Astron. Astrophys. 319, 617-629 (1997)

4. Discussion

4.1. The ions

4.1.1. Spatial distribution of H₂ O and CO

The most remarkable feature in our data set is the different spatial distribution of CO and H₂ O , well seen in Figs. 4 and 5. H₂ O possesses a narrower tail and is more concentrated toward the nucleus. CO has a broader tail and a more diffuse distribution. This difference is well seen in the profiles perpendicular to the projected tail axis presented in Fig. 6. Two cuts across the tail are shown at tailward distances 70 and and 170 10³ km, respectively. The CO column density decreases smoothly with increasing distance from the tail axis. Despite some local variations, resulting from the changing positions of moving features, the overall shape of the CO profiles remains essentially unchanged in the time interval of 1 hour covered by the observations. The H₂ O profiles seem to consist of two components, a conspicuous concentration close to the tail axis ( km) and smoothly decreasing wings at larger distances. The central component of the H₂ O profiles is reminiscent of the plasma sheet revealed by the ICE measurements in the tail of comet Giacobinni-Zinner (McComas et al. 1987). The CO profiles have similar structure but the concentration around the axis is not so strongly expressed as in the case of H₂ O , neither near to the nucleus, nor further in the tail. The shape and the strength of the CO profiles change only slightly with increasing tailward distance. At the same time the central component of the H₂ O profiles decreases by a factor of about 2 from 7 10⁴ to 17 10⁴ km in the tail.

Fig. 6. Cuts perpendicular to the tail axis at distances 7 and 17 104 km tailward from the nucleus. The cuts correspond to 5.3 and 13 104 km projected distance in Figs. 4 and 5. Solid lines: the water ions. Dashes, dash-dots and dots describe the carbon monoxide profiles in the 1st, 2nd and 3rd CO image, respectively.

The considered profiles reveal the spatial distribution of the ions only locally. The overall picture can be analyzed by means of a histogram analysis. We used the spatially matching areas of our CO and H₂ O images to construct the histograms presented in Fig. 7. The right extended wing of the H₂ O histogram, belonging to higher column densities, is formed from a relatively small amount of pixels near the nucleus. The CO histogram indicates a rather smooth transition between the pure cometary signal and the lower column density values around the histogram peak, where the cometary contribution is mixed with background signal.

Fig. 7. Histograms of the final CO and H2 O images. The narrower CO histogram is an illustration for the flatter distribution of CO . The feature in the right wing of the H2 O histogram indicates the existence of higher column density concentrated in a relatively small zone around the nucleus.

The difference between the spatial distribution of CO and H₂ O supports the magneto-hydrodynamic simulations by Wegmann et al. (1987). These authors have modeled the physics and chemistry in the coma of P/Halley in great detail and have evaluated column densities of different ions for comparison with ground-based observations. The CO and H₂ O maps presented by Wegmann et al. (1987) confirm what we observe in comet P/ST, namely that the H₂ O distribution is more concentrated toward the nucleus and around the tail axis.

The different spatial extent of CO and H₂ O can be understood by comparing the channels for their production and the scale lengths of their most probable parents, CO and H₂ O. Photoionization is the main process responsible for the production of both ions. During intermediate solar activity conditions, the photoionization and photodissociation rates of CO are 3.8 10^-7 s^-1 and 2.8 10^-7 s^-1, respectively (Schmidt et al. 1988). Even if the charge exchange mechanism is included, except of extremely enhanced solar wind flux, the scale length of CO will be several hundred thousand kilometers (Fernandez & Jockers 1983). On the contrary, neutral water has a very high photodissociation rate, 1.03 10^-5 s^-1 (Schmidt et al. 1988), which shortens its scalelength to less than 10⁵ km. Cochran & Schleicher (1993) argue in favor of even shorter lifetime of cometary H₂ O. Hence, the difference between the CO and H₂ O spatial distribution is related to the different extent of their sources, the neutral clouds of their parent molecules. Consideration of the photodissociative ionization of CO₂ as an alternative or additional mechanism for CO production would not change this situation because of the low rate coefficient of this reaction. Possible existence of an extended source, as it has been found in Halley's comet (Eberhardt et al. 1987), would further increase the region where CO can be born.

