Pith. sign in

REVIEW 4 major objections 5 minor 4 references

Tungsten isotope evolution during Earth's formation and new constraints on the viability of accretion simulations

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A core-formation model that tracks Hf-W isotopes and metal-silicate equilibration shows only one of six Grand Tack accretion simulations matches Earth's mantle tungsten anomaly, pointing to a late Moon-forming impact.

desk verdict A transparent, well-built core formation model that makes a conditional but genuinely new claim about which accretion simulations survive Hf-W constraints; the central bottleneck is the choice of metal-silicate equilibration volume. read the letter →

arxiv 2411.17681 v1 pith:NZCTORSG submitted 2024-11-26 astro-ph.EP

D.C. Rubie , K.I. Dale , G. Nathan , M. Nakajima , E.S. Jennings , G.J. Golabek , S.A. Jacobson , A. Morbidelli This is my paper · ORCID
classification astro-ph.EP
keywords Hf-Wchronometrytungstenisotopescoreformationmagmaoceanmetal-silicateequilibrationGrandTackaccretionsimulationsMoon-forminggiantimpacttimescales
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to turn the Hf-W isotope system into a sharper test of how Earth formed by coupling it to a multistage core-formation model with physically grounded metal-silicate equilibration. The authors argue that a newly included dispersed fractionation mechanism, in which small metal droplets from planetesimal impacts equilibrate only with a thin basal boundary layer of a global magma ocean, removes the long-standing mismatch between calculated and observed mantle tungsten and molybdenum. They also show that the assumed timescale of magma ocean solidification shifts the final tungsten isotope anomaly, so this parameter must be considered alongside accretion timing. Applied to six Grand Tack N-body simulations, the model leaves only one simulation that satisfies all Earth-mantle constraints, and that simulation predicts a Moon-forming giant impact at 143-183 million years after solar system start. If true, a precise lunar age would decide between competing accretion scenarios.

What carries the argument

The load-bearing object is the dispersed metal-silicate fractionation mechanism (Fig. 1b): metal from planetesimal impacts is assumed to form ~1 cm droplets suspended in a vigorously convecting global magma ocean and to equilibrate only when swept into the mechanical boundary layer at the base (at most 5% of the magma ocean volume), after which it sinks as diapirs without further reaction. This replaces the assumption of a single large-scale equilibration event and is combined with SPH-based estimates of giant-impact melt depths, a tungsten partition coefficient that varies with oxygen fugacity and carbon content, and explicit time evolution of $^{182}\mathrm{Hf}\rightarrow{}^{182}\mathrm{W}$. The mechanism is what allows calculated mantle W and Mo to fall to bulk silicate Earth levels while controlling the final $\varepsilon^{182}\mathrm{W}$.

What would settle it

A fluid-dynamical experiment or simulation showing that ~1 cm metal droplets equilibrate with more than about 10% of the magma ocean volume, or a precise lunar chronology placing the Moon-forming impact before ~100 Myr after solar system start, would contradict the paper's preferred outcome.

Watch

Extended reading notes

Core claim

The central claim is that Earth's mantle $\varepsilon^{182}\mathrm{W}=1.9\pm0.1$ can be reproduced only when metal delivered by planetesimal impacts is dispersed as ~1 cm droplets in a global magma ocean and segregates progressively through a thin basal boundary layer, and when magma oceans solidify rapidly (the instantaneous end-member, with lifetimes less than about 5 Myr). Under those conditions, among the six Grand Tack simulations only simulation 3 also matches bulk silicate Earth W and Mo concentrations and $f_{\mathrm{Hf/W}}$; it predicts the final, Moon-forming giant impact at 143 Myr (instantaneous solidification) or 183 Myr (5 Myr magma ocean lifetime). The other simulations yield $\varepsilon^{182}\mathrm{W}$ values that are too high or too low and can be brought into agreement only by artificially rescaling their accretion timescales, which puts the Moon-forming impact at 53-62 Myr. The paper concludes that the Hf-W system can discriminate viable from non-viable accretion simulations and that the Moon's age is an additional decisive constraint.

