2 Observations

  
2.1 Data base

The sky region examined is confined to the Taurus-Auriga-Perseus area. A detailed description of this region is given in Stelzer et al. (2000) (hereafter SNH00) where we have also presented a sky map showing the ROSAT PSPC observations subject to this and the previous study. The stellar sample investigated in this paper is identical to the one described in SNH00. However, we omit stars from the Perseus clouds IC348 (Preibisch et al. 1996) and NGC1333 (Preibisch 1997), since due to their larger distance the PSPC images are dominated by source confusion. We analyse the X-ray emission of young, late-type stars, represented by TTS and members from the Pleiades and Hyades clusters. The selection of the Pleiades and Hyades as examples of young clusters was motivated by their spatial vicinity to the Taurus-Auriga molecular clouds when projected to the sky. For this reason many Pleiads and Hyads lie in the same ROSAT PSPC fields. Most of the X-ray detected Pleiads and Hyads are zero-age main-sequence (ZAMS) stars. There are also some (higher-mass) post-MS stars which are not studied here, and many PMS brown dwarfs. The sample examined extends down to the latest M-type objects and includes brown dwarf candidates. Most of these are below the detection limit. However, we have detected an M9-type object in Taurus-Auriga, the latest type PMS dwarf seen to emit X-rays so far (see Sect. 2.5). The coolest object detected in the Pleiades has spectral type M5. In the Hyades we detect objects down to M9 (spectral types determined from measurements of B-V). With their different ages the three groups of stars (TTS, Pleiads, and Hyads) allow to examine the evolution of the X-ray luminosity.

We have selected all pointed PSPC observations from the ROSAT Public Data Archive available in October 1998 which contain any TTS in the Taurus-Auriga region, any Pleiad, or any Hyad in the field of view. The TTS in that area of the sky are part of the Taurus-Auriga molecular clouds located at $140~{\rm pc}$ (Elias 1978; Wichmann et al. 1998a). For the distance to the Pleiades cluster we have adopted $116~{\rm pc}$, the value given by Mermilliod et al. (1997). Those Hyades stars for which no individual Hipparcos parallaxes are available are assumed to be at a distance of $46~{\rm pc}$ (Perryman et al. 1998).

A detailed description of the membership lists for TTS, Pleiads, and Hyads is given in SNH00. SNH00 also have presented a complete list of the pointed ROSAT PSPC observations analysed here. In the earlier paper we were dealing with the same observations but have concentrated on large X-ray flares observed on detected stars. Now we discuss the X-ray characteristics of the whole sample, including non-flaring stars and non-detections. Therefore, we also analyse the short exposures and observations with unstable background marked with an asterisk in Table 1 of SNH00, and not considered in that earlier investigation.

  
2.2 Source detection

Source detection is performed based on a maximum likelihood method which combines local and map source detection algorithm (see Cruddace et al. 1988). Sources with a $ML \geq 7.4$ (corresponding to $\sim$3.5 Gaussian $\sigma$and shown to be the best choice by N95) are considered to be a detection. The probability for existence of a source of given ML is given by $P = 1 - \exp{(-ML)}$. For ML = 7.4 the probability is 0.9994, and among the $\sim$800 detected young stars we would expect to find less than one spurious source.

Observations whose center positions are less than $15^{\prime\prime}$apart have been merged to increase the sensitivity for faint detections. The photon extraction radius of the X-ray sources is not well defined if the off-axis positions in individual observations that are merged differ strongly from each other. Therefore, we have analysed observations with less overlap, i.e. more than $15^{\prime\prime}$ separation, separately. The center of the merged image is the center from all individual observations that are added up. The off-axis positions of X-ray sources in merged pointings are computed with respect to this averaged pointing position.

As the positional accuracy of the ROSAT PSPC declines towards the edge of the detector, the identification radius between optical and X-ray position depends on the off-axis angle of the source. We have computed the normalized cumulative number of identifications as a function of the offset between optical and X-ray position, $\Delta_{\rm ox}$, for different ranges of off-axis angles. Following N95, for each of these distributions we have determined the turnover point, $\Delta _{\rm ox,max}$, which corresponds to the value of $\Delta_{\rm ox}$ where wrong identifications begin to contribute significantly to the detected sources. We have then performed a linear fit to the mean off-axis angle as a function of this critical offset $\Delta _{\rm ox,max}$. The fit values of $\Delta _{\rm ox,max}$ for all examined off-axis ranges are listed in Table 1.

