CHANDRA OBSERVATIONS OF COMETS C/2012 S1 (ISON) AND C/2011 L4 (PanSTARRS)

For both comets selected, the Chandra observations were performed using the Advanced CCD Imaging Spectrometer (ACIS). The comet was centered on the S3 chip as it offers the most sensitive low-energy response in the ACIS array, and ACIS was set to very faint mode for all observations to assist in filtering out bad X-ray events, such as cosmic X-rays, from the source events. Each observation was also performed in drift-scan mode where no active guidance is enabled and Chandra’s pointing was only updated to re-center before the comet moved off the chip.

SW proton velocities were extracted from ACE, a satellite located at the L1 Lagrangian point that continuously records SW conditions. SW speed at each comet was calculated through time of flight corrections between ACE and the comet observations, and the resulting SW velocities were found to be consistent with slow SW. However, we note the large discrepancy in heliospheric latitude between PanSTARRS and ACE during our observations. Since PanSTARRS was at high latitude, we infer that it was bombarded with fast SW (Geiss et al. 1995; Schwadron & Cravens 2000). ISON was observed at similar heliospheric latitude to ACE, so SW conditions should be similar between the two. In addition, solar X-ray activity detected by the GOES X-ray satellite indicates that several M-class solar flare events occurred during ISON's observations, while solar activity was average for PanSTARRS. See Table 1 for additional details on the observation parameters for both comets.

Table 1.  Chandra Comet Observation Parameters

			Texp	rc	Δ	Lat⊙	Long⊙	${Q}_{{{\rm{H}}}_{2}{\rm{O}}}$	vp
Comet	Obs. Date	Prop. Num.	(ks)	(AU)	(AU)	(deg)	(deg)	(1028 mol s−1)	(km s−1)
PanSTARRS	2013 Apr 17–23	14108442	45	1.10	1.44	84.16	150.5	5a	377*
ISON	2013 Oct 31–Nov 6	15100583	36	1.18	0.95	1.130	115.0	2b	313

Notes. Observation parameters are listed as follows: Chandra observation date, observation proposal number, exposure time T_exp, comet-Sun distance r_c, comet-Earth distance Δ, Heliospheric Latitude Lat_⊙ and Longitude Long_⊙, H₂O production rate ${Q}_{{{\rm{H}}}_{2}{\rm{O}}}$, and solar wind proton velocity v_p from the ACE-SWEPAM online data archive. Due to the large difference in heliospheric latitude between ACE and PanSTARRS, it is unlikely they experienced similar SW speeds. More likely, PanSTARRS encountered fast SW due to its high altitude. We denote our uncertainty in the observed SW speed value with an asterisk.

^aCombi et al. (2014a). ^bCombi et al. (2014b).

Download table as:  ASCIITypeset image

Since all observations were performed in drift-scan mode, we first convert all images to object-centered coordinates through use of the sso_freeze routine found in the Chandra Interactive Analysis of Observations (CIAO) software package (Fruscione et al. 2006). We then generate our resulting spectra via CIAO's specextract routine and are combined with the combine_spectra routine. All steps are performed with CIAO v4.7. The cumulative cometary and background X-ray emission spectra are shown in Figure 1.

Zoom In Zoom Out Reset image size

In addition to these observations, ISON was also observed with the High Resolution Camera (HRC) on Chandra between the ACIS visits, also in drift-scan mode. HRC observations of a comet had never been performed in conjunction with ACIS observations before. Such observations were proposed for ISON due to HRC's increased sensitivity to soft X-ray emissions over ACIS, see Figure 2, as it makes the two instruments complementary to one another. The resulting images and our discussion of their implications are in Section 4.1.

Zoom In Zoom Out Reset image size

2.2.1. Background Correction

2.2.2. CX Modeling of Cometary X-Ray Emissions

As discussed in Section 1, a primary goal of our work is to develop a CX model from first principles that can provide more accurate diagnostic of the SW plasma interacting with cometary gas.

We begin by expanding upon the CX model outlined in Kharchenko & Dalgarno (2000), Kharchenko & Dalgarno (2001), Krasnopolsky et al. (2002), and Bodewits et al. (2007). The emitted intensity I of the photon flux induced by CX collisions is defined as the total emission resulting from the interaction between k species of cometary atoms/molecules and SW ions l, where l is dependent on both the element and its charge, within the cometary atmosphere. We define it as an integral over the line of sight distance s and the solid viewing angle Ω_s,