4.1.2. Tail structures

Comet P/ST had a very complex structured plasma tail. The most prominent features, labeled as (a), (b-c), (d) , and (e) in the three CO images, shown in Fig. 4, can be considered as markers along four well separated streamers.

The appearance of structures in the H₂ O image differs from the structures in the CO images. In the H₂ O image the two outer elongated condensations (rays) along features (b-c) and (g) dominate the surrounding plasma. In the CO images the condensations are more uniformly distributed across the tail and can be seen farther from the tail axis than in the H₂ O image. Feature (b-c) appears narrower in the H₂ O image than in the CO images. Although the H₂ O image was obtained 21 minutes after the 2nd CO image and 29 minutes before the 3rd one, more similarities can be found between both CO images than between each of them and the H₂ O image. This is an indication that the observed dissimilarities between the structures in the CO and H₂ O images might be present not only temporarily.

The traces of the features marked in Figs. 4 and 5 are presented in Fig. 8. Diamonds, triangles, and squares stand for the positions of condensations in the first, second and third CO image. respectively. The circles represent the location of the features in the H₂ O image. The numbers at the trajectories are the velocities projected in the sky plane (henceforth sky plane velocities spv) in km s^-1). They are calculated by using only the three CO images. The dashed line represents the projection of the comet's prolonged radius-vector. The velocity increases with increasing distance from the nucleus. Moreover, in agreement with in situ measurements (Siscoe et al. 1986) and with MHD modeling (Rauer et al. 1995 and references therein), the velocity is low near to the tail axis and increases to greater lateral distances.

Fig. 8. The location of tail structures in the plane of sky derived from tracing the condensations in the plasma images. The dashed line shows the antisolar vector. The motion of the structures goes from right to left along the full lines (see text). The circles denote the H2 O image, the other symbols the three CO images. The numbers indicate the derived sky plane velocities spv.

The velocity of the structures and their mean column densities are anticorrelated, as it can be seen in Fig. 9. The error bars of the velocities are derived from the tracing process. The errors of the column densities are 1 values in boxes of size 11 11 pixels around each condensation in the third CO image. The full line is a linear fit through the data. The existence of such an anticorrelation was pointed out by Jockers (1985) but no quantitative regression could be derived at that time.

Fig. 9. Relation between sky plane velocities spv of the tail structures and their mean column densities. The features are marked as in Fig. 4. The full line is a linear fit through the data.

Rauer & Jockers (1993) have shown that the apparent motion of structures in the ion tail of a comet is a manifestation of the bulk plasma flow, rather than a wave phenomenon. In the case of comet P/ST we can check this and in this way gain confidence in our velocity measurements by comparing our velocity field with Doppler measurements of other authors. If we assume that the plasma tail orientation is exactly along the antisolar direction, then the line of sight velocities (hereafter lsv), obtained by Doppler measurements, and the velocities in the sky plane (spv), obtained by tracing of plasma structures, are related by the expression , where is the phase angle. In our case, and we should expect apparent velocities in the sky plane about 15% greater than the measured line of sight velocities.

Brown et al. (1993) and Spinrad et al. (1994) (both papers referred to as BSJ93 hereafter) provide radial velocity measurements of H₂ O along the tail of comet P/ST out to a distance of 4 10⁵ km. Unfortunately, in their set of observations there are no measurements on Nov 25, 1992. Therefore we are forced to look if a reasonable information can be extracted from the values obtained on neighboring days. The lsv measurements by BSJ93 do not show any significant temporal variations up to 2 10⁵ km in the tail in the time interval Nov 23 - Dec 1, 1992. So, with some reservations, we could consider a mean value of these measurements as representative for Nov 25, too. Averaging the four measurements from this time interval we obtained a mean value of the lsv = 13 3 km s^-1 at 1.5 10⁵ km projected distance in the tail.

