1: % ; Time-stamp: <2007-08-31 12:25:53 dph>
2: % ;========================================
3: %%%
4:
5: \documentclass[12pt,preprint2]{aastex}
6: % \documentclass[12pt, preprint]{aastex} % for web submission
7:
8: % -------------------------------------------------------------------------
9: % Personal Definitions
10: % -------------------------------------------------------------------------
11: \newcommand{\ea}{{\it et al.}}
12: \newcommand{\eg}{{\it e.g.}}
13: \newcommand{\ie}{{\it i.e.}}
14: \newcommand{\etc}{{\it etc.}}
15: \newcommand{\rmd}{{\mathrm{d}}}
16: \newcommand{\jcomp}{\rmfamily{J. Comp. Phys.}}
17:
18: % aastex defines \sun for \msol to give $M_{\sun}$ Slightly larger sun sign.
19: % \newcommand{\msol}{\rm{M}$_\odot$}
20: \newcommand{\mdot}{\dot{\mathrm{M}}}
21: \newcommand{\msol}{\mathrm{M}_\odot}
22: \newcommand{\msolyr}{\msol~\textrm{yr}^{-1}}
23: \newcommand{\kms}{km~$\rm{s}^{-1}$}
24: \newcommand{\kmsbig}{$^{\textrm{km}}/_{\textrm{s}}$}
25: \newcommand{\kmssmall}{$^{\mathrm{km}}/_{\mathrm{s}}$}
26: \newcommand{\cc}{$\rm{cm}^{-3}$}
27: \newcommand{\arcs}{\char'175}
28: \newcommand{\sfrac}[2]{\,{}^{#1}\!/_{#2}}
29: \newcommand{\beq}{\begin{equation}}
30: \newcommand{\eeq}{\end{equation}}
31: \newcommand{\bdm}{\begin{displaymath}}
32: \newcommand{\edm}{\end{displaymath}}
33:
34: % some local useful abbreviations
35:
36: \newcommand{\aped} {APED}
37: \newcommand{\asca} {{\it ASCA}}
38: \newcommand{\chan} {{\it Chandra}}
39: \newcommand{\xmm} {{\it XMM-Newton}}
40: \newcommand{\rosat} {{\it ROSAT}}
41: \newcommand{\cxo} {CXO}
42: \newcommand{\ciao} {CIAO}
43: \newcommand{\dem} {\mbox{$DEM$}}
44: \newcommand{\emd} {\mbox{$EMD$}}
45: \newcommand{\heg} {HEG}
46: \newcommand{\rgs} {RGS}
47: \newcommand{\hetgs} {HETGS}
48: \newcommand{\letgs} {LETGS}
49: \newcommand{\meg} {MEG}
50:
51: \newcounter{ion}
52: \newcommand{\eli}[2]{\setcounter{ion}{#2}#1{~\sc\roman{ion}}}
53:
54: %%% exponent abbrevs for units:
55:
56: \newcommand{\mone}{^{-1}}
57: \newcommand{\mtwo}{^{-2}}
58: \newcommand{\mthree}{^{-3}}
59: \newcommand{\pone}{^1}
60: \newcommand{\ptwo}{^2}
61: \newcommand{\pthree}{^3}
62: % \newcommand{\kms}{\mathrm{\,km\,s\mone}}
63:
64:
65: %\slugcomment{Accepted for publication in the
66: % Astrophysical Journal (December 20, 2007, v671).}
67:
68: \begin{document}
69:
70: \title{
71: Evidence for Accretion in the High-resolution X-ray Spectrum of
72: the T Tauri Star System Hen~3-600}
73:
74:
75: \author{
76: David P. Huenemoerder\altaffilmark{1},
77: Joel H. Kastner\altaffilmark{2},
78: Paola Testa\altaffilmark{1},
79: Norbert S. Schulz\altaffilmark{1},
80: David A. Weintraub\altaffilmark{3}
81: %
82: {\newline\smallskip{\small
83: (Accepted for publication in the
84: Astrophysical Journal, December 20, 2007, v671.)}}
85: }
86:
87: \altaffiltext{1}{Massachusetts Institute of Technology,
88: Kavli Institute for Astrophysics and Space
89: Research, 70 Vassar
90: St., Cambridge, MA, 02139}
91: \altaffiltext{2}{Chester F. Carlson Center for
92: Imaging Science, Rochester Institute of Technology,
93: Rochester, NY 14623; jhk@cis.rit.edu}
94: \altaffiltext{3}{Dept.\ of Physics and Astronomy, Vanderbilt University,
95: Nashville, TN, 37235}
96:
97:
98:
99: \begin{abstract}
100:
101: We present high-resolution X-ray spectra of the nearby, multiple T
102: Tauri star (TTS) system Hen~3-600, obtained with the High Energy
103: Transmission Grating Spectrograph (\hetgs) aboard the \chan\ X-ray
104: Observatory (\cxo). Both principle binary components of Hen~3-600
105: ($A$ and $B$, separation $1.4''$) were detected in the zeroth-order
106: \cxo/\hetgs\ X-ray image. Hen~3-600-$A$ --- the component with a
107: large mid-infrared excess --- is a factor $\sim2-3$ fainter in
108: X-rays than Hen~3-600-$B$, due to a large flare at Hen~3-600-$B$
109: during our observation. The dispersed X-ray spectra of the two
110: primary components overlap significantly, so spectral analysis was
111: performed primarily on the first-order spectrum of the combined
112: $(A+B)$ system, with analysis of the individual dispersed spectra
113: limited to regions where the contributions of $A$ and $B$ can be
114: disentangled via cross-dispersion profile fitting. This analysis
115: results in two lines of evidence indicating that the X-ray emission
116: from Hen~3-600 $A+B$ is derived, in part, from accretion processes:
117: (1) The line ratios of He-like \eli{O}{7} in the spectrum of
118: Hen~3-600 $A+B$ indicate that the characteristic density of its
119: X-ray-emitting plasma is significantly larger than those of
120: coronally active main sequence and pre-main sequence stars. (2) A
121: significant component of low-temperature (2--3 MK) plasma is present
122: in the Hen~3-600 $A+B$ spectrum; this ``soft excess'' appears
123: somewhat stronger in component $A$. These results for Hen~3-600
124: $A+B$ are consistent with, though less pronounced than, results
125: obtained from X-ray gratings spectroscopy of more rapidly accreting
126: TTS systems. Indeed, all of the emission signatures of Hen~3-600
127: $A+B$ that are potential diagnostics of accretion activity --- from
128: its high-resolution X-ray spectrum, through its UV excess and
129: H$\alpha$ emission line strengths, to its weak near-infrared excess
130: --- suggest that its components (and component $A$ in particular)
131: represent a transition phase between rapidly accreting, classical T
132: Tauri stars and non-accreting, weak-lined T Tauri stars.
133:
134: \end{abstract}
135:
136: \keywords{Xrays: stars --- stars: individual (Hen~3-600, TW~Hya,
137: HD~98800) --- stars: pre-main sequence ---
138: stars: coronae --- accretion, accretion disks}
139:
140:
141:
142: \section{Introduction}\label{sec:intro}
143:
144: X-ray observations of star formation regions have established that
145: low-mass, pre-main sequence (T Tauri) stars are luminous X-ray sources
146: \citep{Feigelson:Montmerle:1999}. Such high-energy radiation may exert
147: a profound influence over the process of planet formation
148: \citep[e.g.][]{Glassgold:al:2004, Feigelson:Getman:al:2005}. The
149: origin of pre-main sequence (pre-MS) X-ray emission is still the
150: subject of debate, however. Although most recently formed, low-mass
151: stars likely emit X-rays as a consequence of solar-like coronal
152: activity
153: \citep{Kastner:Huenemoerder:al:2004,Preibisch:Kim:al:2005,Telleschi:Gudel:Briggs:Audard:Palla:2007}
154: we and others have suggested that the X-ray emission from certain
155: actively accreting pre-main sequence stars may be a direct result of
156: mass accretion from a circumstellar disk onto the forming star
157: \citep{Kastner:02,Kastner:Richmond:al:2004,Kastner:Richmond:al:2006,
158: Stelzer:Schmitt:2004,Schmitt:Robrade:al:2005,Gunther:Liefke:al:2006,
159: Argiroffi:Maggio:Peres:2007,Telleschi:Gudel:Briggs:Audard:Scelsi:2007}.
