EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 699 (July 2025) A&A, 699 (2025) A110 Full HTML
Open Access
Issue
A&A
Volume 699, July 2025
Article Number A110
Number of page(s) 8
Section Planets, planetary systems, and small bodies
DOI https://doi.org/10.1051/0004-6361/202554255
Published online 09 July 2025

© The Authors 2025

Licence Creative CommonsOpen Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This article is published in open access under the Subscribe to Open model. Subscribe to A&A to support open access publication.

1 Introduction

Comets probably formed by the accretion of centimetersized pebbles (Blum et al. 2017) composed of dust particles (Levasseur-Regourd et al. 2018) embedding ices (Fulle et al. 2020). The derived structure of cometary nuclei (Ciarniello et al. 2022) is consistent with the dust ejection from nuclei (Fulle et al. 2020, 2022) and with most of the data collected during the Rosetta mission (Ciarniello et al. 2023; Fornasier et al. 2023).

In this paper we focus on the alkalis Na, K, and Li in comets. While sodium has been observed in many comets (Combi et al. 1997; Rauer et al. 1998; Watanabe et al. 2003; Leblanc et al. 2008; McKay & Cochran 2014), potassium has been measured in the spectra of C/1965 S1 (Preston 1967), C/2011 L4 (Fulle et al. 2013) and C/2012 S1 (McKay & Cochran 2014). The Na/K atomic ratio measured in C/1965 S1 and C/2011 L4 was much higher than the solar Na/K = 15.6 (Lodders 2003).

Lithium has never been detected in the spectra of comets. It is produced in the standard Big Bang nucleosynthesis but the value predicted once the baryon density is taken from the measure of primordial deuterium or from the cosmic background radiation is a factor of 3-4 higher than that measured in the old stars of the Galaxy. The sites for the lithium Galactic production remain elusive, though recently novae have been proposed as the most likely source (cf. Molaro et al. 2023 and references therein). A measure of lithium in comets could provide fresh clues to clarify one or both of these problems. The only ground-based LiI detection occurred during the impact of comet D/1993 F2 Shoemaker-Levy 9 on Jupiter (Roos-Serote et al. 1995). The Na/Li ratio extracted from the spectra of the plume in Jupiter’s atmosphere was consistent with a chondritic ratio Na/Li = 1.04 × 103 (Lodders 2003), although polluted by material from Jupiter’s deep atmosphere (Costa et al. 1997). Mass spectrum analyses of dust ejected by comet 81P/Wild 2 and returned by the Stardust mission allowed Flynn et al. (2006) to determine Li, Na, and K abundances that are higher than the chondritic values.

Fulle et al. (2013) considered the photoionization of the atoms ejected from the dust and flying from the sunward coma to the alkali tails observed in the slit of the spectrograph. This model implies a cross section of the alkali source (e.g., of optical thickness 10−6 for a loss rate of 103 kg s−1 of millimeter-sized dust) much smaller than that of the Moon, Mercury, and the nucleus of giant comets (e.g., C/1995 O1 Hale-Bopp), where sputtering extracts sodium efficiently (Combi et al. 1997).

2 Observations

Comets C/2020 F3 NEOWISE and C/2024 G3 ATLAS were observed with the Multi-Mode Spectrograph at the 0.84 m telescope of the Osservatorio Astronomico Schiaparelli in Campo dei Fiori, Varese, Italy (Munari & Valisa 2014). The collected spectra are the first high-resolution ones of comets observed at rh < 0.4 au since C/1965 S1 (Table 1), and show strong NaI and KI lines (Figure 1). The multi-order Echelle spectra cover the spectral range 424-864 nm at a resolving power λ/Δλ ≈ 104.

Data reduction was performed in IRAF and included all the usual steps for bias, dark, flat, and wavelength calibration. Reference spectra with ArTh Hollow Cathode Lamp were recorded just before and after the science exposures.

Table 1

Log of observations.

thumbnail Fig. 1

Spectra of comet C/2020 F3 NEOWISE (upper panel) taken on 2020 July 10.1 UT and of comet C/2024 G3 ATLAS (lower panel) taken on 2025 January 15.7 UT (observation log in Table 1) shown over the complete observed wavelength range. Prominent features are the C2 and NH2 emission bands (upper panel), the NaI and KI emissions and the O2 telluric absorption bands in the red part of the spectra.

2.1 Comet C/2020 F3 NEOWISE

The spectrum of C/2020 F3 NEOWISE (Figure 1) was summed along the slit and shows a prominent C2 Swan band, and NaI and KI emissions. Spectra of the standard star HR1558 were acquired before the spectra of C/2020 F3, and twilight spectra far from the comet just after. The NaI coma shows a Gaussian profile along the slit. No atmospheric NaI lines were detected in the sky spectrum. The spectrum of the solar light reflected by the cometary dust is much fainter than the twilight spectrum and was not detected. The exceptionally intense NaI λλ 5889.95 and λλ 5895.92 Â lines (Figure 2) were blueshifted by 59 and 58 km s−1 , respectively, consistent with the values in the ephemerides.