As the long slit in the spectroscopic observations of BSJ93 was oriented along the tail, only velocities of those structures should be used in the comparison, which are near to the tail axis in our observations . These are (d), (e), and (g) in Fig. 8. The apparent velocities of these condensations in the sky plane are in agreement with the mean value of Doppler measurements, despite of the lsv values being derived from H₂ O and the spv velocities from CO . This is an indication that the moving structures, mostly observed in the light of CO , allow to derive valid velocities for the CO as well as for the H₂ O ions within the uncertainty of our measurements of km s^-1. These velocities can be used to calculate the ion flux of both types of ions in the plasma tail of comet P/ST.

4.1.3. The ion flux

Once a cometary ion is born in the source region it is picked up by the solar wind and accelerated tailwards. No matter where it has appeared, inevitably it leaves the source region tailwards. If at a certain tail cross-section the density of the ions and their tailward velocities are known it is possible to calculate the ion flux at this cross-section. In what follows we will use a coordinate system originating at the nucleus and with x, y, and z -axis directed, respectively, along the projection of the antisolar direction, perpendicular to x in the sky plane, and along the line of sight. Then the total ion flux through a plane crossing the tail at a distance from the nucleus, is given by:

Here denotes the observed column densities, are the tailward velocities, projected on to the sky plane, and the integration is performed perpendicular to the projected tail axis. At first glance we know all quantities needed to calculate the ion flux, namely the measured column densities and the velocities obtained from the tracing of plasma structures. The problem is that we can associate velocities only with a very limited number of selected, more dense structures in the ion tail, but we do not know the velocities in the ambient plasma. What we need is a velocity profile across the tail. BSJ93 have measured velocities perpendicular to the tail, but unfortunately only a week later than our observations, on Dec 2, 1992. Another problem in the straightforward use of these observations is that they finish at a lateral distance from the tail axis of 4 10⁴ km and 6 10⁴ km, respectively. These distances are in agreement with the extent of the central concentration in the profiles, discussed in the previous subsection, but we observe non-zero column densities at greater distances from the tail axis, too. What to multiply these column densities with?

Taking into account that it is impossible to associate a velocity measurement with each measured column density, we will estimate the ion flux in the tail of comet P/ST using mean values. If we denote the mean flow speed in a cross-section perpendicular to the tail axis as (Bonev & Jockers 1994), the total ion flux is given by:

where is the mean number of ions per unit projected tail length.

Integrating the column densities across the tail in our CO and H₂ O images, we found the spatial distributions of and (H₂ O ), presented in Fig. 10 together with their ratio.

Fig. 10. Ions per unit tail length and the CO /H2 O ratio plotted versus deprojected distance from the nucleus.

The profiles of the ions per unit tail length are significantly different for both ions. As in comet Austin 1990 V (Bonev & Jockers 1994) (H₂ O ) decreases at distances 5 10⁴ km. This trend is not seen in the case of CO . Most probably this again is caused by the different scale lengths of the parents of both ions. Beyond 5 10⁴ km the amount of H₂ O ions added to the plasma flow decreases strongly. As H₂ O is rather stable outside the inner coma it can be assumed that the H₂ O flux remains constant outside of the source region. At the same time the ions are accelerated tailwards. Under these conditions, according to Eq.  5, the number of ions per unit tail length decreases. For CO , the nearly constant -value might be a consequence of the greater scale length of their parents, which feed the plasma flow even far from the nucleus with new ions, enough to compensate for the dilution caused by the acceleration.