Load-bearing premise

The central fragile premise is that all planetesimal metal is suspended as ~1 cm droplets in a global magma ocean and equilibrates only with a thin basal boundary layer (at most 5% of the magma ocean volume); if metal equilibrates with a substantially larger volume of silicate melt, the computed tungsten abundance and $\varepsilon^{182}\mathrm{W}$ change and simulation 3 loses its unique fit.

Editorial extensions

If this is right

  • Earth's mantle tungsten anomaly becomes a selective filter for accretion simulations: simulations must satisfy W, Mo, $f_{\mathrm{Hf/W}}$, and $\varepsilon^{182}\mathrm{W}$ simultaneously, not the isotope ratio alone.
  • The Moon-forming impact in the surviving simulation is late, 143-183 Myr after solar system start, so a firm lunar age would favor either that scenario or the alternative 53-62 Myr scenarios.
  • The timescale of magma ocean solidification, not just the accretion history, shifts $\varepsilon^{182}\mathrm{W}$ by about 0.6 units, so crystallization models are as important as collision histories in Hf-W chronology.
  • Planetesimal metal must be processed by the dispersed fractionation mechanism nearly 100% of the time; any substantial fraction of focused equilibration leaves mantle W and Mo too high.
  • Other Grand Tack simulations can match $\varepsilon^{182}\mathrm{W}$ only after artificial timescale rescaling, meaning their accretion speeds are internally inconsistent with the isotope record.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Inference: if a hydrodynamic model allowing more extensive metal-silicate mixing beyond the thin boundary layer is correct, the uniqueness of simulation 3 would likely not survive, because the tungsten abundance and $\varepsilon^{182}\mathrm{W}$ would shift in the direction of the rejected model.
  • Inference: the artificial timescale rescaling means the 53-62 Myr Moon-forming ages from the other simulations are calibrated outputs rather than self-consistent predictions; they should be read with the rescaling kept explicit.
  • Inference: the same machinery could be applied to non-Grand-Tack accretion scenarios, such as narrow-ring terrestrial planet formation; the paper states this is future work, and the dispersed mechanism's necessity is a testable prediction for those scenarios.
  • Inference: the prediction that a precise Moon age would decide between a 143-183 Myr and a 53-62 Myr lunar formation interval means that improved lunar sample chronology has direct bearing on the physics of metal-silicate equilibration in magma oceans.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper extends the Rubie et al. (2015) accretion/core-formation model with SPH-based estimates of giant-impact melting depth, a new 'dispersed' metal-silicate fractionation mechanism for planetesimal impacts, and a self-consistent 182Hf-182W isotopic evolution model. The authors re-evaluate six Grand Tack N-body simulations and report that only simulation 3 matches the BSE constraints on W, Mo, fHf/W, and ε182W when magma ocean crystallization occurs within <5 Myr. For simulations that do not match, the authors multiply the N-body timescale by an artificial factor t_adjust to force ε182W=1.9, which yields Moon-forming impact times of 53-62 Myr; simulation 3 gives 143-183 Myr depending on the magma ocean model. The paper concludes that an accurate Moon age would provide an additional constraint on the viability of accretion simulations.

Significance. If the discriminating result survives, it would be a valuable step: the paper ties Hf-W chronology to physically motivated metal-silicate equilibration models, quantifies the role of magma ocean solidification timescale, and makes a falsifiable prediction about the Moon-forming impact age. The manuscript is transparent about its main modeling choices, presents all six simulations in Table 1, and explicitly labels the t_adjust procedure as 'completely artificial.' Its strengths include the use of SPH melting depths, the two end-member magma ocean models, and the explicit separation of constraints that are fit from those that are compared. However, the central claim rests on load-bearing assumptions that are not fully tested: the exclusive operation of the dispersed fractionation mechanism, and the circularity of t_adjust-derived ages. The conclusion is therefore more conditional than the abstract suggests, but the issues are addressable within the manuscript's scope.