The KI λλ 7664.90 and λλ 7698.96 Å lines are shown in Figures 3 and 4. The KI λλ 7664.90 Å line shows the presence of the strong telluric O2 line at λλ 7664.73 Å. Once corrected for the atmospheric absorption, the emissions were blueshifted with velocities of 65 and 59 km s−1 in fair agreement with the sodium emission. The blueshift of the KI λλ 7664.90 Å line is 6 km s−1 higher than all the other lines (but only 20% of the spectral resolution) and may be due to the pollution of the telluric O2 line (Figure 3), so that we measured the Na/K ratio on the other KI line (Figure 4). No emission was detected at the LiI λλ 6707.78 Å line (Figure 5).

thumbnail Fig. 2

Spectrum of comet C/2020 F3 in the NaI 5890 Å D line region. The flux has been normalized to 1.9 × 10−11 erg cm−2 s−1 Å−1.
thumbnail Fig. 3

Spectrum of comet C/2020 F3 in the KI 7665 Å line region. The observed spectrum (blue line) has been corrected (black line) for the sky contamination using the twilight spectrum (red line). The flux has been normalized to the twilight continuum of 1.04 × 10−11 erg cm−2 s−1 Å−1.
thumbnail Fig. 4

Spectrum of comet C/2020 F3 in the KI 7700 Å line region. The observed spectrum (blue line) has been corrected (black line) for the sky contamination using the twilight spectrum (red line). The flux has been normalized to the twilight continuum of 1.2 × 10−11 erg cm−2 s−1 Å−1.

2.2 Comet C/2024 G3 ATLAS

Comet C/2024 G3 was observed just after sunset during an extremely clear evening with dry atmosphere (humidity of 25% with 3.7° C). The sky brightness was overwhelming, and the comet was barely perceptible in the pointing and guiding cameras, whose exposure time was decreased to 64 microseconds to avoid saturation. Due to the manual guiding and the poor seeing, the NaI coma uniformly filled the slit. The twilight spectrum was acquired on the next day at exactly the same time and sky location. No spectrum of a standard star was possible at the huge airmass of the comet, so that flux calibration was done by means of archived spectra of standard stars taken with the same spectrograph at the highest airmasses. This procedure did not allow us to evaluate the error affecting the absolute flux, although it did not affect the precision of the intensity ratio between spectral lines.

The spectrum of C/2024 G3 ATLAS (Figure 1) was summed along the slit and was characterized by prominent NaI emissions. The NaI λλ 5889.95 and 5895.92 Å lines (Figure 6) show a P Cygni profile due to a significant absorption at rest by the atmospheric daylight solar spectrum. When this was corrected by means of a twilight spectrum, the D2 and D1 emissions were redshifted by 34.1 and 34.7 km s−1, respectively, lower than the values in the ephemerides by about 20% of the spectral resolution. The KI λλ 7698.96 Å line was also found in emission (Figure 7). When the atmospheric absorption was subtracted, the emission was redshifted by 41.8 km s−1, consistent with the values in the ephemerides. The KI λλ 7664.90 Å line was completely absorbed by a strong telluric O2 line at λλ 7664.73 Å and was too weak to be detected. No emission was detected at the LiI λλ 6707.78 Å line (Figure 8).

thumbnail Fig. 5

Spectrum of comet C/2020 F3 around the Li 6707 Å line. The observed spectrum (blue line) has been corrected (black line, shifted up by 0.05) for the sky contamination using the twilight spectrum (red line). The flux has been normalized to 1.4 × 10−11 erg cm−2 s−1 Å−1.
thumbnail Fig. 6

Spectrum of comet C/2024 G3 in the NaI 5890 Å D line region. The observed spectrum (blue line) has been corrected (black line) for the sky contamination using the twilight spectrum (red line). The flux has been normalized to the twilight of 1.75 × 10−9 erg cm−2 s−1 Å−1.

2.3 Abundances

The intensities or limits of the alkali emissions are reported for comet C/2024 G3 ATLAS (Table 2) and for comet C/2020 F3 NEOWISE (Table 3), shown according to increasing rh values.

The observed abundance of sodium related to an atom x, Na/xobs (Tables 2, 3, and 4) depend on the g-factors gx at 1 au computed by Fulle et al. (2013) and were obtained from the line intensity I by means of the relationship (Na/x)obs=(gx/gNa)(INa/Ix).(Na/x)_{obs} = (g_x/g_{Na}) (I_{Na}/I_x).(1)

thumbnail Fig. 7

Spectrum of comet C/2024 G3 in the KI 7700 Å line region. The observed spectrum (blue line) has been corrected (black line) for the sky contamination using the twilight spectrum (red line). The flux has been normalized to the twilight continuum of 2.84 × 10−9 erg cm−2 s−1 Å−1.
thumbnail Fig. 8

Spectrum of comet C/2024 G3 in the LiI 6707 Å line region. The observed spectrum (blue line) has been corrected (black line) for the sky contamination using the twilight spectrum (red line). The flux has been normalized to the twilight of 3.1 × 10−9 erg cm−2 s−1 Å−1.

3 Potassium and lithium tails in comets

We detected potassium in emission in comets C/2011 L4 PanSTARRS (Fulle et al. 2013), C/2020 F3 NEOWISE, and C/2024 G3 ATLAS, so that we can predict that a potassium tail should be observable in future comets. The ratio of the solar radiation pressure to gravity forces, β, is higher for NaI than for KI, βNa = 71 ± 8, and βk = 53 ± 2 (Fulle et al. 2013), so that the KI tail should lie close to the NaI tail though resolved at a slightly larger distance from the Sun-Comet prolonged radius vector. With respect to sodium, the acceleration is about 3/4 and the photoionization lifetime is 1/4. Therefore, the length of the potassium tail should be ≈ 1/5 of the sodium tail.

