RECURRENT PERIHELION ACTIVITY IN (3200) PHAETHON

The short orbit period (1.43 yr) offers frequent opportunities to observe Phaethon at perihelion. The orbit period of STEREO A is ∼346 days so that three STEREO orbits (2.84 yr) are almost exactly equal to twice the orbit period of Phaethon (2.86 yr). Therefore, observations in 2009 (perihelion June 20) and 2012 (perihelion May 2) share similar perspectives viewed from STEREO A.

We obtained the celestial coordinates of Phaethon from NASA's HORIZONS Web interface and transformed them to pixel coordinates on the HI cameras in the "AZP" Zenithal projection (Calabretta & Greisen 2002; Thompson 2006; B. Thompson 2012, private communication). The sky-plane trajectories of Phaethon are shown in Figure 1. As expected, they follow similar paths in 2009 and 2012. The angular speed of Phaethon in the images is the result of the combined spacecraft parallactic motion and object Keplerian motion. Viewed from STEREO A, Phaethon moved relative to the field center at a speed varying between 0'' hr⁻¹ and 175'' hr⁻¹ from east to west, and at a roughly constant speed 180'' hr⁻¹ (in 2009) and 200'' hr⁻¹ (in 2012) from south to north. Phaethon was readily apparent even in a cursory visual examination of the 2009 and 2012 data.

Zoom In Zoom Out Reset image size

The observing geometry in the intervening orbit in 2010 is completely different from that in 2009 and 2012. Phaethon was in the field of view of STEREO A from November 11 to 19 and from December 15 to 25 (upper panel in Figure 2). It left the HI-1 field of view six days before its perihelion on November 25 and only re-entered the field about 20 days after perihelion in the second time period. Phaethon was not detected in either observing window. During the first time interval, the phase angle was above 120°, making an unfavorable condition for detection. During the second time interval, the phase angle was a modest 10°, but the heliocentric distance had increased to ∼0.8 AU. As a result, the nominal predicted magnitude had fallen to V ∼ 15, making Phaethon too faint to be detected in STEREO data.

Zoom In Zoom Out Reset image size

Phaethon stayed within the field of the STEREO B HI-1 camera for almost three months during its 2010 return (lower panel in Figure 2) but left the field of view six days before perihelion. Between mid-August and mid-November, the phase angle varied from ∼0° to a maximum 30°, while the heliocentric distance decreased from 1.8 to 0.3 AU. The apparent visual magnitude calculated assuming solid body reflection decreased from 18.5 to 12.5. At its predicted brightest, Phaethon should have been marginally within reach of STEREO HI-1 camera detection, but in practice the object was not detected.

Presumably as a result of these different observational circumstances, the anomalous brightening detected in 2009 and 2012 was not found in 2010. Activity specifically at perihelion could not be detected because Phaethon was outside the fields of both the A and B cameras when at perihelion. We conclude that differences in the observational geometry prohibited the detection of Phaethon and its activity at perihelion in 2010.

To supplement the near-Sun photometry from STEREO, we also observed Phaethon when far from the Sun using the Keck 10 m telescope on UT 2012 October 14.5. We used the Low Resolution Imaging Spectrometer (LRIS; see Oke et al. 1995) and a broadband R filter (central wavelength λ_c = 6417 Å, full-width at half maximum, FWHM = 1185 Å) in seeing of FWHM = 10. The LRIS pixel scale is 0135 pixel⁻¹. Flat field images were obtained using an illuminated patch on the inside of the Keck dome. The data were photometrically calibrated using solar-colored standard stars from Landolt (1992). We measured R = 17.25 ± 0.05 at heliocentric and geocentric distances 2.184 AU and 1.346 AU, respectively, and phase angle α = 181. This measurement refers to an unknown rotational phase of Phaethon.

We first consider and reject several mechanisms that might be implicated in the observed anomalous brightening at perihelion.