 

 

Table 1: Maximum offset allowed between optical and X-ray position $\Delta _{\rm ox,max}$ for different off-axis angles $\Omega$. $\Delta _{\rm ox,max}$ is the best-fit value of a linear distribution of offsets found from normalized cumulative numbers of identifications (see text).

Off-axis angle					$\Delta _{\rm ox,max}$
[arcmin]					[arcsec]
		$\Omega$	$\leq$	27.5	40.0
27.5	<	$\Omega$	$\leq$	30	42.4
30	<	$\Omega$	$\leq$	32.5	49.5
32.5	<	$\Omega$	$\leq$	35	56.7
35	<	$\Omega$	$\leq$	37.5	63.8
37.5	<	$\Omega$	$\leq$	40	70.9
40	<	$\Omega$	$\leq$	42.5	78.1
42.5	<	$\Omega$	$\leq$	45	85.2
45	<	$\Omega$	$\leq$	47.5	92.3
47.5	<	$\Omega$	$\leq$	50	99.5

These values are used as identification radii for the cross-correlation of membership lists and X-ray observations. For off-axis angles below $27.5^\prime$ a maximum offset between optical and X-ray position of $40^{\prime\prime}$ is appropriate. Note, that this value agrees with the value found by N95 for the ROSAT All-Sky Survey (RASS). Sources which are located further than $50^\prime$ from the detector center are ignored in the analysis presented here, because at large off-axis angles the point spread function deviates from a Gaussian and can not be adequately modeled by the available software.

We have computed the count rates of detected and undetected sources by integrating all events within a circular region around the source position, i.e. the X-ray position for detections and the optical position for non-detections. We use the 99% quantile of the point spread function at 1 keV as photon extraction radius, except for those few cases where the broad band X-ray image shows that the source obviously exceeds this radius probably due to the energy being different from 1 keV. For these special cases we determine the optimum radius individually by visual inspection of the X-ray image.

The measured counts are background subtracted and divided by the exposure time obtained from the exposure map to determine the count rates. For the background subtraction we have used the information from the background maps. This method is useful in crowded fields where a background annulus around the source may easily be contaminated by adjacent sources.

In the crowded Pleiades region occasionally two or more X-ray sources show significant overlap. In order to separate the contributions from each star we were forced to decrease the photon extraction radius of these sources. This leads to an underestimation of the true count rate, but should not effect our results due to the low number of confused stars (15 versus >200 detections among the Pleiades).

  
2.3 Results of source detection

The result of source detection and identification is summarized in six tables: Tables 2, 3, and 4 contain the X-ray parameters of all detected TTS, Pleiads, and Hyads, and in Tables 5, 6, and 7 the X-ray characteristics of undetected TTS, Pleiads, and Hyads are listed.

In Tables 2-7 the first column contains a number for the observation referring to the numbering in Table 1 in SNH00. (See SNH00 for the ROSAT observation request numbers.) For merged observations we give the numbers of all pointings that have been added up. Column 2 is the designation of the stars. Column 3 contains two flags, one that gives the type of TTS ("W'' - wTTS, "C'' - cTTS) and another one for the multiplicity of the stellar system ("S'' - single, "B'' - binary, "T'' - triple, and "Q'' - quadruple). The distinction between cTTS and wTTS is based mainly on the standard H$\alpha$ equivalent width boundary of 10 Å together with the spectral type of the star (i.e. the H$\alpha$ flux), which is similar to the suggestion by Martín (1997) to use different $W_{\rm H\alpha}$boundaries for different spectral types (GKM). Furthermore, we make use of indications for circumstellar material as revealed from IR and mm observations. SUAur, e.g., is of spectral type G2 and $W_{\rm H\alpha}$ is between 3.5 and 5 Å, but it also has a massive disk and, therefore, clearly is a cTTS. The H$\alpha$ equivalent widths are taken from N95, Kenyon & Hartmann (1995), and Wichmann et al. (1996). The spectral types are shown in Col. 4. The spectral types of Pleiades and Hyades stars were derived from the B-V measurements given in the Open Cluster Data Base compiled by C. Prosser and colleagues (and available at ftp://cfa-ftp.harvard.edu/pub/stauffer/clusters) using the conversion of Schmidt-Kaler (1982). For TTS in Taurus-Auriga we have adopted the spectral types compiled by N95 and König et al. (2001).