$$\begin{eqnarray}&&I({\rm{\hslash }}{\omega }_{j})=\displaystyle \sum _{k,l}\int {n}_{k}{n}_{l}{\sigma }_{k,l}| {{\boldsymbol{v}}}_{k}-{{\boldsymbol{v}}}_{l}| {P}_{k,l}^{(j)}({\rm{\hslash }}{\omega }_{j}){dsd}{{\rm{\Omega }}}_{s},\end{eqnarray} \tag{ 1 }$$

where n_k is the cometary particle density, n_l is the SW ion density at the comet, ${\sigma }_{k,l}$ is the charge transfer cross section for collisions between k neutrals and l ions, v_k is the cometary particle velocity, v_l is the local SW velocity, and ${P}_{k,l}^{(j)}$ is the photon yield for emissions with the energy ${\rm{\hslash }}{\omega }_{j}$ in the collision between k and l species. The total yield of all X-ray photons is normalized to unity, where ${\displaystyle \sum }_{j}{P}_{k,l}^{(j)}({\rm{\hslash }}{\omega }_{j})=1$, per each unique k and l in order to be valid on a per collisional basis.

For our equation's parameters, both n_k and v_k are found from observational data on the comets (Combi et al. 2014a, 2014b). The physical parameters ${P}_{k,l}^{(j)}$ and ${\sigma }_{k,l}$ are obtained from previous lab and theoretical research (Dijkkamp et al. 1985; Janev & Winter 1985; Johnson & Soff 1985; Kelly 1987; Suraud et al. 1991; Cann & Thakkar 1992; Janev 1995; Wiese et al. 1996; Kharchenko et al. 2003; Koutroumpa et al. 2006, 2009). In regards to n_l and v_l, we set them equal to average values taken from previous analyses of the SW plasma (Bochsler 2007).

We initially find that our modeled spectral intensity does not accurately fit the observational data due to the high variance in SW conditions as a function of both time and solar longitude, causing inaccurate modeled values for n_l and v_l. Since we lack any direct observations of these values at the comet, we therefore allow these parameters to vary within physical limits until the best agreement between our relative modeled intensity and the observational intensity over the 0.3–1.0 keV energy range is found.

2.2.3. CX Model Composition

When analyzing cometary X-ray images, it is important to remember that the overall emission morphology is determined by SW interaction with the cometary atmosphere as the majority of the emitted intensity is due to CX. In the collisionally thick case for an active comet, we expect a paraboloid with the comet at the focus where the magnitude of the semimajor axis is dependent on the atmospheric density (Häberli et al. 1997; Wegmann et al. 2004; Lisse et al. 2005; Wegmann & Dennerl 2005). In comparison, the X-ray emission structure is determined by the distribution of gas in the coma for the collisionally thin case. For such a situation, we expect to see regions of enhanced X-ray emission in regions of higher cometary particle density, such as those found along jet structures (Lisse et al. 2013).

We present the ACIS observation images of comet PanSTARRS in Figure 4. All images have been corrected for differences in exposure time and are normalized to the same linear scale. These images show a constant intensity in X-ray emissions in all the observations, as we expect given the constant cometary dust/gas emission rates and SW conditions observed at the time of our observations. We also find that the overall morphology is highly non-uniform, and so we conclude that PanSTARRS was collisionally thin during its observations. This is likely a result of the high dust density present in the cometary atmosphere as dust particles are significantly less efficient in CX X-ray production than molecular gas (Djurić et al. 2005; Wolk et al. 2009; Lisse et al. 2013).

Zoom In Zoom Out Reset image size

The ISON observations were unique as it was the first time HRC observations of a comet were performed in conjunction with ACIS observations. The extracted ACIS and HRC images for the three ISON observations are shown in Figure 5. Each image set, either from ACIS or HRC, has been corrected for differences in exposure time and has been normalized to the same linear scale. The ACIS images demonstrate the expected paraboloid morphology of a collisionally thick case, while the HRC observations depict a more non-uniform emission typical of a collisionally thin case. Given that HRC is more sensitive to soft X-rays than ACIS, this result may indicate that the soft X-ray emissions are due to CX emissions from lighter SW ions with smaller cross sections, such as He²⁺, than the SW ions that emit hard X-rays, such as C⁶⁺ and O⁸⁺. The reduction in cross section will allow deeper penetration into the cometary atmosphere, and it may be substantial enough to generate difference in these image sets. It is also possible that the soft X-ray emissions from ISON are from a different emission mechanism that would not produce the same morphology, such as scattering or fluorescence.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The ISON image sets also demonstrate significant fluctuation in the cometary emission intensity over time and a “see-saw” in intensity between the soft X-ray HRC observations and the hard X-ray ACIS observations, most notably seen on the October 31 and November 03 visits. These intensity fluctuations correlate with increases in SW speed as documented by ACE, where the maximum SW speed was recorded November 03. Such an association between SW and cometary emission intensity is predicted by our CX model as SW speed fluctuations indicate fluctuations in SW ion freeze-in temperatures (Bochsler 2007). Such temperature changes will shift the SW charge state distribution, producing a varying average cometary emission energy based on our normalized photon emission yield function ${P}_{k,l}^{(j)}({\rm{\hslash }}{\omega }_{j})$. As we see a similar shifting of the average cometary emission energy, we therefore assert that CX emissions are the dominant cometary emission mechanism in the soft X-ray region, a fact that will become important in our discussion in Section 4.4.

