X-ray Production by V1647 Ori During Optical Outbursts
William K. Teets, David A. Weintraub, Nicolas Grosso, David Principe, Joel H. Kastner, Kenji Hamaguchi, Michael Richmond
aa r X i v : . [ a s t r o - ph . S R ] A ug X-ray Production by V1647 Ori During Optical Outbursts
William K. Teets , David A. Weintraub , Nicolas Grosso , David Principe ,Joel H. Kastner , Kenji Hamaguchi , Michael Richmond ABSTRACT
The pre-main sequence star V1647 Ori has recently undergone two optical/near-infrared (OIR) outbursts that are associated with dramatic enhancements in the stellaraccretion rate. Our intensive X-ray monitoring of this object affords the opportunityto investigate whether and how the intense X-ray emission is related to pre-MS accre-tion activity. Our analysis of all fourteen Chandra X-ray Observatory observations ofV1647 Ori demonstrate that variations in the X-ray luminosity of V1647 Ori are corre-lated with similar changes in the OIR brightness of this source during both (2003–2005and 2008) eruptions, strongly supporting the hypothesis that accretion is the primarygeneration mechanism for the X-ray outbursts. Furthermore, the Chandra monitoringdemonstrates that the X-ray spectral properties of the second eruption were strikinglysimilar to those of the 2003 eruption. We find that X-ray spectra obtained immedi-ately following the second outburst — during which V1647 Ori exhibited high X-rayluminosities, high hardness ratios, and strong X-ray variability — are well modeled asa heavily absorbed (N H ∼ × cm − ), single-component plasma with characteristictemperatures (kT X ∼ Subject headings: stars: formation — stars: individual (V1647 Ori) — stars: pre-main-sequence — X-rays: stars Department of Physics & Astronomy, Vanderbilt University, Nashville, TN 37235, USA Observatoire Astronomique de Strasbourg, Universit´e de Strasbourg, CNRS, UMR 7550, 11 rue de l’Universit´e,F-67000 Strasbourg, France Goddard Space Flight Center, Greenbelt, MD 20771, USA Rochester Institute of Technology, Rochester, NY 14623-5604, USA
1. Introduction
The young stellar object (YSO) V1647 Ori was first noted to be an erupting source when itbrightened suddenly in November 2003, illuminating a new nebula now known as McNeil’s Neb-ula. This deeply embedded, low-mass YSO is typically considered to be an EX Lupi-type object(an “EXor”) though it shares some spectral characteristics with FU Orionis (Vacca et al. 2004),the prototype of a similar class of erupting pre-main sequence (PMS) stars (“FUors”). EXors areobserved to brighten irregularly at optical wavelengths up to several times per decade; these out-bursts persist for weeks to months (Herbig 2001). FUors are YSOs that erupt less often than EXors,perhaps only once per century, and fade much more slowly, i.e., on timescales of years to decades(Herbig 1977). The general consensus is that these eruptions are the result of massive accretionevents that occur irregularly as young protostars grow; such accretion episodes may be the primaryprocess through which young stars accrete most of their mass (Hartmann & Kenyon 1996). Themechanisms underlying EXor and FUor outbursts remain very poorly understood; however, modelsgenerally invoke the rapid onset of disk instabilities that lead to sudden inward migration of theinner disk truncation radius and dramatic changes in the star/disk magnetic field configuration(e.g., Zhu et al. (2009); K¨onigl et al. (2011)).Soft X-ray ( ∼ α and Br γ emission (Reipurth & Aspin 2004; Brittain et al.2010), confirming that the eruptions of V1647 Ori do resemble those of FUors and EXors.V1647 Ori has been observed in X-rays before and during both of the eruptions (as well asafter the first of the two eruptions) detected in the optical/near-infrared since 2003. Althoughother FUor- or EXor-like YSOs have been observed in X-rays (V1118 Ori (Audard et al. 2010);EX Lup (Grosso et al. 2010); Z CMa (Stelzer et al. 2009); FU Ori (Skinner et al. 2006); V1735Cyg (Skinner et al. 2009)), V1647 Ori is the only eruptive YSO to undergo such extensive X-raymonitoring. During the 2003 eruption, the sudden rise in flux and subsequent decline in both the 3 –optical and near-infrared correlated strongly with a sharp increase and then decline in the X-rayflux, which suggests a common origin (Kastner et al. 2004, 2006). In this paper, we present andanalyze X-ray spectra obtained with the Chandra X-ray Observatory (CXO) during the 2008–2009outburst. In §
2, we describe the observations and data reduction. In §
3, we examine the trendsand patterns seen in the X-ray emission from V1647 Ori over the past two outbursts, presentmodeling results for each of the five recent CXO observations, and compare these results to thoseobtained from previous (2003–2005) CXO and XMM-Newton observations. Finally, in §
2. Observations & Data Reduction
Observations of V1647 Ori made with the Chandra X-ray Observatory (CXO) were triggeredin 2008 September with a 20 ks observation (CXO Cycle 10, PI: D. Weintraub, ObsID 9915), afterV1647 Ori was reported to have undergone a new optical outburst between early 2008 Januaryand late 2008 August (Itagaki et al. 2008). Subsequent 20 ks observations were initiated in 2009January and April (ObsIDs 9916 and 9917, PI: D. Weintraub). In addition, in 2008 November,two observations (ObsIDs 10763 and 8585; PI: N. Calvet) were made of the NGC 2068/2071 regionthat serendipitously included V1647 Ori. Together, these five CXO pointings yield an extendedsequence of observations of this source over a seven-month period immediately following the onsetof the optical outburst in 2008.Fortuitously, a field of view that includes V1647 Ori was observed in 2002 November, a full yearbefore the start of the 2003 outburst (Simon et al. 2004); subsequent CXO observations targetingV1647 Ori were obtained in 2004 March and April, 2005 August and December, and 2006 Mayand August (Kastner et al. 2004, 2006). In addition, targeted XMM-Newton observations wereobtained in 2004 April (Grosso et al. 2005) and 2005 March (Grosso 2006), and a targeted Suzakuobservation was obtained in 2008 October (Hamaguchi et al. 2010). The 2004–2006 observationscover a two-year period during which V1647 Ori more or less faded steadily from a strong YSOX-ray source into a very faint one, albeit with large X-ray flux variability. For direct comparisonwith our analysis of the 2008–2009 epoch datasets, we have re-reduced and analyzed all of the2002–2006 CXO datasets, using the same techniques and software packages that we have used forthe latest (2008–2009) datasets.For all Chandra observations, the Advanced CCD Imaging Spectrometer (ACIS) was used inone of two imaging configurations. ACIS detectors have a pixel size of 0.49 ′′ and each ACIS-I andACIS-S CCD has a field of view of 8.3 ′ × ′ . CXO/ACIS has significant sensitivity over theenergy range 0.3–10 keV, with the soft ( < ′′ radius regions (making sure the aperture was positioned so as to encompass as many sourcephotons as possible, even when V1647 Ori is 4 ′ off-axis) while background spectra were extractedfrom regions near but beyond 2.5 ′′ from the target, on the same chip, using 20 ′′ outer radiusextraction apertures. Inspection of background light curves reveals background levels to be fairlyconstant with no evidence of large fluctuations, such as flares, occurring during the observations.Spectra were also re-extracted from previous Chandra observations using the same apertures toensure that direct comparisons of results from different observation epochs could be performedreliably. For observations in which the primary target was not V1647 Ori (ObsIDs 2539, 10763,and 8585), exposure maps were generated to investigate off-axis position effects. In all three of theseobservations, the net count and mean count rate corrections due to off-axis source positions wereinsignificant, on the order of a few percent. The resultant spectral data points of the five recentobservations were grouped into energy bins with a minimum of five counts per bin before spectralmodeling was done. The count rates were high enough and durations long enough for each of the2008–2009 observations that this bin size yielded PI spectra with good statistics. For the 2002–2006observations, we employed single-count-minimum binning for very low-count observations and five-count-minimum binning for higher-count observations so as to yield spectra suitable for modeling.Because no events with energies less than 0.5 keV and few events with energies greater than 8.0keV were detected from V1647 Ori, we limited our modeling to the 0.5–8.0 keV energy range.Since V1647 Ori was imaged with both front- and back-illuminated CCDs in the exposuresequence under analysis, we generated synthetic spectra and convolved these spectra with the in-strument responses to determine whether mean, broadband (0.5–8.0 keV) count rates were directlycomparable for all observations of V1647 Ori. The simulations (Table 2) showed that, for theplasma temperature regimes considered here, the back-illuminated S3 CCD is ∼
10% more sensitiveto incoming flux than the front-illuminated ACIS-I CCDs. Net counts and mean count rates inthe energy range 0.5–8.0 keV (Table 1) were therefore adjusted downward accordingly for the twoobservations (ObsIDs 5307 and 5308) that used the S3 CCD in order to remove this sensitivitybias.
