arXiv:astro-ph/0003122v1 8 Mar 2000
2and CO+in Comets 122P/1995S1 (deVico) and C/1995O1
Anita L. Cochran, William D. Cochran and Edwin S. Barker
University of Texas at Austin
Accepted for publication in Icarus
We observed comets 122P/1995 S1 (deVico) and C/1995O1 (Hale-Bopp) with high spec-
tral resolving power in order to determine the ratio of N+
clearly detected the CO+in both of these comets, no N+
From these spectra, we derive sensitive upper limits for N+
substantially below other reported detections of N+
the prior N+
2detections and compare them with our observations. The abundance of
N2in comets is important to our understanding of the condensation of ices in the solar
nebula. In addition, N2is a tracer of Ar so study of N2allows an understanding of the
role of comets for delivering volatiles to the terrestrial planets. It appears that many, if
not most, comets are depleted in N2and it will be necessary to search for a mechanism
for depleting this molecule in order to be consistent with current models of the solar
2/CO+in their comae. While we
2was detected in either comet.
2/CO+. These upper limits are
2/CO+in other comets. We discuss
Nitrogen is one of the more abundant elements in the universe and is therefore assumed
to be an important constituent of the solar nebula and of the comets. Nitrogen probably
exists in comets in the form of N2and other nitrogen-bearing molecules, including NH3.
Indeed, the ratio N2/NH3is a sensitive indicator of conditions in the solar nebula. Lewis
and Prinn (1980) point out that at high temperatures and low pressures the dominant
equilibrium species of carbon, oxygen and nitrogen would be N2, CO, and H2O. Only as
temperature and pressure regimes change will CH4and NH3be produced and N2and
CO be depleted. They conclude that the conversion of N2to NH3and CO to CH4would
be sufficiently slow relative to radial mixing in the primitive solar nebula so that only
small amounts of NH3and CH4should be present. Conditions in the circumplanetary
nebulae would be sufficiently different so that jovian planets might have increased NH3
and CH4abundances (Prinn and Fegley 1981).
Comets delivered some of the volatiles that we see today in the atmospheres of the
terrestrial planets, but it is not certain how important a source of volatiles the comets
represent. Owen and Bar-Nun (1995a,b) have pointed out that N2is an important guide
to the volatile abundances of comets because it is trapped and released by amorphous
ice in a manner which is similar to argon (Bar-Nun et al. 1988). Using N2as a guide to
the argon, one can determine the extent to which comets enriched the volatile and noble
gas components of the terrestrial planets. Ices formed at low temperatures will trap gas
from the surrounding nebula, fractionating the original mixture as a function of the local
temperature. Thus, they suggest that comets which formed near Uranus and Neptune,
at temperatures around 50K, would be the source of noble gases for Earth and Mars,
while the higher quantities of neon and argon in the atmosphere of Venus, compared
with Earth, would require comets formed at colder temperatures, such as in the Kuiper
belt, to be the deliverers of some of the volatiles.
We detect such species as NH, NH2and CN in every comet, so evidence of nitrogen
carriers is easily available. Most of these species and their parents are chemically reactive
in the comae of comets. Molecular nitrogen should be less reactive than species such as
NH3or HCN. While spacecraft have flown past comet Halley with mass spectrometers
onboard, measurement of N2is difficult with mass spectrometry since both N2and CO
occupy the mass 28 bin of these instruments (cf. Everhardt et al. 1987 for a discussion
of CO and N2from Giotto observations of Halley). Thus, disentangling the quantity of
N2from the CO is very model dependent.
This leaves the field of ground-based spectroscopy for determining the quantity
of molecular nitrogen. Ground-based studies of molecular nitrogen are very difficult,
however, because of the N2abundance of the Earth’s atmosphere. To circumvent the
difficulty in observing N2, past observations have concentrated on the N2ion, primarily
through observations of the N+
is expected to be seen only in the tails of comets. Care must be taken when observing
this band since N+
2emission is also excited in the atmosphere of the Earth, especially
2(0,0) band at 3914˚ A. This band is extremely weak and
near dusk and dawn, when comets are often observed. Auroral activity will also excite
this band in the terrestrial atmosphere. Additionally, this weak feature can easily be
confused with other, nearby, cometary emissions. Thus, accurate measurement of N+
in cometary spectra requires both good spatial and spectral resolution to separate the
features from that of the Earth and other cometary features.
We observed comets deVico and Hale-Bopp with the 2DCoude spectrograph (Tull et al.
1995) on the 2.7-m Harlan J. Smith telescope of McDonald Observatory. The 2DCoude
has two operating modes. The “lower” resolution mode has a resolving power, R=60,000.
In this mode, spectral coverage is complete from around 3800-5800˚ A and coverage con-
tinues to 1µm with increasing interorder gaps. Typically, 60–65 spectral orders are
observed. Therefore, in the blue, many molecular bands can be observed simultaneously,
regardless of the exact grating setting. In the “high” resolution mode, R=200,000, but
the coverage is much less complete than the lower resolution mode. Typically, high res-
olution covers 10–15 orders of approximately 15˚ A each. Thus, care must be taken to
center key features on a spectral order and many features remain unobserved.