160:
161: To improve our understanding of the physical conditions within X-ray
162: emitting regions of both actively accreting, classical T Tauri stars
163: (cTTS) and apparently non-accreting, weak-lined T Tauri stars (wTTS),
164: we are conducting a high-resolution X-ray spectroscopic study of the
165: TW~Hya Association \citep[TWA;][]{Kastner:Zuckerman:1997,
166: Webb:Zuckerman:al:1999, Zuckerman:Webb:2001,
167: Song:Zuckerman:Bessell:2004, Song:Zuckerman:Bessell:2003} with the
168: \chan\ X-ray Observatory's High Energy Transmission Grating
169: Spectrometer \citep[\hetgs;][]{HETG:2005}. The TWA is uniquely well
170: suited to this study, due to its proximity (typical member distances
171: $D\sim50$ pc) and lack of intervening cloud material, as well as its
172: status as an evolutionary ``missing link'' between embedded pre-MS
173: stars and the zero-age main sequence
174: (\citet{Kastner:Crigger:al:2003,delaReza:al:2006}; also see the review
175: by \citet{Zuckerman:Song:2004}).
176:
177:
178: Observations of TWA members with \hetgs\ provide unique plasma
179: diagnostics that have generated insight into the X-ray emission
180: mechanisms of T Tauri stars. In particular, the density-sensitive line
181: ratios of He-like ions (\eli{O}{7}, \eli{Ne}{9}) detected in the
182: \chan/\hetgs\ X-ray spectrum of TW~Hya indicated plasma densities in
183: the range $11.9 < \log n_e < 12.5$
184: \citep{Kastner:02,Testa:Drake:al:2004b}. This is roughly an order of
185: magnitude larger than all active main sequence stars that have been
186: observed at high spectral resolution
187: \citep{Testa:Drake:al:2004b,Ness:gudel:al:2004}. Similar results for
188: density-sensitive line ratios subsequently have been obtained from
189: X-ray gratings spectroscopy of the cTTS BP Tau
190: \citep{Schmitt:Robrade:al:2005}, V4046 Sgr
191: \citep{Gunther:Liefke:al:2006}, and MP Mus
192: \citep{Argiroffi:Maggio:Peres:2007}. Furthermore, the plasma
193: temperature measured from \hetgs\ and \xmm\ spectra of TW~Hya, $\sim3$
194: MK, is consistent with adiabatic shocks arising in gas at free-fall
195: velocities expected for accretion onto the star \citep{Kastner:02,
196: Stelzer:Schmitt:2004}.
197:
198: In stark contrast, the \hetgs\ spectrum of the wTTS system HD~98800, a
199: member of the TWA and which is apparently non-accreting, closely
200: resembles those of active main sequence stars
201: \citep{Kastner:Huenemoerder:al:2004}. In particular, its He-like
202: \eli{Ne}{9} and \eli{O}{7} triplet line ratios and, hence, plasma
203: densities are well within the range characteristic of ``classical''
204: coronal sources, and its plasma temperature distribution is
205: significantly hotter than that of TW~Hya.
206:
207: Typical signatures of active accretion are strong H$\alpha$ emission,
208: ultraviolet excess from an accretion disk boundary layer, and a
209: near-infrared excess from hot dust in the inner disk. To fully
210: understand the dependence of high energy emission on accretion rate --
211: if any -- we need high resolution X-ray spectra from objects which
212: span a large range in the optical, UV, and IR diagnostics. Here we
213: describe the results of \hetgs\ observations of TWA member Hen~3-600,
214: a multiple pre-MS star system whose principle components are separated
215: by $1.4''$ \citep{Webb:Zuckerman:al:1999,
216: Jayawardhana:Hartmann:al:1999}. Like HD~98800, one primary
217: component of Hen~3-600 evidently is surrounded by a dusty disk, based
218: on its IR spectral distribution \citep{Jayawardhana:Hartmann:al:1999}.
219: The mid-infrared spectrum of Hen~3-600, like that of TW~Hya, displays
220: evidence for crystalline silicate dust grains
221: \citep{Honda:Kataza:al:2003, Uchida:Calvet:al:2004}, suggesting
222: material in the circumstellar disk of Hen~3-600-$A$ has undergone
223: considerable processing and evolution.
224:
225: The presence and strength of H$\alpha$ emission and UV excesses in the
226: spectra of TW~Hya and Hen~3-600-$A$ (Table~\ref{tbl:stars}) indicate
227: that accretion is ongoing in each case. Indeed, apart from TW~Hya
228: itself, Hen~3-600-$A$ is the strongest H$\alpha$ emission source in
229: the TWA \citep{Webb:Zuckerman:al:1999}. Based on H$\alpha$ line
230: profile analysis, both stars are accreting at rates significantly
231: lower than those of $\sim1$ Myr-old TTS that are still found in their
232: nascent dark clouds, with the accretion rate of Hen~3-600-$A$ about an
233: order of magnitude lower than that of TW~Hya
234: \citep{Muzerolle:Calvet:2000}. This is consistent with the relatively
235: small UV excess of Hen~3-600, as UV excess is also widely used as an
236: indicator of TTS accretion rate. TW~Hya is known to have a nearly
237: face-on disk \citep{Krist:Stapelfeldt:2000,Kastner:Zuckerman:1997} and
238: is thus ideal for probing accretion since there is no intervening
239: local X-ray absorption by neutral hydrogen, and because the
240: high-latitude accretion region (if directed by bipolar magnetic
241: fields) is in full view. We have no direct measurement of the
242: inclination of the Hen~3-600 components to our line-of-sight, but we
243: infer that it is also low, given its combination of large mid-IR
244: excess and negligible optical reddening; its measured value of $(B-V)
245: = 1.52$ is nearly that of an unreddened M3 photosphere
246: \citep{Johnson:1966}. The contrasts in accretion diagnostics apparent
247: in Table~\ref{tbl:stars}, and the status of Hen~3-600 as an
248: intermediate case between TW~Hya and HD~98800, make Hen~3-600 a key
249: TWA system for further study with \chan/\hetgs. Both components, $A
250: \& B$, are considered members of the TWA based on photometric,
251: spectroscopic, and kinematic criteria \citep{Torres:Guenther:al:2003,
252: Reid:2003, Webb:Zuckerman:al:1999}.
253: %
254:
255:
256: \section{Observations, Data Reduction, and Results}\label{sec:obs}
257:
258: We observed Hen~3-600 with \chan/\hetgs\ for 101 ks starting on 2004 March
259: 16 (observation identifier 4502) in the default configuration
260: (timed exposure, ACIS-S detector array) and under nominal operating
261: conditions. Data were re-processed with \chan\ Interactive Analysis
262: of Observations software (CIAO, version 3) to apply updated
263: calibrations (CALDB 3.2). Events were also cleaned of the detector
264: artifacts on CCD 8 (``streaks''). The effective exposure time after
265: data reprocessing is 99.4 ks.
266: Spectral responses, including corrections
267: for ACIS contamination, were generated with CIAO.
268:
269: The resulting \chan/\hetgs\ spectral image of the Hen~3-600-$A/B$ binary
270: yielded 883 and 1610 zeroth order counts from $A$ and $B$,
271: respectively, and 3420 combined (\meg\ and \heg) first order counts.
272: The dispersed spectra are not well resolved spatially over the entire
273: spectral range (see \S~\ref{sec:xdisp}). To estimate count rates for
274: the individual components of the Hen~3-600 binary, we extracted
275: separate dispersed spectra within cross-dispersion regions centered on
276: $A$ and $B$, with the corresponding zeroth-order centroids as the
277: wavelength reference. We thus estimate combined \meg\ and \heg\ plus
278: and minus first order count rates of 0.010 s$^{-1}$ and 0.024 s$^{-1}$
279: for components $A$ and $B$, respectively; these estimates are subject
280: to some uncertainty, however, given the overlap between the dispersed
281: spectra of $A$ and $B$ in the cross-dispersion direction.