3. Results3.1. Short-Term and Long-Term Variability
We calculated the median photon energies, mean count rates, and mean hardness ratios forall 14 CXO observations of V1647 Ori. These results are reported in Table 1. In calculating thehardness ratios, the hard (H) X-rays are defined as those in the range from 2.8 to 8.0 keV andthe soft (S) X-rays as those with energies from 0.5 to 2.8 keV (Grosso et al. 2005). The hardnessratio is then HR = (H-S)/(H+S), such that negative values of HR indicate softer spectra, andpositive values indicate harder spectra. The recalculated values and error ranges of the mean count 5 –rates and median photon energies for the 2004–2006 observations differ only marginally from thosereported by Kastner et al. (2006). With data from only three epochs in hand, Kastner et al. (2004)reported that the X-ray flux brightened and hardened during the outburst (2002 November 14 to2004 March 07) and then quickly faded and softened post-outburst (2004 March 07 to 2004 March22). With data from additional epochs available, however, Kastner et al. (2006) reinterpreted thelate 2004 March data as a short-term downward fluctuation, not as a quick end to the originaloutburst. They interpret the observed pattern of changes in the median photon energy as evidencethat the X-ray spectrum hardened during outburst, remained somewhat hard for at least one year,and then softened after 2005 August 27 as both the optical and X-ray flux from V1647 Ori returnedto pre-outburst levels. The results obtained here (for hardness ratio) confirm this general trend.This hardening and softening (the mean hardness ratios (with 1 σ uncertainties) changed from − ± ± − ± X-ray light curves and hardness ratios as a function of time were extracted for each observationto examine the properties of any possible short-term X-ray variability. In particular, these timeseries data permit us to investigate whether the trends in X-ray flux and hardness ratio seen in2002–2006 were repeated during the recent outburst that began around 2008 August. The extractedX-ray light curves and hardness ratio time series for six of the nine observations from 2002–2006 arepresented in Figure 1. X-ray light curves and hardness ratio curves for the five 2008–2009 Chandraobservations are presented in Figures 2–6, which show (from top to bottom) the overall X-ray lightcurve, the soft (0.5–2.8 keV) light curve, the hard (2.8–8.0 keV) light curve, and the hardness ratiotime series. Energy ranges for the soft and hard X-ray bands follow those defined in Grosso et al.(2005), and plotted uncertainties in mean hardness ratios and count rates are 1 σ .During all five of the most recent (2008–2009) observations, V1647 Ori experienced very pro-nounced, short-duration ( ∼ ∼ As of submission of this paper, the outburst appears to be ongoing. ∼ ∼
6. In all fiveepochs, the bulk of the X-ray variability is contributed by the harder portions of the spectra.The X-ray count rate for V1647 Ori was intrinsically lower during most of the 2002-2006observations (the count rate was > − only on 2004 March 07) than in 2008–2009 (the countrate was > − during all five observations). Also, in some cases like ObsIDs 5307 and 5308(both from 2004), the observations were on the order of only a few kiloseconds in duration. Thelow count rates in most of the 2002–2006 data and the short exposure times in two of theseobservations make comparisons with the more recent datasets, in terms of the presence or absenceof 2–10 ks timescale variations, fairly difficult. However, observations 5382 and 5383, obtained in2005 April and August, respectively – more than a year after the initial optical outburst of V1647Ori in 2003 November – yielded enough counts that we can compare these two light curves withthe more recent observations with some confidence. As seen in these two light curves from 2005(Fig. 1, lower-left panel, beginning at ∼
10 ks, and lower-middle panel, beginning at ∼ ∼ − ; one month post-outburst) than the highest level (11.7 ks − ; fourmonths post-outburst) observed during its 2003–2005 outburst (Table 1). In addition, with theexception of the 2004 March 07 observation, the 2008–2009 observations (a period spanning sevenmonths) show V1647 Ori to be consistently brighter than it was during the 2003–2005 outburst.We find that in 2008–2009 the X-ray spectrum stays relatively hard, as is shown by the medianphoton energies, which in 2008–2009 are consistently about twice the energy (3.4 to 3.9 keV) thatwas observed before the first eruption (2.0 keV in 2002) and after V1647 Ori had returned to itsquiescent state in late 2005 (1.7 keV). Also, while the mean hardness ratios were negative bothbefore (2002 November) and after (2005 December) the first eruption, the mean hardness ratios in2008–2009 stay consistently around ∼ − , median photon energy of 3.6 keV, mean hardness ratio of +0.46) to whatappears to be a quiescent state on 2004 March 22 (2.5 cts ks − , median photon energy of 2.0 keV, − − , median photon energy of 3.5 keV, mean hardness ratio of +0.51). Yet by 2005August 27, V1647 Ori was again returning to quiescence (0.9 cts ks − , 3.0 keV median photonenergy, +0.01 mean hardness ratio) and by 2005 December was fully back to quiescence.At least three explanations are plausible for the observed changes we have described in theX-ray emission from V1647 Ori that occurred between 2004 March 07 and 2004 March 22 andbetween 2004 March 22 and 2005 April 11. One: the outbursts generate both hard and soft X-rays;most of the time (2004 March 07 and 2005 April 11), the hard X-ray plasma is dominant in theX-ray spectrum as observed by CXO, but at certain times (2004 March 22) the soft X-ray plasmacomponent is seen more clearly. Two: the first major outburst (2003 November through early2004 March) had ended, at least in the X-ray regime, after only a few months. By 2005 April, thesource was again in an elevated state and by 2005 August, four months later, was again returningto quiescence. Three: the 2004 March 22 observation caught V1647 Ori during a short duration(2–10 ks) quiescent moment during the extended (months long) outburst phase. However, theXMM-Newton observation obtained on 2004 April 4 showed a low X-ray flux level at the beginningof the observation that is consistent with the quiescent level observed by CXO that increased inthe second part of the observation (Grosso et al. 2005). Moreover, the supporting evidence fromobservations made at optical and near-infrared wavelengths for the first outburst ending in 2004mid-March and for a second outburst beginning later that year (explanation “Two”) is weak, atbest. Of the remaining two possibilities, we will argue in § ∼ − ∼ − − • In its quiescent state, V1647 Ori has a low count rate ( < − ), low median photon energy( . • In its elevated state, the count rate is at least 25–50 times greater, the median photon energydoubles, and the hardness ratio becomes strongly positive (hard X-ray dominated).The X-ray evidence shows that when V1647 Ori enters a major optical/near-infrared outburst, theX-ray profile switches from quiescent to elevated, and when the outburst phase ends, the X-rayprofile switches from elevated to quiescent.