For this project, we observed comet Hale-Bopp in the high resolution mode, care-
fully centering the portion of the order containing the N+
observing the CO+(2,0) and (3,0) bands on two other orders. Since the CH+(0,0) band
occurs at a wavelength coincident with the CO+(2,0) band, this ion was also observed.
Comet deVico was observed in the lower resolution mode and the same three ions were
observed. Table I gives the circumstances of the observations. For all observations, the
slit was 8.2arcsec long. For the Hale-Bopp observations, the slit was 0.34arcsec wide,
while it was 1.2arcsec wide for deVico. Each slit width projects to two pixels on the
CCD for the resolving power of the observations. Different positions within the coma
were observed by moving the telescope around the sky under accurate computer control.
2band on the CCD, while also
The data were reduced using the echelle package of iraf. Incandescent lamp
observations were used to determine the flat field; ThAr lamp observations were used for
calculating the dispersion curve. The rms errors of our fits for the dispersion curve are
0.24m˚ A for the Hale-Bopp spectra and 2.5m˚ A for the lower resolution deVico spectra.
The solar spectrum was observed with an identical instrumental setup to that used for
the comets by imaging the Sun through a diffuser on the roof of the spectrograph slit
room and projecting this image through the slit in the same manner as objects observed
through the telescope. Thus, we used an observed solar spectrum in our reductions.
Care was taken to preserve the relative flux levels of the spectra. The spectral
orders were extracted by first tracing the order along the chip and carefully setting the
edges of the apertures. Since the continua of the cometary spectra were rarely of high
enough signal/noise to define well the aperture boundaries, it was assumed that the flat
lamp boundaries were appropriate for the comet and only the position on the chip of
the center of the order was computed for the cometary spectra. Extraction was done
using variance weighting. We used a 1D fit for stars and solar spectra, while a 2D fit was
used for cometary spectra (because of the emission line nature of the cometary spectra).
At the end of the routine reduction, we had files containing n spectra for each initial
spectral image, where n is the number of extracted orders in the image (60 for deVico
and 13 for Hale-Bopp).
The solar spectrum observations were used to remove the underlying continuum
from the cometary spectra. Comet deVico has very little solar continuum, but the con-
tinuum of Hale-Bopp was quite strong. We corrected the comet and the solar spectrum
for the geocentric and heliocentric Doppler shifts so that both were on a common rest
frame. Then, the solar spectrum was carefully weighted to match the continuum level
of the comet in regions away from cometary emissions. Some scattered light might still
remain, but the amount is minimal and was removed when the line intensities were
Figure 1 shows the spectral order of the CO+(2,0) and the CH+(0,0) bands in
the spectrum obtained 100arcsec tailward of the optocenter of Hale-Bopp. For both
these ions, the predicted molecular transitions in the spectrum are marked. Inspection
of this figure shows that only very low J−levels are observed for CH+, while slightly
higher J−levels are observed for CO+. However, even for CO+, J−levels above 10 or 11
are not seen.
Figure 2 shows the same spectral region for observations 100arcsec tailward of the
optocenter of comet deVico. The spectral coverage of an order is longer at the lower
resolving power of the deVico observations. Even with this larger coverage, only one
of the CO+(2,0) ladders is seen in these observations. Both CH+and CO+are again
present, though the ratio of CH+/CO+may be slightly different in these two comets.
Figure 3 shows the Hale-Bopp spectrum obtained 10arcsec tailward of the optocen-
ter in the spectral order which should contain the N+
B2Σ–X2Σ transition and therefore does not have a Q-branch. The P- and R-branch line
positions marked are from Dick et al. (1978) and have an accuracy of 0.01cm−1= 2m˚ A.
Although the solar-subtracted spectrum is somewhat noisy, there are no believable fea-
tures. There might be a spike at 3909˚ A and a broader spike at 3913.5˚ A. Neither of these
is coincident with any of the N+
2line positions. The errors in the wavelengths of our
spectra, coupled with the N+
2laboratory errors would lead us to expect coincidence to
2m˚ A. The most believable feature is the broad feature starting at 3914.5˚ A and degrading
redward. The positions of the C3(0,2,0)-(0,0,0) band transitions are marked underneath
the spectrum. The feature seems to match well the R-branch bandhead of this C3band.
Thus, we conclude that if the feature is real, it is some residual C3 emission. Since
this spectrum was obtained only 10,000km from the optocenter, this does not seem an
unlikely species to observe. Inspection of the spectrum obtained 100arcsec tailward of
the optocenter shows even less evidence for features. We therefore conclude that we did
not detect any N+
2in the spectrum of Hale-Bopp.
