magnitude only.
Circularity Check
No circularity found; all central quantities are measured or fitted and checked against independent estimates.
full rationale
The paper reports a direct measurement of the 7F0–5D0 transition in Sm:SrF2, with the 5D0 lifetime obtained from single-exponential decay fits in the dark and the excitation cross section obtained by fitting the steady-state fluorescence saturation curve to the three-level rate-equation model. The cross section is presented as a measured fit parameter, not as a prediction derived from itself, and the hyperfine-mixing order-of-magnitude estimate is computed from independent nuclear and atomic parameters and compared with, rather than used to produce, the measured value. No load-bearing self-citations appear: references to prior work on Sm:SrF2 and phonon spectra are external and are used only for comparison or to justify a relaxation-rate assumption. The isotope assignment is supported by abundance ratios and linewidth comparisons. The possible rise-time mismatch in extracting S0 from the fluorescence traces is a systematic modeling concern, not a circular reduction: the fitted sigma does not appear in the definition of the fluorescence quantity it is intended to explain, and the lifetime is measured independently. Therefore the derivation chain is self-contained and no circularity score is warranted.
Assumptions & free parameters
free parameters (7)
- Excitation cross-section sigma =
1.9(1) x 10^-18 cm^2
- Collection efficiency times resonant ion number eta_c N =
Not stated numerically; fit returns a value consistent with an independent estimate
- 5D0 excited-state lifetime tau =
12.4(3) ms averaged over spectral classes
- Zero-temperature lifetime tau_1 in Eq. (1) =
Not stated explicitly; approximately 12 ms scale
- Thermally activated state lifetime tau_2 =
200 ns
- Thermal activation energy Delta =
h x 14.2 THz (about 473 cm^-1)
- Nuclear magnetic field Bn for hyperfine estimate =
about 100 G
assumptions (4)
- domain assumption The observed 684 nm lines are the 7F0-5D0 transition of Sm2+ in octahedral SrF2 sites, and the 697 nm fluorescence is the 5D0-7F1 decay.
- domain assumption The three-level rate equations in Eq. (2) describe the population dynamics, with gamma_f >> gamma_e so the 7F1 state relaxes immediately to 7F0.
- domain assumption The two inhomogeneous peaks in Fig. 2 come from 147Sm and 149Sm, matched by natural abundance ratio.
- domain assumption There are no phonon sidebands within 2 THz of the zero-phonon line.
Cite this review
Pith. "Pith review of Direct observation of a highly forbidden optical transition in Sm:SrF$_2$." pith.science (2026). https://pith.science/paper/BWNEEI5D
@misc{pith2026190902633,
author = {Pith},
title = {Pith review of: Direct observation of a highly forbidden optical transition in Sm:SrF$_2$},
year = {2026},
howpublished = {\url{https://pith.science/paper/BWNEEI5D}},
note = {Machine review of arXiv:1909.02633}
}
abstract
The $4f^6$ $^{7}F_0$ $\to 4f^6$ $^{5}D_0$ intra-configuration transition in Sm:SrF$_2$ is forbidden for Sm$^{2+}$ ions in the octahedrally symmetric substitution sites in SrF$_2$. We report the direct observation of this transition using laser-induced fluorescence at cryogenic temperatures, and measurements of the excited state lifetime and the excitation cross section. To the best of our knowledge, this optical transition has the longest lived excited state ever observed in a solid.
Figures
Figures from the paper (2 more)
Reference graph
Works this paper leans on
- [1]
-
[2]
B. G. Wybourne, J. Alloys Compd. 380, 96 (2004)
work page 2004
-
[3]
R. Hull, J. Parisi, R. M. Osgood, H. Warlimont, G. Liu, and B. Jacquier, eds., Spectroscopic Properties of Rare Earths in Optical Materials , Springer Series in Materi- als Science, Vol. 83 (Springer-Verlag, Berlin/Heidelberg, 2005)
work page 2005
- [4]
-
[5]
3: Laser-induced fluorescence measured on the 5D0→ 7F1 line at 697 nm
The lifetime fits well (albeit with a small discrepancy at low temperatures) to a simple model for a thermally- activated decay process [22], 1 τ = 1 τ1 + exp(−∆/kBT ) τ2 , (1) FIG. 3: Laser-induced fluorescence measured on the 5D0→ 7F1 line at 697 nm. The photocathode current is plotted here as a function of time. The probe laser pulse at 684 nm lasts for ...