The Li tail, with βLi = 440 ± 7 (Fulle et al. 2013), may never be observed because a lithium emission line has never been detected. The higher β value for Li would place its tail close to the Sun-Comet prolonged radius vector. With respect to NaI, the acceleration is six times higher and the photoionization lifetime is 1/10 (Fulle et al. 2013). Therefore, the length of the LiI tail should be approximately one-half of the length of the NaI tail.

Precise sky-projected positions of the alkali and neutral iron (Fulle et al. 2007) tails in comet C/2024 G3, assuming that lithium is present, are plotted in Figure 9. Even though FeI lines are not detected in the C/2024 G3 spectrum, due to its low S/N, C/2024 G3 actually shows only the iron tail, probably due to the higher iron abundance, solar Fe/Na = 14.6 (Lodders 2003), and to the slower photoionization rate (Fulle et al. 2007). FeI lines are observed from the UV to the IR spectrum, although the strongest lines are observed at λ < 420 nm. The first FeI tail was detected in C/2006 P1 at 630 < λ < 730 nm (Fulle et al. 2007). No dust particles may reach β ≈ 6 (Fülle et al. 2007). We encourage observers to detect these tails in the next coming bright comets with small perihelia by means of narrowband filters centered on the elemental transitions. The intensities of the KI and LiI tails relative to the NaI tail scale according to Eq. (1).

thumbnail Fig. 9

Left panels: C/2024 G3 ATLAS observed by the SOHO spacecraft on 2025 Jan. 13.446-13.487 UT in four LASCO C3 passbands (Sun at bottom, images 1.2 × 107 km wide). The horizontal trail is visible in the orange passband only (540-640 nm). Right panel: SOHO image taken on 2025 Jan. 14.454 UT in the clear passband. The curved continuous lines are the predicted atomic neutral tails, from left to right: LiI tail (syndyne of β = 440); NaI tail (syndyne of β = 71); KI tail (syndyne of β = 53); FeI tail (syndyne of β = 6). Dotted line: comet orbit fitting the trail.
Table 2

Atomic parameters in comet C/2024 G3 ATLAS, rh = 0.15 au.

Table 3

Atomic parameters in C/2020 F3 NEOWISE, rh = 0.36 au.

Table 4

Atomic parameters in C/2011 L4 PanSTARRS, rh = 0.46 au.

thumbnail Fig. 10

From left to right: C/2024 G3 ATLAS (ESA-NASA, SOHO) in the LASCO C3 clear passband observed on 2025 Jan 11.646, 13.204 and 14.746 UT, respectively. Vertical axis: Antisunward direction in 106 km units. Curved line: FeI tail. Dotted line: comet orbit fitting the trail.

4 Sodium trail

Contrary to the FeI and dust tails, which show a similar brightness in all the SOHO LASCO C3 passbands, the trail of C/2024 G3 (defined as a streamer of debris along the comet orbit, Figure 9) is the first ever observed in an Oort cloud comet and is visible in the orange passband only, along an orbital arch covered by the comet in ΔT ≈ 1 day. The projection of the comet orbit on the sky observed by SOHO fits the trail in all SOHO images of C/2024 G3 (Figure 10). No trail is detected outside 540 < λ < 640 nm, thus excluding solar light scattering by the trail debris and showing that the trail ejects sodium atoms. An ion tail (H2O+ bright band at λ = 620 nm) is excluded by the trail orientation.

The dynamics of the bodies building up the trail depend on the parameter β, which is linked to the diameter of the body d by β=3F⊙8πcGM⊙ρd=Cprρd,\beta = {{3 ~F_\odot} \over {8 \pi ~c ~G M_\odot ~\rho ~d}} = {C_{pr} \over {\rho ~d}},(2)

where F⊙ and M⊙ are the total solar radiation and mass, c is the speed of light, G is the gravitational constant, Cpr = 1.19 × 10−3 kg m−2 (Fulle 2004), and ρ is the bulk density of the trail body (e.g., ρ = 538 kg m−3 for the nucleus of comet 67P/Churyumov-Gerasimenko; Preusker et al. 2017).

Two centuries ago, the osculating orbit of C/2024 G3 had the inbound parameters q = 0.09445554 au and e = 0.99996987. For β → 0 and e > 0, the differences of the orbital parameters of the parent comet and of the body in the trail become Δe=1+e2−1+2(1+e)qrsβ(1−β)2−e=q(1+e)rs−1+e2eβ+O(β2),\Delta e = \sqrt{1 + {{e^2 - 1 + {2 (1 + e) q \over r_s} \beta} \over {(1 - \beta)^2}}} - e = {{q(1 + e) \over r_s} - 1 + e^2 \over e} \beta + O(\beta^2),(3) Δq=q(1+e)(1+e+Δe)(1−β)−q=(1−qrs)qeβ+O(β2),\Delta q = {q(1 + e) \over {(1 + e + \Delta e)(1 - \beta)}} - q = \left(1 - {q \over r_s}\right) {q \over e} ~\beta + O(\beta^2),(4)