The intrinsic brightening is too large (2 mag) and too long-lived (two days) to be plausibly attributed to rotational variation of the projected cross-section of the aspherical nucleus of Phaethon. Lightcurve observations show a variation ⩽0.4 mag and a rotational period of only ∼3.6 hr (Krugly et al. 2002). The possibility that the brightening might be caused by a glint (a specular reflection from a mirror-like patch on the surface) can likewise be rejected on the basis of the longevity of the event. (Separately, the likelihood that the surface of a rocky asteroid could be mirror-like in the optical seems remote.)

As noted in Paper I, the brightening of Phaethon cannot be attributed to comet-like processes driven by the sublimation of near-surface water ice. This is because surface temperatures on Phaethon are far too high for water ice to survive. The deep interior temperature (identified with the blackbody temperature of a body moving with Phaethon's orbitally averaged heliocentric distance) is also too high to permit the survival of buried ice (Paper I). Even if it were present, deeply buried ice would be thermally decoupled from the instantaneous solar insolation (the conduction timescale across the radius is ⩾10⁵ yr), leaving no explanation for why the brightening is correlated with perihelion.

The solar wind kinetic energy flux onto the surface of Phaethon is E_KE = ρv³/2 (W m⁻²), where ρ is the solar wind mass density and v is the solar wind speed. At Earth's orbit, the solar wind number density is about 10⁷ m⁻³. Scaled by the inverse-square law to perihelion at 0.14 AU, the density is about N₁ = 5 × 10⁸ m⁻³. The solar wind speed varies with time and radius, but is of order v = 500 km s⁻¹. Substituting ρ = μm_HN₁, where μ = 1 (for protons) and m_H = 1.67 × 10⁻²⁷ kg, we obtain E_KE ∼ 0.05 W m⁻². This is tiny compared with the solar radiation flux at perihelion, $F_{\odot }/R_{{\rm AU}}^2=70{,}000$ W m⁻². Consequently, we conclude that the solar wind is a negligible source of energy and cannot account for the anomalous brightening by impact fluorescence.

Could part or all of the measured excess optical brightness be thermal emission resulting from Phaethon's high surface temperatures when at perihelion? We obtain a rough lower limit to the temperature by considering the case of an isothermal, spherical blackbody in equilibrium with sunlight, namely, $T_{{\rm BB}} = 278 R_{{\rm au}}^{-1/2}$. At perihelion, R_au = 0.14, we find T_BB = 743 K. A practical upper limit is given by the sub-solar temperature on a non-rotating body (or a rotating one whose rotation axis points at the Sun), namely, $T_{{\rm SS}} = \sqrt{2} T_{{\rm BB}}$, giving T_SS = 1050 K for Phaethon at perihelion. This temperature range is in good agreement with independent estimates (∼800 K ⩽T ⩽ 1100 K) from Ohtsuka et al. (2009).

A detailed calculation of the thermal emission from the surface of Phaethon depends on many unknowns and is beyond the scope of the present work. Instead, we set a strong upper limit to the possible thermally emitted flux density by assuming that the whole surface of Phaethon (not just the sub-solar region) is at T_SS, the maximum possible surface equilibrium temperature. Under this assumption, the flux density in the STEREO bandpass is calculated from

$$\begin{equation} f_\lambda = \frac{\pi r_n^2}{\Delta ^2}\left[\frac{\int _{\lambda _1}^{\lambda _2}B_\lambda (T) d\lambda }{\Delta \lambda }\right] \end{equation} \tag{ 2 }$$

in which B_λ(T) is the Planck function evaluated at T = T_SS, r_n = 2.5 km is the effective circular radius of Phaethon, Δ is the Phaethon-to-observer distance, the integration is taken over the filter transmission from λ₁ = 6300 Å to λ₂ = 7300 Å, and Δλ = λ₂ − λ₁.

For comparison with Equation (2), we convert the apparent magnitudes, m_obs, into flux densities, $f_\lambda ^o$, using $f_\lambda ^o = 3.75\times 10^{-(9+m_{{\rm obs}}/2.5)}$ (erg s⁻¹ cm⁻² Å⁻¹) (Drilling & Landolt 2000). The results are shown in Figure 7, where solutions to Equation (2) (blue) are compared with the observations (red). Evidently, thermal emission even at the peak sub-solar temperature is orders of magnitude too small to account for the anomalous brightening of Phaethon observed in our data.

