Molecular processes from the AGB to the PN stage
aa r X i v : . [ a s t r o - ph . S R ] N ov Planetary Nebulae, an Eye to the FutureProceedings IAU Symposium No. 283, 2012A. Manchado, D. Schonberner, & L. Stanghellini eds. c (cid:13) Molecular processes from the AGB to thePN stage
D. A. Garc´ıa-Hern´andez , Instituto de Astrof´ısica de Canarias, V´ıa L´actea s/n, E-38200 La Laguna, Spainemail: [email protected] Departamento de Astrof´ısica, Universidad de La Laguna (ULL), E-38205 La Laguna, Spain
Abstract.
Many complex organic molecules and inorganic solid-state compounds have beenobserved in the circumstellar shell of stars (both C-rich and O-rich) in the transition phase be-tween Asymptotic Giant Branch (AGB) stars and Planetary Nebulae (PNe). This short ( ∼ -10 years) phase of stellar evolution represents a wonderful laboratory for astrochemistry andprovides severe constraints on any model of gas-phase and solid-state chemistry. One of themajor challenges of present day astrophysics and astrochemistry is to understand the forma-tion pathways of these complex organic molecules and inorganic solid-state compounds (e.g.,polycyclic aromatic hydrocarbons, fullerenes, and graphene in the case of a C-rich chemistryand oxides and crystalline silicates in O-rich environments) in space. In this review, I presentan observational review of the molecular processes in the late stages of stellar evolution with aspecial emphasis on the first detections of fullerenes and graphene in PNe. Keywords.
Astrochemistry, molecular processes, Planetary nebulae, AGB and post-AGB, dust.
1. Introduction
At the end of the Asymptotic Giant Branch (AGB), low- and intermediate-mass starsexperience strong mass loss that efficiently enriches the interstellar medium with gasand dust. The main processes of nucleosynthesis on the AGB are the production ofcarbon and of heavy s-process elements. AGB stars experience a different nucleosynthesisdepending on the progenitor mass and metallicity (e.g., Garc´ıa-Hern´andez et al. 2007aand references therein). In our own Galaxy, low-mass (M < ⊙ ) AGB stars remainO-rich and they probably do not form a Planetary Nebula (PN), while intermediate-mass (1.5 < M < ⊙ ) turn C-rich and they produce s-process elements through the Cneutron source. Finally, high-mass (M > ⊙ ) AGB stars remain O-rich because of theHot Bottom Burning (HBB) activation and they produce s-elements via the Ne neutronsource (Garc´ıa-Hern´andez et al. 2006, 2009). Note that this AGB evolutionary scenariois strongly modulated by metallicity (e.g., 3 rd dredge-up efficiency, HBB activation). Themore massive C-rich and O-rich sources are strongly obscured by the circumstellar dustat the end of the AGB and they experience a phase of total obscuration in their way tobecome PNe, being only accesible in the infrared and millimeter wavelength ranges.More than 60 molecules have been detected in the circumstellar shells around AGBstars (e.g., Herbst & van Dishoeck 2009). Many gas-phase molecules such as inorganics(CO, SiO, SiS, NH , etc.), organics (C H , CH , etc.), radicals (e.g., HCO + ), rings (e.g.,C H ), and chains (e.g., HC N) have been detected around AGB stars. Gas-phase reac-tions cannot explain all of these molecules and solid-state chemistry has to be considered(e.g., molecules can form on dust grains). Indeed, amorphous silicates, weak crystallinesilicates (olivine, pyroxene), refractory oxides (corundum, spinel) have been detectedaround O-rich AGB stars. On the other hand, SiC and amorphous carbon are frequently50 olecular processes AGB-PNe ∼ -10 years) transitionphase between AGB stars and PNe (e.g., Garc´ıa-Lario & Perea-Calder´on 2003; Kwok2004). Thus, post-AGB stars and proto-PNe are wonderful laboratories for astrochem-istry, providing us with crucial information about the formation pathways of complexorganic molecules and inorganic solid-state compounds. Here I present an observationalreview (both in C-rich and O-rich environments) of the molecular processes (i.e., gas-phase molecules and the dust grains that these molecules can form) from AGB stars toPNe as seen by the Infrared Space Observatory (ISO) and the Spitzer Space Telescope,which give us strong constraints to gas-phase and solid-state chemistry. Special attentionis given to the first detections of fullerenes and graphene in PNe.