where rs is the heliocentric distance at which the trail body and the nucleus split (Eqs. (2b) and (2c) in Fulle 1989). The comet orbit fits the trail axis within one-half an image pixel = 2 × 104 km (Figure 9), constraining the trail ejection either at the previous perihelion or at rs≈q+GM⊙(10ΔT/β)23≥25$r_s \approx q + \sqrt[3]{G M_\odot ~(10 ~\Delta T / \beta)^2} \ge 25$ au1 because e → 1, β ≤ 1.4 × 10−3 (Eq. (4)) and Δe ≤ 10−5 (Eq. (3)). The trail width implies ejection velocities ≤0.1 m s−1, making the trail origin inconsistent with a cometary outburst (Jewitt & Luu 2025). The orbital period of C/2024 G3 is T = [q/(1 -e)]3/2 yr. The difference between the trail body orbital period and T is ΔT=(q+Δq)3(1−e−Δe)3(1−β)−T=q32(1−β)[1−e−2qrsβ]32−T=[3(1−e)qrs−1]Tβ+O(β2).\Delta T = \sqrt {(q + \Delta q)^3 \over {(1 - e - \Delta e)^3(1 - \beta)}} - T = {{q^{3 \over 2} (1 - \beta)} \over {\left[1 - e - 2 {q \over r_s} \beta \right]^{3 \over 2}}} - T \\ = \left[{3 \over {(1 -e)}}{q \over r_s} - 1\right] ~T ~\beta + O(\beta^2).(5)

If the trail was ejected at the previous perihelion, then rs = q, i.e., ΔT ≤ 1 kyr if d ≥ 30 m: the trail would be too long and diluted to be detectable. If the trail was ejected at the previous aphelion Q, then rs = Q = q(1 + e)/(1 - e), i.e., ΔT/2 ≤ 1 day as observed if d > 30 m. Q lies in the Oort cloud, where some Galactic tides may have been strong enough to split the nucleus into the trail mininuclei. The perihelion solar tides definitely destroyed the main nucleus (King 2025). At the previous aphelion, a Galactic tide of similar strength overcame the gravity bonds among the sub-kilometer-sized mininuclei, and decelerated C/2024 G3 from the Oort cloud orbital speed GM⊙/Q$\sqrt{G M_\odot / Q}$ to the inbound speed (1−e)GM⊙/Q$\sqrt{(1-e) ~G M_\odot / Q}$ in a few years.

5 Discussion

Several processes are expected to extract alkali atoms from cometary nucleus and dust, namely thermal desorption, photon-stimulated desorption, and solar wind sputtering. From laboratory experiments on cosmic analogs it is expected that the atoms leaving the source maintain their original abundance (Leblanc & Doressoundiram 2011), and that the flux of ejected alkali is proportional to the source cross section. The trail of C/2024 G3 (Figures 9 and 10) suggests that sodium ejection requires high temperatures and also a large volume of the source. In fact, the mininuclei in the trail eject more sodium than the dust tail, which has a much larger total cross section per image pixel.

As for comets C/2011 L4 and C/1965 S1, the observed Na/Kobs ratio for comets C/2020 F3 and C/2024 G3 (Tables 2 and 3) is higher than the solar ratio, i.e. Na/K = 15.6 (Lodders 2003). The alkali tail photoionization model (Fulle et al. 2013) takes into account the alkali photoionization, by the solar UV radiation (Leblanc & Doressoundiram 2011), during their transfer into the alkali tails by solar radiation pressure, and provides Na/Kcor values consistent with the solar ratio (Table 4). However, the Na/Kobs of C/2020 F3 and C/2024 G3 are corrected to Na/Kcor values inconsistent with the solar ratio (Tables 2 and 3).

Regarding comet C/2020 F3, high-resolution spectra allowed Ye et al. (2020) to detect NaI and KI at 0.46 ≤ rh ≤ 0.48 au, and Bischoff & Mugrauer (2021) and Cambianica et al. (2021) to detect NaI at 0.6 ≤ rh ≤ 0.7 au, but not KI. Neither NaI nor KI was detected at 0.8 ≤ rh < 0.9 au. Comet C/1996 B2 Hyakutake showed NaI emission but no KI emission at 0.5 ≤ rh ≤ 0.7 au (Hicks & Fink 1997). Comets 1P/Halley and 153P/Ikeya-Zhang showed NaI emission at rh ≤ 0.8 au only (Combi et al. 1997; Watanabe et al. 2003). Comet C/2012 S1 ISON showed NaI and KI emissions at rh = 0.46 au (McKay & Cochran 2014). These facts, combined with our observations reported in Tables 2, 3, and 4, allow us to conclude that sodium is ejected at rh < 0.8 au, and potassium at rh < 0.5 au, suggesting an ejection process that depends on the different temperatures of sodium and potassium rather than on sputtering or thermal desorption.