Modeling of the spectra employed XSPEC v12.4. For those spectra with single-count binning,XSPEC was set to use the Cash-statistic instead of the χ statistic to assess the goodness of fit. Allmodels at first assumed a thin, single-temperature plasma (APEC model) subject to absorptionby an intervening column of hydrogen (WABS component). From the 2004 April XMM-Newtonobservations, single-component thermal plasma models of the V1647 X-ray spectrum yield a hy-drogen column density, chemical abundance, and plasma temperature of N H = 4.1 × cm − , Z= 0.8 solar, and kT X = 4.2 keV, respectively (Grosso et al. 2005), and recent Suzaku observationsyield similar results (Hamaguchi et al. 2010). Therefore, initially, N H , Z, and kT X were set ap-proximately to the aforementioned values (N H = 4.0 × cm − , Z = 0.8 solar, and kT X = 4.0 9 –keV). As our initial test indicated that Z is poorly constrained by the Chandra data, only N H andkT X were left free to vary during the fitting procedure. For four of the five 2008–2009 observations,the models converged to best-fit solutions with column density and plasma temperature within a90% confidence interval of their initial values; however, the 2008 November 27 spectrum model wasunable to converge to a physically-meaningful fit, so the plasma temperature was fixed at kT X =4.0 keV (Table 3).A similar fitting method was used for the 2002–2006 observations (Table 4). Initially, thefirst five of these spectra spectra were modeled with freely varying hydrogen column densities andplasma temperatures, while the chemical abundance was fixed at 0.8 solar. These spectral models,however, were unable to constrain plasma temperatures or X-ray fluxes and luminosities, so thehydrogen column density was then fixed at N H = 4.1 × cm − , and the models were refit. Theremaining four spectra, obtained when V1647 Ori was reverting to an optical/X-ray quiescent state,were modeled with the hydrogen column density and plasma temperature fixed at N H = 4.1 × cm − and kT X = 0.86 keV, respectively. Given the small number of counts in these four spectra,the associated error ranges for these X-ray fluxes and luminosities were obtained by multiplying theerror ranges for the mean count rates of these observations with an appropriate energy conversionfactor (ECF), where the ECF for each observation epoch was obtained by dividing the derivedX-ray flux by the mean count rate.For the CXO observations of V1647 Ori with sufficient total counts (ObsIDs 9915 and 9917),we performed fits of a two-component thermal plasma model with parameters for hydrogen columndensity set to 4.0 × cm − , plasma temperatures set to kT X = 0.5 keV and 2.0 keV, andchemical abundance fixed at 0.8 solar. Visual inspection of the spectral models shows that thereis a negligible difference between the best-fit single- and two-component models, F-test resultssuggest that there is no statistical improvement in the model fits with the addition of a secondplasma component; i.e., the best-fit parameters for the latter model converge on values such thatthe contribution of the lower-temperature component is negligible. We conclude, therefore, that allof the 2008–2009 CXO data are best fit with a single-component model.Best-fit models for each of the five recent Chandra observations are shown in Figure 9. Theoverall trend of the spectral models is to converge to fits with parameters similar to those of thebest-fit single-component model reported by Grosso et al. (2005). Intervening hydrogen columndensities do vary from model to model but remain in the N H ∼ × cm − range, and plasmatemperatures are kT X ∼ Our initial modeling of the V1647 spectra from 2008 and 2009 suggests the presence of lineemission from near-neutral iron at 6.4 keV as well as from the (unresolved) helium-like iron K α -line triplet at 6.64, 6.67, and 6.70 keV. The neutral iron line emission at 6.4 keV is often seen inaccretion-powered sources and is usually attributed to the fluorescence of cold gas in the presence ofa nearby X-ray continuum emission (Tsujimoto et al. 2005). Thus, the presence of this line in someof the V1647 Ori spectra could be attributed to fluorescence of (neutral) circumstellar disk materialby accretion-generated X-rays. It would not be surprising to detect this emission in the spectra ofV1647 Ori given that the environment of this YSO does appear to contain the necessary ingredientsfor the formation of 6.4 keV emission, namely a strong, relatively hard X-ray source illuminatingcold circumstellar material. Observations of other YSOs, including roughly a half-dozen sources inthe COUP survey, with these environmental components have also shown this feature.Following Grosso et al. (2005), we added a Gaussian component centered at 6.4 keV to thespectral models in order to account for the neutral line component. Models of two observations(2008 September 18 and 2009 April 21) appear to be well fit with the addition of a 6.4 keV linewith equivalent widths of ∼
200 and ∼
500 eV, respectively, while the 2008 November 27-28 and2009 January 23 are fit well without the additional neutral iron line emission component (Fig. 10).On the other hand, it is possible that the 6.4 keV emission is present during all of the 2008–2009observations from CXO but that its spectral signature is muffled by noise in the 2008 Novemberand 2009 January observations. While it appears that we have detected 6.4 keV iron emission inthe 2008 September 18 and 2009 April 21 observations, we cannot conclude definitively that these“detections” are real. We note that the best-fit 6.4 keV equivalent widths are poorly constrainedwith error ranges extending from zero to roughly twice the equivalent width values. Given thisrange of uncertainty, we cannot exclude the possibility that there is no 6.4 keV iron line. We feelconfident, however, that these detections are real given that visual and quantitative comparisonsof our findings with those detections found by Tsujimoto et al. (2005) are very similar.We have compared the intensities/appearances of this line in the various observations of V1647Ori by CXO, XMM-Newton, and Suzaku. The equivalent widths of this emission feature are verysimilar to those found in the spectra of the 2004 April 4 XMM observation (109 eV) and in the2008 October 8 Suzaku observation ( ∼
600 eV). We are unable to clearly determine whether thereis a correlation between the strength and appearance of the 6.4 keV line and any of the associated 11 –plasma characteristics.