2. Molecular evolution from AGB to PNe as seen by ISO and Spitzer.
O-rich chemistry.
ISO spectroscopy has revealed that weak crystalline silicate (olivine,pyroxene) and water ice features are only seen in high mass loss rate O-rich (C/O < ∼
1) show a set of O-rich infrared features thatare different to those seen in OH/IR stars (Hony et al. 2009). Another important resultfrom ISO was the detection of strong crystalline silicates in O-rich post-AGB stars andPNe (e.g., Waters et al. 1996; Molster et al. 2002) as well as the identification of double-dust chemistry objects (e.g., cool WCPNe, Waters et al. 1998). Water ice features (bothamorphous and crystalline) were also detected in some post-AGB stars (e.g., Manteigaet al. 2011 and references therein).With the advent of the Spitzer Space telescope, it has been possible to study O-richAGB stars in other galaxies as well as to study the relation between the dust propertiesand metallicity. It is found that O-rich AGBs are less obscured (a lower dust production)in low metallicity environments and the amorphous silicates are always seen in emission(e.g., Sargent et al. 2010). Another important Spitzer result was the identification of the(heavily obscured) high-mass precursors of PNe (the OHPNe). The latter sources showedunusual crystalline silicate features, likely due to the different nucleosynthesis in the pre-vious AGB phase (Garc´ıa-Hern´andez et al. 2007b). In addition, Spitzer has permitted tostudy the total obscuration phase that could not be done by ISO (e.g., Garc´ıa-Hern´andezet al. 2007c). With regard to O-rich PNe, Spitzer found that O-rich PNe are less commonat low metallicity such as in the Magellanic Clouds (MCs, Stanghellini et al. 2007). Also,many O-rich PNe showing amorphous silicates emission have been detected by Spitzer(e.g., G´orny et al. 2010; Stanghellini et al. 2011 and these proceedings).In short, the evolution of the O-rich dust features proceeds from amorphous (in theAGB phase) to crystalline silicates (in the PNe stage) (Garc´ıa-Lario & Perea-Calder´on2003). Two models for the crystallization of the silicates have been proposed: i) high-temperature crystallization at the end of the AGB as a consequence of the strong massloss (Waters et al. 1996) or ii) low-temperature crystallization in long-lived circumbinarydisks (Molster et al. 1999).
C-rich chemistry.