2(0,0) band. The N+
2transition is a
Figure 4 shows the comparable spectral region for comet deVico, 100arcsec from
the optocenter. There appears to an upward fluctuation between 3913.5 and 3915.1˚ A,
with spikes at 3913.9 and 3914.9˚ A. However, at R=60,000, it is impossible to tell if this
feature is a molecular band and, if so, which way it degrades, or to differentiate whether
it is C3or N+
P-branch should be visible. We do not detect lines at the expected wavelengths, within
the 3m˚ A wavelength uncertainties. On the strength of the Hale-Bopp observation, this
feature could be C3. With the larger spectral coverage of the R=60,000 orders, the blue
end of this order contains the CN (0,0) bandhead (not shown in Fig. 4). Thus, we were
able to use the high signal/noise CN emission lines to confirm that there were no errors
in our wavelength solution or in our Doppler shift corrections. The centers of the CN
lines fell at the correct wavelengths, verifying that any spikes in the N+
also have to be at predicted wavelengths.
2. The N+
2P-branch bandhead occurs at 3914.3˚ A and distinct lines of the
For the deVico observations, our only spectrum off the optocenter was obtained
more than 70,000km into the tail. Thus, it is reasonable to ask about the likelihood that
we observed C3this far from the optocenter. Figure 5 shows the optocenter spectrum
from the same night. We show more of the order to illustrate the abundance of molecular
emissions observed. The CH B–X (0,0) band is clearly detected along with several C3
bands. Indeed, inspection of this plot shows there are several broad, unidentified features
whose structure seems similar to the identified C3bands. Thus, it is likely that this order
is riddled with C3. However, comparison of the strength of the strongest C3band, the
(0,0,0) – (0,0,0) band, in the optocenter and tail spectra makes it unlikely that we
detected the much weaker (0,2,0) – (0,0,0) R-branch bandhead in the tail spectrum.
Thus, it is most likely we detected only noise in the deVico spectrum.
In summary, we do not believe that any N+
Hale-Bopp or deVico. The CO+and CH+were clearly detected in both comet’s spectra.
We are therefore able to place limits on the important ratio of N+
2was detected in the spectra of comets
2/CO+in these two
Limits on N+
For both comets Hale-Bopp and deVico, the CO+(2,0) band was clearly detected. Thus,
we can derive an abundance of CO+from these data for comparison with other comets.
However, coud´ e data can not be easily calibrated into absolute fluxes, so we must work
with band intensities in detector counts.
observations observe all branches of the CO+band, while we only observe the2Π1/2(F2)
branches. We took the simple approach of “integrating” the band by fitting a continuum
and summing the counts in the band above the continuum. We limited our bandpass to
just the region of the detected lines. These values are given in Table II. While a larger
bandpass would be more comparable to prior low-resolution observations of comets, it
is inappropriate for high resolution observations since larger bandpasses would increase
the noise with no increase in signal. Typical low-resolution observations do not return
In addition, typical lower resolving power
to continuum in between the lines of different bands.
We do not believe that we have detected the N+
upper limits by computing how much of a band could be hidden within the noise. We
did this by computing the rms in a bandpass. Then, the upper limit is just 1/2 ×rms×
bandpass, in appropriate units. These are 2σ upper limits. For the same rms, more signal
can be hidden in a large bandpass than a small bandpass. Since we do not know exactly
how many lines would be likely to be detected, we cannot easily define the bandpass.
We assumed a bandpass which would include the complete P-branch of N+
counts, which are 2σ upper limits for what could be hidden in the noise, are listed in
Table II, column 3.
2in either comet. We computed
2. The derived
With the use of a few assumptions and simplifications, we can use our values to
regions, in comparison with the published atlas of Kurucz et al. (1984), shows that
the sensitivity of the N+
sensitivity, we would need to multiply the N+
the calibration of the solar spectrum depends on details such as activity, so this factor
is uncertain. We do have observations of α Lyr with the same instrumental setup as for
deVico, but since the N+
2band occurs in the Balmer decrement, where the α Lyr flux
changes rapidly with wavelength, the α Lyr flux is not calibrated in this region. Assuming
a smooth decrease through the Balmer decrement, we confirm that a correction factor
of 1.5–1.7 for the N+
2counts would be appropriate. We therefore adopt a factor of 1.6.
2/CO+for these two comets. Examination of the solar spectra in these two
2order is lower than that of the CO+order. To match their
2upper limits by a factor of 1.7. However,
Once the band intensity is known, the column density can be computed using
N = L/gν′ν′′
where N is the column density, L is the integrated band intensity and gν′ν′′ is the exci-
tation factor. We used excitation factors of 7.0 × 10−2photonssec−1mol−1for the N+
(0,0) band (Lutz et al. 1993) and 3.55 × 10−3photonssec−1mol−1for the CO+(2,0)
band (the average value from Figure 2 of Magnani and A’Hearn 1986). Then,
For CO+, we observed only one of the two ladders. If we assume the two ladders
are equal strength, we should multiply our CO+intensity by two for the calculation.
We likewise need to multiply the N+
2upper limits by a factor of two since we have only
measured the P-branch and the R-branch should have a similar intensity. In Table II,
column 4, we list our upper limits for N+
for the sensitivity difference of the two orders.
2/CO+, including using a factor of 1.6 to correct
It would be impossible to hide much N+
shows one of the Hale-Bopp spectra, as observed, and, in the lower panel, the same
2in our spectra. Figure 6 (upper panel)
spectrum with a feature added which has enough integrated counts to yield the Halley
this synthetic band is the exact shape that would be present, nor are the “lines” at
exactly the N+
2wavelengths, but it gives an idea of the ease with which we would detect
such a feature. Clearly, no feature this distinctive could be missed in our observations.
