Lithium Dip in Mid-F Dwarfs
- Li-Dip is a narrow temperature range in mid-F dwarfs exhibiting a sharp drop in lithium abundance, challenging standard stellar models.
- Observations across clusters reveal a well-defined temperature boundary with metallicity-dependent mass scaling and distinct hot and cool sides.
- Evidence supports rotation-driven mixing as the primary mechanism behind the Li-Dip, correlating Li depletion with star spin-down and internal shear.
Li-Dip most commonly denotes the stellar lithium dip: a narrow temperature–mass interval on the upper main sequence in which mid-F dwarfs exhibit severe depletion of surface lithium relative to otherwise similar stars on either side of the feature. Lithium abundance is conventionally expressed as . In open clusters, the dip is observed near the mid-$6000$ K regime, with a sharp hot-side wall near the Kraft break and a cooler recovery toward a lithium plateau; its existence is non-standard because shallow surface convection zones in F stars should not, in standard stellar evolution, destroy significant Li during the main sequence (Deliyannis et al., 2019, Baugh et al., 2013). Cluster surveys now show that the dip’s morphology is jointly regulated by age, metallicity, and rotational evolution, while recent subgiant and young-cluster results have sharpened the case for rotation-driven internal mixing as the dominant mechanism (Twarog et al., 2020, Sun et al., 6 Jul 2025, Zhang et al., 7 Feb 2026).
1. Definition and observational phenomenology
The lithium dip was established observationally as a sharp, localized depletion of among mid-F dwarfs. In the Hyades, stars with effective temperatures roughly $6500$–$6850$ K show Li depleted by –$2$ dex compared to stars only $200$–$300$ K cooler or hotter; more generally, it is seen across about $6300$–$6000$0 K in disk clusters (Baugh et al., 2013). A broader synthesis for intermediate-age open clusters similarly places the Li-Dip at approximately $6000$1–$6000$2 K, with the “cool plateau” below it at roughly $6000$3–$6000$4 K (Twarog et al., 2020). In NGC 6819 the dip is described as a narrow temperature–mass domain centered near $6000$5–$6000$6 K, with a hot edge that drops across $6000$7, i.e. only $6000$8 K in $6000$9 (Deliyannis et al., 2019).
The morphology is not symmetric in all clusters, but several features recur. On the hot side, higher-mass stars can preserve near-primordial Li, often with 0–1, until they cross the hot wall and begin to deplete (Anthony-Twarog et al., 29 Oct 2025). Across the dip itself, detections are frequently replaced by upper limits, while on the cool side Li reappears and rises toward a plateau. In NGC 7789, for example, the hot wall lies at 2–3, the dip center at 4–5, the cool wall near 6, and the G-dwarf plateau reaches 7 by 8 (Anthony-Twarog et al., 29 Oct 2025). In NGC 6819, the cool-side plateau is 9, and the observed dispersion exceeds the $6500$0 dex spectroscopic error, implying astrophysical scatter in individual mixing histories (Deliyannis et al., 2019).
A longstanding misconception has been that the Li dip emerges only at ages $6500$1 Myr. That chronology is now challenged by the Snake stellar complex, where a clear Li dip is reported at $6500$2 Myr over $6500$3–$6500$4 K, centered near $6500$5 K, with $6500$6 dex at the minimum and a depletion depth $6500$7 dex relative to a $6500$8 dex reference level (Zhang et al., 7 Feb 2026). This does not erase the classical cluster sequence, but it shifts the earliest known onset to substantially younger ages.
2. Mass scale, metallicity dependence, and cluster calibration
Although the dip is often described in temperature space, open-cluster work shows that its mass scale shifts systematically with metallicity. NGC 2243 provides an explicit relation for the high-mass wall: $6500$9 with a slope of $6850$0 per dex (Anthony-Twarog et al., 2021). The broader open-cluster synthesis gives the same trend qualitatively: the hot boundary mass decreases by $6850$1–$6850$2 per dex drop in $6850$3, while age at fixed metallicity does not move the cliff (Twarog et al., 2020).
