From outburst to quiescence: spectroscopic evolution of V1838 Aql imbedded in a bow-shock nebula
J. V. Hernández Santisteban, J. Echevarría, S. Zharikov, V. Neustroev, G. Tovmassian, V. Chavushyan, R. Napiwotzki, R. Costero, R. Michel, L. J. Sánchez, A. Ruelas-Mayorga, L. Olguín, Ma. T. García-Díaz, D. González-Buitrago, E. de Miguel, E. de la Fuente, R. de Anda, V. Suleimanov
MMNRAS , 1–14 (2018) Preprint 18 March 2019 Compiled using MNRAS L A TEX style file v3.0
From outburst to quiescence: spectroscopic evolution ofV1838 Aql imbedded in a bow-shock nebula
J. V. Hern´andez Santisteban , (cid:63) , J. Echevarr´ıa , , S. Zharikov , V. Neustroev ,G. Tovmassian , V. Chavushyan , R. Napiwotzki , R.Costero , R. Michel ,L. J. S´anchez , A. Ruelas–Mayorga , L. Olgu´ın , Ma. T. Garc´ıa–D´ıaz ,D. Gonz´alez–Buitrago , , E. de Miguel , E. de la Fuente , R. de Anda ,and V. Suleimanov , , Anton Pannekoek Institute for Astronomy, University of Amsterdam, Science Park 904, NL–1098 XH Amsterdam, the Netherlands Instituto de Astronom´ıa, Universidad Nacional Aut´onoma de M´exico, Apartado Postal 70–264,Ciudad Universitaria, M´exico D.F., C.P. 04510, Mexico Instituto de Astronom´ıa, Universidad Nacional Aut´onoma de M´exico, Ensenada, Baja California, C.P. 22830, Mexico Astronomy Research Unit, University of Oulu, PO Box 3000, FIN-90014, Finland Instituto Nacional de Astrof´ısica, ´Optica y Electr´onica, Apartado Postal 51, CP 72000, Puebla, Mexico Centre for Astrophysics Research, Science and Technology Research Institute, University of Hertfordshire, Hatfield AL10 9AB, UK Departamento de Investigaci´on en F´ısica, Universidad de Sonora, Blvd. Rosales y Colosio, 83190 Hermosillo, Sonora, Mexico Department of Physics and Astronomy, University of California, Irvine, California 92697, USA Departamento de Ciencias Integradas, Universidad de Huelva, E–21071 Huelva, Spain Instituto de Astronom´ıa y Meteorolog´ıa, Departamento de F´ısica, CUCEI, Guadalajara, Jalisco, 44100, Mexico Institut f¨ur Astronomie und Astrophysik, Universit¨at T¨ubingen, Sand 1, D-72076 T¨ubingen, Germany Kazan (Volga region) Federal University, Kremlevskaya str. 18, 420008 Kazan, Russia Space Research Institute of the Russian Academy of Sciences, Profsoyuznaya Str. 84/32, Moscow 117997, Russia
Submitted: 18 March 2019, ArXiv: 18 March 2019
ABSTRACT
We analyse new optical spectroscopic, direct-image and X-ray observations of therecently discovered a high proper motion cataclysmic variable V1838 Aql. The datawere obtained during its 2013 superoutburst and its subsequent quiescent state. Anextended emission around the source was observed up to 30 days after the peak of thesuperoutburst, interpreted it as a bow–shock formed by a quasi-continuous outflowfrom the source in quiescence. The head of the bow–shock is coincident with the high–proper motion vector of the source ( v ⊥ = ± km s − ) at a distance of d = ± pc. The object was detected as a weak X-ray source ( . ± . counts s − ) inthe plateau of the superoutburst, and its flux lowered by two times in quiescence(0.007 ± − ). Spectroscopic observations in quiescence we confirmed theorbital period value P orb = . ± . days, consistent with early-superhumpestimates, and the following orbital parameters: γ = − ± km s − and K = ± km s − . The white dwarf is revealed as the system approaches quiescence, which enablesus to infer the effective temperature of the primary T ef f = , ± K. The donortemperature is estimated (cid:46)
K and suggestive of a system approaching the periodminimum. Doppler maps in quiescence show the presence of the hot spot in He I lineat the expected accretion disc-stream shock position and an unusual structure of theaccretion disc in H α . Key words: cataclysmic variables, dwarf novae, white dwarf, stars: individual:V1838 Aql (cid:63) e–mail: [email protected]
Cataclysmic Variables (CVs) are close binary systems wherea white dwarf (WD) accretes from a low–mass star via © a r X i v : . [ a s t r o - ph . S R ] M a r Hern´andez Santisteban et al.
Roche–lobe overflow, often creating an accretion disc (for areview see Warner 1995). The evolution of CVs is driven bythe removal of angular momentum from the system, whichleads to the depletion of the donor and causes the orbital pe-riod ( P orb ) to shrink. This process continues until the donorreaches the sub-stellar regime (i.e. a brown dwarf donor)(Howell et al. 1997), where its internal structure causes thedonor to expand in response to the loss mass, thus leadingto an increase in P orb . This fact causes a sharp cut-off inthe P orb distribution called the period minimum (Paczynski& Sienkiewicz 1981). In addition, the long time-scales asso-ciated at the period minimum leads to the accumulationof most CVs between 1-2 hrs (G¨ansicke et al. 2009), witha large fraction of systems evolving towards longer P orb ,known as period bouncers. Out of this short P orb popula-tion, up to ∼ % (Kolb & Baraffe 1999; Goliasch & Nel-son 2015) should be period bouncers and harbour a sub-stellar donor. Theoretically, the stellar to sub-stellar transi-tion roughly coincides with the period minimum. However,there is little observational evidence of its location giventhe lack of detailed characterisation of systems around theperiod minimum (e.g. Littlefair et al. 2006; Harrison 2016;Hern´andez Santisteban et al. 2016; Neustroev et al. 2017;Pala et al. 2018).Short orbital period systems (often referred as WZ Sge–type objects) are characterised by enhanced brightness, ex-tended outburst duration ( ∼ days) as well as the onsetof superhumps (low-amplitude variability close the orbitalperiod of the system) shortly after maximum brightness.These are often classified as superoutbursts, to distinguishthem from those observed in classical dwarf novae. There-fore, the follow-up and detailed characterisation of these sys-tems is paramount to confirm the nature of the donor andtest against theoretical expectations (Knigge et al. 2011).The discovery of a new transient, V1838 Aql (alsoknown as PNV J19150199+0719471), was initially reportedby Koichi Itagaki on May 31 2013, as a possible Nova reach-ing V ∼
10 mag, who reported also that the object was below15.5 mag on his unfiltered survey image taken on 21.608UT . Kato (vsnet–alert 15776) pointed out that the highproper motion made it a good candidate for a WZ Sge–type nearby star, close to the galactic plane, ( (cid:96) = . ◦ , b = − . ◦ ). This classification was confirmed in subsequentvsnet–alerts, where various stages of superhumps were ob-served as the superoutburst evolved, and a mean superhumpperiod of 0.05803(1) days was reported by Kato (vsnet–alert15931).In this paper we present and discuss CCD direct images,X-ray observations and extensive low and high dispersionspectroscopy during outburst and quiescence of V1838 Aql.We show a most striking result: the detection of a nebulos-ity in emission during the course of the superoutburst. Inaddition, we present time-resolved spectroscopy and deter-mine orbital parameters from the emission lines, as well asDoppler tomography. The presence of a white dwarf revealedat quiescence allowed us to determine the white dwarf (WD)temperature. Finally, we discuss whether the system is ap-proaching or receding from the minimum orbital period. We will present the multi-wavelength observations ofV1838 Aql from its discovery to its quiescent state, spanningover ∼ years. In order to simplify the reference to specificdates or epochs throughout the superoutburst, we make useof truncated Julian Days of the form HJD – 2456000. Observations were obtained with the Echelle spectrographattached to the 2.1m Telescope of the Observatorio As-tron´omico Nacional at San Pedro M´artir, on the nights of2013 June 3 and June 16–19. The Marconi–2, a × detector, was used to obtain a spectral resolutionof R ∼