282:
283: Line positions and fluxes were measured with
284: ISIS\footnote{\raggedright ISIS is available from {\tt
285: http://space.mit.edu/CXC/ISIS}} \citep{Houck:00,Houck:2002} by
286: convolving Gaussian profiles with the instrumental response.
287: Determination of an accurate continuum is important. Even at \hetgs\
288: resolution, there are portions of the spectrum for which the local
289: minima are a poor approximation to the true continuum. Hence, we
290: adopt an iterative procedure whereby we first use a low order
291: polynomial to fit the local continuum knowing it to be in error in
292: some spectral regions. After we obtain a plasma model from line
293: fluxes, we evaluate a global continuum and re-fit the lines and
294: continue to iterate in this fashion if required.
295:
296: % \section{Results}\label{sec:results}
297:
298: \subsection{Zeroth-order image, light curves, and CCD spectra}\label{sec:zodata}
299:
300: %\clearpage
301: \begin{figure*}[t]
302: \epsscale{0.7}
303: % \epsscale{0.3}
304: \centerline{\plotone{f1a.eps}\plotone{f1b.eps}\plotone{f1c.eps}}
305: \caption
306: { \small The zeroth order data for Hen~3-600. The left panel shows the
307: central field in sky pixel coordinates as contours with individual
308: events over-plotted as points. The solid-outlined circles are the
309: 0.7'' (1.4 pixel) radius extraction regions for zero order light
310: curves and spectra. The outer contour is for 1 count, and the
311: interval is 10 counts, with the maximum counts in the brighter
312: source being about 180. The upper left source is component
313: $A$. The middle two-panel plot shows the zero order light curves
314: for $A$ (upper) and $B$ (lower) in 1 ks bins, for the CCD {\tt
315: ENERGY} range of 0.25--7.0 keV; backgrounds are
316: negligible. The right plot shows the zero order CCD spectra, with
317: the brighter source $B$ shaded in black, and source $A$, which is
318: fainter and somewhat softer, in gray. CCD photon pileup was not
319: significant at these low rates.
320: \label{fig:zo}
321: }
322: \end{figure*}
323: %
324:
325: \begin{figure*}[t]
326: \epsscale{2.33}
327: % \epsscale{1}
328: \centerline{\plottwo{f2a.eps}{f2b.eps}}
329: \caption{\small Light curves (top panels) and hardness ratios (bottom
330: panels) for components $A$ (left) and $B$ (right) obtained from
331: the dispersed $\meg+\heg$ spectral events for positive and
332: negative orders 1--3. In the top panels, the lighter (gray)
333: histograms indicate the 1.7--10 \AA\ (``hard'') band, and the
334: darker (black) histograms indicate the 10--26 \AA\ (``soft'')
335: spectral region. The bottom panels display the time
336: history of hardness
337: ratios constructed from the counts in these ``soft'' and
338: ``hard'' bands. Since the dispersed spectra of $A$ and $B$ are
339: not fully resolved, these curves were extracted from narrow
340: regions centered on the component's spectrum; each is contaminated
341: to some degree by the other source.
342: \label{fig:tglc}
343: }
344: \end{figure*}
345: %\clearpage
346:
347: In Figure~\ref{fig:zo} we display the zeroth-order \hetgs\ image,
348: light curves, and CCD energy spectra of the Hen~3-600 system. Both
349: primary components were detected, with separation ($1.4''$) and
350: orientation in excellent agreement with earlier infrared imaging
351: \citep{Jayawardhana:Hartmann:al:1999}. Hence, there is no ambiguity
352: in attributing the X-ray emission to components $A$ and $B$. However,
353: the X-ray flux ratio of the $AB$ pair is the reverse of that in the
354: optical and infrared. In those wavelength regimes, component $A$ is
355: brighter --- indeed, its infrared excess likely dominates that of the
356: Hen~3-600 system \citep{Jayawardhana:Hartmann:al:1999}. During our
357: observations, however, $B$ was the brighter X-ray source.
358:
359: The light curve shown in Figure~\ref{fig:zo} reveals that this X-ray
360: mean flux difference in Hen~3-600 is due to a large flare
361: that occurred on component $B$, $\sim16$ ks after the start of the
362: observation. This flare displayed a steep ($<2$ ks) rise and large
363: amplitude, achieving a count rate a factor $\sim5$ above pre-flare
364: rate. The flare decays abruptly over $\sim10$ ks, but did not return
365: to the pre-flare level, but to one about two times higher.
366:
367: The count rate of component $A$ remained at a low level with small
368: fluctuations throughout our observation. It was actually the brighter
369: of the two main components during the initial $\sim15$ ks of the
370: \hetgs\ observation. A similar reversal in X-ray to optical and IR
371: occurred in HD~98800 in which the component surrounded by a dusty
372: disk, defined by presence of a strong IR excess, was also the weaker
373: and steadier X-ray source of the system
374: \citep{Kastner:Huenemoerder:al:2004}.
375:
376: Component $B$ appears to be the harder of the two components in the
377: zeroth-order CCD spectra, which are also shown in Figure~\ref{fig:zo}.
378: In particular, the zeroth-order spectral energy distributions of $A$
379: and $B$ peak near $\sim0.9$ keV, but component $A$ drops off more
380: rapidly at higher energies. Component $B$ has a weak but significant
381: Fe K$\alpha$ line at $\sim6.7$ keV.
382:
383: There is little evidence that the spectrum of $B$ hardened during the
384: flare. The hardness increased only marginally from 15--40 ks from
385: observation start (Figure~\ref{fig:tglc}). This behavior appears
386: distinct from that of coronal sources, which usually show much more
387: obvious hardening during flares, as in the case of HD~98800
388: \citep{Kastner:Huenemoerder:al:2004} and other coronally active stars
389: \citep[e.g.][]{Gudel:Audard:al:2004, Huenemoerder:Testa:al:2006}. The
390: hardnesses of $A$ and $B$ outside the flare (excluding the 15--40 ks
391: interval) are similar. This suggests that excluding the flare, that
392: the spectral energy distributions are similar. High resolution
393: spectral diagnostics (discussed later) reveal some differences between
394: $A$ and $B$.
395:
396: We did not use the zero order for detailed spectral modeling, but
397: deferred to the high resolution spectrum. We did evaluate the high
398: resolution-derived model for the sum of $A+B$ with the zero order
399: response to examine consistency with the sum of the zero order
400: spectra, but calibration systematic residuals make comparison
401: difficult. We thus only used the zero orders for variability and
402: qualitative assessment of the individual components.
403:
404:
405: \subsection{First Order Spectrum}\label{sec:tgspec}
406:
407: %\clearpage
408: \begin{figure*}[t]
409: \centerline{\includegraphics[scale=1.0]{f3.eps}}
410: \caption{\small The combined $\heg + \meg$ spectrum of $A+B$, heavily
411: binned ($0.08$ \AA/bin). Some prominent or important
412: lines are marked.
413: \label{fig:tgspec} }
414: \end{figure*}
415: %\clearpage
416: %
417: The combined, first-order MEG and HEG spectrum of Hen~3-600 is
418: displayed in Figure~\ref{fig:tgspec}.
419: As is also the case in HETGS spectra of fellow TWA members TW~Hya and
420: HD~98800 \citep{Kastner:02,Kastner:Huenemoerder:al:2004}, the most
421: prominent lines in the HETGS spectrum of Hen~3-600 are those of highly
422: ionized Ne and O, while lines of Fe are relatively weak. We have
423: measured line fluxes by fitting the combined $A+B$ counts
424: (Table~\ref{tbl:lfluxemd}); for stronger lines we have determined the
425: relative contribution of each component (Table~\ref{tbl:abfluxes}).