Gas-phase organic molecules (e.g., C H , HCN, etc.) around theprototype C-rich AGB star IRC +10216 were first detected by ISO (Cernicharo et al.2 Garc´ıa-Hern´andez1999). These organic molecules are usually observed together with a strong dust contin-uum emission and a broad dust feature centered at ∼ µ m (e.g., Yang et al. 2004).The strong dust continuum emission is usually attributed to amorphous carbon while the11.5 µ m feature is believed to be produced by SiC (e.g., Speck et al. 2009). An uniden-tified broad feature at 25-35 µ m (the so-called 30 µ m feature; e.g., Volk et al. 2000) isalready seen during the AGB phase. The 11.5 and 30 µ m features are also observed inthe post-AGB and PNe phases (e.g., Hony et al. 2002; Morisset et al. these proceedings)but the still unidentified 21 µ m feature is genuine of the post-AGB stage (see Sect. 3).The ISO detection of other C-bearing species such as polyynes, benzene, etc. in post-AGB stars may be explained by the polymerization of C H , HCN, and carbon chains(Cernicharo 2004). Indeed, these small hydrocarbon molecules like C H , C H , C H have been suggested to be the building blocks of more complex molecules such as PAHsand that are known to show strong aromatic infrared bands (AIBs at 3.3, 6.2, 7.7, 8.6, and11.3 µ m; e.g., Allamandola et al. 1989) coincident with the unidentified infrared (UIR)emission observed in stars evolving from AGB stars to PNe. However, because of the lowtemperatures of the central stars, it is difficult to believe that the UIRs observed in proto-PNe are due to free-flying gas-phase PAHs (e.g., Kwok et al. 2001; Duley & Williams2011). Interestingly, proto-PNe show aliphatic emission represented by the 8 and 12 µ memission plateaus and the 3.4, 6.9, and 7.3 µ m features together with the classical AIBs(see Figure 1, right panel). This is a strong indication of the coexistence of aliphaticand aromatic structures in the circumstellar shells of these evolved stars. Thus, a betteralternative to the PAH hypothesis is a solid material with a mix of aliphatic and aromaticstructures such as hydrogenated amorphous carbon (HAC; e.g., Duley & Williams 2011),coal (e.g., Guillois et al. 1996), etc. Observationally, it seems clear that the aliphaticcomponent decreases with the evolutionary stage, with the AIBs being stronger in thePN phase (Kwok et al. 2001; Garc´ıa-Lario & Perea-Calder´on 2003). This change fromaliphatic to aromatic structures was attributed to the photochemical processing by theUV photons from the central star (Kwok et al. 2001).Our understanding of the C-rich chemistry in evolved stars has significantly improvedwith the more recent Spitzer data, permitting us to study extragalactic sources in thetransition phase from AGB stars to PNe for the first time. Infrared features from gasphase molecules (C H , HCN, C ) are more common and strong in AGB stars of low Figure 1.
Left panel : Weak crystalline silicate features from pyroxene (P) and olivine (O) foundin OH/IR stars by ISO (Sylvester et al. 1999).
Right panel : ISO spectrum of the proto-PN IRAS22272+5435 showing the presence of aliphatic discrete features (e.g., at 3.4 and 6.9 µ m) andthe 8 and 12 µ m aliphatic plateaus together with AIBs (e.g., at 6.2 and 7.7 µ m). The stillunidentified 21, 26, and 30 µ m features are also seen (Kwok et al. 2001). olecular processes AGB-PNe µ m feature is more common at lowmetallicity. The systematic study of heavily obscured post-AGB stars of our Galaxy showthat aliphatic broad emissions (at ∼ µ m) as well as molecular absorp-tions from small hydrocarbons are typical during the total obscuration phase of the moremassive C-rich post-AGB stars (Garc´ıa-Hern´andez et al. 2007c). As expected from thenucleosynthesis in the previous AGB phase, C-rich PNe were found to be more frequentat low metallicity, showing less processed dust grains (i.e., aliphatic dust dominates andAIBs are rare; Stanghellini et al. 2007).In summary, the C-rich dust features seem to change from aliphatic (in the AGB phase)to mostly aromatic (in the PN stage) although a mix of aliphatic and aromatic featuresare observed in the post-AGB stage and still in some PNe. Several scenarios have beenproposed to explain the aromatization from the AGB phase to the PNe stage: i) dustprocessing (change from aliphatic to aromatic structures) by the UV photons from thecentral star (e.g., Kwok et al. 2001; Garc´ıa-Lario & Perea-Calder´on 2003; Kwok 2004); ii)acetylene (C H ) and its radical derivatives are the precursors of PAHs (e.g., Cernicharo2004). The very recent detections of fullerenes and graphene in post-AGB stars and PNehave provided new valuable information about the dust processing in the circumstellarshells of evolved stars (see Sect. 4). Mixed-chemistry sources.