4. Discussion
Figure 11 suggests that the overall X-ray flux of V1647 Ori is strongly correlated with optical/near-infrared flux. This correlation is revealed more clearly in Figure 12, in which we plot the X-rayluminosity versus the I C -band luminosity. We interpolated I C -band luminosities for the 2002–2005X-ray observation dates. We did not extrapolate I C -band luminosities for the two 2006 X-ray ob-servation dates due to the highly uncertain flux behavior of V1647 Ori. Of the 2008-2009 X-rayobservations, we could only interpolate an I C -band luminosity for ObsID 9916. With these eight in-terpolated luminosities, we derived a correlation coefficient between X-ray luminosity and I C -bandluminosity of 0.65.Dramatic increases in optical/near-infrared flux for YSOs, such as FU Ori, have long beenthought to be associated with enhanced accretion (Hartmann & Kenyon 1996). In such an en-vironment, material is channeled through magnetic funnels from the co-rotation radius of thecircumstellar disk down to the photosphere (Shu et al. 1994). Hence, when the X-ray flux froma pre-main-sequence star or protostar is elevated and the rapid rise in X-ray emission is directlycorrelated with large-scale optical outbursts, the correlation itself strongly suggests that accretionis the mechanism responsible for generating the increase in X-ray output (Kastner et al. 2006).One way in which accretion-generated X-ray emission could be identified observationally wouldbe through the relatively soft X-rays emitted by the plasma when it plunges onto the stellar surfaceat free-fall velocities and is shock heated to temperatures of a few million Kelvin (kT X ∼ X ∼ few to tens of keV) and harderX-rays generated in magnetic reconnection events (Shibata & Yokoyama 2002; Brickhouse et al.2010) within the accretion streams. Whether the soft or hard X-ray generating plasma is observablelikely depends on the observing geometry. If we have an unobscured line of sight to the footprintof the accretion column, our observations should be sensitive to the cooler plasma; if the accretioncolumn obscures our view of the accretion-column footprint, our observations should be sensitiveonly to the hotter plasma in the accretion stream; and if the obscuration of the accretion footprintis partial, we might detect both plasmas. In between periods of dramatically enhanced accretion,previous studies such as COUP (Preibisch et al. 2005; Stassun et al. 2006) and the XMM-NewtonExtended Survey of the Taurus molecular cloud (XEST) (Audard et al. 2007), suggest that theX-ray signature of young stars should be that of normal coronal emission. Such emission wouldbe similar to but much fainter than the hot, hard plasma seen from reconnection events in theaccretion funnel.During quiescence, the X-ray flux of V1647 Ori has a low count rate, low median photon energy,and negative hardness ratio; in contrast, when the X-ray flux is elevated, the count rate increases 12 –by a factor of 25 or greater, the median photon energy doubles, and the hardness ratio becomesstrongly positive. Most of the V1647 Ori light curves reveal that this YSO also experiences whatappear to be short-term (few kilosecond) variability in its X-ray flux. Since most of the CXO observations show this short-duration variability when the X-ray countrate is high, the short-term variations are likely part of the normal behavior for V1647 Ori. Forall 2008–2009 CXO observations, the X-ray variability is almost entirely seen in hard ( > X ∼ cm − would have ∼
98% ofits X-ray flux extinguished. Even given the smallest hydrogen column density found via spectralmodeling of the 2004–2005 V1647 Ori observations when the hydrogen column density was allowedto vary freely ( ∼ × cm − ), such a soft component still has ∼
96% of its X-ray flux absorbed.If the hydrogen column density decreased dramatically — to N H of a few times 10 cm − — theaccretion shock emission could dominate the observed flux. This seems a plausible scenario toexplain the softer plasma detected in late 2004 March, especially if the large intervening hydrogencolumn density inferred at other observing epochs was due mostly to the accretion streams.An alternative explanation is that we observed V1647 Ori when it was in the midst of a largeaccretion episode that pushed the star-disk boundary inward to the point where the accretionbecame non-magnetospheric (Hartmann 1998), effectively reducing the amount of hard X-ray fluxproduced by magnetic reconnection events in the accretion stream.During the 2008–2009 epoch, when the X-ray luminosity of V1647 Ori increased, the spectrumhardened and the emitting plasma increased in temperature (Fig. 7, right panel); also, when theoverall X-ray luminosity decreased, the X-ray spectrum softened and the X-ray generating plasmacooled. These correlations are also seen in the Chandra observations following the 2003 eruption(Fig. 7, left panel). However, between 2008 September and 2008 November, the spectrum appearedto harden slightly as the X-ray luminosity decreased slightly. If we are observing X-rays generatedpredominantly by the ∼ ∼ Our modeling work for the X-ray observations obtained in 2004–2005 and 2008–2009, whenthe mean X-ray count rates were greater than 1 cts ks − yield a best fit value for N H of about 4.1 × cm − , consistent with the results derived by Grosso et al. (2005) and a visual extinction of A V ∼
20 (Vuong et al. 2003). We were unable to fit N H in our modeling work for the low count-rateobservations during the quiescent period in 2005–2006; however, Aspin, Beck, & Reipurth (2008)obtained a best fit value for A V of 19 ±
2, based on their optical, near-infrared, and mid-infraredobservations obtained in February 2007, which was also during the quiescent period (based on theoptical and near-infrared photometry reported by Aspin, Beck, & Reipurth (2008)). The observedvalue of A V for February 2007 lends strong support to our use of N H = 4.1 × cm − for ourmodeling of the quiescent epoch observations. In addition, together these data suggest that N H and, by implication, A V , remained essentially unchanged as inferred from the X-ray observations,whether V1647 Ori was in the quiescent or elevated X-ray state.On the other hand, as seen at longer wavelengths, the extinction toward V1647 Ori haschanged. ´Abrah´am et al. (2004) derive A V = 13 from near-infrared data obtained in 1998 by2MASS. During the outburst in 2004–2005, Brice˜no et al. (2004) found A V = 8–10 on 2004 Febru-ary 18, Reipurth & Aspin (2004) found A J = 1.26, A H = 0.81, and A K = 0.5 on 2004 Feburary18, all of which are consistent with A V of ∼ A V = 11 from measurements of the 3.1 µ m water band on 2004 March 9, and Ojha et al. (2006),who made optical and near-infrared observations from 2004 into very late 2005, reported A V of ∼ A V increased to ∼
10 by the endof 2005, when V1647 Ori had dramatically faded, and Aspin, Beck, & Reipurth (2008) reported A V was as high as 19 by early 2007. Clearly, the extinction, as measured via optical and near-infrared measurements, changed first from quiescence to outburst and then from outburst back toquiescence.These apparently discordant results have a straightforward and consistent explanation in thecontext of an accretion episode. In X-rays, we are essentially detecting V1647 Ori along a direct lineof sight to the stellar photosphere. Our results therefore indicate that the absorption along this di-rect line of sight — which likely includes at least part of a thick circumstellar disk that is tilted about 14 –30 degrees from edge-on (Acosta-Pulido et al. 2007) — does not change significantly as a functionof time, despite the evident changes in X-ray luminosity. The optical and near-infrared photons weobserve, however, emerge from the near-photosphere environment of V1647 Ori along two paths.One path, along our direct line of sight to the photosphere, produces heavily reddened and extinctedlight. The second path takes photons nearly perpendicular to our line of sight, through an evacu-ated polar cavity, where they then scatter into our line of sight (Acosta-Pulido et al. 2007). Thesephotons are bluer and much less heavily extincted than the line-of-sight photons. When V1647 Oriis in the quiescent state (1998, late 2005–2007), we see a faint, reddened, heavily extincted sourcebecause the contribution from scattered light is minimal. During the outburst state (2004–2005;2008–2009), we see a brighter, bluer source because the contribution to the continuum of scatteredphotons is large. Though the plasma temperature of V1647 Ori strongly correlates with