XMM-Newton X-ray Observations of LkCa 15: A T Tauri Star With a Formative Planetary System
aa r X i v : . [ a s t r o - ph . S R ] M a r To appear in the Astrophysical Journal
XMM-Newton X-ray Observations of LkCa 15: A T Tauri StarWith a Formative Planetary System
Stephen L. Skinner and Manuel G¨udel ABSTRACT
High-resolution ground-based images of the T Tauri star LkCa 15 have re-vealed multiple companions that are thought to comprise a formative planetarysystem. The candidate protoplanets orbit at distances ∼
15 - 20 AU within thedust-depleted inner region of the circumstellar disk. Because of its young age( ∼ Chandra observation in 2009. We report here new results obtained from a deeper 37 ks
XMM-Newton observation in 2014. The new data provide better sampling in thetime domain and improved sensitivity at low energies below 1 keV. Spectral fitswith thermal emission models require at least two temperature components atkT cool ≈ hot ≈ hot is about a factor of twoless than inferred from Chandra , suggesting that the hot-component temperatureis variable. The best-fit absorption column density is in good agreement with thatexpected from optical extinction estimates A V ≈ x (0.2 - 10 keV) = 3 × ergs s − . Estimates of the X-rayheating rate of the inner disk and protoplanets are sensitive to the assumed diskgas surface density for which recent ALMA observations give estimates Σ ,gas ∼ g cm − at 1 AU from the star. At such densities, X-ray heating is confinedmainly to the upper disk layers and X-ray penetration through the disk midplaneto the protoplanets at r ≈
15 - 20 AU is negligible.
Subject headings: stars: individual (LkCa 15; NSVS 6777197) — accretion, ac-cretion disks — stars: pre-main sequence — X-rays: stars CASA, Univ. of Colorado, Boulder, CO, USA 80309-0389; [email protected] Dept. of Astronomy, Univ. of Vienna, T¨urkenschanzstr. 17, A-1180 Vienna, Austria;[email protected]
1. Introduction
Although several thousand exoplanets have now been discovered , examples of exoplan-ets orbiting young pre-main sequence (PMS) stars are rare. The identification of planet-hosting young stars of ages a few Myr and study of their circumstellar disks provide valuableinsight into the circumstellar environment in which planets (and planetary systems) formand how the planets affect disk properties such as gas and dust distribution. As such, obser-vational studies of disks and exoplanets around PMS host stars provide crucial constraintson planet-formation models.Perhaps the most striking example to emerge so far of a PMS star hosting a protoplanet,and possibly even a protoplanetary system, is the accreting classical T Tauri star (cTTS)LkCa 15 in the Taurus star-forming region (Table 1). In a remarkable discovery, Kraus &Ireland (2012, hereafter KI12) reported the direct detection using infrared masked apertureinterferometry of a suspected protoplanet at a projected separation of 71.9 ± ∼
16 - 20 AU. Further monitoring in the near-IR has detectedpossible orbital motion of the protoplanet (Ireland & Kraus 2014).Extensive and ongoing studies of the disk surrounding the host star LkCa 15 at infrared,(sub)millimeter, and radio wavelengths show that it is severely depleted of dust inside aradius of ∼
45 - 50 AU (Andrews et al 2011a,b - hereafter A11a,b; Isella et al. 2012, 2014;Espaillat et al. 2008, 2010; Thalmann et al. 2014, 2016). Although the inner region isdust-depleted, there is still gas present as revealed by CO and CO observations (Pi´etuet al. 2007; van der Marel et al. 2015). Also, Isella et al. (2014) detected a compact 7mm continuum source with the Very Large Array (VLA) at the position of the central star.They conclude that the 7 mm emission is not consistent with a stellar photospheric originbut could be due to either millimeter size grains near the star or ionized gas in the vicinityof the star. Plausible explanations for the dust-clearing in the inner disk are variable dustgrain sizes and opacity (Isella et al. 2012) or dynamical clearing by one or more orbitingobjects (A11a,b), with the latter explanation currently favored.The possibility that more than one protoplanet might be present has received somerecent support from new high-contrast near-IR imaging and adaptive optics H α imaging ofLkCa 15 obtained by Sallum et al. (2015). They report the detection of three distinct objectswith best-fit orbital semi-major axes in the range 14.7 - 18.6 AU. One of these objects is For a current catalog of exoplanets, see the
NASA Exoplanet Archive database athttp://http://exoplanetarchive.ipac.caltech.edu. a = 18.6(+2.5, − α detection ofan object orbiting closer to the star at a = 14.7 ( ± a ∼
18 AU was detected at L ′ -band onlyand its properties are not yet well-constrained. Studies of the inner disk and candidateprotoplanets are ongoing. High spatial resolution scattered light images of the inner diskof LkCa 15 based on J-band imaging polarimetry have now been obtained by Thalmannet al. (2016). Their images show structure from scattering material at the positions of thecandidate protoplanets which they argue could be responsible for some of the signal reportedin previous aperture-masking observations.The likely presence of a formative planetary system around LkCa 15 provides an un-precedented opportunity to study the early stages of planet formation in detail at higheffective spatial resolution given the modest distance of ∼
140 pc to the Taurus star-formingregion. Our study presented here focuses on the effects of X-ray irradiation by the centralstar on the gas-dominated inner disk. We detected LkCa 15 as a bright X-ray source ina previous
Chandra observation and provided initial estimates of the X-ray ionization andheating rates in the inner disk (Skinner & G¨udel 2013, hereafter SG13).We report here the results of a more recent X-ray observation of LkCa 15 obtainedwith
XMM-Newton . These observations provide better sensitivity at low energies below 1keV where X-ray absorption due to intervening gas becomes important. The improved low-energy sensitivity provides tighter constraints on the absorption toward the star as measuredby the equivalent neutral hydrogen column density N H , and the intrinsic (unabsorbed) X-rayluminosity (L x ). Fits of the XMM-Newton spectra confirm the earlier
Chandra result thatcool and hot plasma components are present but the temperature of the hot component islower than was measured with
Chandra and variability of the hotter plasma seems likely.X-ray ionization and heating rates of the inner disk are recomputed based on the new
XMM-Newton results and revised inner disk model parameters from recent ALMA observations.The importance of the effects of X-ray and extreme-ultraviolet (EUV) radiation ondisks and protoplanetary systems around young stars has been noted in several previousstudies of which two examples are Igea & Glassgold (1999; hereafter IG99) and Cecchi-Pestellini, Ciaravella, & Micela (2006). X-ray and EUV emission ionizes and heats disk gas(especially in the outer surface layers), affects disk chemistry, accretion, and mass-loss (viaphotoevaporation), and strengthens the coupling between the accretion disk and the stellarmagnetic field. Since X-rays influence mass-loss and disk dissipation they are one of thefactors that constrain the timescale for planet formation. 4 –Table 1. Properties of LkCa 15
Property Value Refs.Sp. type K5 ± ∗ (M ⊙ ) 1.0 4R ∗ (R ⊙ ) 1.6 2T eff (K) 4730 1A V (mag) 0.62; 1.3 - 1.7 1,2 i disk (deg.) 50.5 - 52 5M disk (M ⊙ ) ≈ ∗ (L ⊙ ) 0.74 - 1.2 4,8log L x (ergs s − ) 30.47 [30.40-30.52] 9,10Note. — Refs. (1) Kenyon & Hartmann1995 (2) Espaillat et al. 2010 (3) Kraus & Hil-lenbrand 2009 (4) Simon et al. 2000 (5) Isella etal. 2012 (6) van der Marel et al. 2015 (7) Torreset al. 2009 (8) Andrews et al. 2011b (9) Skinner& G¨udel 2013 (10) this work
2. Previous Chandra Observation
We obtained a 9.8 ks
Chandra observation (ObsId 10999) of LkCa 15 on 27 December2009 using the ACIS-S (Advanced CCD Imaging Spectrometer) array. Results were presentedby SG13 and are briefly summarized here. LkCa 15 was detected with 590 net counts (0.2 - 8keV). No statistically significant variability was present in the X-ray light curve. Acceptablespectral fits were obtained using a two-temperature (2T) apec thermal plasma model andabundances typical of TTS in Taurus. The best-fit model gave an absorption column densityN H = 3.7 [2.4-5.1] × cm − and plasma temperatures kT cool = 0.30 [0.25 - 0.37 keV],kT hot = 5.1 [3.0 - 13.6] keV, where brackets enclose 90% confidence intervals. The intrinsic(unabsorbed) luminosity was log L x (0.3 - 10 keV) = 30.4 ergs s − at an assumed distance of140 pc.
3. XMM-Newton Observation
The
XMM-Newton observation (ObsId 0722340101) began on 20 February 2014 at 19:12UTC and ended on 21 February at 06:02 UTC. Data were acquired with the European PhotonImaging Camera (EPIC) in Full-Frame mode using the Medium optical blocking filter. EPICprovides charge-coupled device (CCD) imaging spectroscopy from the pn camera (Str¨uderet al. 2001) and two nearly identical MOS cameras (MOS1 and MOS2; Turner et al. 2001).The EPIC cameras have energy coverage in the range E ≈ ≈
20 - 50. The MOS cameras provide the best on-axis angular resolution with FWHM ≈ ′′ at 1.5 keV.Data were reduced with the XMM-Newton
Science Analysis System (SAS vers. 15.0)using standard procedures including the filtering of raw event data to select good eventpatterns and removal of data within time intervals of high background radiation. The us-able exposures obtained after removing high background intervals and total exposures (inparentheses) were 27.8 (36.04) ks for pn, 35.8 (37.70) ks for MOS1, and 36.6 (37.67) ks forMOS2. Thus, about 22% of the pn exposure was adversely affected by high background butthe MOS exposures were not severely affected.A circular region centered on LkCa 15 with a radius r = 20 ′′ ( ≈
80% encircled energyat 1.5 keV) was used to extract X-ray light curves and spectra. Background analysis wasconducted on circular source-free regions near the source. The SAS tasks rmfgen and arfgen were used to generate source-specific RMFs and ARFs for spectral analysis. The data wereanalyzed using the HEASOFT
Xanadu software package. 6 –
4. Results
Figure 1 shows the EPIC pn image of LkCa 15 and its surroundings. LkCa 15 isprominently detected at pn centroid position R.A. = 04 h m s .79, decl. = +22 ◦ ′ ′′ .08 (J2000). The previous Chandra observation had somewhat better spatial resolutionand gave an X-ray position R.A. = 04 h m s .787, decl. = +22 ◦ ′ ′′ .28 (SG13). TheseX-ray positions are in good agreement with the HST
GSC v2.3 position R.A. = 04 h m s .787, decl. = +22 ◦ ′ ′′ .26. Evidence for binarity in LkCa 15 has so far not been found(Nguyen et al. 2012).There are no other X-ray sources in the immediate vicinity of LkCa 15. The nearestEPIC source detected by the pipeline processing software lies more than 1 ′ from LkCa 15.The K5 star HD 284589 located 27. ′′ Chandra . The only other bright stellar X-ray source in the EPIC field-of-view is theeclipsing Algol-type binary system NSVS 6777197 (2MASS J04394628+2211503) located ≈ ′ southeast of LkCa 15 (Fig. 1). Its X-ray properties are summarized further below.The EPIC light curves of LkCa 15 are shown in Figure 2. No large-amplitude fluctua-tions or flares are present but there is a slow falloff in the count rate during the observation.Checks for variability on binned background-subtracted broad-band (0.2 - 8 keV) light curvesusing the χ test give a high variability probability. Using 1000 s bins, the probability ofvariability is P var (0.2 - 8 keV) = 0.96 (pn), 0.99 (MOS1), and 0.98 (MOS2).Figure 3 compares the EPIC pn spectrum with the previous Chandra
ACIS-S spectrumand shows an overlay of the two EPIC MOS spectra. A notable difference is that there isno significant emission below ≈ ≈ H from the EPIC pn spectrum since lower-energy photonsare more susceptible to absorption. The only line feature clearly visible is the Ne X Ly α lineat 1.02 keV which is seen in both MOS spectra. There is also a weak feature at 1.86 keV inthe pn spectrum that may be Si XIII.Spectra were fitted using the Astrophysical Plasma Emission Code variable-abundance vapec model in XSPEC version 12.8.2 (Smith et al. 2001). Photoelectric absorption wasapplied using the XSPEC wabs model to determine the equivalent hydrogen column density(N H ). A two-temperature (2T) vapec model was required to obtain acceptable fits. Acomparison of fit results for two vapec models using different abundances is given in Table2. Significant improvement relative to the solar-abundance fit is obtained by allowing theabundances of Ne and Fe to vary (model A in Table 2). Very little further improvement inthe fit is obtained by letting the abundances of other metals to deviate from solar values. 7 –Model B uses typical TTS abundances for the Taurus Molecular Cloud (G¨udel et al. 2007;Scelsi et al. 2007), as were adopted in the Chandra analysis (Model C of SG13). The Neabundance inferred from Model A is Ne = 2.4 [2.07 - 2.71] times solar, but the generic Taurusabundances (Model B) keep the value fixed at Ne = 0.83 times solar.The EPIC fit results for Models A and B are overall quite similar. Cool plasma at kT ≈ Chandra
ACIS-S data (SG13). A hotter component at kT ≈ Chandra , which gavekT ≈ = [1.93 - 2.40] keV (modelB), which does not overlap the corresponding ACIS-S range kT = [3.0 - 13.6] keV. Thissuggests that the temperature of the hot component is variable. Model B associates a largerpercentage of the volume emission measure with the cool plasma component, as determinedby the XSPEC normalization parameter ( norm ).The best-fit absorption column density is consistent with that expected from A V = 1.3- 1.7 (Table 1; Espaillat et al. 2010). Using the N H to A V conversion of Vuong et al. (2003),model A gives A V = 1.4 [1.3 - 1.6] and model B yields A V = 1.8 [1.7 - 1.9]. The Gorenstein(1975) coversion gives A V values that are about 25% lower than above. Thus, there is noconvincing evidence for excess X-ray absorption above that expected from A V . The ratio ofX-ray to stellar luminosity from the EPIC fits is log L x (0.2 - 10 keV)/L ∗ = − − NSVS 6777197 : The star NSVS 6777197 was serendipitously detected as a bright X-raysource by EPIC pn (Fig. 1) and MOS2, but was outside the MOS1 field-of-view. It wasclassified as a detached Algol binary with a 3.928 d period by Drake et al. (2014). Algol-type binaries usually have one component of late spectral type and are often detected asbright coronal X-ray sources (Singh, Drake, & White 1996). Large X-ray flares can occuras seen in the prototype Algol (Schmitt & Favata 1999). But the EPIC X-ray light curvesof NSVS 6777197 show no significant variability. Simultaneous spectral fits of the pn andMOS2 spectra with a 2T vapec thermal plasma model give an absorption column density N H = 1.78 [1.57 - 2.04] × cm − , kT cool = 0.92 [0.79 - 1.04] keV, kT hot = 2.60 [2.23 - 3.26]keV, and iron abundance Fe = 0.22 [0.10 - 0.36] solar, where brackets enclose 90% confidenceintervals. The absorbed (and unabsorbed) fluxes are F x (0.2-10 keV) = 7.2 × − (1.10 × − ) ergs cm − s − . 8 –Table 2. XMM-Newton
Spectral Fits for LkCa 15
ParameterModel A BEmission a Thermal (2T) Thermal (2T)Abundances non-solar c non-solar d N H (10 cm − ) 0.23 [0.21 - 0.25] 0.29 [0.27 - 0.31]kT (keV) 0.41 [0.39 - 0.44] 0.40 [0.39 - 0.42]kT (keV) 2.39 [2.20 - 2.59] 2.08 [1.93 - 2.40]norm (10 − ) b (10 − ) b χ /dof 862.0/634 850.4/636 χ red X (10 − ergs cm − s − ) 0.55 (1.13) 0.54 (1.43)F X , (10 − ergs cm − s − ) 0.23 (0.64) 0.24 (0.92)F X , (10 − ergs cm − s − ) 0.32 (0.49) 0.29 (0.51)log L X (ergs s − ) 30.42 30.52log L X , (ergs s − ) 30.18 30.33log L X , (ergs s − ) 30.06 30.08log [L X /L ∗ ] − − vapec optically plasma models.The tabulated parameters are absorption column density (N H ), plasmaenergy (kT), and XSPEC component normalization (norm). Abun-dances are referenced to the solar values of Anders & Grevesse (1989).Square brackets enclose 90% confidence intervals. The total X-ray flux(F X ) and fluxes associated with each model component (F X , i ) are theabsorbed values in the 0.2 - 10 keV range, followed in parentheses byunabsorbed values. The total X-ray luminosity L X and luminosities ofeach component L X , i are unabsorbed values in the 0.2 - 10 keV rangeand assume a distance of 140 pc. A value L ∗ = 1.0 L ⊙ is adopted basedon an average of values given in the literature. a Models A and B are of form: N H · (kT + kT ) b For thermal apec models, the norm is related to the volume emissionmeasure (EM = n e V) by EM = 4 π d cm × norm, where d cm is thestellar distance in cm. At d = 140 pc this becomes EM = 2.34 × × norm (cm − ). c All abundances were held fixed at their solar values except for Neand Fe which were allowed to vary and converged to Ne = 2.37 [2.07 -2.71] and Fe = 0.45 [0.39 - 0.53] relative to their solar values. d Abundances were held fixed at typical values for TTS in Taurus(G¨udel et al. 2007; Scelsi et al. 2007). These are (relative to solar): H= 1.0, He = 1.0, C = 0.45, N = 0.79, O = 0.43, Ne = 0.83, Mg = 0.26,Al = 0.50, Si = 0.31, S = 0.42, Ar = 0.55, Ca = 0.195, Fe = 0.195, Ni= 0.195.
5. Discussion5.1. X-ray Heating and Ionization
Disk X-ray heating is due to fast electrons ejected by atoms during X-ray ionization.A complete discussion of the methodology for computing the X-ray ionization rates ( ζ ) andheating rates (Γ x ) in the disk can be found in the Chandra study of LkCa 15 (SG13) andthe references cited below. We have recomputed the disk ionization and heating rates forLkCa 15 based on plasma temperature (kT x ) and X-ray luminosity (L x ) values determinedfrom the XMM-Newton spectral fits (Table 2). The revised rates for the cool and hot plasmacomponents are summarized in Table 3.
A disk model must be adopted in order to compute X-ray ionization and heating rates.A key disk parameter is the density of H-nuclei (n H ), which is required to compute the X-rayabsorption and heating rate per unit volume at a given position in the disk. The value of n H isdetermined by the disk gas surface density Σ which is not yet well-constrained observationallyin the inner disk because of limitations on telescope angular resolution, especially at (sub)-mm wavelengths.A cylindrical coordinate system is used to specify positions ( r , z ) in the disk where theradial coordinate r is the distance in the midplane from the center of the star to the specifiedpoint and z is the height above the disk midplane. Azimuthal symmetry is assumed. Weadopt a radial disk temperature profile of the form T( r ) = 400( r /1 AU) − . K based ona stellar effective temperature T eff = 4730 K. Vertical temperature gradients are ignored(IG99). For the above temperature relation and adopted stellar parameters (Table 1) thedisk scale height at r = 1 AU is H ≡ H( r = 1 AU) = 7.1 × cm and scales as H( r ) ∝ r +1 . .We adopt a simple power-law for the gas surface density Σ( r ) = Σ ( r /1 AU) q , normalizedto Σ = 10 g cm − and a power-law exponent q = −
1. This simple power-law ignores theadditional exponential decay term included in the surface density profiles of A11b and vander Marel et al. (2015) since its value is near unity for the inner disk radii r < ∼
15 AUof interest here. The above value of Σ is comparable to that determined from the recentstudies of Manara et al. (2014) and van der Marel et al. (2015) but must be interpreted asan order-of-magnitude estimate due to several factors that limit our knowledge of the gasdistribution in the spatially-unresolved inner disk (Sec. 5.2). We also note that the value of 10 –Σ adopted here is much higher than the value Σ = 10 − g cm − used in SG13 that wasbased on the earlier LkCa 15 disk study of A11b (Sec. 5.2).We adopt the abundance ratio He/H = 0.1 by number and assume that hydrogen inthe disk is predominantly molecular. The mass density at the midplane is ρ ( r , z =0) =0.4Σ( r )/H( r ) which at r = 1 AU gives ρ ≡ ρ ( r =1 AU, z =0) = 5.63 × − g cm − . Thenumber density of H-nuclei is given by n H = ρ /( µ m p ) where m p is the proton mass and µ =1.42 for H-nuclei (Glassgold et al. 2004). Calculation of the X-ray ionization and heating rates in the disk is based on the analyticresults of Glassgold et al. (1997a, hereafter G97a; Glassgold et al. 1997b), IG99, Shang etal. (2002; hereafter S02), and Glassgold et al. (2004). The X-ray emission is modeled asa thermal plasma with a characteristic temperature T x . The ionization rate scales as theinverse-square of the distance from the source according to (eq. [3.9] of S02): ζ ≈ ζ x (cid:20) rR x (cid:21) − (cid:20) kT x ǫ ion (cid:21) I p ( τ x , ξ ) s − (per H nucleus) . (1)In the above, R x fixes the height of the X-ray source above (or below) the disk center tomimic X-ray production in coronal loops. We use R x = 4R ∗ = 6.4 R ⊙ = 4.45 × cm asin previous studies (G97a, S02, SG13) but the results are not very sensitive to the adoptedvalue of R x (IG99). The energy required to produce an ion pair is ǫ ion ≈
37 eV. The term I p ( τ x , ξ ) accounts for X-ray attenuation as a function of optical depth τ x , where the latterdepends on the photon energy E and position in the disk. The term ξ = E /kT x applies alow-energy cutoff E to the X-ray spectrum to account for wind absorption (IG99, S02). Weuse E = 0.1 keV here, as in previous work (S02, SG13). Figure 3 of SG13 shows the effectof decreasing the cutoff energy to E = 0.01 keV.The primary ionization rate ζ x is given by (S02): ζ x = L x σ ( kT x )4 πR x kT x = 1 . × − (cid:20) L x erg s − (cid:21) (cid:20) kT x keV (cid:21) − ( p +1) (cid:20) R x cm (cid:21) − s − (2)where σ ( kT x ) = σ ( E ) is the energy-dependent photoelectric X-ray absorption cross-sectionper H nucleus. It is evaluated using the expression σ ( E ) = σ (E/1 keV) − p cm − where σ = 11 –2.27 × − cm and p = 2.485 for solar-abundance disk plasma (G97a). Numerical valuesof ζ x for the cool and hot plasma components are given in Table 3 notes.The X-ray optical depth used to compute the attenuation factor I p ( τ x , ξ ) is τ x ( r, z, E ) = (cid:20) rR x (cid:21) σ (E)N ⊥ , disk ( r, z ) (3)where the vertically-integrated column density from infinity down to the height z abovethe disk midplane is N ⊥ , disk ( r, z ) = Z ∞ z n H (r , z)dz cm − . (4)In general, τ x will be smaller for the hot component at a given point in the disk, allowingthe harder emission from the hot component to penetrate deeper into the disk. Note thatfor a given energy E and height z , the optical depth τ x is independent of r as the result of acancellation that occurs for the adopted surface density profile Σ( r ) ∝ r − . Also, it is worthemphasizing that the height above the midplane corresponding to τ x = 1 depends sensitivelyon the adopted value of the gas surface density Σ as shown in Figure 4. At a given distancefrom the star, smaller adopted values of Σ correspond to lower scale heights and deeperX-ray penetration into the disk. The ionization rate for the cool and hot components as afunction of τ x is plotted in Figure 5.The X-ray heating rate per unit volume is (G12)Γ x = ζ n H Q (5)where Q is the heating rate per ionization. Several processes can affect the heating rateas discussed by G12. For predominantly molecular disk gas at r ≥ Q = 17 eV = 2.72 × − ergs, asin previous studies (SG13, G12). The X-ray heating rate then becomesΓ x = 2 . × − ζ n H ergs s − cm − (6)The heating rates at r = 1 AU given in Table 3 can be scaled to other radii using thescaling relations for ζ and n H given in Table 3 notes. For the specific model adopted in thisstudy the heating rate scales according to Γ x ∝ r − . as illustrated in Figure 6. As apparentfrom Figure 6 and Table 3, the X-ray heating is mostly due to the cool component as a 12 –result of its larger absorption cross-section and higher X-ray luminosity (L x, ). But for thedisk model adopted here, X-ray ionization and heating are confined mainly to upper layersseveral scale heights above the midplane (Figs. 4 and 5). For the X-ray photon energies E < ∼ XMM-Newton spectrum, gas surface densities Σ < − g cm − would be required in order for photons to penetrate to deeper layers within onescale height of the midplane (Fig. 4). Although such low surface densities were surmised insome early disk models of LkCa 15 (A11b), more recent ALMA observations (van der Marelet al. 2015) suggest higher gas densities Σ ∼ a few hundred g cm −
13 –Table 3. X-ray Ionization and Heating Rates (LkCa 15) r z/H E Σ n H N ⊥ ζ Γ x (AU) (keV) (g cm − ) (cm − ) (cm − ) (s − ) (ergs s − cm − )1 5.0 0.4 100 9.80e07 1.35e19 2.46e-09 6.56e-121 4.1 2.2 100 5.55e09 9.30e20 2.39e-11 3.61e-12Note. — The ionization rate ( ζ ) and heating rate (Γ x ) for the cool (E = 0.4 keV) andhot (E = 2.2 keV) components are based on the inner disk model discussed in the text(Sec. 5.1.1). The rates are computed at r = 1 AU for the specified value z /H , which isthe number of scale-heights above the midplane corresponding to unit X-ray optical depth( τ x = 1). At r = 1 AU the disk scale height is H ≡ H(r = 1 AU) = 7.1 × cm foran assumed disk midplane temperature T( r = 1 AU) = 400 K. A disk gas surface densityΣ ≡ Σ(r = 1 AU) = 100 g cm − is assumed, but the actual value is uncertain and model-dependent. The quantity n H is the number density of H-nuclei at the specified point ( r , z )in the disk. The vertically-integrated H-nuclei column density from infinity down to thespecified height above the midplane is N ⊥ (eq. [4]). At r = 1 AU, values at the midplane ( z = 0) are n H ( r =1 AU, z =0) = 2.38 × cm − and N ⊥ ( r =1 AU, z =0) = 2.1 × cm − .The primary ionization rates (eq. [2]) for the cool and hot plasma components are ζ x, =2.53 × − s − and ζ x, = 4.31 × − s − , respectively. These values are computed usingthe average kT and L x values for the two models in Table 2, namely kT = 0.4 keV, kT = 2.2 keV, log L x , = 30.26 ergs s − , and log L x , = 30.07 ergs s − . The heating rate Γ x = ζ n H Q (eq. [5]) is computed using Q = 17 eV. Scaling Relations: H( r ) ∝ r +1 . T( r ), orH( r ) ∝ r +1 . for an assumed temperature dependence T( r ) = 400( r /1 AU) − . . Scalingswith radius are Σ ∝ r − , n H ∝ r − . , ζ ∝ r − , Γ x ∝ r − . . At a given radius, n H ( r, z ) = n H ( r, z = 0)exp[-0.5( z/H ( r )) ].
14 –
Several different inner disk models have been proposed for LkCa 15 and the modeladopted in our calculations above is only representative. We have assumed Σ( r ) = Σ ( r /1AU) q with q = − = 10 g cm − . The adopted value of Σ issimilar to those determined in recent studies but is about an order-of-magnitude less thanvalues derived in models of the minimum mass solar nebula (Weidenschilling 1977; Hayashi1981). By way of comparison, Manara et al. (2014) found Σ( r = 1 AU) = 83 g cm − forLkCa 15. Also, substitution of the LkCa 15 disk parameters determined in the ALMA studyof van der Marel et al. (2015) into the surface density profile given in their eq. (1) yields Σ( r = 1 AU) = 286 g cm − . This value includes their best-fit correction factor δ gas = 0.1 whichscales down the gas density in the inner disk (1 < r <
45 AU) relative to the better-knowngas distribution at larger radii. But van der Marel et al. note that the value of δ gas is onlyconstrained to within an order-of-magnitude. Several factors contribute to its uncertaintyincluding inadequate spatial resolution to resolve gas in the inner disk, optical depth effectsin the CO line used to map the disk gas with ALMA, and possible asymmetries in thegas distribution. The much smaller surface density Σ( r = 1 AU) = 10 − g cm − obtainedpreviously by A11b reflects differences in observational data and modeling strategy. TheLkCa 15 disk model proposed in A11b was based on 880 µ m dust continuum observationsobtained with the Submillimeter Array (SMA) and assumed an empty cavity devoid of dustand gas at radii 10 ≤ r ≤
50 AU.The value of the gas surface density power-law exponent q is not well-constrained ob-servationally in the inner disk but q = − q = +2.15 in the inner disk. In this model, the surface density increasesin the inner disk and then starts to turn over at r ∼
60 AU to match the observationally-constrained falloff in the outer disk (Fig. 3 of Isella et al. 2012). The smooth viscous modelcontrasts sharply with the discontinuous drop in Σ( r ) at r ∼
50 AU in the model of A11b.The above differences are relevant to X-ray ionization and heating calculations in the innerdisk where the protoplanets are located since the radial dependence of the density of H-nuclei is a function of the power-law index q as n H ( r ) ∝ r q − . for an assumed temperaturedependence T( r ) ∝ r − . .We have adopted a disk temperature profile T( r ) = T ( r /1 AU) β K with the usualexponent β = − = 400 K for a stellar temperature T eff = 4730 K.Other studies have used T = 100 K (A11a), 200 K (Manara et al. 2014), ∼
250 K (Pi´etu etal. 2007, scaled to r = 1 AU), and 334 K (Bergin et al. 2004, scaled to r = 1 AU and T eff
15 –= 4730 K). The temperature inferred from the ALMA study of van der Marel et al. (2015)is T = 424 K, which is scaled from the dust sublimation radius T( r = 0.08 AU) = 1500 Kusing β = − r ) ∝ T( r ) +0 . . This in turn affects the mass-density ρ and H-nuclei number density n H at the midplane (Sec. 5.1.1). At a given radius, a higherdisk midplane temperature equates to a larger scale height H( r ) and smaller n H . But thedependence is weak and scales as n H ( r ) ∝ T( r ) − . , so a factor of two uncertainty in T translates into only a factor of 1.4 uncertainty in n H at r = 1 AU. It is instructive to compare the X-ray heating rates in Table 3 with the rate expectedfrom cosmic rays. The cosmic ray heating rate for a gas consisting of H and H is (Jonkheidet al. 2004) Γ cr = ζ cr [2 . × − n H + 5 . × − n H ] (erg cm − s − ) (7)where the primary cosmic ray ionization rate in the interstellar medium (ISM) is ζ cr ≈ × − s − . We assume that the LkCa 15 disk is predominantly molecular hydrogen andthe second term in the above expression is negligible. At r = 1 AU the molecular hydrogendensity at scale heights corresponding to τ x = 1 are half the n H values given in Table 3,namely n H = 4.9 × cm − ( z = 5.0 H , 0.4 keV photons) and 2.8 × cm − ( z = 4.1H , 2.2 keV photons). We thus obtain Γ cr ( r =1AU, z =5.0 H ) = 6.12 × − ergs cm − s − and Γ cr ( r =1AU, z =4.1 H ) = 3.48 × − ergs cm − s − . These values are ∼ - 10 timesless than the X-ray heating rates. Thus, in the upper disk layers X-ray heating dominatesand cosmic ray heating is negligible by comparison. But at the midplane, X-ray heatingis negligible for the surface density assumed here (Σ = 10 gm cm − ) whereas cosmic rayheating is Γ cr ( r =1AU, z =0) = 1.49 × − ergs cm − s − .The winds and magnetic fields of T Tauri stars can impede penetration of cosmic raysinto the protoplanetary disk, resulting in values of Γ cr that are less than the value appropriatefor the ISM used above. The effects of T Tauri star winds and magnetic fields on cosmicrays were investigated by Cleeves et al. (2013) who concluded that reduced primary cosmicray ionization rates of Γ cr < ∼ − s − are possible in protoplanetary disks, an order-of-magnitude less than rates typically adopted for the ISM. The effect of such lower cosmic rayionization rates would be to reduce cosmic ray heating near the disk midplane and strengthen 16 –the relative importance of X-ray heating in the upper disk layers.
6. X-ray Irradiation of the Protoplanets and Circumplanetary Disks
The analysis above has focused on X-ray irradiation of the LkCa 15 circumstellar diskbut an assessment of X-ray effects on the atmospheres or circumplanetary disks of the proto-planets is also of interest. At present, little is known about the atmospheric properties of theLkCa 15 protoplanets or their disks. But the detection of H α emission from the innermostprotoplanet at r ≈ r = 1 AU from the star is of order Σ ∼ g cm − , as assumed in our calculationsabove, then little if any stellar X-ray emission reaches protoplanets that are near the midplane( z ≈
0) at distances of r ∼
15 - 20 AU. At a distance r from the star the intrinsic stellar X-rayflux F x is reduced to F x e − τ x / r where where τ x → τ x ( r, z, E) is the energy-dependent X-rayoptical depth through the circumstellar disk evaluated at the target point ( r, z ). Assumingas before Σ( r ) = Σ ( r /1 AU) − and Σ = 10 g cm − then we obtain by Equation (4)N ⊥ , disk (r=15 AU,z=0) = 1.4 × cm − . Using Equation (3) the optical depths for thecool and hot plasma components are computed to be τ x ( r =15 AU, z =0,E=0.4 keV) = 1.6 × and τ x ( r =15 AU, z =0,E=2.2 keV) = 2.3 × . T Tauri stars may undergo intermittentbright X-ray flares during which plasma temperatures can briefly reach peak temperaturesT ∼
100 - 250 MK (E ∼ τ x ( r =15 AU, z =0,E=9 keV) = 685 and τ x ( r =15AU, z =0,E=22 keV) = 74. At these large optical optical depths it is obvious that anyprotoplanets located in or near the circumstellar disk midplane at r > ∼
15 AU will beshielded quite effectively from direct stellar X-ray emission.The main uncertainty in the above τ x calculations originates in the assumed surfacedensity profile Σ( r ) = Σ ( r /1 AU) q . As already mentioned (Sec. 5.2), the normalizationΣ and power-law exponent q are not well-constrained by observations for LkCa 15, and thedependence of Σ( r ) with radius could be more complex than a simple power-law. But in 17 –order for stellar X-rays to penetrate through the circumstellar disk midplane to reach theprotoplanets, much lower surface densities than assumed above would be required. Since τ x ∝ Σ for fixed ( r, z, E), values Σ ∼ − g cm − (as in the model of A11b; see also SG13)would be needed to achieve τ x ( r =15 AU, z =0,E) ∼
1. If the gas density is not currentlythat low it will eventually become so as the star-disk system evolves and the inner-disk gasdissipates.
7. Summary
A pointed observation of LkCa 15 with
XMM-Newton confirms that it is a luminousX-ray source, as previously inferred from a shorter
Chandra observation. Although somelow-level count-rate variability is likely present in the
XMM-Newton
X-ray light curves, nolarge-amplitude variability or flares were detected. The X-ray spectrum is characterizedby a two-temperature thermal plasma with temperatures kT cool ≈ hot ≈ Chandra spectrum, suggesting that the hot component is variable as isoften the case in T Tauri stars. Most of the X-ray ionization and heating of the circumstellardisk is due to the cool component. X-ray heating is restricted mainly to the upper disk layersassuming a disk gas surface density at r = 1 AU of Σ ∼ g cm − , as suggested by recentALMA observations. Cosmic-ray heating is negligible compared to X-ray heating in theupper disk layers but few X-ray photons reach the disk midplane, where cosmic ray heatingdominates. Unless the disk gas surface density is much less than the assumed value Σ ∼ g cm − , little or no stellar X-ray emission is able to penetrate the circumstellar diskmidplane to distances of r ≈
15 - 20 AU where the protoplanets are located.This work was supported by NASA Astrophysics Data Analysis Program (ADAP) awardNNX16AL71G. This work was based on observations obtained with
XMM-Newton , an ESAscience mission with instruments and contributions directly funded by ESA member statesand the USA (NASA). This research has made use of the HEASOFT data analysis softwaredeveleoped and maintained by HEASARC at NASA GSFC.
REFERENCES
Anders, E., & Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197Andrews, S.M., Rosenfeld, K.A., Wilner, D.J., & Bremer, M. 2011b, ApJ, 742, L5 (A11a) 18 –Andrews, S.M., Wilner, D.J., Espaillat, C., et al. 2011b, ApJ, 732, 42 (A11b)Balbus, S.A. & Hawley, J.F. 1991, ApJ, 376, 214Bergin, E., Calvet, N., Sitko, M.L. et al. 2004, ApJ, 614, L133Cecchi-Pestellini, C., Ciaravella, A., & Micela, A. 2006, A&A, 458, L13Cleeves, L.I., Adams, F.C., & Bergin, E.A. 2013, ApJ, 772, 5Drake, A.J., Graham, M.J., Djorgovski, S.G. et al. 2014, ApJS, 213, 9Espaillat, C., Calvet, N., Luhman, K.L., Muzerolle, J., & D’Alessio, P. 2008, ApJ, 682, L125Espaillat, C., D’Alessio, P., Hernan´andez, J. et al. 2010, ApJ, 717, 441Favata, F., Flaccomio, E., Reale, F. et al. 2005, ApJS, 160, 469Franciosini, E., Pillitteri, I., Stelzer, B. et al. 2007, A&A, 468, 485Glassgold, A.E., Najita, J., & Igea, J. 1997a, ApJ, 480, 344 (G97a)Glassgold, A.E., Najita, J., & Igea, J. 1997b, ApJ, 485, 920Glassgold, A.E., Najita, J., & Igea, J. 2004, ApJ, 615, 972Gorenstein, P., 1975, ApJ, 198, 95G¨udel, M. et al. 2007, A&A, 468, 353Igea, J. & Glassgold, A.E. 1999, ApJ, 518, 848 (IG99)Hayashi, C. 1981, Supp. Progress Theor. Phys., 70, 35Ireland, M.J. & Kraus, A.L. 2014, in Exploring the Formation and Evolution of PlanetarySystems, Proc. IAU Symposium No. 299, eds. M. Booth, B.C. Matthews, & J.R.Graham (Cambridge U. Press), 199Isella, A., Chandler, C.J., Carpenter, J.M., P´erez, L.M., & Ricci, L. 2014, ApJ, 788, 129Isella, A., P´erez, L.M., & Carpenter, J.M. 2012, ApJ, 747, 136Jonkheid, B., Faas, F.G.A., van Zadelhoff, G.-J., & van Dischoeck, E.F. 2004, A&A, 428,511Kenyon, S.J. & Hartmann, L. 1995, ApJS, 101, 117 19 –Kraus, A.L. & Hillenbrand, L.A. 2009, ApJ, 704, 531Kraus, A.L. & Ireland, M.J. 2012, ApJ, 745, 5 (KI12)Manara, C.F., Testi, L., Natta, A., Rosotti, G., Benisty, M., Ercolano, B., & Ricci, L. 2014,A&A, 568, A18Mulders, G.D., Dominik, C., & Min, M. 2010, A&A, 512, A11Nguyen, D.C., Brandeker, A., van Kerkwijk, M.H., & Jayawardhana, R.J. 2012, ApJ, 745,119Pi´etu, V., Dutrey, A., & Guilloteau, S. 2007, A&A, 467, 163Sallum, S., Follette, K.B., Eisner, J.A. et al. 2015, Nature, 527, 342Scelsi, L., Maggio, A., Micela, G., Briggs, K., & G¨udel, M. 2007, A&A, 473, 589Schmitt, J.H.M.M. & Favata, F. 1999, Nature, 401, 44Shang, H., Glassgold, A.E., Shu, F.H., & Lizano, S. 2002, ApJ, 564, 853 (S02)Simon, M., Dutrey, A., & Guilloteau, S. 2000, ApJ, 545, 1034Singh, K.P., Drake, S.A., & White, N.E. 1996, AJ, 111, 2415Skinner, S.L. & G¨udel, M. 2013, ApJ, 765, 3 (SG13)Smith, R.K., Brickhouse, N.S., Liedahl, D.A., & Raymond, J.C. 2001, ApJ, 556, L91Str¨uder, L. et al., 2001, A&A, 365, L18Telleschi, A., G¨udel, M., Briggs, K.R., Audard, M., & Palla, F. 2007, A&A, 468, 425Thalmann, C., Mulders, G.D., Hodapp, K. et al. 2014, A&A, 566, 51Thalmann, C., Janson, M., Garufi, A. et al. 2016, ApJ, 828, L17Torres, R.M., Loinard, L., Mioduszewski, A.J., & Rodriguez, L.F., 2009, ApJ, 698, 242Turner, M.J.L. et al., 2001, A&A, 365, L27van der Marel, N., van Dishoeck, E.F., Buderer, S., P´erez, L., & Isella, A. 2015, A&A, 579,A106 20 –Vuong, M.H., Montmerle, T., Grosso, N., Feigelson, E.D., Verstraete, L., & Ozawa, H. 2003,A&A, 408, 581Weidenschilling, S.J. 1977, Ap&SS, 51, 153
This preprint was prepared with the AAS L A TEX macros v5.2.
21 – +22d 30m 00s+22d 15m 00s04h 40m 00s 04h 39m 00s
NSVS 6777197
XMM-Newton (EPIC pn) - LkCa 15
Fig. 1.— Broad-band (0.2 - 8 keV) lightly-smoothed EPIC pn image of LkCa 15 and sur-rounding region. The image has been time-filtered to remove intervals affected by backgroundflares. The Algol-type binary NSVS 6777196 is marked near the south edge of the pn field-of-view. The bright source 4.2 ′ north of LkCa 15 is classified as non-stellar in the HST v. 2.3Guide Star Catalog (GSC J043916.97+222515.59). The image is displayed on a log intensityscale with N up and E to left. The coordinates are equinox J2000. 22 – . . . C oun t s / s LkCa 15 (0.2 − 8 keV) PN . . C oun t s / s MOS10 10 . . C oun t s / s Time (s)MOS2 Bin time: 1000. s
Fig. 2.— Background-subtracted EPIC light curves of LkCa 15 in the 0.2-8 keV rangeextracted from a circular region of radius 20 ′′ centered on the source. Binned at 1000 sintervals. Intervals of high background have been removed. Times are relative to 19:00 UTCon 2014 Feb. 20. 23 – . . . . C t s / s / k e V Energy (keV)LkCa 15 EPIC pn [min 30] vs. ACIS−S [min10]|Ne X |Si XIII<− PNACIS−S −> 10.2 0.5 2 5 . . C t s / s / k e V Energy (keV)LkCa 15 EPIC MOS [min10] |Ne X
Fig. 3.— Background-subtracted X-ray spectra of LkCa 15 obtained with
XMM-Newton
EPIC (ObsId 0722340101) and
Chandra
ACIS-S (ObsId 10999). Intervals of high-backgroundemission have been excluded from the EPIC spectra. ACIS-S background is negligible.Top: EPIC pn (blue, 7568 net counts, binned to a minimum of 30 counts per bin) overlaidwith ACIS-S (red, 590 net counts, binned to a minimum of 10 counts per bin). Possible NeX (1.02 keV) and Si XIII (1.86 keV) lines are marked. Bottom: Overlay of EPIC MOS1(black; 2463 net counts) and MOS2 (red; 2743 net counts) spectra, binned to a minimum of10 counts per bin. A spectral feature visible in both MOS1 and MOS2 identified as Ne X(1.02 keV) is marked. 24 – −3 z / H Σ (g cm −2 ) LkCa 15 . k e V . k e V Fig. 4.— Distance above the disk midplane corresponding to unit X-ray optical depth ( τ x =1) for the cool and hot plasma components as a function of the disk gas surface density Σ at r = 1 AU. The height above the midplane is expressed in units of the scale height H ≡ H( r = 1 AU) = 7.1 × cm for an assumed disk midplane temperature T( r = 1 AU) =400 K. Very low surface densities Σ < − are required for the X-rays to penetrateto depths below one scale height. 25 – − − − − − − − ζ ( s − ) τ x Lk Ca 15 . k e V ( A U ) . k e V ( A U ) |z=5.0H o ||z=4.1H o . k e V ( A U ) Fig. 5.— X-ray ionization rate versus X-ray optical depth for the LkCa 15 disk as computedfor the model adopted in this study (Sec. 5; Table 3). The two solid curves show theionization rates for the cool and hot components evaluated at a radial distance r = 1 AUfrom the star. The dashed curve shows the rate for the cool component at r = 15 AU. Theshort vertical lines mark the scale height corresponding to τ x = 1 assuming a disk gas surfacedensity Σ ≡ Σ( r =1 AU) = 10 g cm − . At r = 1 AU the scale height is H = 7.1 × cm (= 0.048 AU) for an assumed disk temperature T( r =1 AU) = 400 K. The ionization ratefor a specific height z scales as ζ ( r ) ∝ r − . 26 – − − − − − − Γ x ( e r g s c m − s − ) r (AU) LkCa 15 . k e V . k e V Fig. 6.— Disk heating rate per unit volume as a function of radial distance from the star forthe cool and hot plasma components. The heating rate is evaluated at the distance abovethe midplane corresponding to τ x = 1 (see Fig. 4), which is z/H = 5.0 (0.4 keV) and z/H = 4.1 (2.2 keV). For the disk model adopted in this study Γ x ∝ r − .25