EVIDENCE OF HOT HIGH VELOCITY PHOTOIONIZED PLASMA FALLING ON ACTIVELY ACCRETING T TAURI STARS

Solar-like pre-main-sequence (PMS) stars and, in general, low mass PMS stars (M_* < 2M_☉), or T Tauri stars (TTSs), are complex dynamical systems made of two basic components, star and accretion disk, as well as a dynamical interface, the stellar magnetosphere (see Gómez de Castro 2013 for a recent review). Magnetic fields of few kG have been detected on the surface of the TTSs (Guenther et al. 1999; Johns-Krull et al. 1999; Johns-Krull 2007). The surface field is not bipolar, but like the Sun's field, it has a rather complex structure (Johns-Krull et al. 2004). Higher-order, multi-polar components fall off more rapidly with radius than the dipolar field. Unfortunately, it is difficult to observationally track the final path followed by matter from the inner disk border to the stellar surface. In addition, the magnetosphere has its own dynamics and force due to the interaction with the disk.

Unfortunately, the characteristics of the TTSs' extended atmosphere and magnetospheres are still escaping diagnosis (see Hartmann 2009). Little is known about them, apart from having a density of ∼10⁹–10¹¹ cm⁻³ and an electron temperature between some few thousand Kelvin and 100,000 K (Gómez de Castro & Verdugo 2007, hereafter GdCV2007). Recent attempts to derive the magnetospheric properties from optical lines have shown that the Hα profile, for instance, is strongly dependent on densities and temperatures assumed inside the magnetosphere and in the disk wind region (Lima et al. 2010). The line widths of typical atmospheric/magnetospheric tracers are about 200–300 km s⁻¹, which far exceed those expected from thermal or rotational broadening even if the lines are assumed to be formed in a magnetosphere that extends to some 4–5 stellar radii and corotates with the star. As of today, it is still unclear whether the broadening is produced by unresolved macroscopic flows or by magnetic waves propagating on the magnetospheric field (Hartmann et al. 1982). In general, the broadening of magnetospheric tracers does not vary significantly in time, indicating that the average motions are rather stable. The combined effect of funnel flows and inclined magnetic rotators as simulated by Romanova et al. (2004, 2012) is the current baseline for the numerical simulation of TTSs magnetospheres. The stellar field is assumed to be anchored in the inner part of the disk, which creates a sheared layer between the rigid body rotation of the star and the Keplerian rotation of the disk. This interaction has a profound influence on the star and the accretion flow and also acts as a dynamo that transforms part of the angular momentum excess in the inner disk into magnetic field amplification. This amplification self-regulates through both quiescent periods of field build up and eruptions when the energy excess is released (see Gómez de Castro & von Rekowsky 2011 for an evaluation of the ultraviolet (UV) output from such an interaction). However, magnetospheric heating processes are poorly known, and accretion shock models are unable to predict the observed line fluxes/broadening (Johns-Krull 2009).

From the observational point of view, the TTSs are split into two groups: stars with noticeable accretion, the classical TTSs (CTTSs), and weak-line TTSs (WTTSs) where accretion signatures in the spectrum are negligible or absent. The atmospheric and magnetospheric energy output in both types is released mainly in the UV. The richness of spectral tracers for a broad range of magnetospheric temperatures and densities is unmatched by any other range. There are several studies of the atmospheric/magnetospheric properties of the TTSs based on low-dispersion UV data (Lemmens et al. 1992; Huélamo et al. 1998; Johns-Krull et al. 2000; Yang et al. 2012; Gómez de Castro & Marcos-Arenal 2012, hereafter GdCMA2012). However, only the Cosmic Origins Spectrograph (COS) (see Green et al. 2012 for a description of the instrument) on board the Hubble Space Telescope (HST) has been sensitive enough to gather high signal-to-noise ratio (S/N) profiles of hot plasma tracers, such as the N v [UV1] resonance multiplet or the He ii Hα transition, with a resolution above 10,000 in late K- and M-type TTSs (see Penston & Lago 1983 for C iv profiles obtained with the IUE; Ardila et al. 2002 for HST profiles obtained with the Goddard High Resolution Spectrograph (GHRS); and Ayres 2005 for the CoolCat set based on HST data obtained with the Space Telescope Imaging Spectrograph (STIS) and compare them with those presented in Section 2). Previous instruments only allowed us to observe bright or nearby sources such as TW Hya (Herczeg et al. 2002).