2/CO+ratio (Wyckoff and Theobald 1989 – discussed below). We do not claim that
Previous Observations of N+
Most comets are not bright enough to be observed with the high spectral resolving
powers that we used for deVico and Hale-Bopp. This was especially true in the past,
when detectors, such as photographic plates, had much lower quantum efficiency than
our current CCD detectors. Therefore, prior observations which have detected N+
cometary spectra have been obtained with lower resolution, often on photographic plates.
2feature is generally weak and is overwhelmed by other molecular emissions
near the optocenter. In addition, since N+
2is an ion, it is entrained in the solar wind
magnetic field and rapidly accelerated into the tail. Thus, spectra of the tail region are
necessary for its definitive detection, yet tail spectra are generally of lower signal/noise
than near-optocenter spectra since the cometary brightness falls with increasing cometo-
centric distance. Despite these difficulties, observations of comets exist which show the
detection of N+
2in the tails of comets.
Only two of the prior reported observations are digitally measured spectra; the rest
are estimates from digital spectra or are photographic spectra. Wyckoff and Theobald
(1989) report a detection of N+
2in the tail of comet Halley at a cometocentric distance
of 3 × 105km tailward. These observations were at much lower resolution than our
observations. They detected a weak emission in the region from 3885–3950˚ A which they
concluded was composed of contributions from the CO+(5,1), CO+
(0,0) bands and an unidentified band. By modeling the combined feature, they were
able to estimate the contribution of N+
2to the mixture. Using this estimate and the
average for the CO+(2,0), (3,0) and (4,0) column densities, they derived a value of
2was not accurate and Wyckoff et al. (1991b) revised the value of the column
density of N+
2using the excitation factors of Lutz (1989–a personal communication).
This excitation factor is the same as that given in Lutz et al. (1993). If we apply the
value from Lutz et al., then N+
CO+is used, then N+
2/CO+= 0.004. However, the excitation factor which Wyckoff and Theobald used
2/CO+= 0.002. If only the (2,0) band column density of
Lutz et al. (1993) reported observations of the tails of two comets obtained at low
resolution (∼ 10˚ A). For comet Halley, they obtained spectra at 2×104and 2×105km from
the optocenter in the tailward direction. They claim to have detected no N+
the Halley tail spectra. However, they also did not detect the CO+emissions in several
Halley spectra. Their derived upper limits for N+
higher than the Wyckoff and Theobald detection.
2/CO+when CO+was detected were
In addition to observations of Halley, Lutz et al. also observed comet C/1987P1
(Bradfield=1987 XXIX). For this comet, spectra were obtained at 2×104and 6×104km
from the optocenter. CO+was detected in both spectra, but N+
the larger cometocentric distance. At their resolution, the N+
the CN (0,0) band. No mention is made of the possible contamination of this feature by
the CO+(5,1) band. Assuming all of their measured band was N+
value of N+
2was only detected at
2feature is on the wing of
2, Lutz et al. derive a
The vast majority of observations of cometary tails were photographic. Not only
were they at lower resolving powers than our observations of deVico and Hale-Bopp, but
photographic plates are even more difficult to calibrate! Non-uniformity in response and
vignetting of the spectrograph slit cause difficulty interpreting these spectra. Still, there
are many fine examples of photographic spectra and these can be used to determine the
is that of Swings and Haser (1956). Examination of the plates in this atlas shows some
comets for which CO+and N+
2are both apparent, while other comets show evidence of
tails (i.e. CO+) but no N+
2. Arpigny examined these and other photographic and digital
spectra at his disposal and estimated the intensity ratio for the N+
band used was the (4,0) band because of its proximity to the N+
communication). Table III lists the 12 comets which he determined had both N+
CO+in these spectra, along with his estimate of the ratio of the intensity of the two
bands (column 2). In order to compare his intensity ratios with other observer’s column
density ratios, it is necessary to multiply by the ratio of the excitation factors, as before.
Since the (2,0) CO+was used in our work and in other published ratios, we converted the
intensity ratios in Table III by using the relationship I(4,0) = 0.6×I(2,0), where I(4,0)
is the intensity of the (4,0) band, I(2,0) is the intensity of the (2,0) band, and the factor
is taken from Table 4 of Magnani and A’Hearn (1986). The resultant column density
ratios are given in column 3 of Table III. Arpigny’s estimates of the intensity ratios are
consistent with the published numbers for comet Bester (Swings and Page 1950) and
comet Humason (Greenstein 1962). It should be noted that Warner and Harding (1963)
also observed comet Humason at a comparable heliocentric distance (however they only
discuss CO+, not N+
been observed in previous spectra of some comets.
2/CO+ratio for some comets. The largest published collection of photographic spectra
2/CO+, where the CO+
2emission (1999, personal
2). However, it is clear from Arpigny’s compilation that N+
In addition, Arpigny reported four comets which had good spectra but for which
no, or only very faint, evidence of a plasma tail existed. These comets are C/1948V1
(Eclipse), C/1963A1 (Ikeya), C/1968N1 (Honda), and C/1975N1 (Kobayashi-Berger-
Milon). Arpigny points out that N+
2emission is always very weak, so we should not
expect to see it when the CO+is weak or non-existent.