The mixed chemistry phenomenon (the simultaneous pres-ence of both C-rich and O-rich chemistry and dust discovered by ISO) was found tobe a common characteristic to cool Wolf-Rayet WCPNe, pointing to the presence of along-lived circumbinary disk (Waters et al. 1998). More recent Spitzer observations showthat the mixed-chemistry phenomenon is more common in the Galactic Bulge, being notrestricted to cool WCPNe (Perea-Calder´on et al. 2009). The binary hypothesis cannot ex-plain the dual-dust chemistry phenomenon in the Galactic Bulge and a very late thermalpulse (Perea-Calder´on et al. 2009) or hydrocarbon chemistry in an UV-irradiated, densetorus (Guzm´an-Ram´ırez et al. these proceedings) have been invoked to explain the highdetection rate of mixed-chemistry sources in the Bulge. In addition, mixed-chemistry hasnow been detected in some post-AGB stars (Cerrigone et al. 2009). Interestingly, thedetection rate of mixed-chemistry PNe is strongly linked to the metallicity (Stanghelliniet al. 2011 and these proceedings).
3. The unidentified 21, 26, and 30 µ m features Apart from the aromatic and aliphatic features mentioned above, there is an interestingset of still UIR features located at 21, 26, and 30 µ m and that are usually observed instars from the AGB to the PN phase. The 21 µ m feature. This feature is only observed in post-AGB stars and their charac-teristics indicate a solid-state carrier with a fragile nature. Many different carriers of the21 µ m feature have been proposed in the literature but it seems clear that the carriershould be a carbonaceous compound. Hydrogenated fullerenes, nanodiamonds, HAC, TiCnanoclusters, etc. are some examples among the proposed carbonaceous species. Amides(urea or thiourea) as the carrier of the 21 µ m emission is the more recent and exoticproposal (Papoular 2011). The 26 and 30 µ m features. The 30 µ m feature, sometimes with substructure at 26 µ m, is seen from the AGB to the PN stage with an important fraction of the total energyoutput. This indicates that the carrier should be very abundant in the circumstellar shell.4 Garc´ıa-Hern´andezThe carrier of the 30 µ m feature is usually assumed to be MgS (Hony et al. 2002) butZhang et al. (2009) argue that MgS is very unlikely the carrier of the 30 µ m emissionseen in C-rich evolved stars. Grishko et al. (2001) also show that HACs can explain the30 µ m feature. The HAC identification can also explain the 21 and 26 µ m features. Morerecently, Papoular (2011) shows that aliphatic chains (CH groups, O bridges and OHgroups) can also explain the 30 µ m emission in sources evolving from AGB to PNe.Finally, it should be noted that other weaker UIR features at 15.8, 16.4, and 17 µ m arealso seen in some proto-PNe (see e.g., Hrivnak et al. 2009).
4. Fullerenes and graphene in circumstellar envelopes.
Fullerenes such as C and C are highly resistant and stable tridimensional moleculesformed exclusively by carbon atoms. Fullerenes and fullerene-related molecules haveattracted much attention since their discovery at laboratory (Kroto et al. 1985) be-cause they may explain certain unidentified astronomical features such as the so-calleddiffuse interstellar bands (DIBs) (see e.g., Luna et al. 2008 for a review on interstel-lar/circumstellar DIBs). The remarkable stability of fullerenes against intense radiation,ionization, etc. reinforced the idea that fullerenes should be present in the interstel-lar medium with important implications for interstellar/circumstellar chemistry. Indeed,fullerenes were found on Earth an on meteorites and several unsuccessful ISO attemptsto look for the mid-IR signatures of fullerenes in circumstellar shells have been previ-ously made, including R Coronae Borealis (RCB) stars (e.g., Lambert et al. 2001) andpost-AGB stars (e.g., Kwok et al. 1999). Discovery of fullerenes in space.