426: In the following sections, we describe these determinations in more
427: detail.
428:
429: %% \clearpage
430: \section{Spectral Analysis}
431:
432: \subsection{Emission measure distribution}\label{sec:emd}
433:
434: To ascertain the ranges of temperature and elemental abundances that
435: characterize the X-ray-emitting plasma of the Hen~3-600 system, we fit
436: its combined ($A+B$) dispersed HETGS spectrum with an emission measure
437: distribution ($EMD$) model. The $EMD$ is a one-dimensional
438: characterization of a multi-thermal plasma. It is a weighting
439: function which relates the line emissivities ($\epsilon$) to flux
440: ($f$) via an integral equation,
441: %
442: $$ f_{jZ} = \frac{A_Z}{4\pi d^2} \int{ \epsilon_{jZ}(T)\, [ n_e n_H
443: \frac{dV}{dT} ]\, dT }$$ \label{eq:emd}
444: %
445: for ion $j$ of element atomic number $Z$ and abundance
446: $A_Z$; $T$ is the plasma (electron) temperature, $V$ the
447: volume, $n_e$ and $n_H$ are the electron and hydrogen
448: densities, and $d$ is source distance. The quantity in
449: square brackets is the $EMD$ (also commonly called the
450: differential emission measure [$DEM$]). We have ignored any
451: density dependence on $\epsilon$, which in principle exists
452: and which would introduce another differential term into the
453: equation. The few strongly density-sensitive lines of the
454: He-like triplets are handled independently. We also assume
455: uniform abundances, such that $A_Z$ appears outside the
456: integral.
457:
458: We applied methods similar to those described in
459: \citet{Huenemoerder:Testa:al:2006} to perform a regularized fit of the
460: emission measure and abundances to line fluxes. There are many
461: caveats to emission measure modeling \citep[see][for discussion and
462: further citations]{Huenemoerder:Testa:al:2006}. Here we have the
463: added complication that components $A$ and $B$ are not fully resolved,
464: so we must fit their summed emission. The results for the $EMD$ of
465: Hen~3-600 hence implicitly rely on the assumption that the abundances
466: are homogeneous and identical in each component. Lines used in the
467: $EMD$ fitting are listed in Table~\ref{tbl:lfluxemd}. Those lines
468: with measured flux but which are known to be density sensitive
469: (He-like triplet intercombination and forbidden lines) or which had
470: large residuals (due to some systematic error, such as
471: misidentification) have their predicted flux enclosed in parentheses.
472:
473: Figure~\ref{fig:emd} shows the resulting $EMD$ for different assumed
474: values of the absorbing column ($N_H$). The upper and lower bounds of
475: the shaded regions result from 100 Monte-Carlo iterations in which the
476: line fluxes were randomly perturbed within their measured
477: uncertainties. Regardless of the assumed value of $N_H$, we find a
478: broad $T$ distribution with strong peaks at $\log T = 6.5$ and $\log T
479: = 7.0$, and a ``hot tail'' above $\log T = 7.4$. The longest
480: wavelength (and characteristically cooler) lines are most sensitive to
481: the assumed $N_H$. Because there are few counts in this
482: long-wavelength regime, we obtain effectively identical solutions for
483: values of $N_H \leq 10^{21}$ cm$^{-2}$. If we assume a value of $N_H
484: = 3\times10^{21}$ cm$^{-2}$, then the $EMD$ rises strongly below $\log
485: T=6.5$; however, in this case the required N abundance is
486: unrealistically large, at 4.7 times Solar. Hence, we adopt a value of
487: $N_H = 10^{20}$ cm$^{-2}$. This value has $<10\%$ effect on line
488: fluxes in the \hetgs\ bandpass, and is consistent with the apparent
489: lack of interstellar reddening toward the TWA
490: \citep{Kastner:Crigger:al:2003} and with the apparently negligible
491: $E(B-V)$ toward and Hen~3-600 in particular. The small inferred value
492: of $N_H$ (see also \S\ref{sec:ovii}) furthermore suggests we observe
493: the dust-disk-enshrouded component, $A$
494: \citep{Jayawardhana:Hartmann:al:1999}, at low inclination
495:
496:
497:
498: The abundances obtained from the $EMD$ fitting under the assumption
499: $N_H=10^{20}$ cm$^{-2}$ are listed in Table~\ref{tbl:abundemd} and
500: displayed in Figure~\ref{fig:abundemd}. We have adopted the
501: abundances of \citet{Anders:89} for our ratios to Solar values.
502: Figure~\ref{fig:emdspec} shows the observed spectrum overlaid with the
503: synthetic spectrum predicted by the emission measure distribution and
504: abundance model.
505:
506: Assuming $N_H = 10^{20}\, \mathrm{cm^{-2}}$, the integrated flux
507: (1.5-25 \AA) obtained from the $EMD$-reconstructed model is
508: $1.2\times10^{-12}\,\mathrm{ergs\,cm^{-2}\,s^{-1}}$
509: ($0.74\times10^{-3}$ $\mathrm{phot\,cm^{-2}\,s^{-1}}$), corresponding
510: to a luminosity $L_x=3.1\times10^{29}\,\mathrm{ergs\,s^{-1}}$ at the
511: estimated distance to Hen~3-600 of 45 pc, the mean of the values
512: derived by \citet{Kastner:Zuckerman:1997} ($39\pm 7$ pc, a
513: ``photometric'' distance) and \citet{Webb:Zuckerman:al:1999} (50 pc,
514: an assumed value, based on Hipparcos distances of 47--67 pc measured
515: for other TWA members).
516: %
517: %\clearpage
518: \begin{figure}[htb]
519: \epsscale{0.9}
520: \centerline{\plotone{f4.eps}}
521: \caption{\small
522: Emission measure distribution derived from line flux fitting to
523: the $A+B$ dispersed spectrum. The darkest shading
524: with the light central line is the preferred solution for a
525: hydrogen column density of $N_H=10^{20}\,\mathrm{cm\mtwo}$, which
526: is effectively a negligible term for the line uncertainties
527: here. The lightest shading seen at higher emission
528: measure at low temperature (with dark central line) is for
529: $N_H=3\times10^{21}$ cm$^{-2}$, which can
530: be excluded because of the implausibly large abundance of
531: nitrogen required to fit the $\lambda24.8$ line. Intermediate
532: shading is for $N_H=1\times10^{21}$ cm$^{-2}$.
533: \label{fig:emd}
534: }
535: \end{figure}
536: %\clearpage
537:
538: %\clearpage
539: \begin{figure}[t]
540: \epsscale{0.9}
541: \centerline{\plotone{f5.eps}}
542: \caption {\small Elemental abundances (relative to Solar) derived
543: from the line-based emission measure fit, plotted against
544: the First Ionization Potential (FIP).
545: % {\bf REMOVE:?} The light circles
546: % show the effect of a higher $N_H=10^{21}$ cm$^{-2}$, which
547: % primarily affects longer wavelength lines.
548: \label{fig:abundemd}
549: }
550: \end{figure}
551:
552: %
553: \clearpage
554: \begin{figure*}[h]
555: \centerline{\includegraphics[scale=0.9]{f6.eps}}
556: \caption{\small The combined $\heg + \meg$ spectrum of $A+B$ (dark
557: histogram), the model (light histogram), and $\delta\chi$
558: residuals (lower panels). Lines measured for emission measure
559: analysis are labeled, and line details can be found in
560: Table~\ref{tbl:lfluxemd}.
561: \label{fig:emdspec} }
562: \end{figure*}
563: \clearpage
564:
565: \subsection{Cross-dispersion decomposition}\label{sec:xdisp}
566:
567: %\clearpage
568: \begin{figure}[h]
569: \centerline{\includegraphics[scale=0.8,angle=0]{f7.eps}}
570: \caption{\small The cross-dispersion distribution of \meg\ first order
571: events. The top panel shows a scatter plot of the events, and the
572: bottom shows their cumulative histogram. The grating
573: cross-dispersion angle has been scaled to the imaging focal
574: length. In detail, this distribution depends upon wavelength. On
575: the average, for an extraction region boundary at about $-1$
576: arcsec, about $7\%$ of the counts on the $A$-side (left) come from
577: $B$, and there is $\sim4\%$ contamination in $B$ from $A$.