Following a previous work where the magnetospheric properties of TTSs were examined based on low dispersion HST data (GdCMA2012), the high resolution profiles of the N v [UV1], O iii], and He ii (1640 Å) transitions are analyzed in this work with two main objectives. First, we intend to determine whether it is feasible to discriminate the contributions to the high energy radiation flux from the high density atmospheric plasma and the accretion flow. Second, we reexamine the unexpected correlation between X-ray flux and the high energy UV tracers found by GdCMA2012 in the light of the kinematic information contained in the high resolution HST/COS profiles. In Section 2, the observations are described and the results are presented in Section 3. Two components are found to contribute to the He ii flux: a low density component (LDC) associated with the accretion flow and a high density component (HDC) of more uncertain origin. The discussion on the possible source of the HDC, its association with accretion shocks, and the properties of the LDC are addressed in Section 4. The work concludes with a brief summary on the relevance of these results.

The He ii profiles of TTSs in the Taurus–Aurigae star-forming region have been retrieved from the HST archive. Only high resolution observations have been considered. Most of them have been obtained with the COS and the gratings G130M and G160M. The resolution is ∼24,000, and each target has been observed three times, with slight offsets in the wavelength range to guarantee that the 18.1 Å gap between the two segments in the FUV detector is covered (see COS Handbook). Moreover, a few TTSs have been observed with the STIS, namely, T Tau, DR Tau, and DF Tau. Only the T Tau profile had a high enough S/N in the He ii line to be considered for this work. The log of the observations is provided in Table 1. Additional information on the HST programs that obtained these observations (programs IDs 11533, 11616, and 8627) can also be found in the table.

The observational strategy allowed us to search for variations in the profiles on time scales of ∼20 minutes, but neither flux nor morphology were found to display significant changes. The one-dimensional spectra produced by the COS calibration pipeline (CALCOS v2.17.3) were aligned and co-added.¹ COS targets were centered to within 0.1–0.2 arcsec to achieve the nominal wavelength accuracy of ±15 km s⁻¹. The (R(3) 1–7) 1489.636 Å and (P(5) 1–7) 1504.845 Å H₂ lines were used to set the zero of the wavelength scale for the targets in Table 1. H₂ emission was dominated by the molecular disk around the stars in most sources (France et al. 2012). The lines were selected to be strong (from Herczeg et al. 2002) and detectable in most of the sources. The profiles of the 1489.636 Å line were plotted in Figure 1 (see also Figure 3 in France et al. 2012) and the He ii profiles were plotted in Figure 2. No shifts were applied to the DN Tau and IP Tau observations because the H₂ emission was too weak to be used for this purpose. Shifts were also not applied to HBC 427, LkCa 19, and LkCa 4 because H₂ emission was not detected. The H₂ profiles are sometimes asymmetric with respect to the rest wavelength, particularly in RW Aur. In this case, the original zero from the CALCOS pipeline has been left. Only a subset of AA Tau observations was used since guide star acquisition failed (see Figure 3). As a result, the first two exposures produced similar profiles while the last one produced a slightly broader and more red-shifted (by ∼0.1 Å or 18 km s⁻¹ at 1640 Å) profile. For this star, the last observation was rejected and only the two first observations were averaged to produce the profiles in Figures 1 and 2.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The He ii profiles can be generically described as composed of a bright and narrow emission feature and a broad, weaker component that differs from one star to another. Notice that the He ii lines are very strong; this fact, together with the strong H₂ emission, contributes to the continuum jump in the low resolution Advanced Camera System on HST reported by GdCMA2012.

Close to the He ii line, there are the O iii]₁₆₆₅ intercombination lines; a doublet with components at λλ1660.802 and 1666.156, which originate under transitions from the level 2s2p^{3 5}S₂ to the term 2s²2p^{2 3}P with J = 1 and J = 2, respectively. The components should have an intensity ratio ≃1:3, equal to the ratio of their transition probabilities (145 s⁻¹ and 426 s⁻¹). The transition is optically thin to a critical density of 3.4 × 10¹⁰ cm⁻³. The O iii] profiles are represented in Figure 4. The two lines of the multiplet are observed in DF Tau, HN Tau, DR Tau, SU Aur, and RW Aur; however, the flux of the weakest line 1660.802 Å, has only been measured for strong sources. In all cases, the flux ratio between the two lines of the multiplet is 1:3 (within the error bars).