The structures in our images, except (f), can be considered as defining a cross-section perpendicular to the axis at about 1.5 10⁵ km in the tail of comet P/ST (see Fig. 8). Using the velocities, weighted by the column density of each structure, we found a mean sky plane velocity spv of 20 km s^-1. This value, associated with the whole cross-section of the tail at 1.5 10⁵ km, is greater than the local velocities measured on the axis at the same distance (structure (e)), which is in agreement with MHD models of the plasma flow in a comet (Wegmann, private communication). With this mean velocity and with the values taken at the same distance we derived fluxes of s^-1 and s^-1 for the H₂ O and CO ions, respectively. The projected distance of 1.5 10⁵ km corresponds to a real, deprojected distance from the nucleus of 2 10⁵ km. The largest contribution to the error comes from the velocity determination. The error of the scale factor is of minor importance, as we are concerned with the large projected distance of 1.5 10⁵ km downstream of the nucleus. In addition we recall that for the H₂ O lines in our filter passband we have somewhat arbitrarily assumed half the g-factor of the 0-8-0 transition.

It is interesting to compare the obtained water ion flux with measurements of the water production rate during this time. On Oct. 15, 1992 Colom et al. (1992) observed P/ST with the Nancay Radio Telescope and estimated an OH production rate of 8 10²⁸ s^-1. Water production rates can be derived from the OH radio lines following the model of Bockelée-Morvan et al. (1990), who argue that Q(H₂ O) = 1.1 Q(OH). Taking into account that in the period from Oct. 15 to Nov. 25 the heliocentric distance of the comet r changed from 1.35 to 1.00 AU, and that the gas production dependence is steeper than r^-2, we derive a lower limit of 1.6 10²⁹ s^-1 for the water production on Nov 25. Tozzi et al. (1994) have inferred a water production rate of s^-1 from IUE observations on November 16. Later, at the beginning of December 1992, Bockelée-Morvan et al. (1994) found that the OH production rate was several 10²⁹ s^-1. Fink & Hicks (1996) measured the emission flux of the [OI] 6300 Å line in P/ST and, thereby, derived a water production rate of 9.5 10²⁹ s^-1 and 8.8 10²⁹ s^-1 on Nov 24 and Nov 26 1992, respectively.

The values of Fink & Hicks are rather high. Nevertheless we adopt them for our discussion, because they were obtained closest to our date of observations. Then, when comet P/ST was at 1 AU heliocentric distance, the ratio H₂ O/H₂ O 220. The obtained ratio of neutral water to water ions is greater than in the MHD model of P/Halley by Wegmann et al. (1987), the model with the most detailed treatment of the physical conditions and chemical reactions in a cometary coma. In this model the production rate of water is 5.5 10²⁹ s^-1 (total gas production rate of 6.9 10²⁹ mol s^-1, 80% of which being water). Integrating the H₂ O column densities (Fig. 7 in Wegmann et al. 1987) perpendicular to tail axis at 10⁵ km tailward from the nucleus we found 10²¹ cm^-1. Using the velocity maps of Wegmann et al. (1987) supplemented by private communication from these authors we obtain the mean velocity = 40 km s^-1. This gives a flux of H₂ O = 5.2 10²⁷ s^-1 and a ratio H₂ O/H₂ O 105.

If only photodissociation and photoionization of water are considered, about 3% of the neutral water should end up as water ions. This is about 3 times more than the value found by Wegmann et al. (1987). In the inner coma, where the neutral water density is sufficiently high, water ions are lost by the reaction
H₂ O + H₂ O H₃ O + OH
Another destructive channel is the dissociative recombination. Schmidt et al. (1988) pointed out that as soon as the water ion becomes dominant the reaction
H₂ O + e OH + H
dominates the other recombination processes. Most probably, these two mechanisms are responsible for the destruction of about two of three created H₂ O ions, explaining thus why in a detailed model about 1% of the neutral water ends up as water ions. Water ion destruction is important only in the inner coma, i. e. our previous discussion on the total ion fluxes in the plasma tail remains valid.

