Two techniques are commonly used to obtain the spatial distribution of coma species. Either maps are built from long slit spectra recorded at various locations in the coma or images are directly obtained through narrow band filters. The first method primarily suffers from the difficulty one meets in exactly locating the position of the slit inside the coma. Most of the time, only emission profiles recorded on lines passing close to the nucleus are used, since then a reasonable assumption about the position of the unseen nucleus can be made before subtracting the continuum signal that lies under all emission lines/bands. All other intensity profiles are difficult to position, even relative to each other. Building a monochromatic image of the coma from a set of slit spectra is a quite difficult, nearly impossible task to achieve. The second method applies well to bright, spectrally isolated, emissions. Unless a bright star is present near the comet, there is no reliable way to position the diverse color images relative to each other. To overcome this difficulty, it is customarily assumed that the brightest part of each image corresponds to the position of the nucleus. This assumption is unfortunately not very robust in some cases, as we will see below. The observed maximum is neither the continuum brightness maximum nor that of the emission line but a combination of the two.
There exists a third way to study the spatial emissions of comet comae that allows one to perform a quasi- perfect subtraction of the continuum emission from the data and to build monochromatic emission maps that are perfectly positioned relative to each other: spectrophotometry with integral field instruments. This technique consists of simultaneously recording spectra at a large number of locations forming a mesh covering the coma from which emission maps are derived. TIGER, an array of microlenses coupled to a grism-based spectrograph, is such an instrument that was operated at the 3.60 m Canada-France-Hawaii telescope until 1997 (Bacon et al., 1995, A&A, 113, 347, 1994). It has never been used before for a comet observation and we present here a high spatial resolution map of the oxygen red forbidden in which the nucleus is accurately positioned.
On 6-8 November 1994, we observed comet 19P/Borrelly, then located 1.4 A.U. from the sun and 0.69 A.U. from the Earth, to achieve a very high spatial resolution in the inner coma. Each observation consisted in simultaneously recording spectra at 572 points separated by about 0.5" from their neighbors. Typical seeing during our observations was in the 0.55-0.65" range, thus matching the maximum spatial resolution achievable with this instrumental setting. Due to small guiding errors, the effective seeing for each night was of order 0.75", which corresponded to a spatial resolution at the comet of 350 km. We concentrated our efforts on the red part of the spectrum where oxygen atoms, C2 and NH2 radicals have strong emissions. The oxygen 1D atoms have a short lifetime (~150 s) and are excellent tracers of their parent species. We mapped the oxygen atom forbidden emission at 6300-64 Å to infer the emission pattern of the water molecules at the nucleus surface, providing us with information on the distribution of active areas, information that, except for a few cases, has been poorly documented. In addition, since we could accurately locate the nucleus position, and since the oxygen red emission is heavily quenched in collisions with ambient neutral molecules, we conducted the first study of the collisional region of the coma of a comet. Figure 6 shows the three maps we obtained that can be used to describe the spatial distribution of the oxygen atoms, the solid particles and the NH2 radicals.
Water molecules are the main source of 1D oxygen atoms and the red line distribution thus reflects their spatial distribution. Because of their high velocity of 2.25 km/s, the oxygen atoms can travel ~400 km before decaying towards the lower 3P level and thus populate the tailward part of the coma, even in the absence of parent species in that part of the coma (Figure 6 also shows an obvious projection effect). Our modeling revealed the need for a strongly anisotropic ejection of water molecules. The continuum emission map is even more anisotropic than that of the oxygen emission: solid particles are restricted to a smaller solid angle. The isophotes of the continuum and oxygen emissions accept the same axis of symmetry, close to the sun- comet vector. This is the first time that the strong coupling that exists between the solid grains and the water molecules is "imaged" with such detail. The oxygen emission has its maximum on the sunward side of the nucleus, while that of NH2 is more circular and almost coincides with that of the continuum. This difference can be explained by the high quenching suffered by the oxygen 1D atoms in the near nucleus region in collisions with ambient molecules.
To interpret the data, we used an atmospheric model in which the parent species are assumed to be produced either by sublimation from a single active region located near the subsolar point or sublimation from the lit part of a spherical nucleus. Figure 7 shows the angular distribution of the parent species in these two cases, based on theoretical calculations by Crifo (1997, private communication) and Keller et al. (1994, PaSS, 42, 5, 367). Details on the assumptions underlying those two limiting cases will be found in Festou et al. (1998, A&A, submitted). Then, parent molecules are photolised and the spatial distribution of the daughter products is calculated using an anisotropic vectorial model. Whereas the water production rate is calculated where the oxygen emission is not perturbed by collisions, the quenching coefficient can be empirically adjusted using the very inner coma data (since no value for the comet conditions is available).
The result of the adjustment of the oxygen model is shown in Figure 8. Four sources of oxygen atoms are taken into account, namely, water, OH, CO and CO2. Only water significantly contributes in the region of the coma explored with this program. No model assuming a single active region near the subsolar point fits the data. Instead, the observations are almost perfectly fitted with the other model, which implies that the lit face of the nucleus outgasses. The observations revealed no morphological changes from night to night, which suggests a nucleus of spherical shape if the rotation axis of the nucleus was not pointing directly at the sun when we observed. The comparison with an unquenched model indicates that the red emission is perturbed by collisions with ambient water molecules up to about 2×103 km from the nucleus. Interestingly, the quenching coefficient found in the literature (Streit et al., J. Chem. Phys., 65, 4761, 1976) does not fit the observations. The coefficient must have a value 2.5 times higher, in qualitative agreement with the fact the kinetic temperature that corresponds to the velocity of the oxygen atoms is very large, in excess of 103 K. The water production rate derived from this study, 3.4×1028 s-1, differs from that derived from nearly contemporaneous International Ultraviolet Explorer observations by only 15%. This indicates that no source other than water and OH can contribute to the production of O 1D atoms in that comet. If derived from observations of the 6300 Å line, the water production rate should be calculated using our value of the quenching coefficient, i.e. 6.1 × 10-10 cm3 s-1. The NH2 data have lower S/N value and can only lead to an evaluation of the ammonia production rate.
Conclusions and prospects
This set of observations shows that the nucleus of comet 19P/Borrelly has very likely a spherical shape. Either the illuminated hemisphere is entirely emissive or its lit side is covered by many active regions of smaller extent that mimic the behavior of an hemispheric outgassing surface. Emission of water molecules occurs on the sunward side of the nucleus. Unless the rotation axis of the comet is pointing towards the sun, the absence of modifications of the oxygen coma implies that the entire surface of the nucleus is capable of emitting water molecules. The anti-solar part of the coma is populated by water molecules injected there by intermolecular collisions. The high velocity of the oxygen atoms relative to that of their parent species further produces an oxygen coma less anisotropic than the initial water coma. However, quenching of the red emission strongly shifts the maximum of 6300 Å emission on the sunward side. Solid particles are entrained via collisions with surrounding water molecules. Since gas and dust rapidly decouple, the dust coma is much more asymmetric than the gas coma. The NH2 coma has the appearance one may expect from a species produced from the dissociation of a parent species (NH3) that has the same spatial distribution as the more numerous water molecules and which emission is not quenched. The excellent agreement between the CFH and IUE determinations of the water production rate of the comet near perihelion times indicate that no molecule other than water and OH significantly contributes to the production of O 1D atoms.