Astron. Astrophys. 337, 945-954 (1998)

2. Observations and preliminary data analysis

The periodic comet 19P/Borrelly passed to within 0.62 AU of the Earth in early December 1994, approximately one month after its perihelion passage on 1 November 1994. During an eleven-hour time interval on 28 November 1994, we performed six observations of the comet using the WFPC2 in Planetary Camera mode with the F675W filter, which has an effective wavelength of 670 nm and bandwidth of 89 nm. Four images were taken during each observation: two had exposure times of 10 sec and two had exposure times of 30 sec, and all were taken within a time span of 7 mins. Table 1 lists the dates of the observations corresponding to the midpoints of the four consecutive images. The heliocentric distance (r), the geocentric distance (), and the solar phase angle () of the comet remained practically constant during these observations at 1.401 AU, 0.622 AU, and , respectively. The WFPC2 pixels are square and are on a side, which projected to a distance of 20.5 km at the comet.

Table 1. Log of observations, photometry and size of the nucleus of comet 19P/Borrelly

Our choice of the F675W filter - instead of the F702W filter used for P/Faye - was motivated by its superior photometric accuracy. As a primary WFPC2 filter, it benefits from detailed, extensive calibrations, and the conversion to standard (Landolt) R magnitudes requires only the V-R color of the object, and not both V-R and V-I, as required for the F702W filter and which introduces additional uncertainty.

All images were processed using the Routine Science Data Processing System at the Space Telescope Institute. Cosmic ray impacts were removed using a standard technique based on image duplication. Careful examination of the four images taken during each of the six visits using the x and y profiles through the pixel of maximum signal revealed no detectable shifts among them. This results from the short time spanning the sequence (7 mins) and the excellent tracking capability of HST which we have now verified on several comets. Consequently, the four images in each visit were co-added to increase the signal-to-noise ratio, giving a total effective exposure time of 80 sec. The only exception is the sixth observation, for which there was a mismatch of a fraction of a pixel between the individual images, resulting from a change of guide stars. Consequently, we separately processed the two long exposures for the sixth observation.

A comparison of the new WFPC2 images of 19P/Borrelly with the past WFPC images of comet 4P/Faye revealed impressive differences, notably the total absence of spherical aberration in the WFPC2 images and the strong anisotropy of the coma of 19P/Borrelly, which had a pronounced dust fan in the solar direction (Fig. 1).

Fig. 1. Isophotal contours of the inner coma of 19P/Borrelly show an impressive dust fan in the sunward direction (isophote levels are in log10(DN)). The arrows indicate the anti-solar direction (prolonged radius vector r), the direction of celestial north (N), and the heliocentric orbital velocity vector (Vorb) of the comet projected onto the sky plane.

The key aspect of the data analysis is to separate correctly the signal of the nucleus from that of the coma. Our previous approach of fitting an appropriate model of the comet to the images, taking into account the Point Spread Function (PSF) of the telescope, remains valid. However, its practical implementation has been radically modified to cope with the new, corrected PSF of the telescope as well as with the complex morphology of the coma of 19P/Borrelly. An accurate knowledge of the PSF is of particular importance. We used version 4.0b of the TinyTIM software written by Krist (1995). This version includes mirror zonal error maps obtained from phase retrieval of on-orbit WFPC2 data and revised aberration specifications, which significantly improve the PSF models compared to previous versions. Using this software, PSFs can be created at finer samplings than the detector pixel, a distinct advantage when dealing with the undersampled WFPC2 data and possible sub-pixel effects. All PSFs used in this study were calculated for the actual locations of the nucleus on the WFPC2 chip.

In order to assess the relative contributions of the nucleus and coma to the observed signals in the central pixels, we ran a preliminary experiment using a simple model in which the coma surface brightness decreases as , where is the projected distance between the line-of-sight and the nucleus, and the light distribution from the nucleus is described by the PSF. This is the situation that should apply when there is steady-state, spherically symmetric outflow of dust from the nucleus with constant velocity. The corresponding total surface brightness distribution is given by:

where is the Dirac function and is the convolution operator.

After making a crude fit of the model to a few images, we were able to reach the following conclusions:

• i) The nucleus is unambiguously detected.

• ii) The signal from the nucleus drops to insignificant levels at a cometocentric distance of 100 km (), which implies that the signal beyond this point is essentially all due to the coma.

• iii) The gradient in the coma surface brightness profile (-1 in this case) is affected by the PSF only within a radius of 120 km.

• iv) Sub-pixel specification of the nucleus significantly improves the fit of the model to the data.

• v) This simple model obviously does not provide an adequate description of the highly anisotropic coma of 19P/Borrelly.

© European Southern Observatory (ESO) 1998

Online publication: August 27, 1998
helpdesk.link@springer.de