major comments (4)
  1. [Section 3.2, Table 1, Fig. 6] The t_adjust procedure is circular for the central timing claim. The recalculated times of the last giant impact are obtained by multiplying all N-body times by a factor chosen to force ε182W=1.9, which the text itself calls 'completely artificial.' This applies to all simulations, including simulation 3 in the 5 Myr magma ocean model, where the supplementary material lists t_adjust=1.35. Consequently, the 53-62 Myr and 183 Myr Moon-forming ages are transformations of the assumed isotope ratio rather than independent predictions. The paper should state explicitly which simulations satisfy ε182W=1.9 without any rescaling, and should present t_adjust-derived ages only as diagnostics, not as model predictions to be compared with lunar age estimates.
  2. [Section 2.2.1-2.2.2, Fig. 3] The dispersed metal-silicate fractionation mechanism is load-bearing for the uniqueness of simulation 3. Fig. 3 shows that the calculated W concentration, χ², fHf/W, and ε182W all vary strongly with the fraction of planetesimal events assigned to this mechanism, and the text concludes that the fraction must be close to 1.0. The rejection of the Landeau et al. (2021) scaling in Section 2.2.1 because it 'leads to poor results' is not a test of the alternative; since that scaling predicts larger equilibrating volumes, the computed W abundance and ε182W for all six simulations would change. The manuscript should run the same evaluation with the Landeau scaling or an equivalent larger-volume model and report whether simulation 3 remains uniquely viable. Without this test, the central discriminative claim is conditional on a single, assumed equilibration geometry.
  3. [Section 3.2, magma ocean lifetime paragraphs] Simulation 3's success depends on near-instantaneous magma ocean solidification. With 5 Myr magma ocean lifetimes, the model gives ε182W=2.6, which is outside the 1.9±0.1 range and can be reconciled only after applying t_adjust. The text also notes that only one inter-giant-impact interval in simulation 3 exceeds 5 Myr, so the '5 Myr lifetime' model is not a well-sampled average. The statement that simulation 3 is consistent with magma ocean lifetimes <5 Myr should be presented as an extreme-case bound, not as a model-supported inference robust to crystallization timescale.
  4. [Table 1 and Section 3.2] The instantaneous-solidification value ε182W=2.0 for simulation 3 sits exactly at the upper edge of the 1.9±0.1 interval. The abstract's phrase 'reproduces ε182W=1.9±0.1' is therefore stronger than what the table shows: the unique-simulation claim has zero margin at the tolerance boundary for this model. Please clarify whether the acceptance criterion is the central BSE value or the full uncertainty interval, and state how the selection of simulation 3 responds to small changes in the assumed BSE ε182W uncertainty.
minor comments (5)
  1. [Section 3.1] The text says '≤4.9 in ε180W units'; this should be ε182W units.
  2. [Section 2.3 and References] The in-text citation 'Jacobson, 2005' should be 'Jacobsen, 2005' to match the reference list entry.
  3. [Fig. 3 caption and Supplementary Fig. S2] There is a typo 'T he' at the start of the Fig. 3 caption, and the supplementary caption refers to 'Grever et al. 2015' where 'Greber et al. 2015' is meant.
  4. [Data availability] The statement 'Additional data and software are available on request' is weaker than current practice for a modeling paper with many free parameters; a permanent repository for the code and input files would substantially improve reproducibility.
  5. [Section 3.2] Minor language errors: 'identical to a n estimated late age' and 'the results of isotopic ... studies are similar to the earlier ages' should be rewritten for clarity.

Circularity Check

2 steps flagged · score 6.0 of 10

Partial circularity: t_adjust timescale factors are fitted to ε182W and then recast as predicted Moon-forming impact ages, while the rejection of the Landeau (2021) equilibration model rests on the authors' own Dale et al. (2023) work, making simulation 3's uniqueness conditional on a self-cited modeling choice.

  1. fitted input called prediction [Section 3.2, Table 1, Fig. 6, and Conclusions]
    "we rescale the timescale of each simulation by multiplying all times by a factor t-adjust, and determine which value of this parameter reproduces the BSE value of ε182W of 1.9±0.1. While this is, of course, completely artificial, it helps to gain insight into the actual Earth accretion timescale, even if the simulations did not reproduce it self-consistently. ... Other simulations, however, predict much earlier Moon-forming impacts occurring 53-62 Myr after the start of the solar system."