19 000 , All observations were carried out with the300 l/mm cross–dispersor, which has a blaze angle at around5500 ˚A. The spectral coverage was about – ˚A. Theexposure time for each spectrum was 300 s for June 3 and900 s for June 16–19. Low dispersion spectroscopy was alsoobtained on 2013 June 5–7 and 9 with the Boller & Chivensspectrograph (B&Ch), with two different gratings. On thefirst three nights, a 400 l/mm grating was used, to obtain ahigh S/N ratio in order to study the spectral energy distri-bution of the system, with a broad wavelength coverage ofabout – ˚A. The exposure time for each spectrumwas 300 s. In addition (on June 9) a 1200 l/mm grating wasused, to obtain a higher spectral resolution around the in-terval – ˚A. In this setup, the exposure time for eachspectrum was 300 s.Further observations were obtained with the B&Chduring the nights of 2013 June 28, 29 and 30, with the1200 l/mm grating to cover the range around H α (June 29)and H β (June 28 and 30). Additional low resolution obser-vations were obtained also with the B&Ch on 2013 August15 ( – ˚A coverage) with an exposure time of 900 sper spectrum, and two years later on 2015 September 17( – ˚A coverage) with an exposure time of 1800 s perspectrum, both with the 400 l/mm grating and when thesystem had already dropped to V ∼
17 mag. All observationswere made with a 1 . (cid:48)(cid:48) λλ = − ˚A with the R2500U(600 s), R2500V (500 s), R2500R (500 s), and R2500I (600 s)volume–phased holographic gratings (see bottom spectra inFig. 3). Phase–resolved spectroscopy was then obtained withthe R2005R grating, in the wavelength interval − ˚A,centred around the H α emission line. The exposure time foreach spectrum was 235 sec, with a total coverage of aroundone and a half orbital cycles. Standard data reduction pro-cedures for all spectroscopic observations were performedusing the iraf software. The log of all spectroscopic obser-vations is shown in Table 1. iraf is distributed by the National Optical Observatories, op-erated by the Association of Universities for Research in Astron-omy, Inc., under cooperative agreement with the National ScienceFoundation. MNRAS000
17 mag. All observationswere made with a 1 . (cid:48)(cid:48) λλ = − ˚A with the R2500U(600 s), R2500V (500 s), R2500R (500 s), and R2500I (600 s)volume–phased holographic gratings (see bottom spectra inFig. 3). Phase–resolved spectroscopy was then obtained withthe R2005R grating, in the wavelength interval − ˚A,centred around the H α emission line. The exposure time foreach spectrum was 235 sec, with a total coverage of aroundone and a half orbital cycles. Standard data reduction pro-cedures for all spectroscopic observations were performedusing the iraf software. The log of all spectroscopic obser-vations is shown in Table 1. iraf is distributed by the National Optical Observatories, op-erated by the Association of Universities for Research in Astron-omy, Inc., under cooperative agreement with the National ScienceFoundation. MNRAS000 , 1–14 (2018) he bow-shock nebula in V1838 Aql
440 460 480 500 520 540 560 580 600101112131415161718 M a gn it ud e C oun t r a t e / s Quiescent valueat JD=2546
JD-2456000
Figure 1.
AAVSO and
Swift /XRT light curves during and after the 2013 superoutburst of V1838 Aql . The observations were taken inV (green dots), R (red squares) and B (blue diamonds) standard filters. The grey vertical marks indicate the epochs of our spectroscopicobservations. The 0.3-10 keV
Swift /XRT count rate is shown in purple stars. Note that the magnitude of the star is still above the minimumderived from the pre-outburst magnitude V=18.6 at the time of the GTC spectral observations on day 1566. The last
Swift/XRT pointis shown as a reference
Direct deep narrow filters images in H α , H α Continuum,[O iii ] 5007 ˚A, and [N ii ] 6583 ˚A were obtained during 2013,June 3, 18 and 29 at the 0.84m Telescope of the ObservatorioAstron´omico Nacional at San Pedro M´artir, Mexico. The logof all imaging observations is also presented in Table 1.We also performed multi-colour optical and near-infrared photometry. On 2016 November 8, we obtainedBVRI images with the Andalucia Faint Object Spectrographand Camera (ALFOSC) at the 2.56 m Nordic Optical Tele-scope (NOT), in the Observatorio de Roque de los Mucha-chos (ORM, La Palma, Spain). The integration times were300 s in each filter. We observed the target again on 2017October 6 with the New Technology Telescope (NTT) atLa Silla Observatory, Chile. The images were captured withthe ESO Faint Object Spectrograph and Camera (EFOSC2– Buzzoni et al. 1984) through the BVRiz filters with ex-posure times of 40, 40, 40, 60 and 120 sec, respectively. Inaddition, we obtained two sets of near-infrared (NIR) obser-vations with the JHK s filters. The observations were per-formed on 2017 March 8 with the NOTcam instrument onthe NOT, and on 2017 Oct 6 with SOFI on the NTT (Moor-wood, Cuby & Lidman 1998). The log of all photometricobservations is also shown in Table 1.Pre-outburst observations were derived from the Pan-STARRS1 database . The mean epoch of the observations is The Pan-STARRS1 Surveys (PS1) and the PS1 public sciencearchive have been made possible through contributions by the In- around day 29 in our notation (i.e on 2012, April 12), morethan a year earlier to the superoutburst. We extracted thedata using a small radius of 0.03 arcmin and obtained thefollowing mean PSF magnitudes in the g, r, i, z, and y filters:18.56 ± ± ± ± ± V1838 Aql was also observed five times with the Neil GehrelsSwift Observatory (Gehrels et al. 2004). These observationswere taken in the middle of the superoutburst plateau stageon 2013 June 11 and 14 (the total exposure time of thisdata subset is 5.35 ks), at the end of the rapid fading stage stitute for Astronomy, the University of Hawaii, the Pan-STARRSProject Office, the Max-Planck Society and its participating in-stitutes, the Max Planck Institute for Astronomy, Heidelberg andthe Max Planck Institute for Extraterrestrial Physics, Garching,The Johns Hopkins University, Durham University, the Univer-sity of Edinburgh, the Queen’s University Belfast, the Harvard-Smithsonian Center for Astrophysics, the Las Cumbres Observa-tory Global Telescope Network Incorporated, the National Cen-tral University of Taiwan, the Space Telescope Science Institute,the National Aeronautics and Space Administration under GrantNo. NNX08AR22G issued through the Planetary Science Divisionof the NASA Science Mission Directorate, the National ScienceFoundation Grant No. AST-1238877, the University of Maryland,Eotvos Lorand University (ELTE), the Los Alamos National Lab-oratory, and the Gordon and Betty Moore Foundation.MNRAS , 1–14 (2018)
Hern´andez Santisteban et al.
Table 1.
