POTASSIUM DETECTION AND LITHIUM DEPLETION IN COMETS C/2011 L4 (PANSTARRS) AND C/1965 S1 (IKEYA-SEKI)

Comets provide unique information on the cosmic abundances of the solar nebula which collapsed to form the solar system. Besides sodium, which is already detected in many comets, very few data are available regarding the alkali content of comet nuclei. In particular, potassium was remotely detected (Preston 1967) in spectra of the comet C/1965 S1 (Ikeya-Seki), where the Na/K ratio was as high as that observed in Mercury (Killen et al. 2010). A detection of the potassium line at 7698.9645 Å in the comet C/1995 O1 (Hale-Bopp) was reported by Fitzsimmons & Cremonese (1997), but no Na/K ratio has yet been extracted from this spectrum. Laboratory analyses of samples collected from the comet 81P/Wild 2 showed that most potassium is in the form of eifelite and K-feldspar grains (Zolensky et al. 2006). Detections of the potassium and lithium lines were also reported from the impacts of the comet D/1993 F2 (Shoemaker-Levy 9) on Jupiter (Roos-Serote et al. 1995). The Na/Li ratio extracted from spectra of the plume in Jupiter's atmosphere was consistent with a meteoritic ratio (Costa et al. 1997), although these transient emissions, unlike resonant fluorescence, could not be readily converted into atomic abundances and were probably contaminated by the alkali of Jupiter's deep atmosphere. In this Letter, we discuss the remote detection of potassium in the comet C/2011 L4 (Panstarrs), which allows us to discuss the Na/K abundance in this comet. In addition to models of Mercury's exosphere (Leblanc & Doressoundiram 2011), we will discuss the transfer of alkali atoms from the parent bodies (mainly the dust grains in the sunward coma) to the beginning of alkali tails (Fulle et al. 2007) where they have been observed. This will allow us to infer estimates of lithium abundances in comets. The physical process relating alkali line intensities to atomic abundance is very simple, namely, resonance fluorescence (Swing 1941). The absorption of solar radiation in the resonance transitions populates atomic upper levels, which, trickling down, give rise to the emission lines. The population in the upper level depends on the energy available at the considered wavelength, i.e., upon whether or not a Fraunhofer line comes in the way of absorption. These Fraunhofer lines as seen from the comet have different Doppler shifts depending on the radial heliocentric velocity v (the so-called Swing effect).

High-resolution (λ/Δλ ≈ 10⁴ in the spectral range 424–864 nm) Echelle spectra (Figure 1) of the comet C/2011 L4 (Panstarrs) were obtained on 21.8 UT 2013 March (mid-exposure time) with the Multi-Mode Spectrograph mounted on the 0.6 m telescope at the Osservatorio Astronomico Schiaparelli located in Campo dei Fiori, Varese, Italy (Ashish et al. 2012). The multi-order Echelle spectra were absolutely flux calibrated against spectra of the standard star HR 464 located on the sky nearby, observed immediately after the comet, and then merged into a one-dimensional continuous spectra. The slit was east–west (E–W) oriented and guided on the brightest coma by means of a monitor of a guiding CCD camera covering 3 × 4 arcmin² on the sky. The slit width projected on the sky was set to 4'', and the slit length to 17''. The total exposure time was 40 minutes. At the observations, the Sun–comet distance was r = 0.46 AU; the Earth–comet distance Δ = 1.19 AU; the comet was receding from the Sun at a speed of 36 km s⁻¹, and from Earth at a speed of 14 km s⁻¹. The Sun–comet–Earth phase angle was 54°. Images of the same dust coma showed that the apex distance (i.e., the distance between the brightest inner coma and its sunward boundary) was about 100'', corresponding to 10⁵ km projected along the Sun–comet vector.