Zoom In Zoom Out Reset image size

To see this a different way, we substituted the measured peak flux densities, $f_{\lambda }^o \sim$ 10⁻¹³ erg s⁻¹ cm⁻² Å⁻¹ (cf. Figure 7), into the left-hand side of Equation (2) and solved for the temperature. We find that values T ∼ 1650 K (2009) and 1700 K (2012) would be needed for thermal emission to account for the anomalous optical brightening. This is far hotter even than the sub-solar temperature on the nucleus at perihelion and, therefore, can be discounted as unphysical. We conclude that the perihelion brightening of Phaethon is not due to thermal emission from the surface.

The passband of the STEREO camera includes the 6300 Å and 6363 Å forbidden lines of oxygen. These are "prompt emissions," formed when oxygen atoms are produced in the excited ¹D state by photodissociation of a parent molecule, for example water (Festou & Feldman 1981). As noted earlier, although water ice cannot survive, water might be bound within hydrated minerals in Phaethon and released by desiccation. In the case of water, the photodissociation timescale at perihelion is about half an hour, so that any water molecules released from Phaethon would be destroyed within a single pixel of the STEREO camera. We estimate the flux density produced by prompt emission in oxygen, averaged over the passband of the camera, from

$$\begin{equation} f_{\lambda }^{[{\rm O\,{\scriptsize I}}]} = \frac{\alpha Q h c }{4\pi \Delta ^2 \lambda \Delta \lambda } \end{equation} \tag{ 3 }$$

in which α ∼ 10% is the fraction of water dissociations leaving oxygen in the excited ¹D state, Q is the production rate of water molecules, h is Planck's constant, c is the speed of light, Δ is the Phaethon to STEREO distance, λ is the wavelength, and Δλ is the filter FWHM, expressed in Angstroms. Setting $f_{\lambda }^{[{\rm O\,{\scriptsize I}}]}$ = $f_{\lambda }^o$, we find that a water production rate Q ∼ 10³⁰ s⁻¹ (3 × 10⁴ kg s⁻¹) would be needed to account for the measured excess flux density of Phaethon. Although the required rate of production (which is similar to that of comet 1P/Halley at perihelion) seems high, we cannot rigorously rule out the possibility that some fraction of the excess perihelion emission is caused by prompt emission from oxygen. However, the observation that the fading of Phaethon after peak brightness follows the phase function of a solid object (Figure 5) suggests that gas is not the dominant cause of the anomalous brightness.

The remaining alternative is that Phaethon has ejected dust particles with a combined cross-section larger than that of the solid nucleus, as earlier concluded in Paper I. In this scenario, the rise in brightness in Figure 3 then corresponds to the ejection of dust, while the subsequent decline in brightness is naturally explained as fading owing to the ever-growing phase dimming, perhaps aided by grain sublimation or disintegration. The natural test of this hypothesis would be to search for coma scattered by the ejected dust. Unfortunately, as also noted in Paper I, the limited angular resolution and high background surface brightness in the STEREO data make the detection of resolved coma difficult.

A temperature-controlled mechanism for the ejection of dust is strongly suggested; the activity is observed at the highest (perihelion) temperatures, and is absent in observations of Phaethon taken at substantially larger heliocentric distances and lower temperatures. The reflection spectrum of Phaethon has been interpreted in terms of thermally modified hydrated minerals (Licandro et al. 2007; Ohtsuka et al. 2009; de León et al. 2010; Yang & Jewitt 2010), while the depletion of sodium in some Geminids provides independent evidence for thermal alteration (Kasuga et al. 2006). The perihelion temperatures on Phaethon exceed those needed to break-down phyllosilicates (Akai 1992), and are sufficient to induce thermal fracture (Paper I; Jewitt 2012). In this sense, thermal disintegration and fracture are plausible sources of the anomalous brightening and Phaethon may be accurately labeled a "rock comet" (Paper I). An additional requirement is that dust must be cleared from the surface in order for these processes to operate. A regolith of fine particles built up in previous orbits will inhibit thermal fracture, since small grains are unable to sustain large temperature differences. Likewise, surface materials dehydrated by baking in previous perihelion passages must be cleared away from the surface in order for dehydration cracking to remain a persistent dust source.