The only comet showing NaI emission at rh > 0.8 au was C/1995 O1 Hale-Bopp, probably because its nucleus of radius −30 km (Lamy et al. 2004) made efficient alkali sputtering from its surface. The bulk of its NaI emission line showed a sunward extension of 2 104 km (Rauer et al. 1998), consistent with the radius of a gas collisional coma (Zakharov et al. 2018). Alkali sputtering at rh - 3 au was detected by ROSINA when Rosetta was less than 10 km from the nucleus surface of 67P/Churyumov-Gerasimenko (Wurz et al. 2015). A sodium source from dust-dust collisions requires dust nanograins (Ip & Jorda 1998), excluded by the Rosetta mission (Levasseur-Regourd et al. 2018). Leblanc et al. (2008) excluded thermal desorption as a significant sodium source by comparing the Na loss rates from C/1995 O1 and C/2006 P1. The sunward extension of C/2006 P1 NaI line of −104 km (Leblanc et al. 2008) is consistent with a gas collisional coma if the nucleus radius is >10 km, and the observed NaI emission along the trailing orbit may suggest a sodium trail much shorter than the C/2024 G3 trail.

The mass spectrum analysis of the craters in the aluminum foil exposed to the dust flux during the Stardust mission allowed Flynn et al. (2006) to determine the elemental composition of silicates and organics ejected by comet 81P/Wild 2; they found lithium, sodium, and potassium in higher abundances with respect to the chondritic values. The fact that lithium was never detected in the spectra of comets (e.g., in comet C/2020 F3 characterized by a Na/Li 34 times higher than the solar value) suggests the need of a new source of the observed alkali atoms consistent with different activation temperatures and relative abundances.

Here we assume that a fraction of sodium and potassium is contained in the aromatic component of cometary dust (Sandford et al. 2006), namely in phenoxides (Kojc̆inović et al. 2024), where sodium and potassium oxides are bonded to a benzene ring. This assumption is consistent with the Na/Fe abundances that are higher than the chondritic value, as measured in dust of comets 1P/Halley by Giotto, in 81P/Wild 2 by Stardust, and in 67P/Churyumov-Gerasimenko by Rosetta (Flynn et al. 2006; Levasseur-Regourd et al. 2018). Giotto and Rosetta, unlike Stardust, found a chondritic abundance of potassium, suggesting that the potassium fraction in phenoxides may be lower than that of sodium by a factor of two to three, directly matching Na/Kobs.

Laboratory experiments have shown that carbon dioxide extracts sodium and potassium atoms from phenoxides forming salicylic acid (Kojc̆inović et al. 2024). In particular, at CO2 pressures higher than 1 MPa, phenoxides absorb carbon dioxide ejecting sodium atoms in the temperature range 350 < T < 450 K, while potassium atoms are ejected for 470 < T < 490 K. The observed increase in Na/Kobs as rh increases (Tables 2, 3, and 4) is consistent with the higher reaction temperature of potassium phenoxides with respect to that of sodium phenoxides. The mininuclei observed in the trail of C/2024 G3 are big enough to maintain CO2 ice below their surface. The NaI line profile of C/2020 F3 along the slit of the spectrograph has a FWHM ≈4 × 103 km, fitting the diameter of a gas collisional coma if the nucleus radius is ≥2 km (Zakharov et al. 2018). The NaI line intensities (Table 3) provide a C/2020 F3 loss rate of ≈3 × 1025 Na s−1, to be compared to the water loss rate of 4.8 × 1029 mol s−1 at rh = 0.34 au (Faggi et al. 2021). The C/2020 F3 Na loss rate is a factor ≥30 smaller than that from C/2006 P1 (Leblanc et al. 2008), fitting the squared ratio of the respective nucleus radii.

The water enriched blocks (WEB) model of cometary nuclei (Fulle et al. 2020; Ciarniello et al. 2022) provides the surface temperatures Ts and thermal gradients ▼T inside the centimetersized pebbles at the nucleus surface (Table 5), which are perfectly consistent with the temperatures required by the reaction between CO2 and alkali phenoxides: potassium phenoxides eject potassium atoms at 470 < T < 490 K (Kojc̆inović et al. 2024), i.e., at rh < 0.5 au (Table 5), and sodium phenoxides eject sodium atoms at 350 < T < 450 K (Kojc̆inović et al. 2024), i.e., at rh < 0.8 au (Table 5), as actually observed. The full temperature range 350 < T < 450 K is reached at rh < 0.5 au (Table 5), directly explaining the steep rh dependence of the sodium loss rate observed at 0.5 < rh < 0.8 au (Watanabe et al. 2003). According to the WEB model (Fulle et al. 2022), most of the carbon dioxide is present in the water-poor matrix embedding the WEBs where the draining of ices D is faster than the nucleus erosion E (Fulle et al. 2020), suggesting that the proposed chemical reaction may last from 45 to 80 minutes (Table 5). At the corresponding reaction temperatures, within these times about half of the phenoxides are converted into salicylic acid (Kojc̆inović et al. 2024). During the C/2024 G3 perihelion week, the erosion E reaches 2 m, exposing the nucleus pristine interior as well as its disintegration does.

At a given temperature, the water vapor pressure is thousands of times lower than that of carbon dioxide, thus excluding chemical reactions between water vapor and alkali phenoxides. On the other hand, alkali phenoxides easily react with terrestrial liquid water (Kojc̆inović et al. 2024), making their presence in CI chondrites improbable, although phenol-derivates in meteorites can be generated through Fischer-Tropsch type reactions in the solar nebula (Hayatsu et al. 1980). No specific laboratory analysys in search of phenoxides was performed on the Stardust samples (Sandford et al. 2006). Theoretical analysis demonstrates the potential presence of phenol-like molecules in the interstellar medium (Ashworth et al. 2021; Ghosh et al. 2022).