Our own examination of the atlas of Swings and Haser (1956) found five comets
for which there was evidence of a tail but no evidence for N+
(1910I) showed only continuum in the tail, so this was presumably a dust tail. Comets
Halley (1910II), Brooks (1911V), Gale (1912II), and Jurlof-Achmarof-Hassel (1939III)
showed evidence of weak CO+emission but no N+
observed by Bobrovnikoff in spectra of Halley obtained in 1910, but these spectra are
2. The Big Comet of 1910
2emission (Arpigny notes that N+
not included in the Swings and Haser Atlas). These would be similar to Hale-Bopp and
deVico in the absence of N+
2while other ions are present. However, with the weakness
of the CO+emissions in these four photographically observed comets, the N+
is most probably below the plate sensitivity.
In this paper, we have presented high resolution observations of two comets with which we
were able to study the relative abundances of N+
understanding the quantity of N2and CO, two of the least chemically reactive cometary
coma species. Conversion from the quantity of the ions to the quantity of the neutrals
is dependent on an understanding of the photodestruction branching ratios which are
not well understood (Wyckoff and Theobald argue you must multiply the ion ratio by a
factor of 2, while Lutz et al. find no factor necessary), so we will continue to discuss these
species in terms of their ions. For both Hale-Bopp and deVico, CO+was easily detected
2appears to be missing from the spectra. Thus, we have put very low upper limits
on the ratio of N+
comets which show CO+but not N+
2for which sensitive upper limits cannot be derived.
2and CO+. These two ions are proxies for
2/CO+. We note, however, that there are previous observations of
The quantity of N2and CO expected in a comet depends on several factors includ-
ing the temperature at which the ice was deposited, when in the history of the formation
of the solar system the gases were trapped in the ice and the orbital history of the comet
itself. Current models of the solar nebula have comets which now reside in the Oort
cloud forming in the Uranus-Neptune region (cf. Weissman 1991; Duncan, Quinn and
Tremaine 1987). The temperature in this region was probably about 50 ± 20K (Boss et
al. 1989). Thus, a first guess to the deposition temperature of cometary ices is 50K. This
is consistent with laboratory experiments described by Owen and Bar-Nun (1995b).
The first direct measure of a deposition temperature for ice came with observations
of deuterium in comet Hale-Bopp. Meier et al. (1998b) reported the detection of HDO
in Hale-Bopp and determined a ratio of D/H=(3.3 ± 0.8) × 10−4in H2O. In addition,
Meier et al. (1998a) detected DCN for the first time and derived a ratio of D/H=(2.3 ±
0.4) × 10−3in HCN. Note that the D/H ratio is different for these two species, with
D/H measured from HCN 7 times higher than from H2O. Since the D/H enrichment for
different molecules is a strong function of temperature, Meier et al. (1998a) were able
to derive a temperature for the cloud fragment in which this comet formed of no colder
than 30 ± 10K.
Bar-Nun et al. (1988) performed laboratory experiments on deposition of various
gases along with H2O ice and showed that CO is trapped 20 times more efficiently
than N2in amorphous ice which formed at 50K, when these two gases are present in
equal abundances with CH4 and Ar. This ratio changes slightly when only CO and
N2 are present in the gas (Notesco and Bar-Nun 1996, Table I) but generally shows
enrichment factors of 15–30. From these laboratory experiments, Owen and Bar-Nun
(1995a) concluded that icy planetesimals formed in the solar nebula at around 50K, the
temperature at which the studied comets should have formed, would have N2/CO≈ 0.06
in the gases trapped in the ice if N2/CO≈ 1 in the nebula. The predicted cometary ratio
of N2/CO is much higher than our upper limits for deVico and Hale-Bopp and is higher
even than the detections of Wyckoff and Theobald (1989) and Lutz et al. (1993), though
some of the estimates might show ratios this high.
Several factors might mitigate this discrepancy.
point out that the gas/water vapor ratio is not necessarily representative of the ratio of
ices in the nucleus. However, the laboratory experiments of Bar-Nun et al. (1988) have
demonstrated that CO and N2should be released simultaneously in the same proportion
as they exist in the ices. There is evidence for a source of CO at around 10,000km
from the nucleus (Eberhardt et al.
1987) which may be attributable to grains.
addition, Krankowsky (1991) points out that H2CO is probably an additional parent for
CO. Thus, there may be additional mechanisms for the production of CO which do not
exist for N2. While these factors might change the predicted ratio for N2/CO, Owen
and Bar-Nun (1995a) go on to make the specific prediction that “future observations of
dynamically new comets will show values of N2/CO systematically higher than those in
well-established short-period comets”. They point out that as comets are continuously
exposed to solar radiation in the inner solar system, they would be expected to lose any
N2in their outer layers in a brief period of time.