- [6]
- [7]
- [8]
Show all 29 references
-
[9]
Dibos, M
A. Dibos, M. Raha, C. Phenicie, and J. Thompson, Phys. Rev. Lett. 120, 243601 (2018)
2018
-
[10]
S. Cook, T. Rosenband, and D. R. Leibrandt, Phys. Rev. Lett. 114, 253902 (2015)
2015
-
[11]
M. J. Thorpe, L. Rippe, T. M. Fortier, M. S. Kirchner, and T. Rosenband, Nat. Photonics 5, 688 (2011)
2011
-
[12]
Siyushev et al
P. Siyushev et al. , Nat. Commun. 5, 3895 (2014)
2014
-
[13]
R. W. Equall, Y. Sun, R. Cone, and R. Macfarlane, Phys. Rev. Lett. 72, 2179 (1994)
1994
-
[14]
D. R. Leibrandt, M. J. Thorpe, C.-W. Chou, T. M. Fortier, S. A. Diddams, and T. Rosenband, Phys. Rev. Lett. 111, 237402 (2013)
2013
-
[15]
K¨ onz, Y
F. K¨ onz, Y. Sun, C. Thiel, R. Cone, R. Equall, R. Hutche- son, and R. Macfarlane, Phys. Rev. B 68, 085109 (2003)
2003
-
[16]
Stoneham, Rev
A. Stoneham, Rev. Mod. Phys. 41, 82 (1969)
1969
-
[17]
Ferrier, B
A. Ferrier, B. Tumino, and P. Goldner, J. Lumin. 170, 406 (2016)
2016
-
[18]
Wasserburg, S
G. Wasserburg, S. Jacobsen, D. DePaolo, M. McCulloch, and T. Wen, Geochimica et Cosmochimica Acta 45, 2311 (1981)
1981
-
[19]
Oswald, M
R. Oswald, M. G. Hansen, E. Wiens, A. Y. Nevsky, and S. Schiller, Phys. Rev. A 98, 062516 (2018)
2018
-
[20]
R. D. Shannon, Acta Crystallographica A 32, 751 (1976)
1976
-
[21]
[23] excited ions up to the 5D0 state in a two-photon configu- ration using a pulsed laser and reported 14603 cm −1 as the transition wavenumber
reported the transition wavenumber as 14616 cm−1 based on a weak emission line, whereas Gˆ aconet al. [23] excited ions up to the 5D0 state in a two-photon configu- ration using a pulsed laser and reported 14603 cm −1 as the transition wavenumber. Macfarlane and Meltzer [ 24] d...
-
[22]
Y. Jia, J. Solid State Chem. 95, 184 (1991)
1991
-
[23]
Wood and W
D. Wood and W. Kaiser, Phys. Rev. 126, 2079 (1962)
1962
-
[24]
A. M. Alam and B. Di Bartolo, Phys. Lett. A 25, 157 (1967)
1967
-
[25]
Gˆ acon, G
J.-C. Gˆ acon, G. Burdick, B. Moine, and H. Bill, Phys. Rev. B 47, 11712 (1993)
1993
-
[26]
Macfarlane and R
R. Macfarlane and R. Meltzer, Opt. Commun. 52, 320 (1985)
1985
-
[27]
P. A. Tanner, Chem. Soc. Rev. 42, 5090 (2013)
2013
-
[28]
R. Yano, M. Mitsunaga, and N. Uesugi, Opt. Lett. 16, 1884 (1991)
1991
-
[29]
Kaiser, W
W. Kaiser, W. Spitzer, R. Kaiser, and L. Howarth, Phys. Rev. 127, 1950 (1962)
1962
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.