Our measured H₂ O /H₂ O ratio of is greater than both the value of derived from the simple comparison of photodissociation and photoionization of water and the value 105 of the MHD model of Wegmann et al. (1987). It is possible that part of the ions remained undetected in our measurements, namely the ions at greater distances from the tail axis, where the signal/noise ratio is insufficient. They can give a considerable contribution to the ion flux, because their velocities are high. The problem of nondetection of fast moving ions has been discussed in Jockers (1991), Rauer and Jockers (1993), and Bonev & Jockers (1994). When the comparison of measured fluxes has been made with values based on the simple photodissociation - photoionization model, the amount of the nondetected ions has been overestimated. Our analysis shows that most probably there remains an underestimation of the H₂ O ion flux but this effect is not as strong as previously believed. The comparison of the H₂ O flux with water production values derived from extrapolated OH measurements yields a lower H₂ O/H₂ O ratio, in better agreement with Wegmann et al. (1987).

It should be noted that, for several reasons, in the calculation of the CO flux a greater underestimation has to be expected as compared to the H₂ O flux. First, our measurement of the ions per unit tail length can not account for ions at larger lateral distances from the tail axis, because of observational restrictions (detector sensitivity and limited field of view). Second, is expected to be greater for CO than for H₂ O , because a larger contribution of CO ions at greater distances from the tail axis. And third, the deprojected distance of 2 10⁵ km is still in the source region of CO ions.

4.1.4. The CO to H₂ O ratio

From the observational point of view different kinds of CO /H₂ O ratios can be defined, the ratio of column densities, ratio of the ions in tail cross-sections, or the ratio of the total ion content. Fig. 6 shows that the ratio of the column densities increases from a mean value of about 0.5 on the axis to 1 at lateral distances 6 - 7 10⁴ km. The ratio between the ions per unit tail length, presented in Fig. 10, increases from about 0.52 at 1 10⁵ km deprojected distance to more than 0.6 at 2 10⁵ km. Because of contamination close to the nucleus we cannot quote a value for the total ion content.

Most suitable for comparison with other observations is the column density ratio measured on the tail axis. Lutz et al. (1993) have measured a mean value CO /H₂ O between 1 and 3 km in the plasma tail of P/Halley. In comet P/ST we find a value of 0.39 close to the axis at a tailward distance of 7 km. At the distance of 17 CO /H₂ O = 0.58. The errors are 7% (9%) at 7 km and 4% (6%) at 17 km. The errors in the brackets include the photometric error in the ratio of the two wavelengths. Apart from photometry the scale factor (Eq.  1) contributes the largest error to this estimate. As the relative contribution of the ion emission to the on-line image increases along the tail, the error decreases with increasing distance into the tail. This higher value of the CO /H₂ O ratio might be an indication that in P/ST the relative abundance of CO and CO₂ with respect to H₂ O is greater than in P/Halley. When P/Halley was observed by Lutz et al. (1993), it was at a heliocentric distance of about 1.3 AU, larger than the distance of 1.0 AU of P/ST during our observations, but sufficiently close to the Sun that water readily sublimes. Therefore, as was shown by A'Hearn et al. (1995), the ratios between gaseous species measured at 1.3 AU can be used at 1.0 AU with little error.

Additional evidence for a greater relative abundance of CO parents with respect to H₂ O in P/ST than in P/Halley comes from the comparison with model results. The column densities, presented in the model of P/Halley by Wegmann et al. (1987), yield a value of about 0.25 for the CO / H₂ O ratio near to the axis, at a tailward distance 1 10⁵ km. Fink & Hicks (1996) have shown that P/ST has a higher gas production rate than P/Halley at the same heliocentric distance. According to the scaling laws in MHD (Wegmann 1995), when all other conditions are unchanged, the model spatial scale expands for comets with higher production rates. At distances 10⁵ km we measured CO /H₂ O 0.5, with a tendency of increasing further tailwards. Thus the CO /H₂ O column density ratio in P/ST is significantly greater than the one derived by MHD modeling of P/Halley. This supports our statement that the abundance of CO-bearing molecules relative to H₂ O is slightly increased in P/ST in comparison to P/Halley. It should be noted, however, that this discussion depends on the proper selection of the fluorescence efficiency factor for the H₂ O ions in our filter, which was rather arbitrarily assumed to be half of the total 0-8-0 transition.