Prialnik and Bar-Nun (1990)
Figure 7 shows all of the values and limits discussed in this paper, plotted as
a function of 1/ao, the original semimajor axis [four of the values for the estimates
are the oscullating 1/a, as noted in Table III; for C/1987P1 (Bradfield) 1/ao= 0.006380
(Marsden and Williams 1995); for Hale-Bopp 1/ao=0.00535 and for deVico 1/aosc=0.057
(Marsden personal communications 1999)]. For the Wyckoff and Theobald Halley ob-
servation, we include the range of derived values.
If one ignores the Hale-Bopp upper limit, there would seem to be an increase
2/CO+with decreasing 1/ao, as was predicted by Owen and Bar-Nun. However,
the trend is based mostly on Arpigny’s estimates, which are approximate. In addition,
the Hale-Bopp upper limit can not be discarded since it is clear from inspection of the
spectrum in Figure 6 that even as much N+
can be questioned for the estimates, inspection of the atlas of Swings and Haser (1956)
and spectra such as the Humason spectrum of Greenstein (1962) clearly show an emission
which is coincident with the location of the N+
2in their spectra. However, at lower spectral resolution, blending of features
may lead to spurious detections of N+
2and wrong estimates of the strength of this band.
The potential for blending, coupled with the weakness of the N+
exists, point to a need for caution in interpretation.
2as would be needed to equal the Halley
2/CO+cannot be hidden in this spectrum. Conversely, while the values for the ratio
2feature. Thus, at least some comets have
2feature even when it
Thus, at this point, we have contradictory evidence for the ratio of N2/CO. At
least in the active outer regions of Hale-Bopp and deVico, these comets appear to be
very depleted in N2relative to CO. The observations of Halley by Wyckoff and Theobald
(1989) also pointed to a depletion of N2for Halley. Indeed, Wyckoff et al. (1991a) derived
a nitrogen depletion for comet Halley of a factor of ∼ 6 relative to the Sun. Owen and
Bar-Nun (1995) have pointed out the strong temperature dependence of trapping of N2
and CO. Comets for which the deposition temperature was greater than 50K could trap
progressively less CO and N2. However, H2CO will continue to be trapped in comparable
quantities to ices deposited at 50K. Thus, there would continue to be a source of CO,
but the N2will be depleted relative to the CO.
Perhaps the solution to the quandary of depleted nitrogen is that our assumption
that the solar nebula preferentially condenses nitrogen into N2instead of NH3is incorrect.
However, observations of dense molecular clouds (Womack et al. 1992) have shown that
N2>> NH3for these potential star-forming sites. Another possibility is that the comets
formed with much more molecular nitrogen but that it was depleted post-formation. It
is certainly true that comets will deplete volatile gases in their outer layers as they pass
close to the Sun, but this cannot be used as an explanation when comparing comets
with similar orbital histories, such as Halley and deVico or Hale-Bopp and Bennett and
C/1987P1 (Bradfield), which show discrepant ratios of N+
(1990) conclude that there might be some post-formational processing of Halley, but not
to any large extent because of the low internal temperatures which are derived from
the spin temperature of H2O (Mumma et al. 1993). They point out, however, that gas
can be trapped in water ice efficiently, but only if the ice is amorphous, such as it is in
the various laboratory experiments. They conclude that codeposition into amorphous
ice is unlikely to be a favorable mechanism in forming cometary ices since the water ice
condenses, for most solar nebula models, at 140–160K, at temperatures where the ice
is likely to be crystalline and to not adsorp volatiles readily. However, the D/H ratios
of H2O and HCN suggest that cometary deposition temperatures were not as warm as
Engel et al. posit.
2/CO+. Indeed, Engel et al.
In summary, in this paper we presented evidence of two comets for which no N+
was detected, along with stringent upper limits, which would indicate that these comets
are depleted in N2relative to CO. These observations are at odds with our understanding
of the formation processes of ices in the solar nebula. Either a mechanism must be
found to deplete the N2ice once formed or we must understand how a gaseous cloud
with N2 >> NH3 formed ices which do not contain much molecular nitrogen. Since
we believe that N2 will be deposited into H2O ice in a manner which is similar to
Ar, understanding this process has important implications for understanding the role of
comets for delivery of noble gases to the terrestrial planets. It is therefore important that
more unambiguous observations of the ion tails of comets be obtained, when possible, to
determine the intrinsic values of N2/CO in comets.
This work was funded by NASA Grant NAG5 4208. We thank Walter Huebner for
encouraging the Hale-Bopp observations and Toby Owen for stimulating our examination
of the data. We especially thank Claude Arpigny for graciously allowing us to use his
estimates of earlier observations and for many helpful discussions.
Bar-Nun, A., I. Kleinfeld, and E. Kochavi 1988. Trapping of gas mixtures by
amorphous water ice. Phys. Rev. B 38, 7749–7754.
Boss, A. P., G. E. Morfill, and W. M. Tscharnuter 1989. Models of the for-
mation and evolution of the solar nebula. In Origin and Evolution of Planetary and
Satellite Atmospheres (S. K. Atreya, J. B. Pollack, and M. S. Matthews, Eds.) pp. 35–
77 Univ. of Arizona Press, Tucson, AZ.
Dick, K. A., W. Benesch, H. M. Crosswihite, S. G. Tilford, R. A. Gottscho,
and R. W. Field 1978. High resolution spectra of bands of the first negative group
of ionized molecular nitrogen (N+
g). J. Mol. Spec 69, 95–108.