578: \label{fig:rddist} }
579: \end{figure}
580: %\clearpage
581: %
582: Though the main components of the Hen~3-600 binary system are well
583: resolved in the zeroth-order image (see Figure~\ref{fig:zo}),
584: the dispersed spectra of the two components are only marginally
585: resolved spatially in the cross-dispersion direction
586: (Figure~\ref{fig:rddist}). The degree of cross-contamination of the
587: spectrum increases with wavelength due to the astigmatism inherent in
588: a Rowland spectrometer. To obtain emission line fluxes and ratios in
589: a consistent manner across the entire spectrum we have modeled the
590: cross-dispersion profiles in the first-order spectra as a function of
591: wavelength to quantify the relative contribution of each source to
592: the brighter emission lines.
593:
594: Templates for cross-dispersion profiles were based on the same MARX
595: simulations from which the \chan\ calibration database line spread
596: fuctions were derived \citep{Marshall:Dewey:Ishibashi:2004}. The
597: cross-dispersion profile of a single source is characterized by the
598: sum of two Gaussian plus two Lorentzian distributions whose parameters
599: depend on the grating type and wavelength.
600:
601: The Hen~3-600 cross-dispersion profiles were fit with the sum of two
602: cross-dispersion calibration profiles, each corresponding to a single
603: source at the wavelength range of interest. These wavelength ranges
604: were selected to be as narrow as feasible --- to extract information
605: on the relative strength of the two sources in individual spectral
606: lines where possible --- but wide enough to include counts sufficient
607: to allow a reliable fit. For each selected wavelength range, the
608: observed cross-dispersion profile generally shows two peaks which
609: correspond to the two main components of the Hen~3-600 binary. The
610: separation between the two centroids was set equal to the known
611: separation between $A$ and $B$ ($1.4''$, or $3.9\times10^{-4}$
612: degrees) and not allowed to vary. The Gaussian and Lorentzian width
613: parameters were also frozen at their known calibration values. The
614: only free interesting parameters of the fit were the relative
615: normalizations of each component. Figure~\ref{fig:Xdisperse} shows a
616: collection of the cross-dispersion fits for several line regions.
617:
618: %\clearpage
619: \begin{figure*}[t]
620: \includegraphics[scale=0.3,angle=-90]{f8a.eps}
621: \includegraphics[scale=0.3,angle=-90]{f8b.eps}
622: \includegraphics[scale=0.3,angle=-90]{f8c.eps}\newline
623: \includegraphics[scale=0.3,angle=-90]{f8d.eps}
624: \includegraphics[scale=0.3,angle=-90]{f8e.eps}
625: \includegraphics[scale=0.3,angle=-90]{f8f.eps}
626: \caption{\small Cross-dispersion profiles
627: of the MEG first-order spectrum of Hen~3-600 $AB$ (thin black
628: histograms) in several spectral regions. A two-component model
629: fit, based on the calibration model MEG cross-dispersion profile,
630: is indicated in each panel by the thick gray histogram. Left to
631: right, top row: \eli{Si}{14} 6 \AA; \eli{Mg}{11}-\eli{Mg}{12}
632: 8--10\AA; and the Fe and continuum region around 11 \AA.
633: Left-to-right, bottom row: \eli{Ne}{10} 12 \AA;
634: \eli{Fe}{17} 15--17 \AA; and \eli{O}{8} 20 \AA.
635: The lower panel in each plot shows the residual $data-model$
636: counts. The zero-point of the x-axis is arbitrary; component $A$
637: is on the left, $B$ is on the right.
638: \label{fig:Xdisperse}}
639: \end{figure*}
640: %\clearpage
641: %
642: To derive the fluxes of individual spectral lines for each of the two
643: stars, we first determined the line fluxes in the combined ($A+B$),
644: dispersed spectrum. We then partitioned the combined line flux on the
645: basis of the weights derived from the best fit to the cross-dispersion
646: profiles in the corresponding wavelength region. The results are
647: listed in Table~\ref{tbl:abfluxes}. For individual bright lines such
648: as \eli{Ne}{10} $\lambda12.13$ and \eli{O}{8}
649: $\lambda18.97$, the cross-dispersion profile fitting
650: can be performed over a narrow wavelength range, and the fluxes are
651: well determined for each binary component. For weaker lines like
652: \eli{Si}{13} or the \eli{Ne}{9} intercombination and forbidden lines,
653: broad wavelength ranges were required for the profile fitting, and we
654: are only able to obtain the mean line and continuum contributions from
655: each component across a line complex. Other line fluxes that we
656: measured but which were used only in the emission measure
657: analysis (\S~\ref{sec:emd}) are listed in Table~\ref{tbl:lfluxemd}.
658:
659: \subsubsection{Temperature sensitive line ratios of H-like and He-like ions}\label{sec:tsens}
660:
661: In Section~\ref{sec:emd} we have used $EMD$ analysis to derive thermal
662: structure and element abundances for the composite spectrum of $A+B$,
663: using the fluxes of a large set of spectral lines. Spectral
664: diagnostics based on a few strong spectral lines can be used as
665: independent checks on these $EMD$ results. For instance, the He-like
666: to H-like Ly$\alpha$ line ratios can be used to derive an estimate of
667: the temperature of the emitting plasma. Even though the accurate
668: temperature determination from these ratios strictly holds only for
669: isothermal plasma, whereas the actual temperature distributions of
670: X-ray emitting plasma in stars are far from isothermal, these
671: diagnostics can provide useful comparisons between sources \citep[see,
672: e.g., ][]{Testa:al:2007b,Ness:gudel:al:2004}.
673:
674: In Table~\ref{tbl:lineratios}, we summarize results for key line
675: ratios that are diagnostic of plasma temperature. Following the
676: cross-dispersion decomposition, a handful of these
677: temperature-sensitive line ratios are available separately for $A$ and
678: $B$. These ratios --- in particular, Ly$\alpha$/He-r line ratios of Ne
679: --- indicate that there were sharp plasma temperature differences
680: between $A$ and $B$ during our observations, with $B$ the hotter
681: component. The ratios of the H-like Ly$\alpha$ to He-like resonance
682: lines of Ne and O are indicative of plasma temperatures of $\sim2-3$
683: MK for component $A$ and $\sim7$ MK for component $B$. This is
684: consistent with the results from cross-dispersion profile analysis of
685: the He-like \eli{Ne}{9} and \eli{O}{7} triplet regions, which indicate
686: (albeit with a high degree of uncertainty) that $A$ is the fainter of
687: the two components in the former region but is comparable to component
688: $B$ in the latter, longer-wavelength spectral region.
689:
690: %% \clearpage
691: \subsubsection{Elemental abundances}\label{sec:abund}
692:
693: Using the cross-dispersion profile analysis, we can estimate the
694: abundances of Ne, Fe, Mg, and Si with respect to O for each stellar
695: component individually --- independently of (and in contrast to) the
696: $EMD$ analysis, in which the derived abundances
697: (Table~\ref{tbl:abundemd}) represent an average of the two components.
698: These diagnostics use the ratio of two combinations of lines of two
699: different elements whose combined emissivity curves have very similar
700: temperature dependence. Therefore their ratio is largely independent
701: of the specific thermal distribution of the plasma and provides an
702: estimate of the ratio of the abundances of the relevant elements
703: \citep[see e.g.][]{Drake:Testa:2005,Garcia-Alvarez:al:2005}.