    The parameter t_adjust is a free factor chosen so that the model output ε182W equals the observed value 1.9±0.1. The 'recalculated time of final giant impact' is then simply the original N-body event time multiplied by this fitted factor (Table 1 lists factors of 0.5, 0.6, 1.35, etc.). Consequently the reported early Moon-forming ages (53-62 Myr) and the 183 Myr case are one-to-one transformations of the assumed isotope ratio, not independent predictions of the accretion simulations. The paper itself labels the procedure 'completely artificial', confirming that using these rescaled impact times as a discriminant between accretion simulations is partially circular: the isotope target is the input that produces the age, not an independent check on it.

  2. self citation load bearing [Section 2.2.1]
    "We have not used the revised hydrodynamic model of Landeau et al. (2021), which includes the effects of impact velocity, because, especially for planetesimal impacts, it predicts much larger volumes of equilibrating mantle than the Deguen et al. model and leads to poor results when fitting Earth's calculated mantle composition to the BSE composition. The application of the Landeau et al. model is discussed in more detail by Dale et al. (2023)."

    The uniqueness of simulation 3 (Section 3.2, Table 1) depends on applying the dispersed metal-silicate mechanism to all planetesimal impacts, which in turn depends on rejecting the Landeau et al. (2021) equilibration-volume model. The decisive quantitative support for that rejection is deferred to Dale et al. (2023), a paper sharing most of the present authors (Dale, Rubie, Nakajima, Jacobson, Nathan, Golabek, Morbidelli). No independent, quantified comparison with the Landeau scaling is given here, so the load-bearing exclusion of an alternative physical model is justified by an internal self-citation chain rather than by an externally established result.

full rationale

The paper's core Hf-W decay and mass-balance equations are standard and externally grounded, and the ε182W value for simulation 3 under the instantaneous solidification model is not directly fitted: element concentrations are fitted to BSE values first, and ε182W is then compared. However, two load-bearing parts of the argument are partially circular. First, for simulations that do not naturally produce ε182W=1.9, the paper multiplies all times by t_adjust, chosen precisely to force ε182W=1.9, and then reports the resulting giant-impact times (53-62 Myr, and the 5 Myr-model 183 Myr age) as predictions; these ages are simply rescaled versions of the fitted isotope target, as the paper itself acknowledges by calling the procedure 'completely artificial'. Second, the conclusion that only simulation 3 satisfies all BSE criteria hinges on adopting the dispersed metal-silicate mechanism for all planetesimal impacts and rejecting the Landeau et al. (2021) model; the rejection is supported mainly by Dale et al. (2023), which overlaps heavily with the present authorship, so the central viability ranking is not fully self-contained. The circularity is partial rather than total: the isotopic evolution model itself is not defined in terms of the output, and simulation 3's instantaneous-model ε182W=2.0 is a genuine model result under the adopted assumptions. Score 6 reflects that some 'predictions' reduce by construction (t_adjust) and that a load-bearing choice is imported from the authors' own prior work, without the entire derivation being equivalent to its inputs.

Assumptions & free parameters 6 free parameters · 8 assumptions · 0 invented entities

The model imports many external inputs (Hf-W system, partition coefficients, SPH scaling laws) and fits six free parameters to BSE chemistry plus a time-rescaling factor. No new physical entities are postulated. The most consequential choices are the post hoc selection of the Deguen over the Landeau mixing model and the requirement that all planetesimal events use the dispersed mechanism. These choices directly control which simulation appears viable.