Duncan, M., T. Quinn, and S. Tremaine 1987. The formation and extent of the
solar system comet cloud. A. J. 94, 1330–1338.
Eberhardt, P., D. Krankowsky, W. Schulte, U. Dolder, P. L¨ ammerzahl,
J. J. Berthelier, J. Woweries, U. Stubbemann, R. R. Hodges, J. H. Hoff-
man, and J. M. Illiano 1987. The CO and N2abundance in comet P/Halley. Astr.
and Ap. 187, 481–484.
Engel, S., J. I. Lunine, and J. S. Lewis 1990. Solar nebula origin for volatile gases
in Halley’s comet. Icarus 85, 380–393.
Greenstein, J. L. 1962. The spectrum of comet Humason (1961e). Ap. J. 136, 688–
Krankowsky, D. 1991. The composition of comets. In Comets in the Post-Halley
Era (R. L. Newburn Jr., M. Neugebauer, and J. Rahe, Eds.) pp. 855–877 Kluwer
Academic Publishers Dordrecht, The Netherlands.
Kurucz, R. L., I.
1300 nm. National Solar Observatory Atlas, Sunspot, New Mexico.
Furenlid, and J. Brault 1984. Solar flux atlas from 296 to
Lewis, J. S. and R. G. Prinn 1980. Kinetic inhibition of CO and N2reduction in the
solar nebula. Ap. J. 238, 357–364.
Lutz, B. L., M. Womack, and R. M. Wagner 1993. Ion abundances and implica-
tions for photochemistry in comets Halley (1986 III) and Bradfield (1987 XXIX). Ap.
J. 407, 402–411.
Magnani, L. and M. F. A’Hearn 1986. CO+fluorescence in comets. Ap. J. 302,
Marsden, B. G. and G. V. Williams 1995. Catalogue of Cometary Orbits. Smith-
sonian Astrophysical Observatory, Cambridge, MA 10th edition.
Meier, R., T. C. Owen, D. C. Jewitt, H. E. Matthews, M. Senay, N. Biver,
D. Bockelee-Morvan, J. Crovisier, and D.
comet C/1995 O1 (Hale-Bopp): Detection of DCN. Science 279, 1707–1710.
Gautier 1998a. Deuterium in
Meier, R., T. C. Owen, H. E. Matthews, D. C. Jewitt, D. Bockelee-Morvan,
N. Biver, J.Crovisier, and D.
HDO/H2O ratio in comet C/1995 O1 (Hale-Bopp). Science 279, 842–844.
Gautier 1998b. A determination of the
Mumma, M. J., P. R. Weissman, and S. A. Stern 1993. Comets and the origin of
the solar system: Reading the Rosetta stone. In Protostars and Planets III (E. Levy
and J. Lunine, Eds.) pp. 1172–1252 The University of Arizona Press Tucson, AZ.
Notesco, G. and A. Bar-Nun 1996. Enrichment of CO over N2by their trapping in
amorphous ice and implications to comet P/Halley. Icarus 122, 118–121.
Owen, T. and A. Bar-Nun 1995a. Comets, impacts and atmospheres. Icarus 116,
Owen, T. and A.
and noble gases.
pp. 123–138 AIP Press, New York (AIP Conf. Proc. No. 341).
Bar-Nun 1995b. Comets, impacts and atmospheres II. Isotopes
In Volatiles in the Earth and Solar System (K. A. Farley, Ed.)
Prinn, R. G. and B. Fegley, Jr. 1981. Kinetic inhibition of CO and N2reduction in
circumplanetary nebulae: Implications for satellite composition. Ap. J. 249, 308–317.
Swings, P. and L. Haser 1956. Atlas of Representative Cometary Spectra. Imprimerie
Ceuterick, Louvain, Belgium.
Swings, P. and T. Page 1950. The spectrum of comet Bester (1947k). Ap. J. 111,
Tull, R. G., P. J. MacQueen, C. Sneden, and D. L. Lambert 1995. The high-
resolution cross-dispersed echelle white-pupil spectrometer of the McDonald Observa-
tory 2.7m telescope. Pub. A.S.P. 107, 251–264.
Warner, B. and G. A. Harding 1963. Spectra of comet Humason (1961e). Observa-
tory 83, 219–223.
Weissman, P. R. 1991. Dynamical history of the Oort cloud. In Comets in the Post-
Halley Era (R. L. Newburn, Jr., M. Neugebauer, and J. Rahe, Eds.) pp. 463–486
Kluwer Academic Press Dordrecht, Netherlands.
Womack, M., S. Wyckoff, and L. M. Ziurys 1992. Observational constraints on
solar nebular nitrogen chemistry: N2/NH3. Ap. J. 401, 728–735.
Wyckoff, S., S. C. Tegler, and L. Engel 1991a. Ammonia abundances in four
comets. Ap. J. 368, 279–286.
