How many Dead Comets are there?

One long-standing question is: how many asteroids – especially near-Earth asteroids (NEAs) – were comets in the past? In the classical conception, comets have surface ices and form tails and comae in the vicinity of the Sun; asteroids are simply rocks in space. In the recent years, the two classifications became less clear than that.

Comets can end up on orbits that bring them close to the Earth. We can observe them. But what happens to them after spending a while there? For some comets, like ISON, being close to the Sun is fatal; they simply disintegrate and disappear. For other comets in near-Earth space, the strong degree of insolation forces them to lose their surface ices quickly and eventually their activity stops – they are dead or dormant comets. Since comet surfaces have very low albedos, these dead comets simply look like low-albedo asteroids and are really hard to identify.

However, not all dead comets are really dead. Don Quixote is a good example, since we discovered activity in this objects that has not been detected in the last three decades. For this reason, dead comets should really be called dormant comets, as it is not clear if they ceased activity for good, or if they only sleep. The real question is: do these objects still harbor ices in their interiors? If so, this can have significant implications for the transport of water and other volatiles to Earth in the past and for the potential resource utilization in space in the future.

In order to identify dead comets that hide in the NEA population, we performed a statistical analysis in two steps. In the first step, we compared the orbits of currently active comets with those of asteroids and identified asteroids on comet-like orbits. In the second step, we identified those asteroids on comet-like orbits with comet-like (i.e., low) albedos. Those are most likely candidates for dead comets. We found a total of 23 dead comet candidates in albedo data provided by ExploreNEOs, NEOWISE, and Akari. Interestingly, only about 50% of all asteroids on comet-like orbits also have comet-like albedos.

In a second analysis, we estimated how many asteroids are likely to be of cometary origin. This question sounds easier than it is. The problem is that most asteroids are discovered with optical telescopes, which are more likely to discover asteroids with high albedos than those with low – and especially comet-like – albedos. Also, dead comets move on orbits that take them far away from the Sun, which makes them even less likely to be discovered by asteroid surveys. Hence, we have probably discovered proportionally fewer dead comets than other asteroids.

In order to resolve this discrepancy and account for this so-called ‘discovery bias’, we used a well-characterized survey that operates in the thermal-infrared, which is less prone to albedo-based discovery bias. NEOWISE is a program that uses the WISE all-sky survey data to search of all kinds of asteroids. Assuming a realistic but synthetic NEA population, we checked, how many of these synthetic NEAs would have been discovered by NEOWISE. With that information and those NEAs actually observed by NEOWISE, we can estimate how many more dead comets there are. We checked two different samples: dead comets larger than 1 km in diameter and dead comets with absolute magnitudes brighter than or equal to 21.  We find that 0.3-3.3% of the NEAs with H <= 21 and 9 (+2/-5)% of the size-limited NEA population are dead comets. These numbers are slightly lower than previously assumed, implying that fewer NEAs than previously assumed have a cometary origin.

All the details of this analysis are available in a paper that is currently in press at AJ, but already available on arxiv.


Detection of Cometary Activity in NEO Don Quixote

Part of the NEO population is considered to consist of so-called dead comets. Dead comets are comets that have spent a long time as NEOs and have been depleted their volatile inventories in numerous, close encounters with the Sun, i.e., they are extinct comets. They can be identified through their distinctive comet-like orbits and their low, comet-like albedos.

One of the dead comet prototypes is NEO (3552) Don Quixote. We observed Don Quixote as part of the ExploreNEOs program, in which ~600 NEOs were observed to derive the diameters and albedos of these objects. Our observations at 3.6 and 4.5 µm were saturated because the target was brighter than expect, revealing something rather unexpected.

donquixote1The subtraction of the Spitzer IRAC point-spread-function (psf) from our observations show some kind of extended emission around the object at 4.5 µm, but not so at 3.6 µm. We went into some effort to show that the extended emission is not an image artifact, it is not a background source, not a latency effect, and not a result of the saturation of the object. The latter is obvious after applying the same psf-subtraction technique to a saturated image of an even brighter calibration star (HD 149661).

donquixote2The nature of the extended emission becomes clear after plotting the radial brightness distribution around the object. At 3.6 µm, the distribution is rather noisy and basically in agreement with a null result, whereas at 4.5 µm, the radial brightness distribution is clearly proportional to the reciprocal distance from the target center. This behavior is typical for cometary comae, consisting of optically thin dust and gas that is ejected from the surface as ices sublimate in the warming sunlight. After subtracting the model for a cometary coma from our observations, even a faint tail becomes obvious.


The fact that we observe cometary activity at 4.5 but not at 3.6 µm provides constraints on the composition of the coma. If significant amounts of dust would be ejected together with the gas, the emission would be visible in both bands, more likely so at 3.6 µm, which is not the case. The most likely explanation is emission from molecular bands: both CO and CO2 have molecular band emission lines that fall into the 4.5 µm bandpass, but not into the 3.6 µm bandpass. Both materials have been found in cometary spectra, but CO2 is usually more likely. The activity we find in Don Quixote is rather minute and comparable to the weakest activities found in active comets. Still, the fact that this object does show activity, which has not been discovered in 30 years after its discovery as an asteroid, is extraordinary.

Another interesting fact we could derive from its taxonomic classification as a D-type asteroid and meteorite analog material is that Don Quixote is likely to hold a large amount of water. Based on the diameter we derived (18.4 km), Don Quixote is likely to hold 100 billion tons of water, which is about the same amount of water as in Lake Tahoe, California.

Our study has been published here and press releases have been issued by several institutions including NASA JPL and the European Planetary Science Congress, where the results have been presented.