At the pebble temperatures listed in Table 5, the CO2 pressure is always higher than 1 MPa (Fulle et al. 2022), without implying a significant dust ejection because D > E. This is consistent with the apparent lack of dust tails associated with the mininuclei in the sodium trail of C/2024 G3. The proposed chemical reaction is a significant CO2 energy and mass sink, which may avoid CO2 pressure overcoming values of a few megapascals. Alkali atoms should be extracted from homogeneous porous aggregates of monomer grains 0.1 μm in size (Mannel et al. 2019), i.e., dust particles that may survive at the megapascal pressures of CO2 inside them (Blum et al. 2017). Fireball observations suggest internal tensile strengths from 0.01 to 10 MPa in meteoroids associated with comet 2P/Encke (Wetherill & ReVelle 1982).

Lithium was not detected in the spectra of four comets where it looked for. The most stringent upper limit was derived for comet C/2020 F3 where Li/Na is found to be a factor of 34 below the solar value. Assuming a solar Na abundance, this ratio implies a lithium abundance of A(Li) < 1.75, where A(Li) = log(Li/H) + 12. This is at variance with the Stardust samples where it is more abundant than in CI chondrites, namely A(Li) > 3.28 (Flynn et al. 2006). The derived upper limit is one order of magnitude lower than the A(Li) ≈ 2.7 predicted by theoretical primordial nucleosynthesis and also lower than the A(Li) ≈ 2.2 observed in the dwarf stars of the Galactic halo (Navas et al. 2024). Lithium in comets may be present either in silicates only, or in phenoxides reacting with carbon dioxide according to different reaction pathways (Markovic et al. 2007) characterized by an Arrhenius factor 104 times lower than the reactions of sodium phenoxide (Staude & Hussain 1971).

Table 5

Nucleus parameters according to the WEB model.

6 Conclusions

  1. We analyzed the first high-resolution spectra of comets observed at heliocentric distances rh < 0.4 au since C/1965 S1, showing the strongest NaI and KI lines ever recorded.

  2. The excess of the Na/K ratio with respect to the solar value and its observed trend with the heliocentric distance are consistent with alkali phenoxides in the aromatic fraction of cometary dust (Sandford et al. 2006) reacting with CO2 to eject sodium and potassium atoms (Kojc̆inović et al. 2024).

  3. Lithium has never been detected in the spectra of comets. The Li upper limit for comet C/2020 F3 is very stringent at a factor of 34 below the solar value, at variance with the Stardust samples where it is more abundant than in CI chondrites (Flynn et al. 2006). The Na/K ratio is always greater than the solar ratio by a factor of two to three. KI is detected at rh < 0.5 au, whereas NaI at rh < 0.8 au. These three facts exclude thermal, photon-stimulated, and solar-wind desorptions, which should maintain cosmic abundances.

  4. Lithium may be present either in silicates only, orin phenoxides reacting with carbon dioxide at a rate 104 times slower than sodium phenoxides (Staude & Hussain 1971).

  5. The widespread chemistry of carbon dioxide with organic dust (Liu et al. 2015) may provide a significant energy and mass sink of carbon dioxide in all comets, also at heliocentric distances >1 au, reconciling models of cometary activity (Attree et al. 2024a,b) with Rosetta CO2 measurements.

  6. For comets observed at heliocentric distances <0.5 au, we predict a potassium neutral tail spatially resolved from the sodium tail, and encourage its observation.

  7. C/2024 G3 is the first Oort cloud comet to show a trail, detected thanks to its sodium emission and possibly composed of sub-kilometer-sized mininuclei fragmented in the Oort cloud by the Galactic tides that decelerated C/2024 G3.

Acknowledgements

We thank an anonymous referee for having significantly improved a previous version of the manuscript. We thank J. Agarwal, R. Ligustri, J. Markkanen and F. Moreno for useful discussions about the C/2024 G3 trail. We are grateful to A. Milani for quickly modifying the pointing camera acquisition software to reduce the exposure time thus avoiding saturation. Part of this work was supported by the ASI-INAF agreement 2023-14-HH.0 project number.