704:
705: The cross-dispersion decomposition results for abundances are listed
706: in Table~\ref{tbl:abundratios}. Consistent with the $EMD$ analysis,
707: we find both components of Hen~3-600 display anomalously high Ne/O
708: ratios and Fe deficiency, with component $A$ slightly less extreme in
709: terms of the latter. The Ne and Fe abundances and ratios to O of both
710: components are very similar to those of HD~98800
711: \citep{Kastner:Huenemoerder:al:2004} or BP~Tau
712: \citep{Robrade:Schmitt:2006}. The Ne/O ratios, though, are not as
713: extreme as in the case of TW~Hya \citep{Kastner:02,
714: Robrade:Schmitt:2006}.
715:
716: Component $A$ displays evidence for enhanced abundances of Mg and Si
717: relative to O. These anomalies, if confirmed, are similar to patterns
718: in TW~Hya, whose high-resolution X-ray spectrum \citep{Kastner:02}
719: reveals a relative overabundance of about 3 (compared to Solar ratios)
720: for both elements relative to oxygen.
721:
722: %
723:
724: \subsection{Density-sensitive line ratios of He-like ions}\label{sec:dsens}
725: %
726: %\clearpage
727: \begin{figure}
728: % \epsscale{0.9}
729: \epsscale{0.8}
730: \centerline{\plotone{f9.eps}}
731: \caption{\small Top: The \eli{O}{7} triplet counts spectrum (dark line) and
732: best fit of Gaussians
733: folded through the instrument response (light-shaded line). No
734: background was subtracted and hence, uncertainties are Poisson.
735: Bottom: confidence contours in the triplet density sensitive $R =
736: f/i$ and temperature sensitive $G = (f+i)/r$ ratios. Inner to
737: outer contours are for 68\%, 90\%, and 99\% confidence levels.
738: The grid \citep[after][]{Smith:01} gives the theoretical contours
739: for constant densities and temperatures as labeled.
740: \label{fig:otripconf}
741: }
742: \end{figure}
743:
744: \begin{figure}[t]
745: % \epsscale{0.9}
746: \epsscale{0.8}
747: \centerline{\plotone{f10.eps}}
748: \caption{\small Top: The \eli{Ne}{9} triplet spectrum and confidence
749: contours. (See Figure~\ref{fig:otripconf} for details.)
750: \label{fig:netrip}
751: }
752: \end{figure}
753: %\clearpage
754: %
755:
756: Ratios of the forbidden ($f$) to intercombination ($i$) line
757: intensities within the triplet line complexes of He-like ions (e.g.,
758: \eli{Mg}{11}, \eli{Ne}{9}, \eli{O}{7}) are potential diagnostics of
759: plasma density. For the combined ($A+B$) spectrum of Hen~3-600, the
760: $f/i$ ratio for the \eli{O}{7} triplet is significantly smaller than
761: the low-density limit of 3.9. The measured value, $f/i \sim 1\pm0.5$,
762: lies between the \eli{O}{7} $f/i$ line ratios displayed by HD~98800
763: ($\sim3.5$) and TW~Hya ($\sim0.1$). It is not possible to infer plasma
764: densities independently and unambiguously for the two main binary
765: components of Hen~3-600, as the relatively poor photon counting
766: statistics limit the cross-dispersion profile fitting to broad
767: wavelength ranges that average over these triplet line complexes.
768: Table~\ref{tbl:abfluxes} and Figures~\ref{fig:zo}
769: and~\ref{fig:Xdisperse} suggest that components $A$ and $B$ contribute
770: roughly equally to the photon count rate in the \eli{O}{7} triplet
771: wavelength range, with $A$ perhaps the brighter component in this line
772: complex.
773:
774: We computed confidence contours in the ratios $R = f/i$ and
775: $G=(f+i)/r$ (where $r$ is the intensity of the triplet resonance line)
776: directly from the Hen~3-600 $A+B$ spectrum
777: (Figure~\ref{fig:otripconf}). The results show that the 90\%
778: confidence limit extends to near the low density limit (middle
779: contour), but is still significantly above this limit. However, the
780: $G$ ratio is unrealistically large, extending to temperatures outside
781: the theoretical range of sensitivity, indicating some peculiarity in
782: the emission, or a problem with the atomic data. If we consider only
783: the $f/i$ ratio ($R$), then $\log n_e [\mathrm{cm\mthree}] \sim
784: 10.75$, with the 90\% confidence interval at $\log T = 6.0$ ranging
785: from 10.25 to 11.25. The \eli{Ne}{9} $f/i$ line ratio indicates an
786: upper limit of $\log n_e [\mathrm{cm}\mthree] \le 11.75$ (90\%
787: confidence), consistent with \eli{O}{7} (Figure~\ref{fig:netrip}).
788: The \eli{Ne}{9} triplet region has Fe line blends, but since Fe is
789: relatively weak in the spectrum of Hen~3-600, the $R$ ratio for
790: \eli{Ne}{9} is essentially uncontaminated. The \eli{Mg}{11} triplet
791: is compromised by weak signal, \eli{Ne}{10} H-Ly$\alpha$ series blends
792: which bracket the intercombination line, and significant continuum.
793:
794: \subsection{\eli{O}{7} ratios:
795: $N_H$, abundances, and binarity}\label{sec:ovii}
796:
797: In principle, given the theoretical ratio of flux in \eli{O}{7}
798: He-Ly$\beta$ (18.63 \AA) to the flux in the triplet complex (22 \AA),
799: we should be able to determine the line-of-sight absorbing column of
800: neutral hydrogen $N_H$. As $N_H$ increases, the ratio of 18.63 \AA\
801: to 22 \AA\ apparent flux should increase above the theoretical value
802: of 0.06, due to preferential absorption of longer-wavelength photons.
803: Indeed, the ratio we observe for He-Ly$\beta$ to the sum of the
804: triplet is about 2.5-3 times the theoretical value of 0.06. The
805: measurement uncertainty is large, however; the 90\% confidence
806: interval reaches into the range 0.04-0.08 in the ratio defined by its
807: temperature dependence over the peak emissivity of \eli{O}{7}
808: (approximately $\log T = 6.2 $ to 6.5).
809:
810: The binarity of the system introduces complications, however, since
811: the dispersed spectra of components $A$ and $B$ are blended. There is
812: no reason {\it a priori} that any local $N_H$ be the same for the two
813: components; indeed, we would expect $N_H$ toward $A$ to be larger than
814: that toward $B$, given that only the former component has a detectable
815: mass of circumstellar dust. In addition, the line fluxes are
816: determined by the product of the emission measure and the oxygen
817: abundance, either of which could also differ between the two stars.
818:
819: One empirical constraint, from the cross-dispersion profile fitting
820: (\S \ref{sec:xdisp}), is that the emergent flux near 20 \AA\ is about
821: equal for $A$ and $B$. Assuming that this reflects similar flux in
822: the \eli{O}{7} transitions from each star, the ratio of \eli{O}{7}
823: fluxes to the ratio of theoretical emissivities is then only a
824: function of $N_H$ for each stellar component. Contours of constant
825: ratios vs.\ $N_H(A)$ and $N_H(B)$ (not shown) indicate that $N_H$ must
826: be $> 2\times10^{21}$ for either component to drive the observed ratio
827: to values $>2$. However, such a large value of $N_H$ is inconsistent
828: with the relatively strong \eli{N}{7} line, as discussed earlier
829: (\ref{sec:emd}). We conclude that the deviation of the observed ratio
830: of 18.63 \AA\ to 22 \AA\ flux from the theoretical value is most
831: likely due to poor photon counting statistics.