The RCB stars have been considered to be the idealastrophysical environments for the formation of fullerenes (Goeres & Sedlmayr 1992).This is because the H-deficiency together with the He and C-rich characters of RCBsresemble the experimental conditions on Earth, facilitating the formation of fullerenes(e.g., Kroto et al. 85). In late 2009 Garc´ıa-Hern´andez, Rao & Lambert looked for C in acomplete sample of 31 RCBs using Spitzer. They got the unexpected result that C wasonly detected in the two RCBs (DY Cen and V854 Cen) with some H in their circumstellarshells. Interestingly, their detection of C around RCBs occurred in conjunction withthe presence of PAHs. In addition, they found that the V854 Cen’s IR spectrum evolvedfrom HACs (ISO 1996’s spectrum) to PAHs and C (Spitzer’s 2007) (see Figure 2, leftpanel). These unique IR spectral variations prompted them to claim that the PAHsand fullerenes evolved from HAC grains. Because of the unexpected result in RCBs, theylooked for fullerenes in ∼
240 PNe pertaining to very different environments and observedby Spitzer. The mid-IR signatures of the C fullerenes were clearly found in five PNewith normal H abundances (including the PN Tc 1). The common presence of PAHfeatures in fullerene-containing PNe confirmed them the unexpected results obtained inRCBs. Meanwhile, Cami et al. (2010) reported in Science the great discovery of the IRdetection of C and C fullerenes in Tc 1 as due to the H-poor conditions in the innercore † . However, neither the central star, nor the inner core and the nebula are H-deficientand current understanding of stellar astrophysics does not allow for Tc 1 being a H-poorPN (see Garc´ıa-Hern´andez et al. 2010; Garc´ıa-Hern´andez, Rao & Lambert 2011). Thus,the detection of fullerenes in PNe and RCB stars have challenged the paradigm regardingfullerene formation in space, showing that, contrary to general expectation, fullerenes are † Garc´ıa-Hern´andez, Rao & Lambert submitted their unexpected RCB results to
Science onApril 19th 2010 but their work was rejected; e.g., one reviewer was opposed to their work becauseit was “in the wrong direction with the least H-deficient stars showing the C features” . olecular processes AGB-PNe has been recentlydetected in the proto-PN IRAS 01005+7910 (Zhang & Kwok 2011) and two binary post-AGB stars (Gielen et al. these proceedings), indicating that fullerene formation can occurjust after the AGB phase. Formation of fullerenes in H-rich ejecta.
The simultaneous detection of PAH-like fea-tures and fullerenes toward C-rich and H-containing PNe indicates that modificationsare needed to our current understanding of the chemistry of large organic moleculesas well as the chemical processing in space (Garc´ıa-Hern´andez et al. 2010). The sug-gestion was made that both fullerenes and PAHs can be formed by the photochemicalprocessing (e.g., as a consequence of UV irradiation or shocks) of HAC, which shouldbe a major constituent in the circumstellar envelope of C-rich evolved stars. This ideais supported by the laboratory experiments on the decomposition of HAC, which showthat the products of destruction of HAC grains are PAHs and fullerenes in the form ofC , C , and C molecules (Scott et al. 1997). More recently, Garc´ıa-Hern´andez et al.(2011) have presented new Spitzer detections of C and C fullerenes in PNe of the MCs(MCPNe), which have permitted an accurate determination of the C and C abun-dances (C /C ∼ /C ∼ and C fullerenes (Iglesias-Groth et al. 2011; Cataldo et al. theseproceedings). In addition, the new MCPNe studies show that neutral fullerenes are likelyin solid-state and collisionaly excited (i.e., they are not excited by the UV photons fromthe central stars) and that they probably evolved from the shock-induced decompositionof small solid particles similar to that of HAC dust (Garc´ıa-Hern´andez et al. 2011). V854 Cen ISOV854 Cen SpitzerHAC
Figure 2.
Left panel : Residual ISO and Spitzer/IRS spectra for the RCB star V854 Cen. Thelaboratory emission spectrum of HAC at 773 K is shown for comparison (Garc´ıa-Hern´andez,Rao & Lambert 2011).