Log of photometric and spectroscopic observationsSpectroscopy Julian Date Range No. of Exposure CommentsDate (2456000+) (˚A) Spectra Time (s)2013 June 3 444 3900–7300 14 300 Echelle2013 June 5–7 446–448 4130–7575 29 300 Boller & Chivens2013 June 9 450 5510–6730 29 300 Boller & Chivens2013 June 16–19 460–464 3900–7300 17 900 Echelle2013 June 28 472 4500–5700 14 420 Boller & Chivens2013 June 29 473 5500–6700 14 420 Boller & Chivens2013 June 30 474 4500–5700 14 420 Boller & Chivens2013 August 15 520 4000–7000 10 900 Boller & Chivens2015 September 17 1282 4000–7500 03 1800 Boller & Chivens2016 June 27 1566 3440–4610 01 600 Osiris – GTC2016 June 27 1566 4500–6000 02 245 Osiris – GTC2016 June 27 1566 5575–7685 02 254 Osiris – GTC2016 June 27 1566 7330–10000 01 600 Osiris – GTC2016 June 27 1566 5575–7685 30 235 Osiris – GTCImaging Julian Date Filter FWHM Exposure CommentsDate (2456000+) (˚A) Time (s)2013 June 3 444 V 980 302013 June 18 462 H α
11 1260 180 s × α
11 1800 180 s ×
10 images2013 June 29 473 Continuum (6650˚A) 70 1800 180 s ×
10 images2013 June 29 473 [OIII] 5007 52 1800 180 s ×
10 images2013 June 29 473 [NII] 6583 10 1800 180 s ×
10 imagesMulticolor Band Photometry Filters Telescope Site Comments2016 November 9 1701 BVRI NOT ORM Section2017 March 8 1820 JHK s NOT ORM 2.22017 October 6 2032 BVRiz NTT La Silla for more2017 October 6 2032 JHK s NTT La Silla details on June 28 ( ∼ ∼ Swift -XRT detected aweak X-ray source with a count-rate of 0.0149 ± − , which is dropped to the level of about 0.0012 ± − at the end of the rapid fading stage. In quies-cence, however, the count-rate has been found at the levelof 0.0071 ± − , that is lower than during thesuperoutburst plateau but is higher than during the rapidfading stage (see Table 2 for the observation log and Fig. 1).This pattern — the outburst flux is higher than in quies-cence with a deep dip during the outburst rapid fading —is in contrast to ordinary dwarf novae which usually showa depression of the X-ray flux during outbursts, but is inagreement with X-ray observations of WZ Sge-type stars(see Neustroev et al. 2018, and references therein).The plateau-, decline-stage, and quiescent spectra ofV1838 Aql consist of only 58, 6, and 12 counts, respectively,therefore no meaningful spectral analysis is possible. Never-theless, assuming that the spectrum of V1838 Aql is similarto other WZ Sge-type stars such as SSS J122221.7 − ∼ × − and ∼ × − erg s − cm − , respectively, with the correspond- Table 2.
Log of
Swift /XRT observationsJulian Date Exp. Time Obs. ID X-ray count rate2456000+ ks count s − + . − . + . − . + . − . + . − . ing luminosity of ∼ × and ∼ × erg s − (adoptingthe distance of 202 pc, see Section 3.1). These luminositiesare in agreement with those found for other WZ Sge-typestars and accreting white dwarfs (Reis et al. 2013; Neustroevet al. 2018). We report the detection of an extended component aroundthe object shortly after maximum light. This nebulosity was
MNRAS000
MNRAS000 , 1–14 (2018) he bow-shock nebula in V1838 Aql first detected in the Echelle spectra on days 460–464, (seeday 460 in Fig. 2, lower panel, two days before the first H α image). Further H α images were taken on day 462 which re-vealed a clear diffuse emission centred on the object, as seenin top two panels of Fig. 2. This explains the asymmetryof the line profile observed in the Echelle spectroscopy, dueto the slit position (east–west) only covering a fraction ofthe extended structure at the west of the object. We used daophot (Stetson 1987) to subtract the point sources, asshown in the middle panel of Fig. 2. Its morphology resem-bles that of a bow shock e.g. BZ Cam (Hollis et al. 1992).Additional images were taken on day 473 with a narrow fil-ters around the continuum near 6650 ˚A and also with twonarrow filters [OIII] 5007 ˚A and [NII] 6563 ˚A. No emissionfrom these forbidden lines was detected. The extended H α emission was present for nearly a month after the peak ofthe outburst with no apparent change in morphology. Nonebular stage was detected later. As mentioned in Section 1, V1838 Aql shows a strong propermotion, initially observed from our field images obtained in2013 and archival data from the Palomar survey (POSS–Iand POSS–II). Recently, the
Gaia
DR2 release (Lindegren etal. 2016; Gaia Collaboration et al. 2018) confirmed the highproper motion as well as provided a parallax for the source .The proper motion of V1838 Aql is µ α cos δ = − . ± . masyr − and µ δ = − . ± . mas yr − . In conjunction with theparallax ( π = . ± . mas), we performed a joint Bayesianinference for the distance and the tangential velocity follow-ing Bailer-Jones et al. (2018) which leads to a tangentialvelocity of v ⊥ = ± km s − and a distance of ± pc.The joint and marginal posterior distributions are shown inAppendix A. The distance inferred is consistent with our ini-tial SED fitting estimates in quiescence presented in Section4.4. Thus, combining the systemic radial velocity measure-ments (see Section 4) with the proper motion, we obtaina space velocity of ± km s − . V1838 Aql is the thirdCV with the highest transverse velocity in the Galaxy, justbelow SDSS 1507+22 ( v ⊥ = ± km s − ), a confirmedmetal–poor halo–binary (Patterson et al. 2008; Uthas et al.2011), and BF Eri, with an estimated value of v ⊥ ∼ kms − (Klemola et al. 2004; Neustroev & Zharikov 2008). Fol-low up H α observations are needed to perform a more de-tailed study of its kinematics and origin, and in particular,far– and near–ultraviolet observations are desired to searchfor anomalous line ratios that might indicate evidence forPopulation II membership.The extended emission depicted in the upper and mid-dle panels of Fig. 2 has a (Balmer) bow shock–like shape,formed by our high space velocity object, which we believeis moving supersonically through the interstellar medium(Wilkin 1996). Its shape and duration is very different fromthe recently discovered nova shells around CVs (Sahman etal. 2015). In fact, the proper motion vector (displayed bythe arrow in Fig. 2), coincides with the main axis of the bow V1838 Aql catalogue ID is
Gaia
DR2 4306244746253355776 The R-based code is available at https://github.com/ehalley/parallax-tutorial-2018
RA (J2000) +7°19'15.0"30.0"45.0"20'00.0"15.0" D ec ( J2000 ) RA (J2000) +7°19'15.0"30.0"45.0"20'00.0"15.0" D ec ( J2000 ) Spatial Dimension / arcsec R a d i a l V e l o c i t y / k m s WestEast Bow Shock
Figure 2.
Deep H α imaging of V1838 Aql. Top:
The centroidposition for the Palomar plates is shown as the circle. The propermotion vector (scaled for clarity) is shown and coincides withthe direction of the diffuse emission form the bow shock. TheEchelle slit is marked as the dotted box.
Middle:
Stars have beensubtracted to show only the extended emission.