832:
833: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
834: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
835: %
836: \section{Discussion}\label{sec:disc}
837:
838: \subsection{He-like triplet line ratios}
839:
840: Both TW~Hya and Hen~3-600 exhibit H$\alpha$ emission line equivalent
841: widths and UV excesses (Table~\ref{tbl:stars}) that are larger than
842: typically observed for chromospherically active stars. Such large
843: H$\alpha$ emission intensities and UV excesses are generally
844: interpreted as evidence for accretion. In
845: Figures~\ref{fig:R_vs_Halpha} and \ref{fig:R_vs_EUB}, respectively, we
846: plot the equivalent widths of H$\alpha$ and ultraviolet excesses
847: $E(U-B)$ vs.\ $f/i$ ratios of triplet lines of He-like ions for
848: pre-main sequence stars and (in Figure~\ref{fig:R_vs_EUB}) for
849: chromospherically active main sequence stars. These figures make
850: clear that:
851: %
852: \begin{enumerate}
853:
854: \item for T Tauri stars, H$\alpha$ equivalent widths and X-ray $f/i$
855: line ratios are well correlated;
856:
857: \item the cTTS systems (TW Hya, BP Tau, V4046 Sgr, and MP Mus) have
858: anomalously small values of $f/i$;
859:
860: \item the properties of the apparently nonaccreting (wTTS) systems AU
861: Mic and HD 98800 more closely resemble those of coronally active
862: main sequence stars than actively accreting pre-main sequence stars
863: \citep[as also noted by][]{Kastner:Huenemoerder:al:2004};
864:
865: \item the weakly accreting Hen~3-600 appears as an intermediate case,
866: lying between the cTTS on the one hand, and wTTS and coronally
867: active main sequence stars on the other.
868:
869: \end{enumerate}
870: %
871:
872: %\clearpage
873: \begin{figure}[t]
874: \includegraphics[scale=0.45,angle=90]{f11.eps}
875: \caption{\small Equivalent width ($EW$) of H$\alpha$ vs.\ the
876: forbidden to intercombination line ratio $f/i$ within
877: triplet lines of the He-like ions \eli{O}{7} (top
878: panel) and \eli{Ne}{9} (bottom panel). The values of
879: H$\alpha$ $EW$ were obtained from \citet{Webb:Zuckerman:al:1999}
880: (Hen 3-600, TW Hya, HD 98800),
881: \citet{GregorioHetem:al:1992} (MP Mus),
882: \citet{Byrne:1986} (V4046 Sgr), \citet[][and
883: references therein]{Schmitt:Robrade:al:2005} (BP Tau), and
884: \citet{BarradoyNavascues:al:1999} (AU Mic).}
885: \label{fig:R_vs_Halpha}
886: \end{figure}
887: \clearpage
888: \begin{figure}[ht]
889: \includegraphics[scale=0.45,angle=90]{f12.eps}
890: \caption{\small Ultraviolet excess $E(U-B)$ vs.\ $f/i$ within triplet lines
891: of the He-like ions \eli{O}{7} (top panel) and \eli{Ne}{9}
892: (bottom panel) as measured for chromospherically active, single
893: main sequence stars (asterisks) and low-mass, pre-main sequence
894: stars (diamonds, cTTS; triangles, wTTS). The values of $E(U-B)$
895: were estimated from $U$ and $V$ photometry listed in
896: \citet{delaReza:al:1989} (Hen~3-600, TW~Hya),
897: \citet{Hutchinson:Evans:al:1990} (V4046 Sgr),
898: \citet{GregorioHetem:al:1992} (HD 98800, MP Mus), and SIMBAD (all
899: other stars) by taking into account intrinsic $(U-B)$ colors
900: appropriate for each star's spectral type \citep{Johnson:1966}.}
901: \label{fig:R_vs_EUB}
902: \end{figure}
903: %\clearpage
904:
905: The presence of intense UV fields generated in accretion shocks that
906: lie in close proximity to the X-ray-emitting regions of cTTS may
907: affect the $f/i$ ratios of such systems, via radiative pumping of ions
908: out of the metastable state that leads to forbidden line emission. If
909: such radiative pumping is important, clearly this would compromise the
910: utility of the $f/i$ ratio as a density diagnostic. However, models
911: suggest that that such UV pumping contributes negligibly to the
912: observed line ratios of He-like ions for TW~Hya and Hen~3-600.
913: Specifically, to produce the $f/i$ ratios observed for Hen~3-600
914: (\eli{O}{7}) and TW~Hya (\eli{Ne}{9}), the UV fields would have to be
915: characterized by temperatures of $\sim10^4$ K and $\sim1.5\times10^4$
916: K, respectively \citep{Porquet:01}. In contrast, the extant UV
917: continuum data for these stars indicates temperatures well below
918: $\sim10^4$ K in both cases
919: \citep{Costa:Lago:al:2000,Muzerolle:Calvet:2000}. Furthermore, there
920: is no apparent correlation between $f/i$ and ultraviolet excess, among
921: cTTS systems (Figure~\ref{fig:R_vs_EUB}). Hence it appears the plasma
922: densities characterizing cTTS systems and Hen~3-600 are anomalously
923: high among X-ray-emitting, late-type stars observed thus far at high
924: spectral resolution. Recent results reported by
925: \citet{Ness:Schmitt:2005} concerning anomalous Fe line ratios in the
926: X-ray spectrum of TW~Hya support this assertion.
927:
928: \subsection{X-ray temperatures and origins in $A$ and $B$}
929:
930: Strong low-$T$ ($\log T\sim 6.5$) emission is apparent in the \hetgs\
931: spectrum of Hen~3-600. This result is obtained both from the \emd\
932: modeling results for the combined ($A+B$) system and from temperature
933: sensitive Ne and O line ratios for component $A$ as derived from
934: cross-dispersion decomposition (Table~\ref{tbl:lineratios}). The
935: similarly low plasma temperatures measured for Hen~3-600-$A$ and TW~Hya
936: point to the possibility that, in both cases, the X-ray emission is
937: derived (at least in part) from shocks. In this regard, we note that
938: the X-ray flux from Hen~3-600-$A$ was relatively constant during our
939: observation, consistent with a shock origin. Furthermore, the
940: temperature derived from \eli{O}{7} line ratios in the spectrum of $A$
941: ($\sim3$ MK) is similar to that measured for other accreting TTS,
942: whereas the \eli{Ne}{9} and \eli{O}{7} line ratio temperatures of $B$
943: ($\sim5-7$ MK) more closely resembles those of non-accreting TTS
944: \citep{Telleschi:Gudel:Briggs:Audard:Scelsi:2007}.
945:
946: There is additional evidence that the X-ray emission from component
947: $B$ is dominated by magnetically heated (as opposed to shock-heated)
948: plasma. Specifically, it displayed a hotter spectrum than $A$ as
949: determined both in the continuum and in line ratios, and it displayed
950: a large flare. The flare was somewhat peculiar, however, in that the
951: spectral hardness showed no strong change as is typical for coronally
952: active stars. This circumstantial evidence may simply be due to
953: statistics; further observations may detect a flare on $A$. In the
954: Taurus molecular cloud, cTTS and wTTS do show similar variability
955: \citep{Stelzer:Flaccomio:al:2007}. This implies that even strongly
956: accreting sources still probably have a significant emission component
957: from magnetically heated plasma. As any accretion-generated emission
958: component weakens, magnetic coronal activity will become more
959: apparent. In this middle regime, high-resolution (line-based)
960: diagnostics of temperatures and abundances are needed.
961:
962:
963:
964: \subsection{Abundances}
965:
966: Whereas TW~Hya displays evidence for metal (O, Fe, Mg, Si) depletion
967: that has been attributed to grain growth in its circumstellar disk
968: \citep{Stelzer:Schmitt:2004, Drake:al:2005}, the abundance patterns
969: found in the X-ray spectrum of Hen~3-600 are ambiguous in this regard:
970: Fe is depleted, but the absolute abundance ratio Ne/O is $\sim0.4$
971: ($\sim2.8$ relative to the solar Ne/O ratio), very much in line with
972: the mean Ne/O ratio of coronally active stars
973: \citep{Drake:Testa:2005}.
974:
975: If, however, we consider only the results from high-resolution X-ray
976: spectroscopy, it is not at all clear that there is such strong
977: evidence for depletion. The Si and Mg abundances in TW~Hya are about
978: Solar according to \cite{Kastner:02}, and while those authors stated
979: problems regarding the \eli{Mg}{11} model, \eli{Si}{14}, \eli{Si}{13},
980: and \eli{Mg}{12} are well described by their adopted emission measure
981: and abundances. Thus, the largest difference between Hen 3-600 and
982: TW~Hya is in their Ne/O ratios which are about 3 for Hen~3-600 (as
983: well as for HD~98800, BP~Tau, and CR~Cha), and about half that of
984: TW~Hya (using the \citet{Anders:89} scale). The abundances of Mg and
985: Si themselves appear to be substantially lower in Hen~3-600 than in
986: TW~Hya.