the X-ray luminosityand hardness ratio (Fig. 7), it is unclear whether we are observing a single-component plasmathat increases or decreases in temperature and thus causes the observed changes in X-ray lumi-nosity or if a second, lower-temperature plasma is also present and whose contribution to thetotal spectrum is overwhelmed by the hotter temperature plasma during outbursts. Other sources,such as V710 Tau (Shukla et al. 2008) and EX Lupi (Grosso et al. 2010) have spectra that can bemodeled as two-temperature plasmas, with one component fading as the star returns to quiescentlevels. Grosso et al. (2005) was able to model the XMM-Newton observation of V1647 Ori in 2004April with a single-temperature plasma but also found a better fit using a two-component model.However, this observation of V1647 Ori was at least twice as long as most of the Chandra ob-servations from 2004–2009. In addition, XMM-Newton is more sensitive to lower-energy photonsthan CXO/ACIS. Therefore, it may be the case that V1647 Ori does have two plasmas that con-tribute to the X-ray spectra but that ACIS was unable to consistently detect the lower-temperaturecomponent due to the inherit limitations of its design, the shorter observation times of the CXOobservations, and the high degree of softer X-ray absorption by the intervening hydrogen column.In order to test whether CXO would be able to detect two distinct plasma temperatures in 5 ksand 20 ks observations, we simulated a two-component plasma (using fakeit ) with a normalizationratio and plasma temperatures identical to those found by Grosso et al. (2005) and suffering extinc-tion by the same hydrogen column density. The simulated spectra were then convolved with theresponses of the front-illuminated ACIS-I3 and back-illuminated ACIS-S3 CCDs. The simulationsincorporated appropriately-scaled normalization parameters so that the resultant count rates werecomparable to those of the 2005 and 2008–2009 CXO observations. Each of the simulated spectrawas first fit with a single-component model and then with a two-component model. All of the20 ks observations were found to be fit better using two-component models with each fit yieldingvalues for plasma temperatures and normalization that were close to the original simulation param- 15 –eters. F-tests also suggested a slight improvement in the model fits if the 75%/25% normalizationratio between the low/high components, as found by Grosso et al. (2005), was forced. When fitwith a two-component model, the simulated 5 ks observation of the more sensitive ACIS-S3 CCDyielded a best fit (using the 75%/25% normalization ratio) that converged to a model with twonearly-identical temperature plasmas, i.e., to a single-temperature plasma, and thus does not leadto a better fit. Thus, it appears that CXO should have the sensitivity to detect a two-componentplasma such as that found by Grosso et al. (2005) if the observations are long enough and thelower-temperature plasma flux is strong enough compared to the higher-temperature plasma flux.The spectral models for each of the five Chandra observations obtained in 2008–2009 consis-tently converge to fits with a single-temperature plasma that, to within the errors, is ∼ Finally, the second (2008) eruption of V1647 Ori has very similar spectral characteristics tothat of the first (2003–2005) eruption. From Tables 1, 3, and 4, the five 2008–2009 observations ofV1647 Ori, which span a period of approximately seven months, show • mean hardness ratios that are consistently as hard (to within the errors) as the hardest valueof any observation following the previous eruption; • median photon energies at levels that are very similar to the greatest median photon energyof any observation from the previous eruption; • mean count rates that are 20–50 times higher than those observed during X-ray quiescenceand that are typically 1–2 times the highest mean count rate of the 2003–2006 observations; • plasma emission measures that are usually 1–2 times greater than any modeled from the2003-2005 observations; • and X-ray luminosities that are consistently greater than 1.5 × ergs s − , 1–2 times theX-ray luminosity of V1647 Ori during any of the 2003-2005 CXO observations.Given that the spectral characteristics of V1647 Ori are so similar in observations in which theX-ray flux is elevated above the X-ray quiescent level, it appears that the same X-ray generationmechanism was at work during both eruptions and that the plasma characteristics were very similarduring the two eruption epochs. The derived X-ray luminosities suggest that the second eruptionwas more energetic than the first, and given that the modeled emission measures are greater for 16 –the observations following the second eruption, it is reasonable to conclude that we observed thesame phenomenon in both eruptions but that a larger mass of X-ray emitting plasma was activeduring the outburst that began in 2008.
5. Summary
Our X-ray monitoring demonstrates that the two optical/near-IR outbursts undergone by theenigmatic V1647 Ori in less than a decade were accompanied by strong X-ray outbursts. Duringthese two outbursts (2003–2005 and 2008–present), we see that the X-ray flux rose to peak lumi-nosity over a span of a few weeks and then remained elevated for approximately two years duringthe first eruption and for at least one year during the second eruption. Given that there is verystrong evidence that the outbursts observed in the optical and near-infrared regimes are driven byaccretion, we conclude that the correlated outbursts in X-rays are also driven by accretion.We find that all of the CXO spectra of V1647 Ori are best modeled with a single, moderate-temperature (2–6 keV) plasma. In almost all cases, the plasma temperature that emerges frommodels of the CXO spectra is too high to be generated via accretion hotspots on the stellar pho-tosphere but is reasonable for a plasma generated via magnetic reconnection events. However,the X-ray-emitting plasma could also be located in a strongly enhanced stellar corona, or at theinner edge of the circumstellar disk. Given that accretion is ongoing, lower-temperature plasmagenerated by shocks at the accretion footprint is very likely present; however, during these CXOobservations, any such soft component contributed much less flux than the moderate-temperatureplasma and so usually did not leave a distinct signature in the X-ray spectra. With the elevatedhard X-ray flux lasting the duration of the 2008–2009 epochs, we conclude that the X-ray flux isnot the result of typical coronal flares generated via reconnection events. We believe that sincethe optical/near-infrared flux remains elevated throughout this observation epoch, we are insteadobserving the X-rays generated from reconnection events in the accretion stream, with the softerX-ray flux possibly being generated by accretion hotspots at the stellar photosphere.We find no significant change in X-ray absorbing column, indicating that varying optical/IRcolor measurements, which have previously been interpreted as evidence for variable reddeningtoward V1647 Ori, may instead be indicative of varying contributions from scattered vs. directphotospheric emission from the YSO. Two of the spectra obtained during the most recent (2008)outburst appear to show the 6.4 keV neutral iron feature indicating fluorescence from cold (pre-sumably circumstellar disk) gas surrounding V1647 Ori.With V1647 Ori being observed intensely during both outbursts at X-ray, optical, and infraredwavelengths, this objects stands as one of the best characterized systems that exhibits such a closecorrespondence between X-ray output and accretion rate. As a result of intense monitoring atX-ray, optical, and infrared wavelengths during two successive accretion-driven outbursts, V1647Ori stands as the best characterized YSO in terms of the correspondence between X-ray output 17 –and accretion rate. We have shown, furthermore, that this YSO exhibited strikingly similar X-raybehavior and spectral properties during its recent accretion bursts. These results underscore theneed for X-ray monitoring of additional eruptive YSOs, so as to evaluate whether the remarkableconsistency of V1647 Ori is the exception or the norm.We thank Nuria Calvet for providing early access to the data in CXO ObsIDs 10763 and8585. This research was supported via awards numbers GO8-9016X and GO9-0006X to VanderbiltUniversity issued by the Chandra X-ray Observatory Center, which is operated by the SmithsonianAstrophysical Observatory for and on behalf of NASA under contract NAS8-03060.