Right panel : Residual Spitzer spectra of fullerene-MCPNe. The bandpositions of C (dashed) and planar C (dotted) are marked (Garc´ıa-Hern´andez et al. 2011). First detection of graphene in space.
Interestingly, some of the fullerene-containing MCPNedisplay very unusual emission features at ∼ µ m (see Figure 2, right panel)coincident with the theoretical transitions of planar C (Garc´ıa-Hern´andez et al. 2011).Planar C is a very stable molecule (more than the C fullerene) and can be viewedas a small fragment of a graphene sheet. The detection of these very unusual featuresin fullerene sources represents the first possible detection of graphene in space. How-ever, a definitive confirmation has to wait for laboratory infrared spectroscopy of C ,which is extremely difficult because of the high reactivity of C . The possible detectionof graphene precursors (C ) opens the possibility of detecting other complex forms ofcarbon (e.g., carbon nanotubes, carbon onions, etc.) in space. UIR emissions in fullerene sources.
Remarkably, fullerenes and graphene have beendetected in PNe whose IR spectra are dominated by aliphatic C-rich dust, representedby broad emissions such as those at 6-9, 10-15, 15-20, and 25-35 µ m. The 6-9 µ m featuremay be attributed to HACs or large PAH clusters. On the other hand, the broad 10-15 µ m (the 11.5 µ m feature) and the 25-35 µ m emission (the 30 µ m feature) are usuallyattributed to SiC (e.g., Speck et al. 2009) and MgS (e.g., Hony et al. 2002), respectively.However, the observed variability of these broad features is quite consistent with thevariable properties of HACs (a material with mixed aliphatic and aromatic structures),which are able to provide a wide range of different spectra (e.g., the relative strength andposition of the IR features) depending on their physical and chemical properties (e.g., size,shape, hydrogenation; e.g., Grishko et al. 2001). Indeed, these aliphatic emissions in low-metallicity extragalactic post-AGB stars and PNe are stronger and much more commonthan in the higher metallicity Galactic counterparts (Volk et al. 2011; Stanghellini et al.2007, 2011). Because of the lower metal content (Si, Mg) of MC post-AGBs and PNe, theopposite is expected. Thus, it is very unlikely that the carriers of the broad 11.5 and 30 µ m emissions - usually seen in fullerene-containing PNe - are related with SiC and MgS,respectively, as it was suggested in the past. The carriers of the broad 11.5 and 30 µ mfeatures (also the bump at 15-20 µ m and possibly the so-called 21 and 26 µ m features)are more likely related with other decomposition products of HACs or a similar material(e.g., fullerene and graphene precursors or intermediate products not yet identified) buta definitive answer requires further laboratory efforts.