For all detected stars (Tables 2-4) we list the X-ray position (Cols. 5 and 6), the offset $\Delta$ between optical and X-ray position (Col. 7), the off-axis angle (Col. 8), and the maximum likelihood (Col. 9) of existence. We give the X-ray hardness ratios HR1 and HR2 in Cols. 10 and 11. The PSPC hardness ratio HR1 is defined as follows:

$$HR1 = \frac{H - S}{H + S}$$ (1)

where H is the hard band count rate between 0.5-2.0 keV, and S is the count rate in the soft band (between 0.1-0.4 keV). HR2 is given by:

$$HR2 = \frac{H2 - H1}{H2 + H1}$$ (2)

where H2 and H1 are the count rates in the upper and lower part of the hard band between 0.5-0.9 keV (H1) and 0.9-2.0 keV (H2), respectively. In cases where no counts are observed in any one energy band, and therefore HR1 or HR2 are either +1.0 (no soft counts) or -1.0 (no hard counts) we have computed upper limits to the hardness from the counts in the background. Column 12 gives the exposure time and Col. 13 the X-ray luminosity.

In order to determine the count-to-energy-conversion-factor CECF for the compilation of luminosities we have used the hardness criterion given by Fleming et al. (1995): $CECF = (8.31 + 5.30 \cdot HR1) \times 10^{-12}~{\rm erg\,cm^{-2}\,cts^{-1}}$. Since the soft band in HR1 is sensitive to A_V, this way we implicitly take account of the extinction. It should be noted that HR1 "saturates'' for extinctions ${A_{V} >} \sim$0.5. High extinctions are however rare in the Taurus region, and do not play a role for the Pleiades and Hyades. But to ensure that no systematic errors are introduced by this method of count-to-energy conversion we have compared the resulting distribution of X-ray luminosities with those directly derived from the available A_V measurements (see Sect. 4.1).

The values of the luminosity given in Tables 2-4 have been derived dividing the count rate by the multiplicity of the stellar system. This means we assume that each of the components in the system contributes the same level of X-ray emission (see König et al. 2001 and Sect. 4.3). The mean value of the CECF is $1.00 \pm 0.25 \times 10^{-11}~{\rm erg\,cm^{-2}\,cts^{-1}}$. This value was used to obtain the luminosity in cases where HR1 is a upper/lower limit, and therefore Fleming's relation cannot be applied. Uncertainties in $\log{L_{\rm x}}$ are derived from the statistical errors without taking account of systematic uncertainties in the distance estimate.

X-ray parameters for non-detections are summarized in Tables 5-7. The meaning of Cols. 1 to 4 in Tables 5-7 is the same as in Tables 2-4. In Cols. 5 and 6 we list the optical position. The off-axis angle of the undetected stars is given in Col. 7. Column 8 contains the upper limits to the source counts, Col. 9 the exposure time, and Col. 10 the X-ray luminosity. We have used the mean value of the CECF for the compilation of an upper limit to $L_{\rm x}$ in the case of non-detections. The X-ray luminosity was divided by the number of stellar components.

Multiple stellar systems are represented by a single entry in Tables 2-7, but the designations and if known the spectral types of all components are given. Whenever more than one star lies in the X-ray-to-optical identification radius we list the designations of all possible counterparts.

If a star was detected in both unmerged and merged observations we list only the result from the merged observations. The same applies to stars which are detected neither in the merged nor in the unmerged observations. Here, we list only the upper limit from the merged observation. In a few cases a star was detected in a single but not in the merged observations. This can occur if the source is not within the inner $50^\prime$ of the merged observation due to the shift in pointing centers during the merging process, or if the source or background is variable.

Stars which have shown an X-ray flare (discussed by SNH00) are represented by their quiescent emission, i.e. the flare has been removed from the data. We have marked the flare observations in Tables 2-7 by a label after the observation ID.

  
2.4 Detection rates

The aim of this study is to examine the X-ray emission from magnetically active stars. Stars of spectral types earlier than $\sim$F5 are not expected to show dynamo activity because they have no or only shallow convection zones (Walter 1983). We are not interested in the X-ray emission of these stars because they obey a different emission mechanism. Therefore, we restrict the following analysis to stars with spectral types G and later.

An overview over the detection rates for stars from the different stellar groups is given in Table 8.