The empirical calibration across representative clusters is compactly summarized below.
| Cluster | Age and $6850$4 | Li-dip calibration |
|---|---|---|
| Hyades/Praesepe | $6850$5 Gyr, $6850$6 | $6850$7 (Anthony-Twarog et al., 2021) |
| NGC 7789 | $6850$8 Gyr, $6850$9 adopted | Li-wall at 0 K; dip-center mass 1 (Anthony-Twarog et al., 29 Oct 2025) |
| NGC 6819 | 2–3 Gyr, 4 | Dip center 5 to 6; high-mass boundary 7 (Deliyannis et al., 2019) |
| NGC 2204 | 8 Gyr, 9 | Blue wall at predicted mass $2$0 (Anthony-Twarog et al., 2024) |
| NGC 2243 | $2$1 Gyr, $2$2 | Dip spans $2$3 to $2$4; center $2$5 (Anthony-Twarog et al., 2021) |
These calibrations support two complementary statements. First, the hot wall is a robust inter-cluster marker. Second, the dip is better regarded as tied to a near-constant main-sequence $2$6 regime than to a fixed stellar mass. NGC 6819 is especially instructive: when its turnoff stars are mapped back to their main-sequence $2$7 at the Hyades age, the evolved-forward profile matches Hyades/Praesepe after a modest alignment offset of $2$8 K with VR isochrones or $2$9 K with $200$0 isochrones, including the hot-edge wall, the depth, and the cool-side rise (Deliyannis et al., 2019). This strongly supports a temperature-defined boundary whose mass coordinate shifts with $200$1.
3. Rotation, the Kraft break, and time-dependent formation
The central observational result of modern Li-dip work is that Li depletion tracks rotational evolution. In the Hyades/Praesepe, the Li minimum lies just below $200$2 K, where $200$3 begins its steep decline toward cooler stars: from $200$4 km s$200$5 near $200$6 K to $200$7 km s$200$8 by $200$9 K (Deliyannis et al., 2019). NGC 7789 identifies the Li-wall at $300$0 K as the true hot boundary of the Kraft break, with the onset of rapid spin-down and mixing occurring precisely there (Anthony-Twarog et al., 29 Oct 2025).
The rotational structure across the turnoff in NGC 7789 is particularly explicit. Above the Li-wall, $300$1 spans $300$2–$300$3 km s$300$4, with median $300$5 km s$300$6 for the brightest MSTO stars. Within the Li-dip regime, rotation contracts to $300$7 km s$300$8, with many stars between $300$9 and $6300$0 km s$6300$1. Cooler than the dip, $6300$2 declines to $6300$3–$6300$4 km s$6300$5 by $6300$6, and then to $6300$7 km s$6300$8 well below the plateau (Anthony-Twarog et al., 29 Oct 2025). The co-evolution of $6300$9 and $6000$00 across these three regimes is the observational basis for treating the dip as a spindown-governed phenomenon.
Cross-cluster comparisons extend this to older ages. The sequence Hyades/Praesepe $6000$01 NGC 7789/752 $6000$02 NGC 3680/2506 $6000$03 NGC 6819 shows progressive contraction of the hot-side rotation distribution and a growing fraction of Li-poor stars above the dip (Deliyannis et al., 2019, Anthony-Twarog et al., 29 Oct 2025). In the youngest sample, hotter stars are fast rotators and largely Li-rich; by the age of NGC 6819, the mean and dispersion of $6000$04 above the dip are as low as within the dip, and many hot-side stars exhibit Li upper limits comparable to dip members (Deliyannis et al., 2019). This has been described as a later-time second dip at higher mass, produced when A/F turnoff stars evolve through the red hook and hydrogen-exhaustion phase, slow substantially, and develop deeper surface convection zones that couple to previously mixed layers (Deliyannis et al., 2019).