Zoom In Zoom Out Reset image size

To complete the view on the distribution of hot plasma in the TTSs environment, the profiles of the resonance UV 1 multiplet of the N v have also been retrieved from the HST archive (see Table 2). In the blue edge of the 1238.8 Å line, the strongest in the doublet, there are some narrow emission lines produced by molecular hydrogen (lines λ1237.918 1–2 P(8) and λ1237.589 2–2 R(11)) that blur the profile somewhat. The zero of the wavelength scale has been set again, resourcing to H₂ emission lines. The (P(2) 0–4) 1338.63 Å line has been used for this purpose since it is strong in most of the stars and not blended with other features (see Figure 5 with the H₂ profiles). As mentioned above, the original zero of the wavelength scale has not been shifted for DN Tau, IP Tau, HBC 427, LkCa 19, and LkCa 4 because the H₂ lines were either very weak or absent. The N v profiles can be most generally described by a single component that ranges from being narrow in stars like HBC 427, LkCa 19, and LkCa 4 to being broad and asymmetric in AA Tau or GM Aur (see Figure 6).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Some relevant properties of the TTSs used in Section 3 are gathered in Table 3. Notice that there are wide variations in published values of important parameters such as the stellar luminosity or the extinction (also see comments in GdCMA2012); data in Table 3 are gathered for reference for other researchers. The X-ray fluxes have been retrieved from the XMM-Newton extended survey of the Taurus (XEST) molecular cloud (Güdel et al. 2007). The He ii, 1238.82 Å N v, and the O iii] line fluxes have been measured after subtracting the local continuum. Also the fluxes for the 1489.636 and 1338.63 H₂ lines have been measured (see Section 5). The line fluxes are provided in Table 4; they are not extinction corrected. For some sources, the measurement of the 1238.82 Å N v flux has required subtracting the nearby H₂ features. In such cases, the H₂ flux has been subtracted by linear interpolation in the N v profile. However, there are a few profiles where this interpolation was uncertain (see quality flags in Table 4).

From Figures 2, 4, and 6, a generic trend can be inferred.

• 1.  
  The weak-line TTSs (WTTSs) in the sample, i.e., evolved TTSs with no evidence of mass infall (LkCa 19, LkCa 4, HBC 427), display neither H₂ emission nor nebular O iii] emission. They have only rather narrow He ii and N v lines.
• 2.  
  The classical TTSs (CTTSs) emit in all these tracers (see below).
• 3.  
  Some intermediate objects (TTSs), such as IP Tau, DN Tau, and V836 Tau, have weak H₂ and no nebular O iii] emission. Both He ii and N v lines have narrow emission profiles.

CTTSs cover a broad range of profile morphologies. Strong H₂ emission is detected in all of them, and the lines are narrow except in RW Aur A (see France et al. 2012). Nebular O iii] emission is detected in all of them except DM Tau, UX Tau A, and AA Tau. O iii] is especially strong in T Tau, DF Tau, HN Tau, DR Tau, and RW Aur. SU Aur and DE Tau seem to be intermediate objects. Hints of a possible O iii] emission at 1666 Å are seen in GM Aur, though unfortunately the S/N is low and the weakest component of the multiplet is not detected. A low S/N feature is detected at 1666 AA in DN Tau, IP Tau, AA Tau, and UX Tau spectra. Observations of the C iii] intercombination transition line are only available for three of the stars in the sample, namely, DE Tau, T Tau, and RW Aur, and no significant differences have been found between the O iii] and the C iii] profiles obtained either with the GHRS (compare Figure 4 with Figure 1 in Gómez de Castro & Verdugo 2001) or STIS (compare Figure 4 with Figure 1 in Gómez de Castro et al. 2003).

In general, the He ii profiles can be described as having a narrow emission component superimposed on a broader contribution that mimics the observed in the O iii] nebular lines when high enough S/N data profiles are observed. This effect is clearly apparent in HN Tau and RW Aur (see Figure 7). Note that in DR Tau and DF Tau, the overlap is lost at blueward-shifted velocities; there is a blueward-shifted O iii] emission with no He ii counterpart. This blueward shifted excess could be caused by the contribution of an unresolved jet to the line emission. As shown for RY Tau by GdCV2003, the semiforbidden emission has two components, one associated with the jet and another with the accretion flow, that were disentangled for this source due to their variability. The N v profiles however, do not follow the same morphological trend. A rather narrow symmetric profile is observed in DE Tau that becomes wider in DM Tau, DF Tau, UX Tau A, GM Aur, and SU Aur, all of which are sources with absent or weak O iii] nebular emission. DR Tau, HN Tau, RW Aur, and AA Tau display very peculiar profiles that clearly indicate that N v emission is not produced in the stellar atmosphere but in another dynamical component. In particular, the N v profiles of RW Aur, HN Tau, and AA Tau are asymmetric, extending from redward-shifted velocities to peak at blueward-shifted velocities; moreover, the RW Aur profile peaks at the velocity of the optical jet (Hirth et al. 1997) that was also detected in the C iii] intercombination line by Gómez de Castro & Verdugo (2003). This type of profile asymmetry has also been detected in the C iii] profile of RY Tau in GdCV2007, which pointed out that the line could be formed in a PMS analog of the solar wind. In this context, it is worth remarking that the H₂ profile of RW Aur has a completely different asymmetry. The flux peaks to the red of the line, suggesting the presence of infalling cold molecular gas similar to that observed in the PMS close binary AK Sco (see also Gómez de Castro et al. 2013).