free parameters (6)
  • Average planetesimal metal-silicate equilibration pressure (pf, fraction of CMB pressure) = 0.33 average for simulation 3
    Fitted to reproduce BSE major and trace element concentrations; sets depth of magma ocean at which planetesimal-delivered metal equilibrates (Section 2.3, Supplementary Table).
  • Oxygen content gradient parameters (four values) = Not explicitly listed
    Four parameters define oxidation state of starting bodies as a function of heliocentric distance; refined by least squares to match BSE chemistry (Section 2.3).
  • Fraction of embryo cores that equilibrates during giant impacts = 0.7-1.0 depending on simulation; 0.9 for simulation 1
    Chosen after observing that chi squared improves until ~0.6 then plateaus; value selected per simulation (Section 3.1, Table 1).
  • Fraction of planetesimal metal-silicate events using dispersed mechanism = ≈1.0
    Required to match W and Mo abundances and fHf/W; varied in Fig. 3, set to 1.0 for main results (Section 3.2).
  • t_adjust timescale rescaling factor = 0.4 to 2.1 depending on simulation and magma ocean model
    Multiplies all N-body times to force ε182W = 1.9; used to estimate Moon-forming impact times (Table 1, Section 3.2). Paper labels this 'completely artificial'.
  • Refractory element enhancement factor and radial extent = 30% for bodies originating within ~1 au
    Changed from 22% applied to all bodies in Rubie et al. (2015); chosen to match primitive mantle refractory abundances (Section 2.1).
assumptions (8)
  • domain assumption Hf-W isotopic system initial values (182Hf/183W=2.836e-4, 182W/183W=1.850664, 180Hf/183W=2.633) and 182Hf decay constant 0.0778 Myr^-1.
    Adopted from Vockenhuber et al. (2004) and Kleine et al. (2004, 2009); anchors the chronology (Section 2.3).
  • domain assumption Core-mantle differentiation of embryos and differentiated planetesimals occurred 2 Myr after solar system formation.
    Based on 26Al heating timescales; sensitivity to 0.5 and 2.0 Myr shown in Fig. 3d (Section 2.3).
  • domain assumption N-body simulations start 5 Myr after solar system formation, near gas disk dispersal.
    Assumed start time; variants at 2.0 and 8.0 Myr tested (Section 2.3).
  • domain assumption The tungsten metal-silicate partition coefficient parameterization of Jennings et al. (2021) is valid across the full P-T-fO2-carbon range of the model.
    Used for every core-forming event; DW varies from ~12000 to ~70 during accretion (Section 2.1, Fig. 5).
  • ad hoc to paper The Deguen et al. (2011, 2014) hydrodynamic model describes focused metal-silicate mixing for embryo impacts; the Landeau et al. (2021) model is rejected.
    Section 2.2.1: Landeau model 'leads to poor results when fitting Earth's calculated mantle composition to the BSE composition'. Selection is based on fit quality, not independent validation.
  • domain assumption Dispersed metal droplets are ~1 cm, settle at ~0.5 m/s, remain suspended in convection up to 40 m/s, and segregate only in a basal boundary layer with volume fraction <=0.05; segregation is fast relative to 182Hf decay.
    Section 2.2.2; based on Rubie et al. (2003) and Martin and Nokes (1988); boundary layer fraction set to 0.05 for computational convenience.
  • ad hoc to paper Essentially all planetesimal metal-silicate fractionation events occur by the dispersed mechanism.
    Section 3.2: adopted because lower fractions fail W, Mo, and fHf/W constraints (Fig. 3). This is a fitted requirement, not an independently measured process rate.
  • ad hoc to paper For simulation 3, the instantaneous magma ocean solidification model is the most relevant, and magma ocean lifetimes <5 Myr are inferred.
    Section 3.2: simulation 3's ε182W is 2.0 for instantaneous solidification but 2.6 for 5 Myr lifetime; the viability conclusion depends on preferring the instantaneous model.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Tungsten isotope evolution during Earth's formation and new constraints on the viability of accretion simulations." pith.science (2026). https://pith.science/paper/NZCTORSG

@misc{pith2026241117681,
  author       = {Pith},
  title        = {Pith review of: Tungsten isotope evolution during Earth's formation and new constraints on the viability of accretion simulations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NZCTORSG}},
  note         = {Machine review of arXiv:2411.17681}
}
read the original abstract