Having found a satisfactory agreement in the comparison of our data with both other observations and model results, it is tempting to check if a simplified consideration of the same subject would give similar results. CO is produced mainly by photoionization of carbon monoxide and, near to the nucleus, by photodissociative ionization of carbon dioxide. However, theoretical modeling by Huebner and Giguere (1980) shows that only about 10% of the total CO production comes from CO₂. Thus, in a first approximation, the observed CO densities can be considered as reflecting mainly the CO abundance in a comet.

According to Huebner (1985), the branching ratio of the reaction
CO + h CO + e
is about 0.5. Several processes are contributing to the production of H₂ O but the dominant one is the photoionization of neutral water:
H₂ O + h H₂ O + e
with a branching ratio of about 0.03. The consideration of only these two channels of ion production yields following relation between the parent species and their daughters:
CO / H₂ O = 0.06 CO / H₂ O
With CO / H₂ O = 0.3, the mean value measured by Lutz et al.(1993) in the tail of P/Halley, we obtain a CO/ H₂ O ratio of 0.02, much less than the value derived from in situ measurements of P/Halley (Eberhardt et al. 1987). This contradiction shows how the result would be biased in the case of a simplified treatment of measured CO and H₂ O ion densities. At the same time, the agreement with the detailed MHD model by Wegmann et al. (1987) illustrates the importance of accounting for all known chemical and physical processes in the inner coma of a comet. It also indicates that within the observational uncertainties the Wegmann et al. model is complete.

4.2. The dust

4.2.1. The dust color

The scaling factors, obtained in the process of continuum subtraction, allow to derive the color of the dust in the spectral range of our observations. A good useful measure of the continuum color is the normalized gradient of the reflectivity, (A'Hearn et al. 1984; Jewitt 1991):

where R is the reflectivity at given wavelength, , and is the mean reflectivity within a wavelength range . In our case the ratio between the reflectivities at both considered wavelengths is equivalent to the ratio between expected response of the instrumentation to solar continuum radiation and the empirically obtained continuum scaling factor. Applying the definition of the normalized gradient of the reflectivity (6) to the data in Table 3, we find the following colors for the dust in comet P/ST:

The reflectivity gradients in (7) are reduced to 1000 Å. The subscript j stays for the cases in which a good removal of the jet was achieved, i.e. these are the colors of the dust particles residing in the jet. Positive values correspond to reddening of the dust. Negative values would indicate that the dust is bluer than the solar spectrum. The dust in comet P/ST is redder in the spectral region between 4260 Å and 6420 Å, and this reddening is stronger for the particles residing in the jet. A stronger reddening in a jet as compared to the diffuse dust coma was found by Hoban et al. (1989) in P/Halley. In a recent paper Kolokolova et al. (1996) show that the dust color is mainly determined by the grain size distribution. Hence, the stronger reddening observed in cometary dust jets might indicate that the spatially isolated eruptive event on the nucleus causing the jet produces relatively more grains with greater sizes than the process of continuous release of dust particles from the cometary surface. More detailed conclusions can possibly be drawn about the dust properties in comet P/ST if our results are considered in combination with polarization and infrared measurements. Recently, Eaton et al. (1995) found that the polarization is up to 4% higher in the jet than in the surrounding coma in this comet. Goldberg and Brosh (1995) argue that the polarization in the jet region is higher than values typical for `dusty' comets.

The accuracy of our measurements does not allow to be confident about the color in the spectral region around 6300 Å.

4.2.2. Radial profiles and dust production

In the case of a steady state radial outflow from a central source one would observe a surface brightness decrease of the dust coma , where is the projected distance from the source. In comet P/ST, which is well known for its activity, to a large extent the dust coma and tail are replenished by a number of spatially localized outbursts, rather then by steady release of dust particles from the whole nucleus. Our single continuum image, presented in Fig. 3, reveals a highly inhomogeneous dust distribution in the inner coma. A detailed description of this very complex coma would require elaborate modeling, accounting for each structure individually, e.g. in a way as it has been done by Sekanina (1981). We will concentrate here only on some more general features of the dust distribution.