Zoom In Zoom Out Reset image size

These groups can be cleanly recognized in velocity dispersion diagrams. The characterization of the underlying velocity field and thermal properties of the line emission region is complex in the TTSs environment. Instead of using the standard fitting to Gaussian or Voight profiles, it is preferable to characterize the profile in terms of the standard Pearson statistics moments and measures, i.e., mean or centroid, dispersion, kurtosis, and skewness. They provide a quantitative measurement of the deviation of the profile from the expected for thermal plasmas, i.e., a normal distribution convolved with the line spread function of COS (Kriss 2011). Note that in this approach, the profile is assumed to be formed by the contribution to the line flux of independent gas parcels, i.e., the (background subtracted) profile is treated as a histogram of the flux emitted per parcel in the wavelength (velocity) space; a similar approach was followed by GdCV2007 to compare the observed C iii] and Si iii] profiles of RY Tau with the theoretical predictions. This treatment permits the characterization of the profiles of optically thin lines formed in complex velocity fields. In Table 5, the dispersions of the He ii and N v lines are provided together with those obtained for two control lines, the H₂ transitions at 1339 Å and 1489 Å. RW Aur, HN Tau, and AA Tau are not included because their N v profiles are peculiar. As shown in the bottom panel of Figure 8, the average dispersion of the H₂ profiles is 31 ± 5 km s⁻¹ and 43 ± 13 km s⁻¹ for the 1339 Å and 1489 Å lines, respectively. There is no correlation between the broadening of these two H₂ lines, pointing out that the scattering of the dispersions in the diagram is related to random effects associated with the measurement process. A different trend is drawn from the N v and He ii dispersions. The dispersions of the WTTSs and DN Tau are comparable to those measured in the H₂ line. In intermediate objects, like V836Tau and IP Tau, the dispersions in the He ii line are comparable to those measured in the H₂ lines, but the N v lines are significantly broader. Finally in the CTTSs both σ(N v) and σ(He ii) are larger than σ(H₂). Note that stellar rotation may contribute to the line dispersion; SU Aur, the fastest rotator in the sample, also has the largest dispersion. However, a large dispersion can also be produced by profile asymmetry. For instance, DM Tau has dispersions comparable to those measured in SU Aur and it is one of the slowest rotators in the sample (see Table 3). Two objects do not follow this trend: DR Tau and DE Tau, both of which have intermediate dispersions in the N v lines and large dispersions in the He ii lines caused by the broad emission component. A quick inspection in the summary of the TTSs properties (see Table 3) indicates that the only possible cause of this discrepancy is the high accretion rate, as is expected.

Zoom In Zoom Out Reset image size

3.1. Two Hot Plasma Components in the TTSs

From Figures 2, 4, and 6, it is clearly inferred that there are at least two different plasmas contributing to the spectral lines under study:

• 1.  
  A low-density component (LDC) that it is most conspicuously traced by the O iii] line. The LDC also produces the He ii broad emission component observed in RW Aur, HN Tau, DR Tau, and DF Tau. The critical density of the O iii] sets an upper limit to the electron density² of the plasma in the LDC of ≃ 3.4 × 10¹⁰ cm⁻³ (also see Section 4.2). The LDC profiles display a non-thermal broadening and draw a complex velocity field around the stars. In fact, the LDC profiles seem to trace some kind of complex magnetospheric infalling pattern high above the stellar surface. Also, in some cases, they could be associated with unresolved wind structures (Gómez de Castro & Ferro-Fontán 2005), as the reported for RY Tau (GdCV2007). Notice that the co-existence of O iii] and He ii radiation from the same kinematic structure would point to unrealistically high electron temperatures for the line emission region (log T_e (K) ∼ 5.4), if collisional equilibrium at a single temperature is assumed and electron densities below the O iii] critical density are considered.³ In fact, its UV spectrum is reminiscent of that observed in photoionized nebulae (also see Section 4). To the current sensitivity, the contribution of this component to the N v flux is negligible.
• 2.  
  A high density/temperature component (HDC) that dominates the N v emission. O iii] profiles are very different from N v profiles, suggesting that the density of the N v formation region is higher than the O iii] critical density.