 

 

Table 8: X-ray statistics of TTS, Pleiads, and Hyads observed and detected in pointed ROSAT PSPC observations. $N_{\rm D}$ - Number of detections, $N_{\rm N}$ - Number of non-detections, $N_{\rm stars}$ - Number of different stars observed, $N_{\rm mult. FOV}$ - Number of stars in the FOV of more than one observation, $N_{\rm mult. D}$ - Number of stars detected in more than one observation.

Sp.Type	$N_{\rm D}$	$N_{\rm N}$	$N_{\rm stars}$	$N_{\rm mult. FOV}$	$N_{\rm mult. D}$
Taurus-Auriga TTS
G	28	19	17	12	11
K	66	30	59	23	19
M	74	88	98	44	33
Pleiades
G	41	82	41	31	20
K	118	231	112	87	59
M	52	139	65	47	31
Hyades
G	29	2	22	8	8
K	71	20	54	29	26
M	84	69	99	46	33

We have split up each sample according to the spectral types of its members. The column labeled "$N_{\rm D}$'' gives the sum of all detections, and "$N_{\rm N}$'' is the number of non-detections. The number of observed stars (column $N_{\rm stars}$) is smaller than $N_{\rm D} + N_{\rm N}$ due to multiple observations of a given sky position. The columns labeled " $N_{\rm mult. FOV}$'' and " $N_{\rm mult. D}$'' denote the total number of stars observed/detected in more than one observation. Note, that for most multiple stars only the spectral type of the primary is known, and therefore the stellar system has only one entry in Table 8.

Histograms of the distribution of spectral types in the different stellar samples are displayed in Fig. 1.

Figure 1: Spectral type distribution of the observed late-type stars: a) single stars, b) multiple stellar systems. For multiples the spectral type of the primary is plotted, and the secondary is not taken into account. Solid lines denote the total number of stellar systems in any field of the ROSAT PSPC observations from Table 1 in SNH00. The hatched areas represent the number of detected systems. The numbers given in each panel represent the total number of observed systems (" $N_{\rm FOV}$'') and detected systems ("$N_{\rm D}$''). Note, that individual stellar systems may have been detected in more than one observation. The fraction of detected stars depends on distance, integration time, possible flaring activity, line-of-sight absorption, and stellar parameters such as age, mass, and rotation.

We show separate histograms for single and multiple star systems. In the latter sample only the primary is considered. (For most secondaries the spectral type is unknown). The empty histogram bins give the number of stars in the ROSAT PSPC field of any observation and the hatched histograms the subgroup of detected stars. The total number of stars displayed in the figure is also given and labeled " $N_{\rm FOV}$'' (all observed stars) and "$N_{\rm D}$'' (all detected stars) respectively.

As seen in Fig. 1, the detection rate is higher for the Hyades than for the Pleiades or TTS, although the Hyades are older. This is probably due to their shorter distance. The relative number of detections is larger for TTS than for the Pleiades presumably because TTS are young and more active. Throughout all spectral types the detection rate is higher for unresolved binaries as compared to single stars. This could indicate that all stars in multiple systems contribute to the X-ray emission. The actual detection rate is a complicated function of many influencing factors, such as distance, integration time, absorption, age and mass. A detailed analysis of the X-ray emission levels of the different groups of stars is given in the following sections.

  
2.5 X-ray detection of very-late type dwarfs

A number of very low-mass dwarfs with spectral types between M5 and M9 have been detected. In particular, we report on the detection of LH0429+17, to date the latest PMS dwarf with X-ray emission. This object was listed as a candidate member of the Hyades in a photometric study by Leggett & Hawkins (1989). In the course of a spectroscopic survey for brown dwarfs in the Hyades Reid & Hawley (1999) have detected strong H$\alpha$ emission but weak absorption in the gravity sensitive Na I line, which is an indication for young age. Taking into account its location on the sky, LH0429+17 can, therefore, be considered as member of the Taurus star forming region.

X-rays from young brown dwarfs and brown dwarf candidates in the Chamaeleon, Taurus-Auriga and $\rho$Ophiuchus star forming regions have first been observed by Neuhäuser & Comerón (1998) and Neuhäuser et al. (1999). These objects have spectral types between M6 and M8. Note, that we confirm here the detection of all brown dwarfs and brown dwarf candidates in the Taurus region which have been listed in Neuhäuser et al. (1999).