Facilities:
CXO (ACIS)
REFERENCES ´Abrah´am, P., K´osp´al, A., Csizmadia, S., Mo´or, A., Kun, M., & Stringfellow, G. 2004, A&A, 419,L39Acosta-Pulido, J. A., et al. 2007, AJ, 133, 2020Anthony-Twarog, B.J. The H-beta distance scale for B stars - The Orion association, AJ, 87,1213-1222 (1982)Aspin, C., Beck, T. L., & Reipurth, B. 2008, ApJ, 135, 423Aspin, C., et al. 2009, ApJ, 692, L67Audard, M., Briggs, K. R., Grosso, N., G¨udel, M., Scelsi, L., Bouvier, J., & Telleschi, A. 2007,A&A, 468, 379Audard, M., et al. 2010, A&A, 511, A63Becklin, E. E., Matthews, K., Neugebauer, G., & Willner, S. P. 1978, ApJ, 220, 831Brice˜no, C., et al. 2004, ApJ, 606, L123Brickhouse, N. S., Cranmer, S. R., Dupree, A. K., Luna, G. J. M., & Wolk, S. 2010, ApJ, 710, 1835Brittain, S., Rettig, T., Simon, T., Gibb, E., & Liskowsky, J. 2010, ApJ, 708, 109Fedele, D., van den Ancker, M. E., Petr-Gotzens, M. G., Ageorges, N., & Rafanelli, P. 2007, A&A,472, 199Grosso, N., Kastner, J. H., Ozawa, H., Richmond, M., Simon, T., Weintraub, D. A., Hamaguchi,K., & Frank, A. 2005, A&A, 438, 159 18 –Grosso, N. 2006, in ESA Special Publication, Vol. 604, The X-ray Universe 2005, ed. A. Wilson,51-56Grosso, N., Hamaguchi, K., Kastner, J. H., Richmond, M., & Weintraub, D. A. 2010, A&A, 522,A56G¨udel, M. & Naz´e, Y. 2009, A&A Rev., 17, 309Hamaguchi, K., Grosso, N., Kastner, J. H., Weintraub, D. A., & Richmond, M. 2010, ApJ, 714,L16Hartmann, L., & Kenyon, S. J. 1996, ARA&A, 34, 207Hartmann, L. 1998 Accretion Processes in Star Formation (Cambridge, Cambridge UniversityPress)Herbig, G. H. 1977, ApJ, 217, 693Herbig, G. H., Aspin, C., Gilmore, A. C., Imhoff, C. L., & Jones, A. F. 2001, PASP, 113, 1547Isobe, H., Shibata, K., Yokoyama, T., & Imanishi, K. 2003, PASJ, 55, 967Itagaki, K., Nakano, S., & Yamaoka, H. 2008, IAU Circ., 8968, 2Kastner, J. H., Huenemoerder, D. P., Schulz, N. S., Canizares, C. R., & Weintraub, D. A. 2002,ApJ, 567, 434Kastner, J. H., et al. 2004, Nature, 430, 429Kastner, J. H., et al. 2006, ApJ, 648, L43K¨onigl, A., Romanova, M. M., & Lovelace, R. V. E. 2011, arXiv:1106.2356McGehee, P. M., Smith, J. A., Hendon, A. A., Richmond, M. W., Knapp, G. R., Finkbeiner, D.P., Ivezi´c, ˇZ., & Brinkmann, J. 2004, ApJ, 616, 1058McNeil, J. W., Reipurth, B., & Meech, K. 2004, IAU Circ., 8284, 1Ojha, D. K., et al. 2005, PASJ, 57, 203Ojha, D. K., et al. 2006, MNRAS, 368, 825Ojha, D. K., Ghosh, S. K., Kaurav, S. S., Bhatt, B. C., Sahu, D. K., & Tej, A. 2008, IAU Circ.,9006, 1Porquet, D., Dubau, J., & Grosso, N. 2010, Space Sci. Rev., 157, 103Preibisch, T., et al. 2005, ApJS, 160, 401 19 –Raassen, A. J. J. 2009, A&A, 505, 755Reipurth, B. & Aspin, C. 2004, ApJ, 606, L119Sakamoto, T., Wallace, C. A., Donato, D., Gehrels, N., Okajima, T., & Ukwatta, T. N. 2011, Adv.Space Res., 47, 1444Semkov, E. 2004, Inf. Bull. Variable Stars 5578, 1Semkov, E. 2006, Inf. Bull. Variable Stars 5683, 1Shibata, K. and Yokoyama, T. 2002, ApJ, 577, 422Shu, F. H., Najita, J., Ostriker, E., Wilkin, F., Ruden, S., & Lizano, S. 1994, ApJ, 429, 781Shukla, S. J., Weintraub, D. A., & Kastner, J. H. 2008, ApJ, 683, 893Simon, T., Andrews, S. M., Rayner, J. T., & Drake, S. A. 2004, ApJ, 611, 940Skinner, S., Briggs, K., & G¨udel, M. 2006, ApJ, 643, 995Skinner, S., Sokal, K., G¨udel, M., & Briggs, K. 2009, ApJ, 696, 766Stassun, K. G., van den Berg, M., Feigelson, E., & Flaccomio, E. 2006, ApJ, 649, 914Stelzer, B. & Schmitt, J. H. 2004, A&A, 418, 687Stelzer, B., Hubrig, S., Orlando, S., Micela, G., Mikul´aˇsek, Z., & Sch¨oller, M. 2009, A&A, 499, 529Tsujimoto, M., Feigelson, E. D., Grosso, N., Micela, G., Tsuboi, Y., Favata, F., Shang, H., &Kastner, J. H. 2005, ApJS, 160, 503Vacca, W. D., Cushing, M. C., & Simon, T. 2004, ApJ, 609, L29Venkat, V. & Anandarao, B. G. 2006, IAU Circ., 8694, 2Venkat, V. & Anandarao, B. G. 2009, CBET, 2104Venkat, V. & Anandarao, B. G. 2011, CBET, 2647Vuong, M. H., Montmerle, T., Grosso, N., Feigelson, E. D., Verstraete, L., & Ozawa, H. 2003,A&A, 408, 581Zhu, Z., Hartmann, L., Gammie, C., & McKinney, J. C. 2009, ApJ, 701, 620
This preprint was prepared with the AAS L A TEX macros v5.2.
20 –Table 1. 2002–2009 ACIS Observations.
ObsID Observation JD ACIS Exposure Net Counts Mean Count Median Photon Mean Hardness a Hardness Ratio b Date Chip (ks) (0.5–8.0 keV) Rate (ks − ) Energy (keV) Ratio of Total Counts2539 c ± ± − ± − ± d ± ± ± ± d ± ± − ± − ± ± ± ± ± ± ± ± ± ± ± ± − ± ± ± − ± − ± ± ± ± ± ± ± − ± − ± ± ± ± ± c ± ± ± ± c ± ± ± ± ± ± ± ± ± ± ± ± a Average of hardness ratios computed from 2 ks light curve data bins (ObsID 2539 uses 10 ks light curve data bins). b Hardness ratio computed using the total numbers of hard and soft X-ray photons from the entire observation. c V1647 Ori was not the target of the observation, and given values have been adjusted to compensate for the 4 ′ off-axis position ofV1647 Ori. Exposure maps indicated net counts and mean count rates for ObsID 2539 should be increased by 8% while those for ObsIDs10763 and 8585 required a 3% increase. d The back-illuminated ACIS-S3 CCD is more sensitive to X-rays from plasma in the temperature regime characteristic of V1647 Orithan are the front-illuminated ACIS CCDs. Values displayed for net counts and mean count rates, and their associated errors, for ObsIDs5307 and 5308 have been adjusted downward by 10%, based on spectral simulations in order to make net counts and mean count ratesdirectly comparable between all chips.Note. — All errors are 1 σ . The net counts for each observation are the total number of counts within the 0.5–8.0 keV range. Medianphoton energy uncertainties were calculated via the half-sample method used in Kastner et al. (2006). Mean count rates were determinedby dividing the net counts by exposure times. Uncertainties for mean count rates and hardness ratios of total counts follow Poissonstatistics. The uncertainty for the hardness ratio of the total counts for ObsID 5384 could not be calculated due to the even distributionof the very low number of counts
21 –Table 2. Comparison of Simulated Front- and Back-Illuminated Chip Exposures
Plasma S3/I3 S3 I3 HRTemperature (kev) Count Ratio HR HR Difference2 1.21 0.02 0.12 0.105 1.14 0.30 0.38 0.087 1.11 0.40 0.48 0.08Note. — Net count and hardness ratio (of total counts) compar-isons for simulated spectra convolved with instrument responses fromObsIDs 5307 and 9915, which used the back-illuminated S3 CCD andfront-illuminated I3 CCD, respectively. Each of the 1-ks spectra weresimulated using the fakeit command in XSPEC with the interveninghydrogen column density and chemical abundance fixed at 4.0 × cm − and 0.8 solar, respectively. The front-illuminated chip detectsroughly 85% of the net counts of the back-illuminated chip, with agreater drop in sensitivity for soft versus hard X-rays. Table 3. Model Fits for 2008–2009 Chandra Observations.