Though the kinematics of the N v emission region are clearly different from that traced by the LDC, the N v flux is correlated with the He ii flux as shown in Figure 9. The Spearman rank correlation coefficient is r_s = 0.87 (with significance level, α = 0.001, see Sachs 1982 for details), and

$$\begin{equation*} \log \left(F({\rm He\,\scriptsize{II}})\right) = (0.8 \pm 0.1) \log \left(F({\rm N\,\scriptsize{V}}) \right) - (2.5 \pm 1.7), \end{equation*}$$

with rms = 0.29 (see bottom panel in Figure 9). Also the fluxes normalized to the stellar surface are correlated with r_s = 0.82, with α = 0.002 (see top panel in Figure 9), and

$$\begin{equation*} \log \left(\frac{F(L({\rm He\,\scriptsize{II}})}{F_{\rm bol}} \right) = (0.9 \pm 0.1) \log \left(\frac{F({\rm N\,\scriptsize{V}})}{F_{\rm bol}} \right) + (0.0 \pm 0.6), \end{equation*}$$

with an rms = 0.31. The normalized flux is defined as the rate $F_{{\rm He\,\scriptsize{II}}}/F_{\rm bol,*}$ or $F_{{\rm N\,\scriptsize{V}}}/F_{\rm bol,*}$ and was introduced by GdCMA2012 to provide a measure of the line emissivity weighted over an unknown thickness but corrected for stellar radii and surface temperature. In this manner, the normalized fluxes compensate for scaling effects associated with the broad range of mass, luminosity, and stellar radius covered by the TTSs. Stars whose He ii flux has a significant contribution from the LDC are marked in the plot. Notice that they are evenly distributed in the figure, which suggests that He ii and N v fluxes are correlated regardless of whether the He ii flux is dominated by the narrow emission component.

Zoom In Zoom Out Reset image size

3.2. The Connection between UV and the X-Ray Radiation from the TTSs

Based on low resolution observations, GdCMA2012 pointed out that the normalized He ii flux anti-correlates with the strength of the X-ray flux as derived in the XEST survey (Güdel et al. 2007) carried out with the XMM-Newton telescope in the 0.3–10.0 keV band. In Figure 10, the normalized He ii fluxes from the high resolution COS/HST observation (see Table 3) are represented against the normalized X-ray luminosities as derived from the XEST survey. Also, the Chandra/ACIS observations of LkCa 4, DE Tau, and GM Aur from Yang et al. (2012) are used; they are integrated X-ray luminosities in the 0.3–10 keV range. The He ii flux is extinction corrected according to Valencic et al. (2004), assuming R = 3.1 and the extinctions in Table 3. The low dispersion GdCMA2012 data also have been plotted for those sources with no available high dispersion data. A first inspection of Figure 10, shows three groups. The WTTSs (HBC 427, LkCa 4, HD283472), UX Tau, and the fast rotator SU Aur are very close to the main sequence stars regression line. The CTTSs follow the generic trend pointed out by GdCMA2012: the He ii flux increases as the X-ray flux decreases. Notice that the sources with strong LDCs are far from this trend, excluding the fast rotator SU Aur. In fact, the Spearman rank correlation coefficient r_s = −0.2622 has α = 0.294, with α being the level of significance, i.e., there is a probability of α = 0.294 that the X-ray and He ii normalized luminosities are not correlated (see Sachs 1982 for a detailed description of α for normal and non-normal error distributions and the most usual statistical tests, including the Spearman rank correlation). However, if HN Tau, DE Tau, and DM Tau are excluded, the Spearman rank correlation coefficient is r_s = −0.591, with α = 0.033, and the least square fit is

$$\begin{equation*} \log \left(\frac{F(L_X)}{F_{\rm bol}} \right) = (-0.20 \pm 0.05) \log \left(\frac{F({\rm He\,\scriptsize{II}})}{F_{\rm bol}} \right) - (4.3 \pm 0.2), \end{equation*}$$

with rms = 0.15.