Bottom:
2D H α profile obtained with the Echelle spectrograph. The wing extend-ing towards the west coincides with the bow shock observed atlater times with photometric observations.MNRAS , 1–14 (2018) Hern´andez Santisteban et al. shock. The head of the nebula is located about ∼ (cid:48)(cid:48) from theobject and the size of the bow shock cone is ≈ (cid:48)(cid:48) along thesky plane with respect to the nebular symmetry axis. Thecone open angle is about ∼
60 degrees with its width andbrightness quickly decreasing with increasing distance fromthe bow shock head.Similar bow shocks are frequently detected around OB–runaway stars (van Buren & McCray 1988) and high–velocity pulsars (Yoon & Heinz 2017, and references therein).However, such bow shocks have also been observed in high-mass transfer CVs (e.g. BZ Cam and V341 Arae Hollis etal. 1992; Greiner et al. 2001; Bond & Miszalski 2018, re-spectively). It is generally accepted that the interstellar gas,compressed in bow shocks, is heated and ionised by the in-tense stellar radiation/wind and produces emission as a re-sult of the collisional and/or the charge–exchange excitationof neutral hydrogen atoms in the post–shock flows, with asubsequent emission produced via bound–bound transitions(Chevalier et al. 1980).The time-scale associated with the presence of the bowshock suggests it was illuminated by the outburst and notmaterial expelled in the outburst itself. Assuming very-fastoutflows speeds for the ejecta ( v ∼ . c), the time-scale isof the order of ∼ days to reach the observed distance( × cm). This time-scale is larger than the limit im-posed by the first detection of the bow shock ( < days).Although high velocity outflows can be achieved in ener-getic novae events (e.g. Metzger et al. 2014), these velocitiesseem unlikely scenario for an outburst of short orbital periodsystem (which implies low-mass transfer rates Breedt et al.2014).Furthermore, the high proper motion of the system(153 mas yr − ) implies that, within an average recurrencetime between outbursts ( ∼ − yr for WZ Sge-type CVs),the system would be displaced from the origin of a previoussuperoutburst by ∼ . arcsec. If this is the case, then theillumination of the apparent nebula would be centred on thelocation of the previous outburst, contrary to what is ob-served during the 2013 outburst. Moreover, the morphologyof such expanding nebula would differ from a bow shock asshown in other outbursting systems like nova events (Sah-man et al. 2015). However, we do not observe any change inthe illuminated nebula for over a month.Therefore, in order to have a standing bow shock, theV1838 Aql must have a quasi-continuous mass outflow. Frommomentum balance arguments (Weaver et al. 1977; van Bu-ren & McCray 1988), we can estimate a mass outflow rateof ∼ − M (cid:12) yr − to reproduce the observed bow shock(assuming a ISM density of n = . cm − , e.g. Hollis et al.1992). This value is roughly on the same order of magni-tude as the putative mass transfer rate from the donor fromevolutionary models (Knigge et al. 2011), which suggests anoutflow mechanism capable of expelling a significant per-centage of the in-falling material with velocities of ∼ kms − . Both requirements highly suggest that either a wind orjet-like outflow might be working in the system during thequiescent state. We took a series of optical spectra during the evolution ofthe superoutburst of V1838 Aql from nearly maximum lightdown to a quiescent level (see Table 1). The spectral evo-lution of the emission and absorption lines as well as theshape of the continuum is summarised in the next two sub-sections. We also show in Fig. 3 a comprehensive graph of allthe flux calibrated spectra (excluding the Echelle spectra),which illustrate the overall behaviour during this complexevent.
The first 14 outburst spectra obtained with the Echelle spec-trograph, taken only three days (day 444) after the first re-port of the eruption (Itagaki 2013), covered an interval of 2.5hr. These high–resolution spectra show very strong double–peaked H α emission and a mixed emission and absorptioncomponent at the H β line. Both emission line componentshave a FWHM ∼ km s − and are superimposed on a verybroad absorption component ( ± − ). The broadcomponent in H α is much weaker than in H β . No H γ ispresent at all. Other features present on the spectra (notshown here) are weak He i i α line suggests an origin on the outer edges of the disc.On days 446–448 the B&Ch spectra show a substan-tial change, as shown in the co–added spectra for the threenights in Fig. 3. The broad absorption components dominatethe Balmer series as well as He i lines, except H α which showa single peak weak emission. We measured the radial veloc-ity shifts of the centre of this emission line (including night450) using a single Gaussian with a FWHM ∼ ˚A. The firstthree nights show clear Doppler variations and a drift of thesystemic velocity (from (cid:104) γ (cid:105) ∼ + to − km s − ). However,our orbital coverage during these nights are very poor andno coherent modulation is observed. On day 450, the radialvelocity measurements show a clear modulation with an ap-parent period coinciding with the early super–hump/orbitalperiod. Due the short time–span of the data, we were notable to find a period value with enough reliability.The inversion of Balmer emission lines to absorption iscommon for dwarf novae in outburst (e.g. Clarke & Bowyer1984; Neustroev et al. 2017). This contrasts to what wasobserved during the superoutburst of the bounce–back can-didate V455 And (Tovmassian et al. 2011). In the latter, theBalmer lines, after an initial switch from emission to broadabsorption, they suddenly reversed their course and shootup back into emission. From the analysis of the width oflines and their radial velocities, Tovmassian et al. (2011)concluded that this is evidence of evaporation and discwind. This is noteworthy, because the observed bow shockof V1838 Aql supposes some kind of outflow from the object(see discussion in Section 3.1). However, no similar wind oroutflow features are seen at the beginning of the superout-burst in this case.As the outburst progressed, the broad absorption linecomponents shrink and eventually disappear. On the con-trary, the H α line in emission broadens throughout the de-cline to its quiescent value of FWHM ∼ km s − (see MNRAS000
The first 14 outburst spectra obtained with the Echelle spec-trograph, taken only three days (day 444) after the first re-port of the eruption (Itagaki 2013), covered an interval of 2.5hr. These high–resolution spectra show very strong double–peaked H α emission and a mixed emission and absorptioncomponent at the H β line. Both emission line componentshave a FWHM ∼ km s − and are superimposed on a verybroad absorption component ( ± − ). The broadcomponent in H α is much weaker than in H β . No H γ ispresent at all. Other features present on the spectra (notshown here) are weak He i i α line suggests an origin on the outer edges of the disc.On days 446–448 the B&Ch spectra show a substan-tial change, as shown in the co–added spectra for the threenights in Fig. 3. The broad absorption components dominatethe Balmer series as well as He i lines, except H α which showa single peak weak emission. We measured the radial veloc-ity shifts of the centre of this emission line (including night450) using a single Gaussian with a FWHM ∼ ˚A. The firstthree nights show clear Doppler variations and a drift of thesystemic velocity (from (cid:104) γ (cid:105) ∼ + to − km s − ). However,our orbital coverage during these nights are very poor andno coherent modulation is observed. On day 450, the radialvelocity measurements show a clear modulation with an ap-parent period coinciding with the early super–hump/orbitalperiod. Due the short time–span of the data, we were notable to find a period value with enough reliability.The inversion of Balmer emission lines to absorption iscommon for dwarf novae in outburst (e.g. Clarke & Bowyer1984; Neustroev et al. 2017). This contrasts to what wasobserved during the superoutburst of the bounce–back can-didate V455 And (Tovmassian et al. 2011). In the latter, theBalmer lines, after an initial switch from emission to broadabsorption, they suddenly reversed their course and shootup back into emission. From the analysis of the width oflines and their radial velocities, Tovmassian et al. (2011)concluded that this is evidence of evaporation and discwind. This is noteworthy, because the observed bow shockof V1838 Aql supposes some kind of outflow from the object(see discussion in Section 3.1). However, no similar wind oroutflow features are seen at the beginning of the superout-burst in this case.As the outburst progressed, the broad absorption linecomponents shrink and eventually disappear. On the con-trary, the H α line in emission broadens throughout the de-cline to its quiescent value of FWHM ∼ km s − (see MNRAS000 , 1–14 (2018) he bow-shock nebula in V1838 Aql Section 4.2). This is explicitly shown in the inset of Fig. 3.On days 472–474 additional B&Ch spectra were obtained,H α and H β became again strong with double peak emission. The system returned to its quiescence state after 500 days,thus presenting an opportunity to analyse the system inits quiescent state. The full optical spectrum, taken over1000 