Prior to the observation of comet PanSTARRS, there was much speculation if its high dust-to-gas ratio would significantly affect its X-ray CX emission intensity as it is more favorable to produce Auger electrons instead of X-rays when undergoing CX with dust particles (Djurić et al. 2005; Wolk et al. 2009; Lisse et al. 2013). We therefore make sure to note any X-ray spectral irregularities within our results and, if so, their possibility of being due to dust particles.

Utilizing our CX model, we are able to successfully characterize PanSTARRS’ spectrum without making any adjustments to our CX scenario. We find a unique solution for the emission spectrum that fits well to the observations up to 1.0 keV. Above 1.0 keV, the uncertainty in the observations becomes too great to distinguish between noise and emission peaks. Analysis of SW composition through the use of our model shows a lower than average amount of highly charged ions, such as ${{\rm{O}}}^{8+}$ and Ne⁹⁺, with an increase in their lower energy variants, like O⁶⁺ and Ne⁸⁺. This result agrees with our previous assessment that PanSTARRS was observed at fast, polar SW. Beyond this irregularity, PanSTARRS’ spectrum possesses no additional traits that would classify it different from any other comet X-ray spectrum.

Although we cannot infer from our spectral analysis how PanSTARRS’ large dust quantities may have impacted other emissions mechanisms present within the cometary spectra, our results indicate that it had little to no observable impact on the comet's CX X-ray emissions. As such, any differences present are more likely attributed to the SW flux density and ionization state at the time of observation.

Despite being one of the brightest comets in recent years, the Chandra observations we analyzed were taken slightly prior to ISON's drastic increase in gas production rate starting on 2013 November 13. Fluctuations in SW speeds, as confirmed by ACE, and several M-class solar flares, as reported by GOES, were also observed during ISON's Chandra visits. These highly volatile SW conditions may significantly impact ISON's emission spectra.

Using the ACIS observations and applying the same method as done for PanSTARRS, we are able to model ISON's spectrum as CX below 1 keV and extract SW composition ratios. ISON's ratios confirm the above average SW speed with an overabundance of highly charged SW ions, like O⁸⁺ and Ne⁹⁺, that produces a distinct plateau in the spectrum from 0.75–1.00 keV. The O⁷⁺ ratio is twice that seen from PanSTARRS, best visualized via the emission peak at 0.6 keV.

In addition to these results, ISON exhibits some possible peak structures in its emission spectrum at energies above 1 keV: one peak at 1.35 keV and another at 1.85 keV, as seen in Figure 1. Such peaks have been previously seen in Chandra's observations of Comet 153P (Ikeya-Zhang; Ewing et al. 2013), another comet viewed during volatile SW conditions. Our model is presently unable to accurately calculate theoretical CX emissions in this energy range due to the lack of information about the presence of such highly charged ions in the solar wind plasma, but we may comment on the possible emission candidates.

Comparison to atomic emission line tables from NIST indicate that the most probable ions for each emission is Mg xi (1s^{2 1}S–1s2p ${}^{\mathrm{1,3}}$P) for 1.35 keV and either Si xiii (1s^{2 1}S–1s2p ${}^{\mathrm{1,3}}$P) or Mg xii (1 s ²S–4p ²P) for 1.85 keV (Kramida et al. 2014). However, it is unclear if these peaks could be a result from CX as these exotic candidates have not been detected via in situ observations of SW ion composition (von Steiger et al. 2000; Lepri et al. 2013). Furthermore, theoretical models describing the charge abundance of heavy SW ions predict an extremely low probability of finding these ions because of the inability to reach such high freezing-in temperatures in regular SW and coronal mass ejections (Bochsler 2007). On the other hand, these spectral lines are clearly presented in the spectra of the solar X-ray flares as well as in a regular X-ray emission from the Sun (McKenzie et al. 1985; Dere et al. 1997; Landi et al. 2013). It therefore may be possible these peaks are due to a different mechanism whose emissions are increased by solar flare activity, such as scattering of solar X-rays (Krasnopolsky 1997; Snios et al. 2014). At present, we cannot conclude what is the primary source of these exotic emissions detected from ISON as further analysis of ISON's spectrum with a revised CX model, and possibly also a scattering emission model, is required.