Wyckoff, S., S. C. Tegler, and L. Engel 1991b. Nitrogen abundance in comet
Halley. Ap. J. 367, 641–648.
Wyckoff, S. and J. Theobald 1989. Molecular ions in comets. Adv. Space Res. 9,
Table I: Observational Parameters
deVico3 Oct 1995 0.661.00 -3.4 -14.30
4 Oct 19950.66 0.99-2.3 -12.9
Hale-Bopp 6 Apr 1997 0.921.402.9 18.40
Table II: Results
a3910.9–3914.33˚ A bandpass
bCorrected for order sensitivity (see text)
c10 arcsec tailward
d100 arcsec tailward
3.0 × 10−4
9.9 × 10−5
6.5 × 10−5
Table III: Estimates from Earlier Spectra
Mrkos ( 1957V)
aOld designations listed in parentheses besides name
bClaude Arpigny – personal communication
cFrom Marsden and Williams (1995)
1/a (osculating) for Arend-Roland, Bennett, Halley and Wilson
dpg=photgraphic;FTS=Fourier Transform spectrometer; IT=Image tube;
CCD=Charge Coupled Device; Ret=Reticon
Wilson (1987VII)0.8 0.07-0.000260
Figure 1: The spectral region of the CO+(2,0) and CH+(0,0) bands for Hale-Bopp.
The positions of lines within this bandpass are marked, though not all marked lines
are present. The pattern of detected vs. non-detected lines can be explained by the
excitation levels of these molecules.
Figure 2: The spectral region of the CO+(2,0) and CH+(0,0) bands for deVico. The
spectral orders are longer for the R=60,000 mode so more CH+lines were detected. Note
that CH+is stronger relative to CO+in deVico than in Hale-Bopp.
Figure 3: The spectral region of the N+
(0,2,0)–(0,0,0) band is in this spectral region. The expected positions of lines for both of
these bands are marked. There appears to be a feature at 3914.5˚ A, which we tentatively
attribute to C3, not N+
2(0,0) band for Hale-Bopp. In addition, the C3
Figure 4: The spectral region of the N+
there appears to be a feature at 3914.5˚ A. However, we believe this feature is just noise.
2(0,0) band for deVico. As with Hale-Bopp,
Figure 5: An optocenter spectrum of deVico, including the region of the (0,0) band of
2. Several C3bands and the CH B–X (0,0) band are identified. Additional, unidenti-
fied, C3bands are probably present. We show this spectrum to show the abundance of
C3in this comet.
Figure 6: Simulated N+
Bopp. The positions of the potential N+
shows this same spectrum but with a “fake” band replacing the data between 3913 and
3914˚ A. The fake band would have enough counts so that the ratio of N+
equal the value detected for Halley. In the real data, there are some upward excursions
which do not correspond to any N+
2lines. Even integrating just these upward excursions
yields only 1/4 the necessary counts.
2data. The upper panel shows an actual spectrum of Hale-
2lines are marked beneath it. The lower panel
Figure 7: Values for N+
plotted in order to examine whether a trend exists with dynamical age of the comets.
The prediction for this ratio of Owen and Bar-Nun (1985a) is shown as a dotted line.
The open symbols are values derived from estimates of features in photographic and
digital spectra, while the closed symbols represent measured digital spectra. See text for
2/CO+as a function of 1/ao. The various ratios and limits are
4248 4250 4252
456789 10 11
CO+ (2,0) R2 + Q21
CO+ (2,0) Q2 + P21
CO+ (2,0) P2
1112 13 1415 16
CO+ (2,0) R21
P(2) Q(5)R(11) P(3)Q(6)
Figure 1: Cochran, Cochran and Barker
42304235 42404245 42504255
456 7 8 9 10 11
CO+ (2,0) R2 + Q21
01 2 3 4 56
CO+ (2,0) Q2 + P21
CO+ (2,0) P2
111213 14 15 16
CO+ (2,0) R21
R(2,6)R(1,7)R(0)R(8) R(9) Q(1)Q(2) Q(3) R(10)Q(4) P(2) Q(5) R(11)P(3)Q(6)
Figure 2: Cochran, Cochran and Barker
3906 3908 39103912
5432101 2 3 4 5 6
R branch P branch
Figure 3: Cochran, Cochran and Barker
5432101 2 3 4 5 6
R branchP branch
Figure 4: Cochran, Cochran and Barker
3890 3900 391039203930
220.127.116.11 4.5 18.104.22.168 9.5 10.5
0.51.5 2.5 22.214.171.124.57.5 8.5 9.510.5
1.5 2.5 3.54.5
CH B-X (0,0)
Figure 5: Cochran, Cochran and Barker
Actual data (Hale-Bopp) 10 arcsec tailward
R branch P branch
3908 390939103911 3912 39133914 3915 39163917
Synthetic data with N2
+ feature added to match
Figure 6: Cochran, Cochran and Barker
25 Download full-text
-.02 -.010 .01.02
deVico and Hale-Bopp (This Work)
Halley (Wyckoff and Theobald 1989)
Bradfield (Lutz et al. 1993)
Owen and Bar-Nun
Figure 7: Cochran, Cochran and Barker