days after the outburst, shows a clear presence of ac-cretion disc emission lines superimposed on the WD broadabsorption lines, as shown in Fig. 3. Time-resolved spec-troscopy however, was focused on the region around the H α line, where the least contribution of the WD broad absorp-tion lines is observed. Hence, we were able to apply a twoGaussian method (Shafter et al. 1986; Horne et al. 1986)to determine the radial velocity of the primary. We madean interactive search for the optimal width and separationof the Gaussians in a grid between a FHWM of 500–1000km s − in 50 km s − steps and 500–2600 km s − in 100km s − steps, respectively (for further explanation of thismethod see Hern´andez Santisteban et al. 2017). At everycombination, we performed a fit of the radial velocities, V ( t ) ,to a circular orbit: V ( t ) = K em sin [ π ( φ − φ )] + γ , (1)where K em is the semi–amplitude, φ is the phase offset be-tween the spectroscopy and photometric ephemeris, and γ isthe systemic velocity. We used the minimum in the diagnos-tic quantity σ K em / K em , where σ K em is the 1 σ uncertaintyon the semi–amplitude, to determine the optimal solution(Shafter et al. 1986).Initially, we have used an arbitrary zero-point close tothe observations to calculate the phase offset and have as-sumed the orbital period of the early superhumps. We ob-tain an orbital solution with P orb = . ± . days,K em = ± km s − and γ = − ± km s − as shown inFig. 4. This spectroscopic orbital period is consistent withthe early superhump period found during the superoutburst( P = . ± . d, Kato et al. 2014a; Echevarr´ıa et al.2019).After correcting for the phase offset to fix the zero pointof the inferior conjunction of the secondary, we can constructthe spectroscopic ephemeris: T ( H JD ) = . ( ) + . ( ) E . (2) The quiescent spectrum of V1838 Aql contains emission linesthat allow us to study the structure of the accretion flow inH α and He I 5875 ˚A and 6678 ˚A, as shown in Fig. 5. We haveused Tom Marsh’s molly software package to normaliseand subtract the continuum of each individual spectrum, aswell as re–bin in equal velocity bins. All three lines show adistinct double–peak profile which suggests the presence of an accretion disc, shown in the mean profiles in the top panelof Fig. 5. In addition, the lines contain an extra componentwith a semi–amplitude of ∼ km s − , which modifies thesymmetry of the median profile (most evident in the He linesand are clearly seen in the trail spectra (mid–panels, Fig. 5).However, we note that the extra component in H α seems tobe shifted in orbital phase with respect to both He i lines.The time–resolved spectroscopic data allow us to studythe structure of the accretion disc via Doppler tomography(Horne & Marsh 1986). We have used the ephemeris ob-tained via the H α wings (see Section 4.2) to calculate thetomograms for each individual emission line. We employedthe Doppler tomography package trm-doppler to producethe maps shown in the bottom panels of Fig. 5. The H α to-mogram reveals a clear accretion disc as well as an additionalcomponent at V x , V y ≈ [ , − ] km s − . Surprisingly, we finda lack of distinct emission of the hot spot in H α , commonlyobserved in most CVs. The He i tomograms produce a si-nusoidal contribution superimposed on the weaker accretiondisc. The location of the emission in He i coincides with theexpected position in velocity space of the hot spot, wherethe ballistic trajectory of material ejected from the L1 pointintersects the outer edge of the accretion disc.The fact that the ephemeris is obtained via the wings ofthe H α line and simultaneously providing a consistent posi-tion for He I hot spots, indicates that the H α position mightbe real and not an error of the zero–point used. However,we need more radial velocity observations of V1838 Aql inquiescence to lock down the real zero point. In quiescence, the optical spectrum of V1838 Aql is domi-nated by the broad absorption hydrogen lines arising fromthe atmosphere of the WD. However, even at these low lu-minosities, the accretion disc can contribute a significantfraction of the continuum and line emission at optical wave-lengths (e.g. Aviles et al. 2010; Hern´andez Santisteban et al.2016). Thus, in order to retrieve the physical parameters ofthe WD, we have modelled the SED in quiescence, F q , asa combination of a WD atmosphere model F wd broadenedto the instrumental resolution ( R ∼ ) and a power-lawcomponent: F q ( λ ) = R wd d F wd + A · ( λ ) Γ , (3)where R wd is the radius of the WD, d the distance, A is thenormalisation factor of the power law and Γ is the powerlaw index. We have masked the cores of the Balmer linesand fitted the range between 3900 – 7500 ˚A. We fixed thedistance from the Gaia
DR2 estimate, R W D = . R (cid:12) (ra-dius for a 0.8 M (cid:12) and log ( g ) = . WD star) and left Γ andthe normalisation as free parameters. We then performed agrid search over a set of WD atmosphere models made with tlusty/synspec (Hubeny & Lanz 1995, 2017). This gridconsisted of a single value of surface gravity log ( g ) = . which corresponds to the mean WD mass for CVs (Zoro-tovic et al. 2011) and a range of effective temperatures,T ef f , 8000–30000 in steps of 100 K. The best fit model is Available at https://github.com/trmrsh/trm-doppler
MNRAS , 1–14 (2018)
Hern´andez Santisteban et al.
Wavelength / Å F / er g s c m Å v / km s Figure 3.
Optical spectral evolution throughout the superoutburst. We show the average spectrum for the corresponding epochs,labelled by our date notation, defined in Section 3. The GTC spectrum has been smoothed for clarity.
Inset:
The line profile evolutionof H α . R a d i a l V e l o c i t y / k m s Figure 4.
Radial Velocity curve of H α in quiescence for theGTC spectra. Radial velocities were obtained via 2–Gaussiantechnique. The best fit is shown as the blue line and 1 σ errorbars have been scaled so χ ν = . shown in Fig. 6, where we show the broad absorption wingsof the Balmer series. We used this relation and obtainedT ef f = , ± K for the WD temperature and a powerlaw index of Γ = − . ± . . The accretion disc contributes41 ± % of the optical light in this region, similar to othershort orbital period systems (Aviles et al. 2010; Zharikov et al. 2013). The contribution of the disc is lower at later epochs( ∼ %), when the system reaches its pre-outburst flux level,assuming the WD temperature does not cool down signifi-cantly during this period. However, the lack of spectroscopyat later date prevents us to confirm this scenario.We can also approximate the observed quiescent accre-tion disc spectrum using a spectrum of the optically thinslab which mimics the radiation of the quiescent disc. Weused simplified one-zone approximation for the spectrum cal-culation, considering homogeneous slab. The spectrum wascomputed using the well known solution for the homoge-neous slab F λ = π B Λ ( − exp (− κ λ ( T , ρ ) ∗ ρ ∗ z )) , where z isthe geometrical slab thickness, ρ is the matter density inthe slab, and T is the slab temperature. The true opacity κ λ was computed in LTE approximations using correspond-ing subroutines from Kurucz’s code atlas (Kurucz 1970,1993) adopted by V. Suleimanov (Suleymanov 1992; Ibragi-mov et al. 2003; Suleimanov & Werner 2007). The true opac-ity means that we considered only bound-bound, bound-free, and free-free transitions and ignored electron scatter-ing. The slab parameters were tuned by hand to find theones for which the model shows the best agreement with theobservations. The accepted parameters are T =
28 4000 K, z = × cm, ρ = . × − g cm − , and V sin i = km s − .The derived size of the slab is R sl √ cos i = . × cm. Weassumed that the slab has the solar chemical composition.We note that the calculated slab spectrum is in agreement MNRAS000
28 4000 K, z = × cm, ρ = . × − g cm − , and V sin i = km s − .The derived size of the slab is R sl √ cos i = . × cm. Weassumed that the slab has the solar chemical composition.We note that the calculated slab spectrum is in agreement MNRAS000 , 1–14 (2018) he bow-shock nebula in V1838 Aql Wavelength (Å)
I/I c -1000 -500 0 500 1000V (km/s)00.511.52 P h a s e -1000 -500 0 500 1000 V x (km/s) -1000-50005001000 V Y ( k m / s ) -1000 -500 0 500 1000-1000-50005001000 Wavelength (Å)
I/I c -1000 -500 0 500 1000V (km/s)00.511.52 P h a s e -1000 -500 0 500 1000 V x (km/s) -1000-50005001000 V Y ( k m / s ) -1000 -500 0 500 1000-1000-50005001000 Wavelength (Å)
I/I c -1000 -500 0 500 1000V (km/s)00.511.52 P h a s e -1000 -500 0 500 1000 V x (km/s) -1000-50005001000 V Y ( k m / s ) -1000 -500 0 500 1000-1000-50005001000 H α He I 5875 He I 6678H α He I 5875 He I 6678
Figure 5.
Time-resolved spectra of V1838 Aql during quiescence for H α , He I 5875˚A and 6678˚A. For each line we show the medianline profile ( top ), trail spectra ( middle ), and Doppler tomogram reconstructed for each line ( bottom ). The Roche lobe surface for thedonor (solid line) and primary (dotted line) were calculated using following orbital parameters: i = ◦ , q = . and γ = − km s − .The Keplerian and ballistic trajectories in the figure are marked as the upper and lower curves, respectively. The crosses are the velocity(from top to bottom) of the secondary star, the centre of mass and the primary star.MNRAS , 1–14 (2018) Hern´andez Santisteban et al.