References

  1. Ashworth, E. K., Anstöter, C. S., Verlet, J. R. R., & Bull, J. N. 2021, Phys. Chem Chem. Phys., 23, 5817 [Google Scholar]
  2. Attree, N., Gutièrrez, P., Groussin, O., et al. 2024a, A&A, 690, A82 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  3. Attree, N., Schuckart, C., Bischoff, D., Gundlach, B., & Blum, J. 2024b, MNRAS, 535, 65 [Google Scholar]
  4. Bischoff, R., & Mugrauer, M. 2021, Astron. Nach., 342, 833 [Google Scholar]
  5. Blum, J., Gundlach, B., Krause, M., et al. 2017, MNRAS, 469, S755 [Google Scholar]
  6. Cambianica, P., Cremonese, G., Munaretto, G., et al. 2021, A&A, 656, A160 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  7. Ciarniello, M., Fulle, M., Raponi, A., et al. 2022, Nat. Astr., 6, 546 [Google Scholar]
  8. Ciarniello, M., Fulle, M., Tosi, F., et al. 2023, MNRAS, 523, 5841 [CrossRef] [Google Scholar]
  9. Combi, M. R., Di Santi, M. A., & Fink, U. 1997, Icarus, 130, 336 [NASA ADS] [CrossRef] [Google Scholar]
  10. Costa, R. D. D., de Freitas Pacheco, J. A., Singh, P. D., de Almeida, A. A., & Codina-Landaberry, S. J. 1997, ApJ, 485, 380 [Google Scholar]
  11. Faggi, S., Lippi, M., Camarca, M., et al. 2021, AJ, 162, 178 [NASA ADS] [CrossRef] [Google Scholar]
  12. Flynn, G. J., Bleuet, P., Borg, J., et al. 2006, Science, 314, 1731 [Google Scholar]
  13. Fornasier, S., Hoang, H. V., Fulle, M., Quirico, E., & Ciarniello, M. 2023, A&A, 672, A136 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  14. Fulle, M. 1989, A&A, 217, 283 [NASA ADS] [Google Scholar]
  15. Fulle, M. 2004, in Comets II, eds. M. C. Festou, H. U. Keller, & H. A. Weaver (Tucson: University of Arizona Press), 565 [Google Scholar]
  16. Fulle, M., Leblanc, F., Harrison, R. A., et al. 2007, ApJ, 661, L93 [Google Scholar]
  17. Fulle, M., Molaro, P., Buzzi, L., & Valisa, P. 2013, ApJ, 771, L21 [Google Scholar]
  18. Fulle, M., Blum, J., Rotundi, A., et al. 2020, MNRAS, 493, 4039 [CrossRef] [Google Scholar]
  19. Fulle, M., Lazzarin, M., La Forgia, F., et al. 2022, MNRAS, 513, 5377 [NASA ADS] [Google Scholar]
  20. Ghosh, R., Sil, M., Kumar, M. S., et al. 2022, Res. Astron. Astroph., 22, 065021 [Google Scholar]
  21. Hayatsu, R., Winans, R. E., Scott, R. G., et al. 1980, Science, 207, 1202 [Google Scholar]
  22. Hicks, M. D., & Fink, U. 1997, Icarus, 127, 307 [Google Scholar]
  23. Ip, W.-H., & Jorda, L. 1998, ApJ, 496, L47 [Google Scholar]
  24. Jewitt, D., & Luu, J. 2025, AJ, 169, 338 [Google Scholar]
  25. King, B. 2025, in Astronomy & Observing News, ed. AAS, Sky & Telescope, 1 [Google Scholar]
  26. Kojc̆inović, A., Likozar, B., & Grilc, M. 2024, Int. J. Mol. Sci., 25, 12923 [Google Scholar]
  27. Lamy, P., Toth, I., Fernandez, Y., & Weaver, H. 2004, in Comets II, eds. M. C. Festou, H. U. Keller, & H. A. Weaver (Tucson: University of Arizona Press), 223 [Google Scholar]
  28. Leblanc, F., & Doressoundiram, A. 2011, Icarus, 211, 10 [Google Scholar]
  29. Leblanc, F., Fulle, M., López Ariste, A., et al. 2008, A&A, 482, 293 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  30. Levasseur-Regourd, A.-C., Agarwal, J., Cottin, H., et al. 2018, Space Sci. Rev., 214, 64 [NASA ADS] [CrossRef] [Google Scholar]
  31. Liu, Q., Wu, L., Jackstell, R., & Beller, M. 2015, Nat. Comm., 6, 5933 [Google Scholar]
  32. Lodders, K. 2003, ApJ, 591, 1220 [Google Scholar]
  33. Mannel, T., Bentley, M. S., Boakes, P. D., et al. 2019, A&A, 630, A26 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  34. Markovic, S., Markovic, Z., Begovic, N., & Manojlovic, N. 2007, Russian J. Phys. Chem. A, 81, 1392 [Google Scholar]
  35. McKay, A. J., & Cochran, A. L. 2014, Lunar Planet. Sci. Conf., 1777, 2303 [Google Scholar]
  36. Molaro, P., Aguado, D. S., Caffau, E., et al. 2023, A&A, 679, A72 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  37. Munari, U., & Valisa, P. 2014, Contrib. Astron. Observ. Skalnaté Pleso, 43, 174 [Google Scholar]
  38. Navas, S., Amsler, C., Gutsche, T., et al. 2024, Phys. Rev. D, 110, 030001 [Google Scholar]
  39. Preston, G. W. 1967, ApJ, 147, 718 [Google Scholar]
  40. Preusker, F., Scholten, F., Matz, K. D., et al. 2017, A&A, 607, L1 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  41. Rauer, H., Arpigny, C., Manfroid, J., Cremonese, G., & Lemme, C. 1998, A&A, 334, L61 [NASA ADS] [Google Scholar]
  42. Roos-Serote, M., Barucci, A., Crovisier, J., et al. 1995, Geophys. Res. Lett., 22, 1621 [Google Scholar]
  43. Sandford, S. A., Alèon, J., Alexander, C. M. O. D., et al. 2006, Science, 314, 1720 [Google Scholar]
  44. Staude, E., & Hussain, A. 1971, Polymer J., 2, 468 [Google Scholar]
  45. Watanabe, J.-i., Kawakita, H., Furusho, R., & Fujii, M. 2003, ApJ, 585, L159 [Google Scholar]
  46. Wetherill, G. W., & ReVelle, D. O. 1982, in Comets, ed. L. L. Wilkening (Tucson: University of Arizona Press), 297 [Google Scholar]
  47. Wurz, P., Rubin, M., Altwegg, K., et al. 2015, A&A, 583, A22 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  48. Ye, Q., Zhang, Q., Brewer, J., Knight, M., & Kelley, M. 2020, AAS Div. Planet. Sci. Meeting, 52, 111.02 [Google Scholar]
  49. Zakharov, V. V., Ivanovski, S. L., Crifo, J. F., et al. 2018, Icarus, 312, 121 [NASA ADS] [CrossRef] [Google Scholar]
1