Zoom In Zoom Out Reset image size

The cometary spectrum shows a number of emission features with very prominent Na i, but also with the K i lines clearly detected. Other identified emissions include the C₂ Swan band dv = 0 and the satellite dv = 1e and dv = −1, NH2 (7–0) and NH2 (9–0), and the [O i] at λλ6300 and 6363 Å redshifted of 14 km s⁻¹ (Figure 1). The alkali lines are shown in Figures 2 and 3, where the reconstructed profile is estimated after accounting for the contamination of the telluric O₂ absorption bands and of the solar spectrum reflected by the cometary dust. These corrections are computed by means of a twilight spectrum which was also recorded with the same spectrograph setup. The K i line at λλ7664.8991 Å (binned thick black line in the left panel of Figure 2) is blended with two telluric O₂ lines at λλ7664.73 and 7665.79 Å, and with the K i solar absorption produced by comet dust reflection of the solar light and redshifted by 50 km s⁻¹. In order to reconstruct the cometary K i emission, we used a twilight spectrum (continuous red line in the same panel) from which we subtracted the solar K i absorption at rest (left dashed red line in the same panel) and added the K i absorption redshifted by 50 km s⁻¹ (right dashed red line in the same panel). The obtained local continuum (dashed blue line in the same panel) well reproduces the wings of the absorption. The resulting emission K i line (continuous blue line in the same panel) is the difference between the original spectrum (binned thick black line in the same panel) and the reconstructed local continuum (dashed blue line in the left panel of Figure 2). The other K i emission line at λλ7698.9645 Å (right panel of Figure 2) and the Na i lines at λλ5895.9242 and 5889.9510 Å (Figure 3) are not contaminated by telluric bands, but are partially eroded by the corresponding solar absorption line reflected by the comet dust. In these cases, the twilight feature (continuous red line) has been redshifted by 50 km s⁻¹ to match the red wing and to reconstruct the true emission profile (continuous blue line in Figures 2 and 3). The intensities of the corrected alkali lines are reported in Table 1. No emission is detected at the position of the Li i λλ6707.78 Å line (Figure 4) and the 3σ upper limit for the Li i emission is shown in Table 1.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The abundance Na/x of sodium related to an atom x (Table 1) is extracted from the line intensity I (Table 1) by means of the relationship Na/x = (g_x/g_Na)(I_Na/I_x), where the g factors g_x at 1 AU are computed (Table 1) as a function of the heliocentric radial velocity v (thus taking into account the Swing effect) using the high-resolution visible solar flux (Kurucz et al. 1984) and the oscillator strengths for the observed resonant lines (Morton 2003, 2004). Regarding g_K at λ = 7664.8991 Å, in the range 35 ⩽ v ⩽ 45 km s⁻¹ the solar spectrum is dominated by the strong absorption of telluric oxygen, so that we had to assume the same mean value obtained at smaller and larger v values. The intensity ratio between the two Na and K lines (1.7 ± 0.1 for Na and 1.6 ± 0.3 for K, respectively) matches the corresponding g-factor ratio (2.0 ± 0.4 for Na and 1.9 ± 0.1 for K, respectively), so that we can exclude significant optical thickness in the lines.