The Hf-W isotopic system is the reference chronometer for determining the chronology of Earth's accretion and differentiation. However, its results depend strongly on uncertain parameters, including the extent of metal-silicate equilibration and the siderophility of tungsten. Here we show that a multistage core-formation model based on N-body accretion simulations, element mass balance and metal-silicate partitioning, largely eliminates these uncertainties. We modified the original model of Rubie et al. (2015) by including (1) smoothed particle hydrodynamics estimates of the depth of melting caused by giant impacts and (2) the isotopic evolution of 182W. We applied two metal-silicate fractionation mechanisms: one when the metal delivered by the cores of large impactors equilibrates with only a small fraction of the impact-induced magma pond and the other when metal delivered by small impactors emulsifies in global magma oceans before undergoing progressive segregation. The latter is crucial for fitting the W abundance and 182W anomaly of Earth's mantle. In addition, we show, for the first time, that the duration of magma ocean solidification has a major effect on Earth's tungsten isotope anomaly. We re-evaluate the six Grand Tack N-body simulations of Rubie et al. (2015). Only one reproduces epsilon182W=1.9+/-0.1 of Earth's mantle, otherwise accretion is either too fast or too slow. Depending on the characteristics of the giant impacts, results predict that the Moon formed either 143-183 Myr or 53-62 Myr after the start of the solar system. Thus, independent evaluations of the Moon's age provide an additional constraint on the validity of accretion simulations.

Figures

Figures reproduced from arXiv: 2411.17681 by the authors.

Figure 1
Figure 1. Metal-silicate fractionation models. a) Focused metal-silicate fractionation. When a large impacting body is differentiated, as is the case here, its metallic core sinks in the impact-induced silicate-magma melt pool (or global magma ocean), and progressively emulsifies and entrains increasing amounts of the surrounding silicate liquid. The result is a high-density plume of molten metal + molten silicate that expand… view at source ↗
Figure 2
Figure 2. The effects on (a) chi squared and (b) ε182W of varying the average fraction of embryo cores that equilibrates with silicate liquid at the base of giant impact-induced melt pools in simulation 1. Reduced chi squared is determined from fitting calculated mantle concentrations of MgO, FeO, SiO2, Ni, Co, Nb, Ta, V, Cr, W and Mo to BSE values (Palme and O’Neill, 2014; Greber et al., 2015). Results are shown for fraction… view at source ↗
Figure 3
Figure 3. Effects of the fraction of metal-silicate fractionation events that occur by the dispersed mechanism, following planetesimal impacts, on the results of simulation 1 (4:1-0.5-8) accretion/core formation model. (a) Calculated W concentration compared with the BSE value reported by König et al. (2011) and Palme and O’Neill (2014). The error bar is based on propagated uncertainties in the metal-silicate partition coeffi… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

4 extracted references · 1 canonical work pages

  1. [8]

    continuous magma ocean model

    (Rubie et al., 2015). Mass accreted, normalized to one Earth mass, is plotted against time and each symbol represents an accretional event. The black symbols/lines show the original N-body time scale assuming that accretion starts 5 Myr after the start of the solar system, which results in ε182W = 1.3 when accreted metal equilibrates at the time of each i...

    work page 2015

  2. [2011]

    The Earth’s tungsten budget during mantle melting and crust formation. Geochim. Cosmochim. Acta 75, 2119-2136. Maller, A., Landeau, M., Allibert, L., Charnoz, S., 2024. Condition for metal fragmentation during Earth-forming collisions. Phys. Earth Planet. Int. 352, 107199. Maurice, M., Tosi, N., Schwinger, S., Breuer, D., Kleine, T., 2020. A long-lived ma...

    2024

  3. [2016]

    8” is the run number. The prefix “i

    Highly siderophile elements were stripped from Earth’s mantle by iron sulfide segregation. Science 353, 1141-1144. Rudge, J.F., Kleine, T., Bourdon, B., 2010. Broad bounds on Earth’s accretion and core formation constrained by geochemical models. Nature Geoscience 3, 439-443. Solomatov, V. S., 2015. Magma oceans and primordial mantle differentiation. In T...

    arXiv 2011

  4. [2018]

    Icarus 302, 27-43

    Impact splash chondrule formation during planetesimal recycling. Icarus 302, 27-43. Martin, D., Nokes, R., 1988. Crystal settling in a vigorously convecting magma chamber. Nature 332, 534-536. Michel, A., van der Marel, N. , Matthews, B.C., 2021. Bridging the Gap between Protoplanetary and Debris Disks: Separate Evolution of Millimeter and Micrometer-size...

    1988

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.