ObsID Observation Reduced Degrees of N H kT X EM Observed F X Observed L X Date χ Freedom ( × cm − ) (keV) ( × cm − ) ( × − ergs cm − s − ) ( × ergs s − )9915 2008 Sep 18 0.85 77 4.1 +0 . − . +2 . − . +6 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +2 . − . +1 . − . +9 . − . +0 . − . +0 . − . +1 . − . +7 . − . +3 . − . +0 . − . +0 . − . +1 . − . +2 . − . +4 . − . +0 . − . +0 . − . Note. — Uncertainties given for hydrogen column density, plasma temperature, and chemical abundance correspond to the 90% confidenceintervals, whereas the observed X-ray luminosities are given with their corresponding 68% confidence intervals. Chemical abundance wasfixed at 0.8 solar. For ObsID 10763, in order to constrain the observed X-ray flux and luminosity, the hydrogen column density and plasmatemperature were fixed at the given values. Emission measures and luminosities assume a distance of 400 pc to V1647 Ori.
22 –Table 4. Model Fits for 2002–2006 Chandra Observations.
ObsID Observation Reduced Degrees of kT X EM Observed F X Observed L X Date Statistic Freedom (keV) ( × cm − ) ( × − ergs cm − s − ) ( × ergs s − )2539 2002 Nov 14 1.37 13 0.9 +1 . − . +1 . − . +0 . − . +0 . − . a
10 4.0 (F) 5.4 ± ± ± +1 . − . +6 . − . ± b ± a
14 3.4 +3 . − . +1 . − . +0 . − . +0 . − . +10 . − . +0 . − . +0 . − . +0 . − . +0 . − . ± b ± +0 . − . ± b ± +0 . − . ± b ± +0 . − . ± b ± a Value is the reduced- χ value. b Observed X-ray flux errors and corresponding X-ray luminosity errors were derived by multiplying the mean count rateerrors from Table 1 by an energy conversion factor (ECF).Note. — Uncertainties given for hydrogen column density and plasma temperature correspond to the 90% confidenceintervals, whereas the observed X-ray luminosities are given with their corresponding 68% confidence intervals. All models usea fixed hydrogen column density of N H = 4.1 × cm − and a chemical abundance fixed at 0.8 solar. For ObsID 5307, theplasma temperature was fixed (“F”) at a typical post-outburst plasma temperature of kT X = 4.0 keV in order to constrain theremaining model parameters. For ObsIDs 5384, 6413, 6414, and 6415, the very low number of counts did not permit spectralfitting with freely-varying plasma temperature, so the plasma temperature was fixed at a pre-outburst value as discussed inKastner et al. (2004). Observed fluxes and emission measures were derived from the resulting model fits, an energy conversionfactor (ECF) was calculated for each observation, and the mean count rate uncertainties were multiplied with the ECF tobetter constrain the source fluxes displayed. All models, unless otherwise noted, use the Cash statistic to determine goodness offit, and the reduced statistic is the Cash-statistic or χ value divided by the number of degrees of freedom. Emission measuresand observed luminosities assume a distance of 400 pc to V1647 Ori.
23 –
Elapsed Observation Time (ks) C oun t R a t e ( c t s / s ) H a r dne ss R a t i o ( H − S / H + S ) −2−10 Elapsed Observation Time (ks) C oun t R a t e ( c t s / s ) H a r dne ss R a t i o ( H − S / H + S ) Elapsed Observation Time (ks) C oun t R a t e ( c t s / s ) H a r dne ss R a t i o ( H − S / H + S ) −2−10 Elapsed Observation Time (ks) C oun t R a t e ( c t s / s ) H a r dne ss R a t i o ( H − S / H + S ) −0.500.511.5 Elapsed Observation Time (ks) C oun t R a t e ( c t s / s ) H a r dne ss R a t i o ( H − S / H + S ) −1.5−1−0.500.511.5 Elapsed Observation Time (ks) C oun t R a t e ( c t s / s ) H a r dne ss R a t i o ( H − S / H + S ) −1.5−1−0.500.51 Fig. 1.—: Background-subtracted X-ray light curves for 2002-2006 epoch observations. Time binsare 2 ks for each observation (except 2002 November 14, which uses 10 ks bins due to the very lowcount rate) and contain counts (0.5–8.0 keV) per total time associated with each bin, not the totaltime associated with the observation. Observations for 2005 December 09, 2006 May 01, and 2006August 07 do not detect V1647 Ori, so their light curves are not presented. Uncertainties in meanhardness ratios and count rates are 1 σ . 24 – Elapsed Observation Time (ks) C oun t R a t e ( c t s / s ) C oun t R a t e ( c t s / s ) C oun t R a t e ( c t s / s ) H a r dne ss R a t i o ( H − S / H + S ) −0.500.51 E ne r g y R ange ( . − . k e V ) E ne r g y R ange ( . − . k e V ) E ne r g y R ange ( . − . k e V ) Fig. 2.—: X-ray light curves and hardness ratio of V1647 Ori during observation 9915. The toppanel light curve covers the 0.5–8.0 keV energy range, the second panel down spans the lowerenergy range of 0.5–2.8 keV, and the third panel down covers the higher 2.8–8.0 keV energies.The bottom panel displays the hardness ratio for each of the bins. Counts were binned into 2 ksbins. Uncertainties in mean hardness ratios and count rates are 1 σ . Apparent “X-ray bright” timeintervals are indicated by the shaded regions in the figure (4–8 ks and 16–20 ks) and in Figures3–6. 25 – Elapsed Observation Time (ks) C oun t R a t e ( c t s / s ) C oun t R a t e ( c t s / s ) C oun t R a t e ( c t s / s ) H a r dne ss R a t i o ( H − S / H + S ) −0.500.511.5 E ne r g y R ange ( . − . k e V ) E ne r g y R ange ( . − . k e V ) E ne r g y R ange ( . − . k e V ) Fig. 3.—: X-ray light curves and hardness ratio of V1647 Ori during observation 10763. SeeFigure 2 for description of panels. Counts were binned into 2 ks bins. Uncertainties in meanhardness ratios and count rates are 1 σ . V1647 Ori had a slightly elevated X-ray flux, seen best inthe hard X-ray band, from 8–10 ks. 26 – Elapsed Observation Time (ks) C oun t R a t e ( c t s / s ) C oun t R a t e ( c t s / s ) C oun t R a t e ( c t s / s ) H a r dne ss R a t i o ( H − S / H + S ) −1−0.500.511.5 E ne r g y R ange ( . − . k e V ) E ne r g y R ange ( . − . k e V ) E ne r g y R ange ( . − . k e V ) Fig. 4.—: X-ray light curves and hardness ratio of V1647 Ori during observation 8585. Countswere binned into 2 ks bins. Uncertainties in mean hardness ratios and count rates are 1 σ . Thehard and broad-band X-ray light curves show that the hard X-ray flux from V1647 Ori increasedsharply at 20 ks into the observation and remained elevated for ∼