Five processes are expected to extract alkali atoms from the parent body, namely, (1) thermal desorption, (2) photon-stimulated desorption, (3) solar wind sputtering, (4) micro-meteoroid vaporization, and (5) photodissociation of parent molecules. Given the atomic parameters of alkali and the results of laboratory experiments on cosmic analogues, it is expected that the cloud of atoms leaving the parent surface maintains its original (presumably solar) abundance (Leblanc & Doressoundiram 2011). Hereafter, when we refer to solar lithium abundance, we mean that measured in meteorites (Asplund et al. 2009). After the atom release, many phenomena must be taken into account to infer the Na/K ratio in Mercury's exosphere (Leblanc & Doressoundiram 2011), and the main one also occurring in comets is photoionization by solar UV radiation. All alkali atoms are pushed in the anti-sunward direction by solar radiation pressure, with accelerations a = GM_☉(∑_λβ_λ)r⁻² in the comet reference frame. Here G is the gravitational constant, M_☉ is the Sun mass, r is the Sun–comet distance, and β is the ratio between solar gravity and radiation pressure forces. From the g factors we compute (Table 1) β_λ = g_λ(1 AU)²/(cmGM_☉), where c is the velocity of light and m is the mass of the atom (Fulle 2004). As it happens for the ejection of iron atoms (Fulle et al. 2007), alkali atoms should mainly be ejected from dust, which should be mostly ejected from the comet nucleus into the sunward sector of the coma. Before reaching the observation slit, the atoms must cover at most 10⁵ km (i.e., the apex distance) in the anti-sunward direction, which requires the flight times Δt_Na = 9.510³ s and Δt_K = 1.210⁴ s, respectively. Taking into account the photoionization lifetimes listed in Table 1, the original Na/K ratio would increase by a factor three, so that we cannot exclude an original solar ratio Na/K = 15.5 (Asplund et al. 2009). In other words, what we actually observed was the beginning of the alkali tail, where potassium is depleted versus sodium by its shorter photoionization lifetime. In order to cover the same distance of 10⁵ km, lithium atoms need a flight time Δt_Li = 4.110³ s. Taking into account its photoionization lifetime (Table 1), we get an increase of the original Na/Li ratio by a factor of two only, very far from the factor of eight required to match the solar ratio Na/Li = 10³ (Asplund et al. 2009) to the observed one. The required factor of eight would be observed at 710⁵ km from the nucleus, corresponding to 12', a distance impossible to accept since the slit tracking was done on the brightest inner coma with a mean diameter of 3'. We conclude that lithium in the comet C/2011 L4 is depleted by a factor ⩾4 with respect to the solar ratio.

Sodium and potassium were detected in the spectra of the comet C/1965 S1 (Ikeya-Seki) at a distance r = 0.14 AU, when the comet was receding from the Sun at a velocity of 110 km s⁻¹ (Preston 1967). We consider the observed intensities of the alkali lines Na λλ5889.9510 and K i λλ7698.9645 measured by Preston (1967), I_Na/I_K = 80 and I_Na/I_Li ⩾ 1.310⁴, respectively. By means of the g factors reported in Table 1, we obtain Na/K = 50 ± 11 and Na/Li ⩾3.310⁴ which are values in close agreement with those measured in C/2011 L4. Preston (1967) obtained even higher values assuming excitation mechanisms more complex than fluorescence. However, the resulting excitation temperatures are quite different among different atoms, and their physical interpretation is difficult. Due to the intensity of sodium lines, these were saturated in the photographic spectra obtained, so that we cannot exclude that the sodium lines were optically thick. However, since sodium is by far the most abundant among alkali atoms, we can assume that optical thickness affected potassium (and lithium if any) much less than sodium. This would further increase the ratios listed above. We assume that the alkali atoms must cover an anti-sunward distance of 410⁴ km (exactly matching the slit length of 1' used in the spectrograph setup) after they are ejected by the dust in the sunward coma. In order to cover such a distance, the atoms need flight times Δt_Na = 1.910³ s and Δt_K = 2.210³ s, respectively. Taking into account the photoionization lifetimes listed in Table 1, the original Na/K ratio would increase by a factor of eight, more than required to match an original solar Na/K ratio. Therefore, an original solar Na/K ratio would remain consistent with observations even if the Na/K ratio in the alkali tail had been a factor of three higher due to optically thick sodium lines. On the other hand, lithium atoms need a flight time Δt_Li = 7.810² s. Taking into account its photoionization lifetime, we get an increase of the original Na/Li ratio by a factor of 4 only, very far from the factor of 33 required to match the solar Na/Li ratio to the observed one. We conclude that lithium was depleted with respect to the solar abundance in comet C/1965 S1 too, by a factor ⩾8, even more constraining than in the comet C/2011 L4. The actual detection of lithium and potassium tails in future bright comets may help to understand if lithium depletion is original in comet nuclei, or is due to a less efficient process of extraction with respect to other alkali atoms.

We thank Jacques Crovisier for his comments which significantly improved the original manuscript, and Pierluigi Selvelli for useful discussion on atomic emission processes.