100 50 0 50 100 / Å N o r m a li s e d F l ux + C on s t a n t HHHH
Figure 6.
WD atmosphere model best fit (red) of the Balmerseries wings. The emission core originating from the accretiondisc has been masked. An arbitrary offset has been applied forclarity. with the power law fit obtained above, but shows less con-tribution to the total system flux in the NIR wavelengths.Initially, we independently calculated the distance toV1838 Aql previous to the
Gaia
DR2 release. Usingthe mean value of the absolute magnitude of 340 WhiteDwarfs in binary systems from the catalogue of McCook–Sion (McCook & Sion 1999), we obtained a mean equalto (cid:104) M (cid:105) = . . We assumed that the brightness of the ac-cretion disc of a period bouncer system contributes from ∼ − % of the brightness of the primary star (Aviles etal. 2010; Zharikov et al. 2013). The faintest magnitude ob-served for V1838 Aql is ( V = . ), combined with the meanabsolute magnitude obtained from the McCook–Sion cata-logue, we find a distance interval ± pc, consistent withthe distance measured afterwards by Gaia .The temperature of the WD is both consistent withtheoretical predictions (Townsley & Bildsten 2003; Kniggeet al. 2011) and observational estimates for systems close tothe period minimum (Zharikov & Tovmassian 2015; Pala etal. 2017). Furthermore, the measured T ef f lies within theinstability strip, where pulsations are observed in isolated(Gianninas et al. 2006) and accreting WDs (e.g. GW Lib,Szkody et al. 2010). Future high-time resolution photometryin the ultraviolet or blue optical bands might provide a newcandidate to study non-radial pulsations in WDs (e.g. Uthaset al. 2012). Both theoretical predictions and observations of the CV pop-ulation show a sharp cut–off in the orbital period distribu-tion at about 80 min, the so-called period minimum (Ritter& Kolb 1998; G¨ansicke et al. 2009). V1838 Aql, with an or-bital period about ∼ min (Kato et al. 2014a; Echevarr´ıaet al. 2019), is located within the spread of systems aroundthe period minimum and therefore makes it difficult to es-tablish if the system is approaching or leaving it (i.e. periodbouncer).The light curve of its first (and only) recorded superout-burst (with no normal outbursts detected) can help addressthis question. Its morphology (e.g. amplitude, duration), hassimilar features observed in other WZ Sge–type systems.We observe a sudden drop in flux (both in optical and X-rays, see Fig. 1), followed by a gradual decrease back to itsquiescent level (likely from the steady cooling of the WD,e.g. Neustroev et al. 2017). In particular, we note the lackof rebrightenings in V1838 Aql which are present in mostWZ Sge–type stars.Following the empirical morphological classification ofsuperoutburst light curves proposed by Imada et al. (2006),V1838 Aql belongs to the type D morphology (i.e. no re-brightenings and sudden flux drop). Kato (2015) ascribesthis morphology classification to an evolutionary sequencefrom pre- to post- period minimum systems (C:D:A:B:E)and points out that type
D might be closely associated withsystems around the period minimum, but still in the upperbranch of the q − P orb branch. This is consistent with thelarger contribution in the NIR by the donor (see Section 5.2),which suggests V1838 Aql to be a pre-bounce system.Another way to discern if a system is a period bounceris to look for distinct features in quiescence such as perma-nent double-hump light curve as well as a spiral arm struc-ture in their Doppler tomography (Zharikov & Tovmassian2015). Echevarr´ıa et al. (2019) presented two light curves ofV1838 Aql obtained in 2018 which do not show a double-hump modulation. However, the lack of a precise ephemerisprevents to further explore the presence of more subtle mod-ulations. With regards to a spiral pattern, we observed adistinct feature in the accretion disc in a peculiar positionof the velocity space (see H α tomogram in Section 4.3). Thisfeature however, is different to the ones observed in good pe-riod bouncer candidates as V406 Vir (Zharikov et al. 2008)or EZ Lyn (Zharikov et al. 2013), where a dual-emitting com-ponent was associated to spiral patterns in the disc. On theother hand, it resembles more to those found in HT Cas (e.g.Neustroev et al. 2016), as discussed further in Section 4.3.In any case, until the real phasing of the system is found nodefinitive conclusion on the origin of these structure can bedrawn.Period bouncers have small mass ratios ( q (cid:46) . ). Inconsequence, the more massive WD should present smallsemi-amplitudes, K , in their radial velocity curves. In orderto test this for V1838 Aql it is necessary to estimate the in-clination angle of the system. The quiescent spectrum clearlyshows a double–peak emission on the Balmer series as wellin the He i which is observed in accretion discs with inclina-tion angles i (cid:38) ◦ (Horne & Marsh 1986). Also, the lack of MNRAS000
D might be closely associated withsystems around the period minimum, but still in the upperbranch of the q − P orb branch. This is consistent with thelarger contribution in the NIR by the donor (see Section 5.2),which suggests V1838 Aql to be a pre-bounce system.Another way to discern if a system is a period bounceris to look for distinct features in quiescence such as perma-nent double-hump light curve as well as a spiral arm struc-ture in their Doppler tomography (Zharikov & Tovmassian2015). Echevarr´ıa et al. (2019) presented two light curves ofV1838 Aql obtained in 2018 which do not show a double-hump modulation. However, the lack of a precise ephemerisprevents to further explore the presence of more subtle mod-ulations. With regards to a spiral pattern, we observed adistinct feature in the accretion disc in a peculiar positionof the velocity space (see H α tomogram in Section 4.3). Thisfeature however, is different to the ones observed in good pe-riod bouncer candidates as V406 Vir (Zharikov et al. 2008)or EZ Lyn (Zharikov et al. 2013), where a dual-emitting com-ponent was associated to spiral patterns in the disc. On theother hand, it resembles more to those found in HT Cas (e.g.Neustroev et al. 2016), as discussed further in Section 4.3.In any case, until the real phasing of the system is found nodefinitive conclusion on the origin of these structure can bedrawn.Period bouncers have small mass ratios ( q (cid:46) . ). Inconsequence, the more massive WD should present smallsemi-amplitudes, K , in their radial velocity curves. In orderto test this for V1838 Aql it is necessary to estimate the in-clination angle of the system. The quiescent spectrum clearlyshows a double–peak emission on the Balmer series as wellin the He i which is observed in accretion discs with inclina-tion angles i (cid:38) ◦ (Horne & Marsh 1986). Also, the lack of MNRAS000 , 1–14 (2018) he bow-shock nebula in V1838 Aql eclipses in the photometry imposes an upper limit, for rea-sonable mass ratios of period bouncers, i (cid:46) ◦ (for q (cid:39) . ,Bailey 1990). In addition, V1838 Aql shows small peak-to-peak separations ( ∼ − km s − for H α ) in contrastto eclipsing systems at similar orbital periods (e.g. ∼ km s − for SDSS J1433+0038, Tulloch et al. 2009). Com-paring this with other eclipsing short–periods CVs, we con-clude that the inclination angle is relatively small ( i ∼ ◦ )Thus, the real not projected semi-amplitude is very high( K ∼ km s − ). Assuming the q determination from su-perhumps (Echevarr´ıa et al. 2019), the donor in V1838 Aqlwould lie above the sub-stellar threshold. Again, this wouldargue for a pre-bounce system. The large wavelength range of the GTC spectrum allowedus to search for evidence of the donor (e.g. absorption fea-tures) as shown in Fig. 7. In particular, we searched in theregion between 0.7–1 µ m where the SED of the donor shouldstart to contribute a significant percentage of the system’slight. However, we find a wide range of the spectrum to bemostly dominated by broad emission Paschen lines. In thenon–contaminated regions, we find no absorption featuresassociated with the donor.Despite the lack of evidence of features of the donorstar in the optical region, the donor continuum should con-tribute as significant fraction at longer wavelengths. We usedthe multi-coloured broadband photometry in a true quies-cent state to illustrate the type of donor expected from itsNIR properties . Theoretically, the orbital period suggeststhat the donor is likely to have a spectral type later thanM6 (T (cid:46) K, Knigge 2006; Knigge et al. 2011) de-pending whether the system lies before or after the periodminimum. We show in Fig. 7 two models using the fit donein the optical region (where the donor contribution is mini-mal) scaling down the power-law contribution to match thephotometry and adding a low-mass star atmosphere modelof
K and
K (Baraffe et al. 2015) appropriate fora pre- and post-period minimum system at P orb ∼ min,respectively. No fit has been performed and only presentedhere as reference.The photometry shows a flux increase starting at ∼ We note that at the time of the GTC spectrum (day 1566), thesystem is slightly brighter than its pre-outburst state taken byPAN-STARRS1 in 2012 and a later observation taken with theNTT (day 2032). Unfortunately, we do not have information ofthe NIR flux close to the GTC epoch. This suggests that even3 years after the superoutburst, either the WD is still cooling orthe disc remains brighter. nature. This is particularly important since few systems havebeen characterised close to the period minimum, where thetransition from stellar to sub-stellar regime is expected.