Zoom In Zoom Out Reset image size

Given the uncertainties in the A_V values for the TTSs, it is worth evaluating whether the deviation from the general trend of the stars is caused by the uncertainties in the A_V values. To examine this effect, the same diagram has been plotted in the bottom panel of Figure 10, though this figure uses stellar luminosities, X-ray luminosities, and A_V values from the recent compilation by Yang et al. (2012). As shown, the three groups are not cleanly separated, and the regression line is more clear now. Moreover, only DE Tau is far from the trend. As a result, r_s = −0.561, with α = 0.036, and

$$\begin{equation*} \log \left(\frac{F(L_X)}{F_{\rm bol}} \right) = (-0.17 \pm 0.07) \log \left(\frac{F({\rm He\,\scriptsize{II}})}{F_{\rm bol}} \right) - (4.1 \pm 0.1), \end{equation*}$$

with rms = 0.15 (excluding DE Tau for these calculations).

In summary, though the clean separation between WTTSs and accreting TTSs may be an extinction associated effect, the trend to release preferentially the high energy excess in the UV rather than in the X-ray channel in accreting objects holds.

The current observations do not show a statistically meaningful deviation of the sources with strong LDCs from the main trend. Note also that the sample is small; there are not X-ray measurements of DR Tau or DF Tau, and RW Aur was detected to have a low soft X-ray flux only with the Einstein satellite (Damiani et al. 1995). If present, such a trend could indicate that X-ray radiation is dominated by different components in sources with strong LDCs (strong nebular component) and in sources with weak or absent LDCs. The X-ray energy distribution of the TTSs is often modeled by two components: a soft component at T_s ≃ (2–5)10⁶ K and a hard component at T_h ≃ (1.5–3)10⁷ K (see, i.e., Glassgold et al. 2000). The hard X-ray component is thought to be associated with magnetic energy release in the stellar coronae. The nature of the soft X-ray component is more uncertain, and it has often been hypothesized that it could be formed in accretion shocks (Lamzin 1998; Gullbring et al. 1998). Unfortunately, only three stars in our sample, namely, T Tau, SU Aur, and HBC 427, have a high enough count rate to allow a spectral fitting to two different optically thin plasmas. Only non-conclusive results could be derived from the fits (see Table 6 with the two-components fit to the X-ray spectrum of these sources—which is from Table 6 in Güdel et al. 2007). The X-ray spectrum of SU Aur, a CTTS, is dominated by the low temperature component with T = 5.22 MK. However, both the soft and hard X-ray components have similar emission measures in HBC 427, a non-accreting WTTSs. Moreover, the hard X-ray component dominates the X-ray spectrum of the CTTS, T Tau.

Table 6. Two Temperature Fits to the XMM-Newton Spectra from XEST

Star	T1	T2	EM1	EM2	EM1/EM2
	(MK)	(MK)	1052	1052
T Tau	4.52	23.65	33.86	57.85	0.585
SU Aur	5.22	23.30	87.01	32.45	2.681
HBC 427	9.04	28.06	14.98	15.07	0.994

Download table as:  ASCIITypeset image

TTSs are complex objects. They are convective PMS stars where a solar-like dynamo begins to set in while the fossil magnetic field is still diffusing. TTSs are surrounded by an external dynamo that powers and makes the stellar magnetosphere rise to the inner border of the molecular disk (see Romanova et al. 2012 for recent simulations). Matter from the disk slides down onto the star along the magnetospheric field lines to end, free-falling onto the open holes of the magnetic configuration. In this environment, hot plasma radiating in the UV tracers studied in this work can be located in the magnetosphere, in the atmosphere, in the accretion shocks, and also in the outflow (either solar-like or driven from the star–disk magnetic interface or the disk). Within the current paradigm, both the magnetosphere and the outflow have significantly lower densities than the stellar atmosphere or the accretion shocks. As a result, spectral line radiation is dominated by radiative de-excitation processes, and forbidden and semiforbidden transitions are strong from this plasma. In the dense atmosphere, collisional de-excitation is relevant, and forbidden transitions are expected to be quenched. UV semiforbidden transitions cannot be observed from the accretion shock itself because it is too hot and dense. However, the soft X-ray radiation produced in the shock front photoionizes the preshock gas, which has a density similar to that of the stellar magnetosphere and may produce forbidden line radiation (Gómez de Castro & Lamzin 1999). Unfortunately, the high column density prevents the UV radiation from the photoionization cascade to escape easily from the accretion column. Also, the profiles of some tracers, from the infrared He i transition (Beristain et al. 2001; Fischer et al. 2008) to the UV lines, do not agree with the predictions of accretion shock models (Johns-Krull 2009). Within this context, the data presented in Section 3 provide some amazing results:

• 1.  
  The He ii, O iii], and N v fluxes do not depend on the spectral type. This confirms that the line emission is not dominated by main sequence like atmospheric magnetic activity, i.e., with the release of the magnetic energy produced by the stellar dynamo, since this one depends on the spectral type (see, e.g., Ayres et al. 1995; GdCMA2012).
• 2.  
  The He ii and N v fluxes correlate well regardless of whether the line profiles are very different, e.g., independently of whether the line emission is produced in the same physical structure. This confirms that all processes (accretion, atmospheric emission, and outflow) are coupled, as expected (see Gómez de Castro 2013 for a recent review). In turn, it makes difficult to obtain specific tracers of individual processes without kinematical information, i.e., without high resolution spectroscopy.
• 3.  
  The high resolution profiles of the N v line show a symmetric line broadening that increases from non-accreting to accreting stars, being significantly suprathermal in these last sources (see Table 5). The profile shape and the density of the line formation region suggests an atmospheric origin (with the exceptions already mentioned in Section 3). The connection between line broadening and accretion suggests that the density and extent of the high atmospheric layers depends on the accretion rate, i.e., on the evolutionary state, as otherwise predicted from the theoretical models (D'Antona & Mazzitelli 1997; Siess et al. 2000). Transport of magnetic energy from the stellar interior to the surface is expected to occur at a different pace in accreting sources than in WTTSs. Moreover, the extended magnetosphere powered by the disk–star magnetic locking, must affect the stellar atmosphere introducing new sources of stirring and turbulence (see, e.g., Kivelson & Russell 1995).
• 4.  
  The physical source of the narrow emission component in the He ii profile, however, remains uncertain. It could either be associated with accretion shocks or with atmospheric features.

In this section, the possible source of the He ii narrow emission component is analyzed, and some constraints on the extent of the magnetosphere are inferred from the semiforbidden line radiation.

4.1. On the Source of the Narrow Component of the He ii Line: Accretion Shocks or Bulk Atmospheric Phenomena?

The kinematics of the region where the narrow component forms, is clearly distinct from that of the N v or the O iii] lines formation region (see Figure 8). The dispersion of the narrow emission component of the He ii ranges from ∼20 km s⁻¹ to ∼60 km s⁻¹, while the dispersion of the N v line varies from ∼40 km s⁻¹ to ∼130 km s⁻¹ for the same sources (see Figure 11). The dispersion of the narrow emission component in the He ii profile has been evaluated as above (see Section 3) but for setting an upper wavelength cut-off to reject the contribution of the LDC. Note that even in setting this cut-off, there is an unknown contribution from the LDC to the flux.

Zoom In Zoom Out Reset image size

Moreover, the narrow component seems to be slightly redshifted in all sources, from 24 km s⁻¹ in UX Tau to 37 km s⁻¹ in DE Tau (see the Appendix and Table 7). Hence, it would be tempting to suggest that the line is produced in accretion shocks.

Table 7. Shift of the He ii Narrow Emission Component

Stars	Δλ = λpeak − λ0
	(Å)
AA Tau	0.175
DE Tau	0.202
DF Tau	0.188
DM Tau	0.166
DN Tau	0.164
DR Tau	0.115
GM Aur	0.186
IP Tau	0.151
UX Tau	0.13
V836 Tau	0.166

Download table as:  ASCIITypeset image

Accretion shocks are produced by the impact of the free-falling material from the disk onto the stellar surface. The kinetic energy of the infalling matter, with typical free-fall speeds of ∼300 km s⁻¹, is involved in gas heating at the shock front that reaches temperatures of about 1 MK. The soft X-ray radiation from the shock front is expected to photoionize (pre-ionize) the infalling gas column (see, i.e., Lamzin 1998; Gómez de Castro & Lamzin 1999; Muzerolle et al. 2001; Orlando et al. 2009). Also, the shock front could back illuminate the stellar surface, becoming a source of atmospheric photoionization to be added to the coronal X-ray radiation. In this context, the narrow component of the He ii line could be produced very close to the shock front. The slight, but small redshift, could be interpreted as an indication of the line being formed in postshock material, and the small width could be caused by thermal broadening. The thermal velocity of fully ionized, solar abundance plasma at 50,000 K is 23 km s⁻¹. However, this cannot be concluded from these data alone. Firstly, the apparent redshift could be caused by the blending of the narrow component with the broad component, which is asymmetric in most sources, e.g., with very low or absent flux at blueward-shifted velocities. The line broadening is affected by the same problem, though no so dramatically given the relative strengths of the narrow and broad components. Unfortunately, unless very high S/N profiles (∼100) of the semiforbidden and the He ii lines are obtained, any fitting is hampered by these uncertainties. Finally, it is also, intriguing that the broadening of the He ii emission line in non-accreting TTSs, such as LkCa 4, LkCa 19, and HBC 427, is comparable to that observed in UX Tau or in V836 Tau, which are TTSs with low accretion rates. In this respect, it is worth noticing that the correlation between the He ii flux and the accretion luminosity, as derived from the U-band excess (Ingleby et al. 2009; Gullbring et al. 1998), is mild,⁴ as shown in Figure 12 (see also GdCMA2012).

Zoom In Zoom Out Reset image size

Taking into account all these facts, as well as the good correlation between the He ii and the N v fluxes (Figure 9) and the convergence of the He ii and N v line broadenings toward the non-accreting WTTSs, a contribution to the narrow component of the He ii line from the stellar atmosphere cannot be neglected. For all the reasons mentioned above, it is most probable that both physical components, accretion shocks and the stellar atmosphere, contribute to the flux.

4.2. Properties of the LDC

The UV radiation from N v, He ii, and O iii] is dominated by the contribution of three main components of the TTSs atmospheric/magnetospheric environment: the magnetospheric flow of infalling matter, the disturbed upper atmospheric layers, and the accretion shock.

The diffuse magnetospheric plasma (LDC) is best traced by the O iii] line: it produces asymmetric profiles, preferentially redshifted. The He ii emission is observed from the same kinematical structures that radiate in O iii]. There are indications that photoionization processes could be significant in the excitation of this plasma.

The hot, dense layers of the cool stars' atmospheres are best traced by the N v line. However, the line dispersion increases steadily from WTTSs to CTTSs, suggesting a connection between the line formation region and the accretion flow.

The He ii flux is dominated by a narrow emission component of uncertain origin. In this work, we have presented evidence indicating that it may be formed in hot postshock material in accretion shocks, but we have also presented evidence of its connection with atmospheric tracers. Both accretion shocks, and the upper atmospheric layers seem to contribute to the narrow component of the He ii profile.

The anticorrelation between X-ray and UV flux found by GdCMA2012 has been confirmed, suggesting that the dissipation of magnetic energy proceeds in the TTSs differently than in main sequence stars. The denser environment produced by mass accretion (see, i.e., Petrov et al. 2011) seems to favor the UV channel for the dissipation of the magnetic energy excess.

All the observations indicate that the UV radiation field during PMS evolution is much harder than that usually implemented in the modeling of protostellar disk evolution. An example of its effect in the dust grain charging and charging profile can be found in Pedersen & Gómez de Castro (2011). Theoretical modeling of protostellar disk chemistry and life generation environments should take this fact into account.

Kevin France brought to my attention the HST/COS data corresponding to AA Tau. The analysis of the data pointed out that the 1640 Å jump in the HST/ACS data was dominated by an unexpectedly strong He ii line. This article has grown from extending this analysis to the TTSs observed with HST/COS. I would like to thank an anonymous referee for suggesting the use of the H₂ lines to set the zero of the wavelength scale. This work has been partly funded by the Ministerio de Economia y Competitividad of Spain through grant AYA2011-29754-C03-C01.

Facilities: HST (COS, STIS) - Hubble Space Telescope satellite, XMM (EPIC) - Newton X-Ray Multimirror Mission satellite

The narrow emission component of the He ii line seems to be systematically redshifted with respect to the H₂ lines. The shifts are small (∼0.16 Å or 29 km s⁻¹), as shown in Table 7. The zero of the wavelength scale has been set with the (R(3) 1–7) 1489.636 Å and (P(5) 1–7) 1504.845 Å H₂ lines because they were observed in most of the stars. These lines are far from the 1640 Å He ii line, and in principle, small uncertainties in the wavelength calibration could drive to these shifts. However, as shown in Figures 13 and 14, this is not the case. The P(17)3–9 line is plotted for the whole sample in Figure 13, though only is clearly observed in DM Tau, UX Tau A, AA Tau, GM Aur, and DR Tau (the H₂ emission from RW Aur cannot be used for this purpose). In all cases, the H₂ transition is at rest. In Figure 14, the C i [uv1] multiplet is plotted. It is observed as narrow emission lines in DN Tau, DM Tau, UX Tau A, and DF Tau. In all cases, the lines are at rest wavelength. This redshift is observed both in CTTSs and intermediate objects (it cannot be measured in WTTSs because of the lack of H₂ emission). Thus, unless the P(14) 3–10 H₂ transition, which is blended with the He ii line, is unusually strong, it must be concluded that the narrow emission component of the He ii line is redshifted in most sources. However, note that the blending with the broad component could produce an apparent redshift if this broad component is asymmetric.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size