While examining the ACIS spectra from ISON, we found a peak-like feature located at 0.2 keV. This feature was also found in the PanSTARRS observations, but with a lower relative spectral intensity. After removing the portion of the spectra caused by the carbon K-shell detector contamination at 0.284 keV, we plot the resulting emission spectra for each of the three ISON visits in Figure 6. Our plots show a soft X-ray region at 0.2 keV that the detector is sensitive to after our corrections, even showing fluctuations that agree with the soft X-ray emission fluctuations detected by HRC (see Figure 5). The overall shape of this feature is likely due to the ACIS effective area function abruptly decaying toward zero at 0.18 keV and is not due to any specific emission line. Also, the fluctuations between the visits exceed the spectral intensity uncertainty in this region, which is 0.08 counts s⁻¹ keV⁻¹, and so we believe these features to be physical.

Zoom In Zoom Out Reset image size

Based on our HRC results from Section 4.1, we assume CX emissions to be the most likely cause of this feature. We therefore extend our CX model down to this soft X-ray region and plot the results with the observational data. See the dotted lines in Figure 7 for our predicted CX model for each observation. Calculation of a unique solution of SW ratios required to produce such intensities is not possible due to the abundance of over 200 unique lines from over 15 different SW ion types that fall within ACIS’ resolution of this soft X-ray feature. We therefore choose to leave our model at average SW abundances in this region. We note that fixing these parameters produces no difference in the spectral fit over the 0.3–1.0 keV energy range.

Zoom In Zoom Out Reset image size

Our results show that our average CX model is not capable of producing the necessary intensity to match the observations for either the ISON emissions or the PanSTARRS emissions, which are not shown. Furthermore, the SW abundances that our model would demand to match these features in intensity far exceed their physical boundaries, with most abundances requiring an increase by an order of magnitude. Although such exotic SW compositions are not impossible given its constantly fluctuating nature, the consistent presence of these soft X-ray features during both fast and slow SW indicate these should be generated under average SW conditions.

Although our current CX model does not agree with the soft X-ray intensities detected, we only consider a single electron capture event per incoming SW ion. Sequential capture events may occur for an ion if the cometary atmosphere is collisionally thick, increasing the amount of soft X-rays emitted from the system as the ion charge state decreases (${{\rm{O}}}^{8+}\;\to$ ${{\rm{O}}}^{7+}\;\to$ ${{\rm{O}}}^{6+}\;\to$ …). We therefore modify our CX model to include these sequential capture events per ion as it may solve our soft X-ray intensity deficit.

For our analysis, we calculate an upper limit on the increase to soft X-ray CX emissions from sequential capture events by assuming all SW ions are neutralized through interaction with the cometary atmosphere. Our results are presented in Figure 7, and they show that the additional CX events are not sufficient to equal the observed soft X-ray intensities. We find that the upper limit of CX emissions only increases the total soft X-ray intensity ∼50%, which is not enough to account for the factors of three to six between the model and the observations. Furthermore, we stress that the actual rate of sequential capture events present in these cometary systems is lower than this upper limit, so the actual emission intensities will reside between our model and the upper limit. We therefore find it unlikely that sequential CX events could account for these soft X-ray features.

As we are confident that our CX model's resulting SW composition and photon yield emission rates are both accurate, we therefore consider two possible explanations for the soft X-ray discrepancy.

• 1.  
  Since the CX model does not match the observational intensities, it is possible we lack the SW ion type required to produce this feature. He²⁺ CX emissions, currently not included in our analysis, would be detectable in this soft X-ray region due to the low resolution of ACIS, and its high abundance may provide the required order-of-magnitude increase in intensity (Kharchenko & Dalgarno 2001; Bodewits et al. 2004). Such SW ions would also have deeper penetration in the cometary atmosphere, which might also explain the collisionally thin appearance of the HRC morphology discussed in Section 4.1. Future iterations of our model should include this ion and compare the modified results to the soft X-ray emissions from ACIS.
• 2.  
  The soft X-ray feature may be a result of previously undocumented detector contamination or degradation that sharply decays below 0.2 keV, producing a peak in observed spectrum. Examination of similar comets observed at different stages of ACIS’ lifetime would show if such a soft X-ray feature is always present, indicating cometary origins, or if this feature has manifested itself over time, indicating a detector issue.

The required analysis for each of these possibilities is beyond the scope of this article, but we believe that any future work on these soft X-ray features from ACIS should provide a thorough analysis of each possible explanation to determine the cause of these unique findings.