The radial profiles along the projected sunward and antisunward directions are well described by a power law with exponents of -1.73 and -1.25, respectively. In deriving both values we excluded the innermost part of the image ( 7 - 8 arcsec) which is blurred by a systematic tracking error. Outside this region the obtained exponents are constant throughout the whole field of view.

For the study of the mean dust distribution we have averaged the continuum image azimuthally around the nucleus. The result is shown in the upper panel of Fig. 11. The nearly constant region in the innermost part is an artifact caused by the above mentioned tracking error. The logarithmic gradient of the linear section between 5 10³ and 4 10⁴ km is -1.4. The faster decrease of the surface brightness might be caused by radiation pressure acceleration of the dust particles. Another explanation could be a gradual fading of the grains in their outward motion. Most probably both effects contribute to the fact that the brightness decreases faster than expected in the case of simple radial outflow with constant velocity. A transition of the logarithmic gradient from -1.4 to -1.9 is seen at about 4 10⁴ km. This is approximately the distance to which the spiral jet extends. We could speculate that the particles in the jet are not so strongly influenced by the radiation pressure as the ambient dust. Therefore their contribution to the brightness reduces the slope in the inner part of the coma. Such an explanation is in agreement with our previous argument that the grains in the jet should have larger sizes.

Fig. 11. Upper panel: The product between geometric albedo, phase function and filling factor as a function of projected distance to the nucleus . Lower panel: The product between geometric albedo, phase function, filling factor, and projected distance.

As a measure of the dust production, A'Hearn et al. (1984) have introduced the quantity , the product of albedo, filling factor of grains within the field of view, and the linear radius of the field of view at the comet. Based on the assumption of a simple radial outflow this quantity is independent of the field of view, giving thus the opportunity to compare measurements obtained under different geometrical conditions. In a similar way we have calculated the quantity , presented in the lower panel of Fig. 11. Recall that the geometric albedo, p, used in our calculation is one quarter of the Bond albedo, A, used by A'Hearn et al. (1984). Moreover, our local filling factor is one half of the aperture integrated value of A'Hearn et al. (1984) ¹. Thus, in order to compare our quantity with values, the latter should be divided by 8.

The outward increase of near to the nucleus reflects simply the blurring of the inner coma by incorrect guiding. If in reality there was a constant section of in this inner most region, it should appear at a level below the value of the artificial maximum. In order to check if the averaged radial profile in the inner most region of the dust coma can be described by an brightness distribution we modeled numerically the smearing imposed on our continuum image. The result of this simulation is shown in both panels of Fig. 11. The dotted line represents the unsmeared model and the dashed line is obtained by smearing this model in the same way as it happened during the observations. The good fit of this rather simple model to the measured data allows to deduce a value of cm for P/ST. The quoted error derives from the known amount of incorrect guiding and the photometric error.

Osip et al. (1992) provide a value of for comet Halley during the Giotto encounter. At this time the phase angle of P/Halley was for the ground-based observer. We observed P/ST at a phase angle . Both values are in a range where the phase function of cometary dust shows little variation (Ney 1982), thus no correction is needed for the phase function. Dividing the value of Osip et al. (1992) by 8 we obtain cm for P/Halley during the Giotto encounter. The extrapolation of this value by application of a heliocentric dependence of the dust production, yields a value cm for P/Halley at 1 AU, the distance at which we observed P/ST. Osip et al. (1992) claim that the dependence is typical for the dust of most comets in their data base. Thus, if the physical characteristics, chemical composition, and the outflow speeds of the dust particles in both comets were the same, comet P/ST had about 1.6 times greater dust production than P/Halley at the same heliocentric distance. As the gas productions of both comets differ by almost the same factor the dust to gas ratio should be comparable in comets P/Halley and P/ST.

© European Southern Observatory (ESO) 1997

Online publication: July 3, 1998
helpdesk@link.springer.de