We have presented a long–term photometric and spectro-scopic study of the 2013 superoutburst of V1838 Aql fromits peak to quiescence. The morphology of the outburstin combination with previous photometric estimates (Katoet al. 2014a; Echevarr´ıa et al. 2019) allow us to confirmV1838 Aql as a short orbital period system and provideupdated ephemeris. A few days after the peak of the out-burst, we discovered extended emission around the object, asobserved in the high-resolution spectroscopy. Further deepH α revealed an illuminated bow shock consistent with theproper motion of the object ( ± km s − ). Althoughthe origin of the material that creates the bow-shock is un-clear, we conclude that a quasi-continuous outflow of mate-rial ( ∼ km s − ) is required to sustain a standing bowshock with the ISM.In quiescence, we obtained time–resolved spectroscopywhich allowed us to determine a semi–amplitude of the pri-mary K = ± km s − . Doppler tomography in H α re-vealed an emission component inconsistent with the ballis-tic trajectory of the accretion stream observed in the He i lines. Further observations and refinement of the ephemerisare needed to discern the origin of this emitting component.The broad band spectroscopy, allowed us to infer the ef-fective temperature of the primary T ef f = , ± K.This is consistent with theoretical expectations (Townsley &Bildsten 2003) as well as observational constrains on similarsystems (Pala et al. 2017).A discussion is made on the possible period-bouncernature of the object. The broadband optical and NIR pho-tometry suggests that V1838 Aql is approaching the pe-riod minimum limit and hosts a > K donor. How-ever, we point out that simultaneous optical and infraredtime-resolved spectroscopy (or photometry) needs to be per-formed, in order to measure the radial velocity curves of bothcomponents and to determine the spectral type of the donorstar and its possible sub-stellar nature.
ACKNOWLEDGEMENTS
The authors are indebted to DGAPA (Universidad Na-cional Aut´onoma de M´exico) for financial support, PA-PIIT projects IN111713, IN122409, IN100617, IN102517,IN102617, IN108316 and IN114917. JVHS is supported by aVidi grant awarded to N. Degenaar by the Netherlands Or-ganization for Scientific Research (NWO) and acknowledgestravel support from DGAPA/UNAM. JE acknowledges sup-port from a LKBF travel grant to visit the API at UvA. VNacknowledges the financial support from the visitor and mo-bility program of the Finnish Centre for Astronomy withESO (FINCA), funded by the Academy of Finland grantNo. 306531. GT acknowledges CONACyT grant 166376. E.de la F. wishes to thank CGCI–UdeG staff for mobility sup-port. VS thanks Deutsche Forschungsgemeinschaft (DFG)for financial support (grant WE 1312/51-1). His work was
MNRAS , 1–14 (2018) Hern´andez Santisteban et al.
Wavelength / m F / × e r g s c m Å T =2200 KT =1900 KWD, T eff = 11, 600 KH slab Power law, = 1.4PAN-STARRS1 2012NOT 2017-04-08NTT 2017-10-06 Figure 7.
Broadband SED of V1838 Aql in quiescence. We show the multi-band photometry before (PAN-STARRS) and post-outburst(NOT and NTT). Two models have been scaled to compare the NIR contribution of the donor using low-mass star atmospheres of 2200K (red) and 1900 K (orange), corresponding to a pre and post-bounce system (Knigge et al. 2011) respectively. The WD and power lawindex are taken from the Balmer wing fit. The power law contribution has been scaled by hand to match the quiescent level ( ∼ %contribution). The thin slab model is also scaled down and show as reference. also funded by the subsidy allocated to Kazan Federal Uni-versity for the state assignment in the sphere of scientificactivities (3.9780.2017/8.9).We thank Tom Marsh for the use of molly . We ac-knowledge with thanks the variable star observations fromthe AAVSO International Database contributed by observersworldwide and used in this research. We acknowledge theuse of public data from the Swift data archive. This re-search made use of astropy , a community–developed core python package for Astronomy (Astropy Collaboration etal. 2013), matplotlib (Hunter 2007) and aplpy (Ro-bitaille & Bressert 2012). Based (partly) on observationsmade with the Gran Telescopio Canarias (GTC), installedin the Spanish Observatorio del Roque de los Muchachosof the Instituto de Astrof´ısica de Canarias, in the island ofLa Palma (GTC7-16AMEX). Partly based on observationsmade with the Nordic Optical Telescope, operated by theNordic Optical Telescope Scientific Association at the Ob-servatorio del Roque de los Muchachos, La Palma, Spain,of the Instituto de Astrofisica de Canarias. The data pre-sented here were obtained in part with ALFOSC, which isprovided by the Instituto de Astrofisica de Andalucia (IAA)under a joint agreement with the University of Copenhagenand NOTSA. The results presented in this paper are basedon observations collected at the European Southern Obser-vatory under programme ID 0100.D-0932. We thank theday and night–time support staff at the OAN–SPM for fa-cilitating and helping obtain our observations. This workhas made use of data from the European Space Agency(ESA) mission
Gaia ( ),processed by the Gaia
Data Processing and Analysis Con-sortium (DPAC, ). Funding for the DPAC has been pro- vided by national institutions, in particular the institutionsparticipating in the
Gaia
Multilateral Agreement. We thankJ. van den Eijnden for help on
Swift’s
DDT proposal.