Recent work on the Snake clusters strengthens the rotational interpretation at much younger ages. Within the $6000$05–$6000$06 K dip window, fast rotators with $6000$07 km s$6000$08 have $6000$09 dex, whereas slow rotators with $6000$10 km s$6000$11 have $6000$12 dex, with the binwise difference persisting across the dip (Zhang et al., 7 Feb 2026). The paper interprets this as stronger rotational shear at the convective–radiative boundary in faster rotators, enhancing turbulent mixing and accelerating Li destruction.
4. Subgiants, giants, and the post-main-sequence imprint
The Li dip is not only a main-sequence anomaly; it also sets the initial conditions for post-main-sequence Li evolution. In NGC 6819, only $6000$13 of single subgiants and giants have measurable Li, and most upper limits in post-turnoff stars are $6000$14 (Deliyannis et al., 2019). In NGC 2243, stars at the RGB base typically have only upper limits around $6000$15, while the upper RGB and red clump show no detections, with upper limits $6000$16 (Anthony-Twarog et al., 2021). In NGC 2204, $6000$17 is below $6000$18 at the base of the RGB and declines to a detectable value of $6000$19 at the tip; six probable AGB stars and all but one red clump star show only upper limits (Anthony-Twarog et al., 2024).
Younger clusters preserve a somewhat less depleted giant-branch record. In NGC 7789, probable first-ascent giants begin near $6000$20–$6000$21 and decline toward the RGB tip, reaching $6000$22 by $6000$23; clump and AGB stars show a preponderance of upper limits, typically $6000$24 where sensitivity is highest (Anthony-Twarog et al., 29 Oct 2025). The 1–3 Gyr cluster synthesis generalizes this sequence: in younger clusters such as NGC 7789 and NGC 2506, giants at the RGB base reach $6000$25, decline weakly until the RGB bump, then approach $6000$26; in older clusters such as NGC 6819 and NGC 2243, upper limits dominate even among subgiants and giants (Twarog et al., 2020).
The most discriminating post-main-sequence evidence comes from subgiants in NGC 188. Their deepening surface convection zones dredge up the subsurface Li profile carved during the main sequence, and the observed $6000$27–$6000$28 decline is shallow and monotonic rather than flat, steeply plunging, or initially rising (Sun et al., 6 Jul 2025). In that framework, standard stellar evolution would give a nearly flat preservation region, mass loss would give an extremely steep drop immediately below the shallow convection zone, and diffusion would produce a Li pile-up that should first raise $6000$29 as dredge-up begins. The NGC 188 subgiants instead reveal a gradual monotonic decline of Li with depth, the profile expected from rotationally induced shear mixing (Sun et al., 6 Jul 2025).
5. Physical interpretation, competing mechanisms, and model space
Standard stellar evolution fails because the surface convection zones of hot F dwarfs are too shallow to transport Li down to $6000$30 K, where Li is destroyed (Deliyannis et al., 2019, Baugh et al., 2013). The physical problem is therefore one of non-standard transport below the surface convection zone. Proposed mechanisms have included rotational mixing, magnetic effects, atomic diffusion, gravity-wave mixing, and mass loss (Sun et al., 6 Jul 2025).
Several empirical tests favor rotation-driven mixing. A decisive binary test is provided by V505 Per, a $6000$31 d short-period eclipsing binary whose components have $6000$32 and $6000$33 at $6000$34 K and $6000$35 K, respectively, placing both stars within the Li dip by ZAMS $6000$36 yet at Li levels larger by factors of $6000$37–$6000$38 than comparable single stars in NGC 752, NGC 3680, Hyades, and Praesepe (Baugh et al., 2013). The interpretation is that early tidal synchronization suppressed main-sequence angular momentum loss and therefore reduced rotationally induced mixing.