In both 2009 and 2012, the apparent brightness increased by ∼2 mag relative to the nominal phase curve (Figures 5 and 6), corresponding to a factor of ∼6. Given that the cross section of the nucleus of Phaethon is C_n = $\pi r_n^2 \,{\sim}\,$20 km², the cross section of added dust is then C = 100 km², in both years. The mass in spherical particles of mean radius $\overline{a}$ having cross section, C, is

$$\begin{equation} M_d \sim \frac{4 \rho \overline{a} C}{3}, \end{equation} \tag{ 4 }$$

where ρ is the grain density. With ρ = 3000 kg m⁻³, we find M_d ∼ 4 × 10⁸ a_mm kg, where a_mm is the grain radius expressed in millimeters. The mass of the nucleus, represented as a 2.5 km radius sphere of the same density, is a much larger 2 × 10¹⁴ kg.

The Geminid stream mass is 10¹² kg ⩽ M_s ⩽ 10¹³ kg (Hughes & McBride 1989; Jenniskens 1994), while the Geminid stream lifetime estimated on dynamical grounds is τ  ∼  1000 yr (Gustafson 1989; Ryabova 2007). Accordingly, to explain the entire mass in the Geminid stream through events like those observed here would require a number of similar events per orbit, N, given by

$$\begin{equation} N \sim \left(\frac{M_s}{\tau }\right) \left(\frac{P}{M_d}\right), \end{equation} \tag{ 5 }$$

where P = 1.4 yr is the orbital period. Substituting, we obtain N ∼ 4$a_{{\rm mm}}^{-1}$ to 40$a_{{\rm mm}}^{-1}$. If the particles are millimeter-sized, a_mm = 1, then the Geminids could be supplied in steady state by 4 ⩽N ⩽ 40 outbursts like the one observed, each orbit.

However, there is no compelling physical reason to assume that Geminid stream production is in steady state. While a_mm = 1 may approximately represent the radii of the Geminid meteors, much larger examples up to 5 kg in mass (equivalent radius ∼7 cm) have been inferred from Lunar night-side impact flashes (Yanagisawa et al. 2008). Moreover, Phaethon is but part of the PGC, which includes at least two other kilometer-scale bodies caused by fragmentation on an all-together much larger and longer (10⁶ yr?) scale. We conclude that, while continuing mass-loss near perihelion may contribute to actively replenishing the smaller Geminids, the PGC complex as a whole is likely the product of a more ancient and catastrophic breakup (cf. Jewitt & Hsieh 2006; Kasuga & Jewitt 2008).

Many puzzles remain in understanding the anomalous brightening of Phaethon and its relation to the ejection of dust and to the Geminids. We list the following key questions.

• 1.  
  What is the value of the effective dust radius, a_mm? This radius directly affects estimates of the ejected mass through Equation (4) and so determines the extent to which on-going activity in Phaethon contributes to the Geminid stream.
• 2.  
  What is the origin of the ∼0.5 day phase lag between perihelion and the anomalous brightening in both 2009 and 2012? Is the brightening always lagged relative to perihelion by this amount?
• 3.  
  How many brightening events occur per orbit and are these events always of the same amplitude? The nature of the STEREO data curtails our ability to detect brightening events outside a limited window of accessibility.
• 4.  
  Can evidence for mass loss be detected when Phaethon is far from perihelion, despite past, failed attempts? Thermal fracture and mineral decomposition are likely only at the extreme temperatures found near perihelion. However, slow-moving dust may linger in the post-perihelion months, making these the prime time for future attempted dust observations.
• 5.  
  Can spectra be obtained at perihelion in order to search for the forbidden lines of oxygen?
• 6.  
  What is the mineralogical composition of Phaethon and are hydrated silicates present on its surface?
• 7.  
  Can meteoroids ejected during recent perihelion events be detected and distinguished from older Geminids? Ryabova (2012) reports that this will be difficult.