5. Concluding remarks
In summary, the most likely explanation for the formation of fullerenes and graphene inH-rich environments is that these molecular species may be formed from the destruction ofa carbonaceous compound with a mixture of aromatic and aliphatic structures - e.g., HAC- which should be widespread in space. In this context, the coexistence of a large varietyof molecular species in H-rich circumstellar environments is supportive of a model inwhich non-equilibrium IR emission occurs from small solid particles containing aromatic,aliphatic, fullerene, and graphene structures similar to that of HAC dust. Indeed, Duley& Williams (2011) have recently suggested a model for the heating of HAC dust via therelease of chemical energy that gives a natural explanation for the astronomical aromaticemission at 3.3 (and other UIR wavelengths) usually attributed to PAHs. Finally, itshould be noted here that instead of HACs, other materials with a complex mix ofaromatic and aliphatic structures (e.g., coal, kerogen, petroleum fractions, soot, quenchedcarbonaceous composites, carbon nanoparticles, etc.) can be present in astrophysicalenvironments and they should be widely explored at laboratory. For example, coal andpetroleum fractions may explain the great diversity of spectral features (aromatic andaliphatic) seen in proto-PNe (Guillois et al. 1996; Cataldo et al. 2002). olecular processes AGB-PNe References
Allamandola, L. J. et al. 1989,
ApJS , 71, 733Bernard-Salas, J. et al. 2006,
ApJ , 625, L29Cami, J. et al. 2010,
Science , 329, 1180Cataldo, F., Keheyan, Y., & Heymann, D. 2002,
IJA , 1, 79Cernicharo, J. et al. 1999,
ApJ , 526, L41Cernicharo, J. 2004,
ApJ , 608, L41Cerrigone, L. et al. 2009,
ApJ , 703, 585Duley, W. W., & Williams, D. A. 2011,
ApJ , 737, L44Garc´ıa–Lario, P. & Perea-Calder´on, J. V. 2003,
ESA Publication Series , 511, 97Garc´ıa-Hern´andez, D. A. et al. 2006,
Science , 314, 1751Garc´ıa-Hern´andez, D. A. et al. 2007a,
A&A , 462, 711Garc´ıa-Hern´andez, D. A. et al. 2007b,
ApJ , 666, L33Garc´ıa-Hern´andez, D. A. et al. 2007c,
APNIV , 46Garc´ıa-Hern´andez, D. A. et al. 2009,
ApJ , 705, L31Garc´ıa-Hern´andez, D. A. et al. 2010,
ApJ , 724, L39Garc´ıa-Hern´andez, D. A. et al. 2011,
ApJ , 737, L30Garc´ıa-Hern´andez, D. A., Rao, N. K., & Lambert, D. L. 2011,
ApJ , 729, 126Goeres, A. & Sedlmayr, E. 1992,
A&A , 265, 216G´orny, S. K. et al. 2010,
A&A , 516, 39Guillois, O. et al. 1996,
ApJ , 464, 810Grishko, V. I. et al. 2001,
ApJ , 558, L129Herbst,E. & vanDishoeck,E.F. 2009,
ARA&A , 47, 427Hony, S. et al. 2002,
A&A , 390, 533Hony, S. et al. 2009,
A&A , 501, 609Hrivnak, B. J. et al. 2009,
ApJ , 694, 1147Iglesias-Groth, S. et al. 2011,
MNRAS , 413, 213Kroto, H. W. et al. 1985,
Nature , 318, 162Kwok, S. et al. 1999,
A&A , 350, L35Kwok, S. et al. 2001,
ApJ , 554, L87Kwok, S. 2004,
Nature , 430, 985Lambert, D. L. et al. 2001,
ApJ , 555, 925Lagadec, E. et al. 2007,
MNRAS , 376, 1270Luna, R. et al. 2008,
A&A , 480, 133Manteiga, M. et al. 2011, AJ , 141, 80Molster, F. J. et al. 1999, Nature , 401, 563Molster, F. J. et al. 2002,
A&A , 382, 184Papoular, R. 2011,
MNRAS , 415, 494Perea-Calder´on, J. V. et al. 2009,
A&A , 495, L5Sargent, B. A. et al. 2010,
ApJ , 716, 878Scott, A. et al. 1997,
ApJ , 489, L123Speck, A. K. et al. 2009,
ApJ , 691, 1202Stanghellini, L. et al. 2007,
ApJ , 671, 1669Stanghellini, L. et al. 2011,
ApJ , (submitted)Sylvester, R. J. et al. 1999,
A&A , 352, 587Volk, K. et al. 2000,
ApJ , 530, 408Volk, K. et al. 2011,
ApJ , 735, 127Waters, L. B. F. M. et al. 1996,
A&A , 315, L361Waters, L. B. F. M. et al. 1998,
A&A , 331, L61Yang, X. et al. 2004,
A&A , 414, 1049Zhang, K. et al. 2009,
ApJ , 702, 680Zhang, Y. & Kwok, S. 2011,