If e = 1, rs = q(1 + f2) and ΔT=kβq3/GM⊙(f+13f3)+O(β2)$\Delta T = k \beta \sqrt{q^3/G M_\odot} (f + {1 \over 3} f^3) + O(\beta^2)$: f is the tangent of half true anomaly, f≪13f3$f \ll {1 \over 3} f^3$ for 5 ≤ rs ≤ 500 au, where, if k = 0.3, rs differs by < 5% from precise computations done at e < 1.

All Tables

Table 1

Log of observations.

In the text
Table 2

Atomic parameters in comet C/2024 G3 ATLAS, rh = 0.15 au.

In the text
Table 3

Atomic parameters in C/2020 F3 NEOWISE, rh = 0.36 au.

In the text
Table 4

Atomic parameters in C/2011 L4 PanSTARRS, rh = 0.46 au.

In the text
Table 5

Nucleus parameters according to the WEB model.

In the text

All Figures

thumbnail Fig. 1

Spectra of comet C/2020 F3 NEOWISE (upper panel) taken on 2020 July 10.1 UT and of comet C/2024 G3 ATLAS (lower panel) taken on 2025 January 15.7 UT (observation log in Table 1) shown over the complete observed wavelength range. Prominent features are the C2 and NH2 emission bands (upper panel), the NaI and KI emissions and the O2 telluric absorption bands in the red part of the spectra.
In the text
thumbnail Fig. 2

Spectrum of comet C/2020 F3 in the NaI 5890 Å D line region. The flux has been normalized to 1.9 × 10−11 erg cm−2 s−1 Å−1.
In the text
thumbnail Fig. 3

Spectrum of comet C/2020 F3 in the KI 7665 Å line region. The observed spectrum (blue line) has been corrected (black line) for the sky contamination using the twilight spectrum (red line). The flux has been normalized to the twilight continuum of 1.04 × 10−11 erg cm−2 s−1 Å−1.
In the text
thumbnail Fig. 4

Spectrum of comet C/2020 F3 in the KI 7700 Å line region. The observed spectrum (blue line) has been corrected (black line) for the sky contamination using the twilight spectrum (red line). The flux has been normalized to the twilight continuum of 1.2 × 10−11 erg cm−2 s−1 Å−1.
In the text
thumbnail Fig. 5

Spectrum of comet C/2020 F3 around the Li 6707 Å line. The observed spectrum (blue line) has been corrected (black line, shifted up by 0.05) for the sky contamination using the twilight spectrum (red line). The flux has been normalized to 1.4 × 10−11 erg cm−2 s−1 Å−1.
In the text
thumbnail Fig. 6

Spectrum of comet C/2024 G3 in the NaI 5890 Å D line region. The observed spectrum (blue line) has been corrected (black line) for the sky contamination using the twilight spectrum (red line). The flux has been normalized to the twilight of 1.75 × 10−9 erg cm−2 s−1 Å−1.
In the text
thumbnail Fig. 7

Spectrum of comet C/2024 G3 in the KI 7700 Å line region. The observed spectrum (blue line) has been corrected (black line) for the sky contamination using the twilight spectrum (red line). The flux has been normalized to the twilight continuum of 2.84 × 10−9 erg cm−2 s−1 Å−1.
In the text
thumbnail Fig. 8

Spectrum of comet C/2024 G3 in the LiI 6707 Å line region. The observed spectrum (blue line) has been corrected (black line) for the sky contamination using the twilight spectrum (red line). The flux has been normalized to the twilight of 3.1 × 10−9 erg cm−2 s−1 Å−1.
In the text
thumbnail Fig. 9

Left panels: C/2024 G3 ATLAS observed by the SOHO spacecraft on 2025 Jan. 13.446-13.487 UT in four LASCO C3 passbands (Sun at bottom, images 1.2 × 107 km wide). The horizontal trail is visible in the orange passband only (540-640 nm). Right panel: SOHO image taken on 2025 Jan. 14.454 UT in the clear passband. The curved continuous lines are the predicted atomic neutral tails, from left to right: LiI tail (syndyne of β = 440); NaI tail (syndyne of β = 71); KI tail (syndyne of β = 53); FeI tail (syndyne of β = 6). Dotted line: comet orbit fitting the trail.
In the text
thumbnail Fig. 10

From left to right: C/2024 G3 ATLAS (ESA-NASA, SOHO) in the LASCO C3 clear passband observed on 2025 Jan 11.646, 13.204 and 14.746 UT, respectively. Vertical axis: Antisunward direction in 106 km units. Curved line: FeI tail. Dotted line: comet orbit fitting the trail.
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.