987:
988: The pattern for Hen~3-600-$AB$ is similar to that found in other
989: active stars from high resolution spectra, but which are also seen in
990: the much younger Orion Nebula Cluster stars from analysis of
991: low-resolution CCD spectra by \citet{Maggio:Flaccomio:al:2007}. The
992: large sample of Orion stars allowed those authors to apply independent
993: statistical tests. Our values for Hen~3-600-$AB$ fall within the
994: $1\sigma$ boxes shown in their Figure~12, which in turn overlap values
995: for other stars.
996:
997: Considering the decomposition by component star, the Si and Mg
998: abundances relative to O in component $A$ are, if anything, enhanced
999: with respect to Solar ratios, and are more like TW~Hya than is
1000: component $B$. Indeed, in light of the fact that the two most rapidly
1001: accreting stars in the TWA, TW~Hya and Hen~3-600, both display
1002: evidence for grain processing \citep{Uchida:Calvet:al:2004}, that only
1003: TW~Hya displays an anomalous Ne/O ratio, and that both TW~Hya and
1004: Hen~3-600-$A$ have high Si/O and Mg/O values, it appears that
1005: circumstellar dust grain evolution may play an important role in
1006: determining X-ray abundance patterns, but these patterns do not seem
1007: to be what was previously expected for Si and Mg. It is clear that
1008: better high-resolution spectra are required to provide much better
1009: statistics and models for the important Mg and Si features. Our
1010: uncertainties on the $A$-$B$ decomposition are quite large due
1011: primarily to the overall signal level, but are not limited by the
1012: spatial separation of the two stars.
1013:
1014: \section{Conclusions}
1015:
1016: The results presented here establish that Hen~3-600, like TW~Hya,
1017: stands out from classical coronal sources, in terms of both its
1018: \eli{O}{7} triplet spectrum and its soft X-ray (low temperature
1019: emission measure) excess. Specifically, among the several dozen stars
1020: with \eli{O}{7} triplet line ratios thus far measured by \chan/\hetgs\
1021: or \xmm/\rgs, Hen~3-600, TW~Hya, BP Tau, V4046~Sgr, and MP~Mus -- all
1022: cTTS systems --- are the only stars with $R=f/i$ less than $\sim1.5$
1023: \citep{Kastner:02,Testa:Drake:al:2004b, Ness:Schmitt:2005,
1024: Robrade:Schmitt:2006, Gunther:Liefke:al:2006,
1025: Argiroffi:Maggio:Peres:2007}. Meanwhile, soft X-ray excesses have
1026: been detected in the spectra of a number of accreting TTS systems and
1027: are not observed in non-accreting systems
1028: \citep{Telleschi:Gudel:Briggs:Audard:Scelsi:2007}. While the
1029: \eli{O}{7} $f/i$ ratio and the $EMD$ peak at $\log{T_X} \sim 6.5$
1030: measured for Hen~3-600 are not as extreme as those found for TW~Hya
1031: \citep{Kastner:02}, it is notable that these two objects --- the only
1032: TWA stars known to be actively accreting --- share these unusual X-ray
1033: spectral properties.
1034:
1035: Hence, it seems that the soft X-ray emission from Hen~3-600-$A$, like
1036: that from TW Hya and few other cTTS observed to date at high spectral
1037: resolution in X-rays, may be directly attributable to accretion shocks
1038: \citep{Kastner:02, Stelzer:Schmitt:2004, Gunther:Liefke:al:2006}.
1039: Indeed --- while conclusions concerning Hen~3-600 are somewhat
1040: tentative at present, due to the poor photon counting statistics in
1041: our \hetgs\ spectroscopy --- the X-ray-derived plasma densities of
1042: TW~Hya and Hen~3-600 ($\log{n_e} \approx 12.5$ and $\log{n_e} \approx
1043: 11$, respectively) appear to scale with their accretion rates as
1044: derived from H$\alpha$ line profiles and UV continuum fluxes:
1045: $\sim5\times10^{-10} M_\odot$ yr$^{-1}$ and $\sim10^{-11} M_\odot$
1046: yr$^{-1}$, respectively \citep{Muzerolle:Calvet:2000}. Alternatively,
1047: this low-$T$ emission perhaps may be the result of shocks in
1048: collimated, disk-driven outflows from these stars. Such soft,
1049: constant X-ray emission components have recently been detected in
1050: \chan\ and \xmm\ CCD spectroscopy of a handful of cTTS disk/jet
1051: systems, such as the Beehive proplyd in Orion
1052: \citep{Kastner:Franz:al:2005} and DG~Tau
1053: \citep{Gudel:Telleschi:al:2007,Gudel:Skinner:al:2005}.
1054:
1055: All of the emission signatures of Hen~3-600-$A$ that are potential
1056: diagnostics of accretion activity --- from its X-ray line ratios,
1057: through its UV excess and H$\alpha$ emission line strengths, to its
1058: weak near-infrared excess --- are fully consistent with its status as
1059: a transition object, placed roughly midway between rapidly accreting,
1060: classical T Tauri stars and non-accreting, weak-lined T Tauri stars.
1061: These results speak to the need for additional high-resolution X-ray
1062: spectroscopy of T Tauri stars spanning a wide range of accretion
1063: states.
1064:
1065: \acknowledgements
1066: Support for this research was provided by NASA/CXO
1067: grant GO4--5012X to RIT, and by NASA through the Smithsonian
1068: Astrophysical Observatory (SAO) contract SV3-73016 to MIT for the
1069: \chan\ X-Ray Center and Science Instruments. We thank Dr.~K.
1070: Ishibashi for assistance with the \hetgs\ LSF calibration data.
1071: This research has made use of the SIMBAD database, operated at CDS,
1072: Strasbourg, France.
1073:
1074: {\it Facilities:} \facility{CXO(HETG)}.
1075:
1076: % for submission, comment out the next line (of two commands) and
1077: % insert <filename>.bbl
1078: %
1079: %\bibliographystyle{jwapjbib} \bibliography{mnemonic,jw_abbrv,apj_abbrv,rsc}
1080: \input{ms.bbl}
1081:
1082: %\clearpage
1083: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1084: \begin{deluxetable}{ccccc}
1085: \tabletypesize{\footnotesize}
1086: \tablewidth{0pt}
1087: \tablecaption{Properties of Hen~3-600, HD~98800, and TW~Hya}
1088: \tablehead{
1089: \colhead{}&
1090: \colhead{}&
1091: \colhead{EW(H$\alpha$)}&
1092: \colhead{$U-B$}&
1093: \colhead{$E(U-B)$}\\
1094: \colhead{Object}&
1095: \colhead{Sp.\ Type}&
1096: \colhead{(\AA)}&
1097: \colhead{(mag)}&
1098: \colhead{(mag)}
1099: }
1100: %
1101: \startdata
1102: HD~98800 & K5\tablenotemark{a} & 0 & 1.1 & 0.0 \\
1103: Hen~3-600 & M3\tablenotemark{b},M3.5 & $-22,-7$ & 0.7 & $-0.5$ \\
1104: TW~Hya & K7 & $-220$ & $-0.3$ & $-1.4$ \\
1105: \enddata
1106: %
1107: \tablenotetext{a}{Hierarchical quadruple system.}
1108: \tablenotetext{b}{Spectroscopic binary.}
1109: \label{tbl:stars}
1110: \end{deluxetable}
1111: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1112:
1113: %%\clearpage
1114: \input{tab2}
1115:
1116: \clearpage
1117: \input{tab3}
1118:
1119: \clearpage
1120: \input{tab5}
1121:
1122: \input{tab4}
1123:
1124: \input{tab6}
1125:
1126:
1127: \end{document}
1128: