Limits on the Gas Disk Content of Two "Evolved" T Tauri Stars
aa r X i v : . [ a s t r o - ph . S R ] D ec Limits on the Gas Disk Content of Two “Evolved” T Tauri Stars
Alan G. Aversa
Steward Observatory, 933 N Cherry Ave., Tucson AZ 85721
ABSTRACT
We derived upper limits of the circumstellar gas disk masses around theT Tauri stars St 34 and RX J0432.8+1735 in order to place constraints ontheories of planet formation and to explore the evolution of the gas-to-dust ratioduring the epoch of disk dissipation around young sun-like stars. Since sub-millimeter lines of CO trace of the cold, outer regions of circumstellar disks,we observed CO J = 2 − ± COexcitation temperature of 20 K, a CO line-width of 5 km s − , and optically-thin emission, we derive 3 σ upper limits on the H circumstellar disk mass forSt 34 and RX J0432.8+1735 to be < .
20 M ⊕ for both disks. Placing theseresults in the context of other studies, we discuss their implications on planetformation models.
1. Introduction
Circumstellar disks of gas and dust, a natural result of the conservation of angularmomentum, are a common outcome of the star formation process. Kenyon & Hartmann(1995) find that over half the low mass ( < ⊙ ) pre–main-sequence T Tauri stars in theTaurus-Auriga star formation region have more infrared emission than expected from anormal stellar photosphere, indicating the presence of a dusty circumstellar disk heated bythe parent star as well as active accretion. T Tauri stars fall into two categories: Weak-Line T Tauri Stars (WTTSs), characterized by low H α equivalent widths, and ClassicalT Tauri Stars (CTTSs), with higher H α equivalent widths indicative of ongoing gas ac-cretion. These circumstellar disks generally have the following properties (Dutrey et al.2007; Andrews et al. 2009): mass surface densities Σ( r ) ∝ r − . , surface temperatures T ( r ) ∝ r − . (depending on disk flaring), and Keplerian rotational velocities V ( r ) ∝ r . . 2 –Gas giant planet formation depends upon the gas content of the circumstellar disk fromwhich they form. Primordial inner disks traced by hot dust disappear after about 3 Myrwith a range of 1 −
10 Myr (Meyer et al. 2007). If gas content and dust content in disksdissipate through similar mechanisms (Damjanov et al. 2007), then we would expect gasto disappear on these same timescales. To understand the timescales of planet formation,we must understand how long a gas disk persists around its parent star. Accretion rates(higher in CTTSs than WTTSs) also trace the time evolution of gas content as gas mustbe present to accrete. Durisen et al. (2007) suggest that gravitational instabilities in gasdisks could account for their rapid ( < ⊙ star due tophoto-evaporation indicate that gas disks disappear in about 6 Myr (Alexander et al. 2006;Dullemond et al. 2007). Because there are various theoretical results for the gas dispersalmechanisms and associated timescales, we might expect to observe diverse properties forgas disks around T Tauri stars.Ideally, sub-millimeter interferometric images of T Tauri stars would yield the mostinformation about circumstellar disks, including their orientation, geometry, and gas con-tent. Yet observations of this sort have only been published for a few nearby stars (e.g.,Sub-Millimeter Array (SMA) observations of TW Hya; Qi et al. 2004). Infrared photomet-ric campaigns with Spitzer , such the Cores to Disks (Evans et al. 2003) and the Formationand Evolution of Planetary Systems (FEPS; Meyer et al. 2006)
Spitzer
Legacy projects,provide knowledge of the dust circumstellar disks based upon IR excesses at many wave-lengths including 24 µ m, which trace dust within a few AU of the parent star, and at 70 µ m,which trace cooler dust at larger radii. Silverstone et al. (2006) surveyed 74 stars with agesbetween 3-30 Myr finding no stars with mid–IR excess that were not gas rich accreting TTauri stars. Padgett et al. (2006) however identified a handful of WTTS (lacking signa-tures of accretion) with evidence for mid– and far–IR excess emission. Photometric IRobservations do not, however, constrain the total gas content in circumstellar disks be-cause of a potentially highly variable gas-to-dust ratio. Observing the rotational energytransitions of CO, a proxy for molecular hydrogen (H ), and assuming the ISM abun-dance ratio [H / CO] ≈ enables one to trace the majority of cold gas in disks out toradii many times larger. Is the timescale for gas dissipation similar to the timescale fordust dissipation around T Tauri stars? This is the central question we address with ournew observations.Expanding on previous works (e.g. Pascucci et al. 2006; van Kempen et al. 2007), wesearch for CO J = 2 − ± . ± . §
2, we describeour selection of sources and observations, then in § § §
2. Observations2.1. Selection of Sources
We observed the CO J = 2 − µ m excesses (Kenyon & Hartmann1995; Padgett et al. 2006, see Table 1). In order to detect gas in their circumstellar disks,we must be able to distinguish emission from the disk and surrounding molecular cloud.From available candidate “evolved” (in age or shape of spectral energy distribution) T Tauristars, we selected those least likely to be contamined by the ambient CO emission from theparent molecular cloud based on the Dame et al. (1987) CO survey. The radial velocitiesof our sources differ from the systemic velocities of any CO emission in the vicinity of oursources by ∼ − − . St 34 (HBC 425) is a binary system of two CTTSs, separated by . .
78 AU (White & Hillenbrand2005) based on the orbital solution for the system (Downes & Keyes 1988) in the Taurus-Auriga T association (Kenyon & Hartmann 1995). Both components of the spectroscopicbinary have roughly equal mass and spectral types of M3 (White & Hillenbrand 2005).White & Hillenbrand (2005) observed St 34 in the optical with the HIRES spectrographat Keck and derived an isochronal age of 8 ± Li) in the spectrum, St 34 must have reached—assuming the stars are completely convective—an internal temperature > × K andsince depleted all of its lithium. St 34 has a low accretion rate of 2 . × − M ⊙ yr − ,and the maximum radial velocity difference between the two binary components of St 34 is58.4 km s − (White & Hillenbrand 2005). St 34, being one of the oldest known pre–main-sequence (PMS) star still accreting from a proto-planetary disk, also has a low dust massof ∼ × − M ⊙ for radii . . Table 1. Candidate Source Summary
Object α δ
Heliocentric RV Source V LSRa
Cloud V LSRb T eff log L ⋆ Mass Age General(J2000) (J2000) (km s − ) (km s − ) (km s − ) (K) (log L ⊙ ) ( M ⊙ ) (Myr) ReferenceRX J0432.8+1735 04 32 53.23 17 35 33.68 18 . c ∼ − . × − · · · ± . d aSt 34 04 54 23.70 17 09 54.00 17 . ± . . ± .
06 0 . ± . e ± e ,f bReferences. — For RVs, see Wichmann et al. (2000); White & Hillenbrand (2005). For general references, see (a) Padgett et al. (2006) and (b) White & Hillenbrand(2005). a We corrected the heliocentric radial velocities (RVs) from the literature into local standard of rest ( V LSR ) radial velocities assuming the sun moves toward J2000( α = 18 . h , δ = 30 . ◦
0) at 20 km s − . b Determined by the CO J = 1 − c Wichmann et al. (2000) d D’Antona & Mazzitelli (1997) e For both binary components f Isochronal age is given. The Li depletion age for both binary components is >
25 Myr. g Estimated by integrating the spectral energy distribution (SED) of RX J0432.8+1735 in Padgett et al. (2006)
RX J0432.8+1735 is a WTTS of spectral type M2 (Mart´ın & Magazz`u 1999). Basedon the PMS tracks of D’Antona & Mazzitelli (1997) RX J0432.8+1735 is estimated tobe 1 . ± . Spitzer andnoticed that its 24 µ m flux is in excess of the expected photospheric value by a factor of 3.Its lack of IR excess ≤ µ m suggests there may be a large inner hole in the disk. Basedon ROSAT observations, Carkner et al. (1996) discovered that RX J0432.8+1735 is alsoan X-ray source. As RX J0432.8+1735 is classified as a WTTS star with no estimates ofits accretion rate, we assume it is not accreting.
On 26-27 November 2007, we observed the CO J = 2 − CO J = 2 − CO J = 2 −
1. We used theForbes Filter Bank (FFB) backend in 4 IF mode, an upper and lower sideband each with1 MHz and 250 kHz of spectral resolution, respectively. The channel width, ∆ ν ch , of ourspectrometer was 0 .
33 km s − . The 1 MHz resolution data were used to determine mainbeam efficiencies η mb , and the 250 kHz resolution data were used to measure the COline. Using CLASS in the GILDAS data reduction package, we estimated the main beamefficiencies by observing the planets shown in Table 2. Typical sideband rejections, ignoredin the calibration, were >
10 dB. The main beam efficiency η was computed followingMangum (1993) and corrected for single-sideband observations: η = T ⋆A (planet) J ( ν s , T planet ) − J ( ν s , T cmb ) × (cid:20) − exp (cid:18) − ln (2) θ eq θ pol θ (cid:19)(cid:21) − , (1)where J ( ν, T b ) = hν/ke hνkT mb − T b and frequency ν , T ⋆A is the single-sideband antenna temperature of the planet, T planet is the planet’s observed brightnesstemperature, T cmb = 2 .
73 K, θ eq and θ pol are respectively the planet’s equatorial and 6 –poloidal diameters in arcseconds, and θ mb = 33 ′′ at ν = 230 GHz. We adopted an averageVenus brightness temperature T b from Kuznetsov et al. (1982) of 287 ±
20 K. For all otherplanets’ T b , we used the JCMT online database . We derived a ratio η Vpol /η Hpol of thetwo IF’s mean main beam efficiencies for both nights of 1 . ± .
04. We used this ratioto scale the Hpol polarization’s antenna temperature up to match the level of the Vpolpolarization’s antenna temperature. After fitting a baseline to each spectrum, we averagedthe sum of the scaled Hpol brightness temperatures and the Vpol brightness temperatures: h T ⋆A (Hpol , scaled) + T ⋆A (Vpol) i = T ⋆A (sum). Thus we computed the corrected main beamtemperature as T mb = T ⋆A (sum) η Vpol . (3)The average beam efficiencies were h η Hpol i = 0 . ± .
01 and h η Vpol i = 0 . ± .
01 for bothnights.Since we had null detections for our two sources, we must assume a line-width tocalculate upper limits on the integrated intensity. We assumed typcial a line-width of∆ ν = 10 km s − (= 7 .
69 MHz). If we assume the CO line is well described by a Gaussianline shape, then the uncertainty in the integrated intensity is given by σ I = σ T mb s ν ch ∆ ν √ ln 2 , (4)where σ I and ∆ v are the CO line fluxes and the the full width at half maximum (FWHM)and ∆ v ch is the channel spacing 0.33 km s − ; see Appendix I of Schlingman et al. (in prep.).The observations are summarized in Table 3.
3. Results & Analysis
The main-beam corrected spectra of our observations are shown in Figure 1.While 9 ′ is a rough spatial scale for comparison to the 33 ′′ beam of the SMT, and sincewe did not detect a CO line in any of our sources, the on-cloud results from SMT areconsistent with Dame et al. (1987). It is unlikely that high spatial frequency variations of33 ′′ scales over 9 ′ regions have systemic velocity shifts of 2-3 km − Since we did not detect any CO line, we convert our 3 σ noise into an upper limit onthe flux. A knowledge of the flux will enable us to estimate upper limits on gas disk mass. η Planet η Vpol η Hpol η Vpol /η Hpol
Mars a . ± .
05 0 . ± .
04 1 . ± . . ± .
05 0 . ± .
04 1 . ± . . ± .
04 0 . ± .
03 1 . ± . . ± .
04 0 . ± .
03 1 . ± . . ± .
03 0 . ± .
03 1 . ± . . ± .
03 0 . ± .
03 1 . ± . . ± .
05 0 . ± .
04 1 . ± . b . ± .
05 0 . ± .
04 1 . ± . c . ± .
05 0 . ± .
04 1 . ± . . ± .
05 0 . ± .
04 1 . ± . . ± .
04 0 . ± .
04 1 . ± . . ± .
05 0 . ± .
04 1 . ± . . ± .
03 0 . ± .
03 1 . ± . a All Mars brightness temperature errors assumed tobe 5% b End of first night c Beginning of second night C l oud S t s - -
20 0 20 40 - V LSR H km s - L T m b H K L C l oud R X J . + -
20 0 20 40 - - V LSR H km s - L T m b H K L Fig. 1.— Spectra of St 34 ( left ) and RX J0432.8+1735 ( right ). The vertical line labeled“Cloud” represents the estimated background cloud velocity determined by the Dame et al.(1987) CO J = 1 − V LSR . For St 34, we show with an arrow the difference between the radialvelocities of its binary components with an arrow centered on the systemic velocity. Table 3. Observational Summary
Source Integration Time σ T mb σ I log ( F ν ) N ( T ex = 10 , ,
100 K) M H ( T ex = 10 , ,
100 K)(sec) (K) (K km s − ) (log (W cm − )) (10 cm − ) (M ⊕ )RX J0432.8+1735 3600 0.019 0.046 < − . < . , . , . < . , . , . < − . < . , . , . < . , . , . T mb is the main beam corrected brightness temperature and I is the corresponding intensity assuming a line width of 5 km s − . F isthe 3 σ line flux upper limit.Note. — That σ T mb for both objects is the same is a fluke. I ν over frequency ν and solid angle Ω: F = Z Z I ν dν d Ω =
Z Z kν T b c dv d Ω . (5)Assuming the brightness temperature does not vary substantially over the telescope beamand that the line is a Gaussian with line-width ∆ v , then the upper limit on the 3 σ lineflux F is F < kν (3 σ T mb ) c πθ r π × ∆ v = (5 . × − ) σ T mb erg s − cm − . (6)The 3 σ upper limits on F are listed in Table 3.Similarly, we can derive the column density in the optically thin limit to be N ( T ex ) = 8 πkν hc g u A ul F ( T ex , E u , ν ) Z T mb dv, (7)where F ( T ex , E u , ν ) ≡ J ν ( T ex ) Q ( T ex ) exp (cid:16) E u kT ex (cid:17) J ν ( T ex ) − J ν ( T cmb ) , (8) A ul is the Einstein A coefficient (spontaneous emission) and has units of s − .Similar to the analysis of Pascucci et al. (2006), we assume an excitation temperature T ex ≈
20 K. Then in our case for CO J = 2 − F ( T ex , E u , ν ) ≈ .
00. The Einstein A ul = 6 . × − s − and partition function Q (20 K) = 15 . ).We compute and tabulate in Table 3 the CO number densities and H gas massesin the optically thin limit. Gas disk masses were derived from Scoville et al. (1986), M H < N ( T ex ) × ( (cid:20) H CO (cid:21) µ G m H πθ d ) g = 110 σ T mb M ⊕ , (9)where [H / CO] ≈ , µ G = 1 .
36 is the mean molecular weight, m H is the mass of an H molecule, and d ≈
140 pc is the distance to Taurus.
4. Discussion
Theories of gaseous planet formation require knowledge of (1) how much gas thereis in circumstellar disks initially and (2) the rate at which gas is depleted over time. To
10 –answer the first question, upper limits on the amount of gas in very young protostellar disksystems puts limits directly on the amount of mass available for gaseous planet formationin the systems observed. Answers to the second question, requiring large samples over awide range of ages, are also necessary because of the competition between the timescale forplanet formation and gas disk dispersion timescales (for a recent review, see Meyer 2009).There are several ways of constraining gas disk masses, each with its own advantagesand disadvantages. To understand gas disk timescales, one can also analyze H α emissionline profiles and determine gas accretion rates and thus constrain the gas mass surfacedensity at the inner edge of the disk tracing perhaps global disk evolution (e.g. Fedele et al.2009). Using UV tracers of gas emission, Ingleby et al. (2009) describe HST observationssearching for evidence of hot gas in emission finding no evidence for H emission for WTTSin their sample. Ultraviolet absorption line from a continuum source (e.g. Roberge et al.2005) can help constrain cold mass in disks, but it requires a continuum source that isbright in the far UV and an edge-on geometry; therefore, it is observationally feasible onlyin special circumstances. Near IR fluorescent H traces gas with excitation temperatures T ex > . µ m and 17 µ m Spitzer bands,trace gas up to 50-200 K. Our observations of rotational lines of CO trace cooler gas atlarger orbital radii. The disadvantage of such measurements is that CO can freeze outat the coldest temperatures corresponding to the outer limits of the disk & ± h ˙ M i = 8 . × − M ⊙ yr − , so after 1 Myr muchmore gas than our upper limit of 4 .
20 M ⊕ would accrete ( ∼
276 M ⊕ ). Perhaps St 34recently lost its outer disk through photoevaporation (e.g. Gorti & Hollenbach 2009) andwe are witnessing the “last gasp” of accretion onto the star. This is possible but not likely,as it requires current observations to be taking place at a very special time.Conversely, RX J0432.8+1735—being a much younger, 1 . ± . M ⊕ could still exist in its outer disk. That RX J0432.8+1735is relatively young and does not have detectable gas content could be significant considering 11 –that there are older systems, such as Hen 3-600, a binary system at between 1 −
10 Myr ofage with apparent WTTS and CTTS components (Jayawardhana et al. 1999); TW Hya,a CTTS at 8-10 Myr (Webb et al. 1999); and DM Tau at ∼ ≤
30 Myr from Pascucci et al. (2006) and with the gas massdeterminations ( solid circles ) of BP Tau ( CO J = 2 −
1; Dutrey et al. 2003); DL Tau,DO Tau ( CO J = 2 −
1; Koerner & Sargent 1995); and DM Tau, DR Tau, GG Tau a,GM Aur, GO Tau, LkCa 15, RY Tau ( CO J = 3 − CO J = 3 −
2; Thi et al. 2001).We note that this is not an exhaustive compilation from the literature, but representative ofrecent results. Assuming a 1:10 gas-to-dust ratio (D’Alessio et al. 2005), we would expectRX J0432.8+1735 and St 34 to have at most 0.420 M ⊕ and 0.420 M ⊕ of dust, respectively,for T ex = 20 K. For St 34, Hartmann et al. (2005) estimates a disk mass of 665 M ⊕ located in a circumbinary disks between the “wall” (the region defined to surround the twocomponents of the St 34 binary; ∼ . <
10 AU),optically-thick CO disk with a gas-to-dust ratio of ∼ µ m (Hartmann et al. 2005)and RX J0432.8+1735 has infrared excess for wavelengths longer than 24 µ m (Padgett et al.2006), yet St 34 is a binary CTTS with an accreting inner disk and RX J0432.8+1735is a WTTS. Binaries tend to disrupt inner gas disks (Jensen 1996) and may decreasedisk lifetimes (Monin et al. 2007). However, Armitage & Clarke (1996) have argued thatclose binaries affect angular momentum exchange in the natural evolution of accretiondisks resulting in longer lived outer disks. Indeed Th´ebault et al. (2004) find that planetformation around binaries might require a long-lived but massive disk. Since circumbinarydisks allow for long gas disk lifetimes, St 34 might have had more time to form planets.
5. Conclusions
Assuming optically thin disks ( τ ≪ T ex = 20 K, anda line-width ∆ v = 10 km s − , we do not detect significant amounts of gas around threeT Tauri stars: < .
20 M ⊕ for the PMS binary St 34 and < .
20 M ⊕ for RX J0432.8+1735.St 34, a CTTS, is still accreting gas although it is 8 ± Fig. 2.— Gas circumstellar disk mass versus age of selected sources: our RX J0432.8+1735and St 34 upper limits ( labeled ); upper limits from the Pascucci et al. (2006) sample ( un-labeled upper limits ) with ages ≤
30 Myr (the Kelvin-Hemholtz contraction timescale fora 1 M ⊙ star); and exact mass determinations ( solid circles ) of BP Tau ( CO J = 2 − CO J = 2 −
1; Koerner & Sargent 1995); andDM Tau, DR Tau, GG Tau a, GM Aur, GO Tau, LkCa 15, RY Tau ( CO J = 3 − CO J = 3 −
2; Thi et al. 2001). We assume errors in stellar ages to be 50%. 13 –RX J0432.8+1735, a WTTS of 1 . ± . REFERENCES
Alexander, R. D., Clarke, C. J., & Pringle, J. E. 2006, MNRAS, 369, 229Andrews, S. M., Wilner, D. J., Hughes, A. M., Qi, C., & Dullemond, C. P. 2009, ApJ, 700,1502Armitage, P. J., & Clarke, C. J. 1996, MNRAS, 280, 458Bary, J. S., Weintraub, D. A., & Kastner, J. H. 2003, ApJ, 586, 1136Carkner, L., Feigelson, E. D., Koyama, K., Montmerle, T., & Reid, I. N. 1996, ApJ, 464,286D’Alessio, P., Mer´ın, B., Calvet, N., Hartmann, L., & Montesinos, B. 2005, Revista Mexi-cana de Astronomia y Astrofisica, 41, 61Dame, T. M., et al. 1987, ApJ, 322, 706Damjanov, I., Jayawardhana, R., Scholz, A., Ahmic, M., Nguyen, D. C., Brandeker, A., &van Kerkwijk, M. H. 2007, ApJ, 670, 1337D’Antona, F., & Mazzitelli, I. 1997, Memorie della Societa Astronomica Italiana, 68, 807Downes, R. A., & Keyes, C. D. 1988, AJ, 96, 777Dullemond, C. P., Hollenbach, D., Kamp, I., & D’Alessio, P. 2007, Protostars and PlanetsV, 555Durisen, R. H., Boss, A. P., Mayer, L., Nelson, A. F., Quinn, T., & Rice, W. K. M. 2007,Protostars and Planets V, 607Dutrey, A., Guilloteau, S., & Simon, M. 2003, A&A, 402, 1003Dutrey, A., Guilloteau, S., & Ho, P. 2007, Protostars and Planets V, 495Evans, N. J., II, et al. 2003, PASP, 115, 965 14 –Fedele, D., van den Ancker, M. E., Henning, T., Jayawardhana, R., & Oliveira, J. M. 2009,arXiv:0911.3320Gorti, U., & Hollenbach, D. 2009, ApJ, 690, 1539Guilloteau, S., & Dutrey, A. 1994, A&A, 291, L23Hartmann, L., et al. 2005, ApJ, 628, L147Hughes, A. M., et al. 2010, arXiv:1007.3267Ingleby, L., et al. 2009, ApJ, 703, L137Jayawardhana, R., Hartmann, L., Fazio, G., Fisher, R. S., Telesco, C. M., & Pi˜na, R. K.1999, ApJ, 520, L41Jensen, E. L. N. 1996, Ph.D. Thesis,Kenyon, S. J., & Hartmann, L. 1995, ApJS, 101, 117Koerner, D. W., & Sargent, A. I. 1995, AJ, 109, 2138Kuznetsov, I. V., Fedoseev, L. I., & Shvetsov, A. A. 1982, Radiofizika, 25, 247Mangum, J. G. 1993, PASP, 105, 117Mart´ın, E. L., & Magazz`u, A. 1999, A&A, 342, 173Meyer, M. R., et al. 2006, PASP, 118, 1690Meyer, M. R., Backman, D. E., Weinberger, A. J., & Wyatt, M. C. 2007, Protostars andPlanets V, 573Meyer, M. R. 2009, IAU Symposium, 258, 111Monin, J.-L., Clarke, C. J., Prato, L., & McCabe, C. 2007, Protostars and Planets V, 395Najita, J. R., Carr, J. S., Glassgold, A. E., & Valenti, J. A. 2007, Protostars and PlanetsV, 507Padgett, D. L., et al. 2006, ApJ, 645, 1283Pascucci, I., et al. 2006, ApJ, 651, 1177Qi, C., et al. 2004, ApJ, 616, L11 15 –Roberge, A., Weinberger, A. J., Redfield, S., & Feldman, P. D. 2005, ApJ, 626, L105Schlingman et al., in prep.Scoville, N. Z., Sargent, A. I., Sanders, D. B., Claussen, M. J., Masson, C. R., Lo, K. Y.,& Phillips, T. G. 1986, ApJ, 303, 416Silverstone, M. D., et al. 2006, ApJ, 639, 1138Skrutskie, M. F., et al. 1993, ApJ, 409, 422Th´ebault, P., Marzari, F., Scholl, H., Turrini, D., & Barbieri, M. 2004, A&A, 427, 1097Thi, W. F., et al. 2001, ApJ, 561, 1074van Kempen, T. A., van Dishoeck, E. F., Brinch, C., & Hogerheijde, M. R. 2007, A&A,461, 983Webb, R. A., Zuckerman, B., Platais, I., Patience, J., White, R. J., Schwartz, M. J., &McCarthy, C. 1999, ApJ, 512, L63White, R. J., & Hillenbrand, L. A. 2005, ApJ, 621, L65Wichmann, R., et al. 2000, A&A, 359, 181