10 ks. During this period ofelevated X-ray flux, the soft-band flux remained more or less constant. 27 –
Elapsed Observation Time (ks) C oun t R a t e ( c t s / s ) C oun t R a t e ( c t s / s ) C oun t R a t e ( c t s / s ) H a r dne ss R a t i o ( H − S / H + S ) −0.500.51 E ne r g y R ange ( . − . k e V ) E ne r g y R ange ( . − . k e V ) E ne r g y R ange ( . − . k e V ) Fig. 5.—: X-ray light curves and hardness ratio of V1647 Ori during observation 9916. Countswere binned into 2 ks bins. Uncertainties in mean hardness ratios and count rates are 1 σ . TheX-ray light curve shows that V1647 Ori began increasing in hard-band and, possibly, soft-band fluxabout 8 ks into the observation, and the flux remained elevated for ∼
12 ks. 28 –
Elapsed Observation Time (ks) C oun t R a t e ( c t s / s ) C oun t R a t e ( c t s / s ) C oun t R a t e ( c t s / s ) H a r dne ss R a t i o ( H − S / H + S ) −1−0.500.51 E ne r g y R ange ( . − . k e V ) E ne r g y R ange ( . − . k e V ) E ne r g y R ange ( . − . k e V ) Fig. 6.—: X-ray light curves and hardness ratio of V1647 Ori during observation 9917. Countswere binned into 2 ks bins. Uncertainties in mean hardness ratios and count rates are 1 σ . V1647Ori appeared to be in an elevated hard X-ray flux state at the onset of the observation, with theX-ray count rate remaining elevated for ∼ Julian Date − 24500002750 3000 3250 3500 3750 4000 Lu m i no s i t y ( e r g s s − ) H a r dne ss R a t i o ( H − S / H + S ) −1−0.500.51 P l a s m a T e m pe r a t u r e ( k e V ) Julian Date − 24500004700 4750 4800 4850 4900 4950 Lu m i no s i t y ( e r g s s − ) H a r dne ss R a t i o ( H − S / H + S ) −1−0.500.51 P l a s m a T e m pe r a t u r e ( k e V ) Fig. 7.—: Time series of observed X-ray luminosity (top panels), mean hardness ratio (middlepanels), and plasma temperature (bottom panels) for V1647 Ori. Crosses represent data obtainedwith ACIS front-illuminated CCDs, and squares represent data obtained with the ACIS back-illuminated S3 CCD. Plotted uncertainties for the hardness ratios and luminosities represent the68% confidence interval (1 σ ) and plotted uncertainties for the plasma temperatures represent the90% confidence interval (1.6 σ ). The modeled plasma temperatures for the 2005 December through2006 May observations are not shown because this parameter was not well constrained by spectralfitting. 30 – Change in Luminosity (ergs s −1 ) −1e+31 −5e+30 0 5e+30 1e+31 C hange i n M ean H a r dne ss R a t i o −1−0.500.51 Fig. 8.—: Two-parameter plot for all CXO observations of V1647 Ori showing the correlationbetween the changes in mean hardness ratios of individual observations and the changes in observedX-ray luminosities. Crosses represent values corresponding to data obtained with ACIS front-illuminated CCDs, squares represent values involving data obtained with ACIS front-illuminatedCCDs and the back-illuminated S3 CCD, and the triangle represents a value that used back-illuminated S3 CCD data only. The correlation coefficient for the changes in mean hardness ratioand changes in observed X-ray luminosity is 0.44. 31 – −4 −3 no r m a li z ed c oun t s s − k e V − ∆ S χ Energy (keV) −4 −3 no r m a li z ed c oun t s s − k e V − ∆ S χ Energy (keV) −4 −3 no r m a li z ed c oun t s s − k e V − ∆ S χ Energy (keV)
Fig. 9.— 32 – −4 −3 no r m a li z ed c oun t s s − k e V − ∆ S χ Energy (keV) −4 −3 no r m a li z ed c oun t s s − k e V − ∆ S χ Energy (keV)
Fig. 9.—: Best-fit XSPEC models of spectra obtained from 2008 September to 2009 April, withobservations in chronological order from top to bottom. The top frame in each panel displays thedata (binned to five-count-minimum bins) in black overlaid with the model in blue. 33 –Fig. 10.—: Spectrum of the 2009 April 21 observation (black) overlaid with the single-componentplasma model (blue) and the model with the additional Gaussian component (red) centered at 6.4keV (equivalent width of 510 eV) added to account for neutral iron emission. The entire modeledwavelength range (0.5–8.0 keV) is shown in the top panel, and the bottom panel displays the 5.0–8.0keV energy range to more easily show the iron emission and the Gaussian component. Spectraldata points employ a three-count-minimum bin size. 34 –
V1647 Ori X−ray, Optical, and Infrared Variability
Julian Date − 2450000
Log O b s e r v ed F l u x ( e r g s c m − s − ) −10−11−12−13−14−15 Log O b s e r v ed Lu m i no s i t y ( e r g s s − ) Fig. 11.—: Near-infrared and X-ray light curves of V1647 Ori. Chandra X-ray data areshown as crosses; the I c -band data are shown as squares, and the H-band data are shown ascircles. I-band Goddard Robotic Telescope (GRT) (Sakamoto et al. 2011) data (Michael Rich-mond, personal communication) were obtained in 2009. For the published I- and H-banddata, error bar size is on the order of the data point size (data are from: Brice˜no et al.(2004), McGehee et al. (2004), Reipurth & Aspin (2004), Semkov (2004), Ojha et al. (2005),Ojha et al. (2006), Semkov (2006), Venkat & Anandarao (2006), Acosta-Pulido et al. (2007),Fedele et al. (2007), Aspin, Beck, & Reipurth (2008), Ojha et al. (2008), Aspin et al. (2009), andVenkat & Anandarao (2011)). The uncertainties for the GRT data and CXO data are 1 σ . Calendaryear is indicated along the top horizontal axis. 35 – Observed X−ray Luminosity (ergs s −1 ) O b s e r v ed I C − B and Lu m i no s i t y ( e r g s s − ) Fig. 12.—: Two-parameter plot for all CXO observations of V1647 Ori showing the correlationbetween the observed I c -band near-infrared and X-ray luminosities. Crosses represent data obtainedwith ACIS front-illuminated CCDs, and squares represent data obtained with the ACIS back-illuminated S3 CCD. Errors for I c -band luminosities that were interpolated at X-ray observationdates were calculated by averaging the errors of the I c -band luminosities of the five observationsmade nearest to the interpolation date. Circles represent four observations (ObsIDs 9915, 10763,8585, and 9917) in which the interpolated I c -band luminosity of ObsID 9916 was assumed to be thecorrelated I c -band luminosity. The correlation coefficient for the I cc