Multi-element depletion patterns point the same way. In the Hyades, Li, Be, and B are destroyed at progressively higher temperatures—$6000$39, $6000$40, and $6000$41 K—and the Li–Be–B dip shows the expected ordered depletion, with Li depleted by $6000$42 dex, Be by about an order of magnitude, and B by about a factor of $6000$43–$6000$44 across the mid-F regime (Boesgaard et al., 2016). The Hyades Be–B correlation has slope $6000$45 over $6000$46–$6000$47 K, while earlier Li–Be work yielded slopes near $6000$48–$6000$49 (Boesgaard et al., 2016). These depth-dependent depletions are reproduced by rotational-mixing models and are difficult to reconcile with pure diffusion or simple mass loss.
Theoretical work has expanded the rotational picture rather than replacing it. The gyroscopic-pumping model treats differential rotation maintained by convective stresses as the driver of large-scale meridional circulation in stars with both an outer convective envelope and an inner convective core. For $6000$50–$6000$51 stars, gyroscopically pumped flows can circulate material between the surface and Li/Be-burning layers; by themselves they overestimate hot-side depletion, but when diffusion of chemical species back into the surface convection zone is included, the model fits the Hyades Li and Be profiles for reasonable parameters (Garaud et al., 2010). A later meridional-circulation framework likewise reproduces the cool side of the Li dip in most clusters and the hot side in systems such as the Hyades, but explicitly notes limitations in fully reproducing dip morphology because of the rotation-velocity distribution of sample stars in this temperature range (Li et al., 27 Aug 2025).
The strongest recent claim is that rotationally induced mixing driven by angular momentum loss is the unique mechanism consistent with the subgiant dredge-up signature in NGC 188 (Sun et al., 6 Jul 2025). That conclusion is narrower than a general dismissal of all other transport processes; it means that, for the main-sequence Li-Dip, alternative primary explanations such as diffusion and mass loss fail the subsurface-profile test. A plausible implication is that secondary processes may still modulate the hot side or the exact depth of depletion, but not define the phenomenon’s basic morphology.
6. Terminological extensions and unrelated uses
Outside stellar astrophysics, the string “Li-Dip” has been used for unrelated lithium-centered phenomena. These usages are homonymous rather than conceptually connected.
In two-dimensional materials chemistry, “Li-Dip” denotes the enhancement of CO$6000$52 capture by Li adsorption on beryllonitrene. Pristine beryllonitrene adsorbs CO$6000$53 only weakly, with best-site $6000$54 eV, whereas Li-doped beryllonitrene binds CO$6000$55 much more strongly, with $6000$56 to $6000$57 eV and bent O–C–O geometries characteristic of chemisorption (Pu et al., 2021).
In detector chemistry, “Li-Dip” has also designated waterless, direct loading of $6000$58Li into diisopropylnaphthalene liquid scintillators. At $6000$59 cm scale, these formulations were reported to achieve light output up to $6000$60 and effective attenuation length up to $6000$61 relative to undoped EJ-309, while preserving robust pulse-shape discrimination (Zaitseva et al., 2023).
In battery materials, the “Li-Dip” in $6000$62 refers to the pronounced minimum in Li-ion diffusivity centered at $6000$63. Atomistic calculations attribute it to zig-zag ordering in half-filled Li layers, which raises the migration barrier parallel to the zig-zag chains to $6000$64 eV while leaving the perpendicular barrier at $6000$65 eV, thereby throttling domain-averaged transport (Bunjaku et al., 2015).
In condensed-matter NMR, the shorthand can refer to the Li-site magnetic dipole field in $6000$66. There the component parallel to a $6000$67 T external field reaches a maximum value of $6000$68 G at a specific orientation, with an inferred effective Mo moment of $6000$69 per ion (Wu et al., 2014).
In current astronomical usage, however, Li-Dip overwhelmingly denotes the lithium dip in stars: a metallicity-dependent mass feature anchored to a nearly fixed $6000$70 regime, formed by non-standard mixing below shallow convection zones, and most strongly linked by present evidence to rotational spindown and the shear-driven transport that follows (Deliyannis et al., 2019, Sun et al., 6 Jul 2025).