REFERENCES
Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., et al.2013, A&A, 558, A33Aviles, A., Zharikov, S., Tovmassian, G., et al. 2010, ApJ, 711,389Bailer-Jones, C. A. L., Rybizki, J., Fouesneau, M., Mantelet, G.,& Andrae, R. 2018, AJ, 156, 58Bailey, J. 1990, MNRAS, 243, 57Baraffe, I., Homeier, D., Allard, F., & Chabrier, G. 2015, A&A,577, A42Bond, H. E., & Miszalski, B. 2018, PASP, 130, 094201Breedt, E., G¨ansicke, B. T., Drake, A. J., et al. 2014, MNRAS,443, 3174Buzzoni B., et al., 1984, Msngr, 38, 9Chevalier, R. A., Kirshner, R. P., & Raymond, J. C. 1980, ApJ,235, 186Clarke, J. T., & Bowyer, S. 1984, A&A, 140, 345Coppejans, D. L., K¨ording, E. G., Knigge, C., et al. 2016, MN-RAS, 456, 4441de Miguel, E., Patterson, J., Cejudo, D., et al. 2016, MNRAS,457, 1447Echevarr´ıa, J., de la Fuente, E. & Costero, R. 2007, AJ, 197, 565Echevarr´ıa, J., de Miguel, E., Hern´andez Santisteban, J. V., et al.2019, Rev. Mex. Astron. Astrofis., 55, 1. (arXiv:1810.09864)Foreman-Mackey, D. 2017, Astrophysics Source Code Library,ascl:1702.002Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018,A&A, 616, A1G¨ansicke, B. T., Dillon, M., Southworth, J., et al. 2009, MNRAS,397, 2170 MNRAS000
Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., et al.2013, A&A, 558, A33Aviles, A., Zharikov, S., Tovmassian, G., et al. 2010, ApJ, 711,389Bailer-Jones, C. A. L., Rybizki, J., Fouesneau, M., Mantelet, G.,& Andrae, R. 2018, AJ, 156, 58Bailey, J. 1990, MNRAS, 243, 57Baraffe, I., Homeier, D., Allard, F., & Chabrier, G. 2015, A&A,577, A42Bond, H. E., & Miszalski, B. 2018, PASP, 130, 094201Breedt, E., G¨ansicke, B. T., Drake, A. J., et al. 2014, MNRAS,443, 3174Buzzoni B., et al., 1984, Msngr, 38, 9Chevalier, R. A., Kirshner, R. P., & Raymond, J. C. 1980, ApJ,235, 186Clarke, J. T., & Bowyer, S. 1984, A&A, 140, 345Coppejans, D. L., K¨ording, E. G., Knigge, C., et al. 2016, MN-RAS, 456, 4441de Miguel, E., Patterson, J., Cejudo, D., et al. 2016, MNRAS,457, 1447Echevarr´ıa, J., de la Fuente, E. & Costero, R. 2007, AJ, 197, 565Echevarr´ıa, J., de Miguel, E., Hern´andez Santisteban, J. V., et al.2019, Rev. Mex. Astron. Astrofis., 55, 1. (arXiv:1810.09864)Foreman-Mackey, D. 2017, Astrophysics Source Code Library,ascl:1702.002Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018,A&A, 616, A1G¨ansicke, B. T., Dillon, M., Southworth, J., et al. 2009, MNRAS,397, 2170 MNRAS000 , 1–14 (2018) he bow-shock nebula in V1838 Aql Gehrels, N., Chincarini, G., Giommi, P., et al. 2004, ApJ, 611,1005Gianninas, A., Bergeron, P., & Fontaine, G. 2006, AJ, 132, 831Goliasch, J., & Nelson, L. 2015, ApJ, 809, 80Greiner J., et al. 2001, A&A, 376, 1031Harrison, T. E. 2016, ApJ, 816, 4Hellier, C. 2001, in
Cataclysmic Variable Stars , Springer, PraxisPublishing, Chichester, UK, p. 165Hern´andez Santisteban, J. V., Knigge, C., Littlefair, S. P., et al.2016, Nature, 533, 366Hern´andez Santisteban, J. V., Echevarr´ıa, J., Michel, R., & Cos-tero, R. 2017, MNRAS, 464, 104Hollis, J. M., Oliversen, R. J., Wagner, R. M., & Feibelman, W. A.1992, ApJ, 393, 217Horne, K., Wade, R. A., & Szkody, P. 1986, MNRAS, 219, 791Horne, K. and Marsh, T. R. 1986, MNRAS, 218, 761Howell, S. B., Rappaport, S., & Politano, M. 1997, MNRAS, 287,929Howell, S. B., Hoard, D. W., Brinkworth, C., Kafka, S., Walen-tosky, M.J., Walter, F. M. & Rector, T.A. 2008, ApJ, 685,418Hubeny, I., & Lanz, T. 1995, ApJ, 439, 875Hubeny, I., & Lanz, T. 2017, arXiv:1706.01859Hunter, J. D. 2007, Computing In Science & Engineering, 9, 90Hurst G. 2013, The Astronomer Electronic Circular, 2919Ibragimov A. A., Suleimanov V. F., Vikhlinin A., SakhibullinN. A., 2003, ARep, 47, 186Imada, A., Kubota, K., Kato, T., Nogami, D., Maehara, H., Naka-jima, K., Uemura, M., & Ishioka, R. 2006, PASJ, 58, 23Itagaki, K. et al. 2013, CBET, 3554, 1Kato T. 2015, PASJ, 67, 108Kato, T., & Osaki, Y. 2013, PASJ, 65, 115Kato T. et al. 2009, PASJ, 61, 395Kato T. et al. 2014a, PASJ, 66, 90Kato, T., et al. 2014b, PASJ, 66, 30Klemola, A. R., Hanson, R. B., Jones, B. F., & Monet, D. G.2004, VizieR Online Data Catalog, 1293,Knigge, C. 2006, MNRAS, 373, 484Knigge, C., Baraffe, I., & Patterson, J. 2011, ApJS, 194, 28Kolb, U. & Baraffe, I. 1999, MNRAS, 309, 1034Kurucz R. L., 1970, SAOSR, 309, 309Kurucz R., 1993, KurCD, 13Lang, D., Hogg, D. W., Mierle, K., Blanton, M., & Roweis, S.2010, AJ, 139, 1782Lenz P., & Breger M. 2005, Commun. Asteroseismol., 146, 53Lindegren, L., Lammers, U., Bastian, U., et al. 2016, A&A, 595,A4Littlefair, S. P., Dhillon, V. S., Marsh, T. R., et al. 2006, Science,314, 1578Marsh, T. R. 1988, MNRAS, 231, 1117McCook, G. P. & Sion, E. M. 1999, ApJS, 121, 1McLaughlin, D.B. 1960, in
Stallar Atmospheres , ed. J.L. Green-stein, Univ. Chicago Press, Chicago, 585Metzger, B. D., Hasco¨et, R., Vurm, I., et al. 2014, MNRAS, 442,713Moorwood A., Cuby J.-G., Lidman C., 1998, Msngr, 91, 9Neustroev, V. V., & Zharikov, S. 2008, MNRAS, 386, 1366Neustroev, V. V., Zharikov, S. V. & Borisov, N.V. 2016, A&A,586, 10Neustroev, V. V., Marsh, T. R., Zharikov, S. V., et al. 2017,MNRAS, 467, 597Neustroev V. V., et al., 2018, A&A, 611, A13Osaki, Y. 1974, PASJ, 26, 429Osaki, Y. 1994, NATO Advanced Science Institutes (ASI) SeriesC, 417, 93Osterbrock, D.E., Fulbright, J.P., Martel, A.R., Keane, M.J.&Trager S.C. 1996, PASP, 108, 277 Otulakowska–Hypka, M., Olech, A., & Patterson, J. 2016, MN-RAS, 460, 2526Paczynski, B., & Sienkiewicz, R. 1981, ApJ, 248, 27Pala, A. F., G¨ansicke, B. T., Townsley, D., et al. 2017, MNRAS,466, 2855Pala, A. F., Schmidtobreick, L., Tappert, C., G¨ansicke, B. T., &Mehner, A. 2018, MNRAS,Patterson, J. 1988, PASP, 110, 1132Patterson, J. 2011, MNRAS, 411, 2695Patterson, J. et al. 1996, PASP, 108, 748Patterson, J. et al. 2005, PASP, 117, 1204Patterson, J., Thorstensen, J. R., & Knigge, C. 2008, PASP, 120,510Paunzen, E., & Vanmunster, T. 2016, AstronomischeNachrichten, 337, 239Reis R. C., Wheatley P. J., G¨ansicke B. T., Osborne J. P., 2013,MNRAS, 430, 1994Ritter, H. & Kolb, U. 1998, A&A, 128, 83Robitaille, T., & Bressert, E. 2012, Astrophysics Source CodeLibrary, ascl:1208.017Sahman, D. I., Dihllon, V. S., Knigge, C. & Marsh, T. R. 2015,MNRAS, 451, 2863Shafter, A. W., Szkody, P., & Thorstensen, J. R. 1986, ApJ, 308,765Skillman D.R., Patterson J. 1993, ApJ, 417, 298Smak, J. 1993, Acta Astron., 43, 101Stellingwerf, R. F. 1978, ApJ, 224, 953Stetson, P. B. 1987, PASP, 99, 191Suleymanov V. F., 1992, SvAL, 18, 104Suleimanov V., Werner K., 2007, A&A, 466, 661Szkody, P., Mukadam, A., G¨ansicke, B. T., et al. 2010, ApJ, 710,64Tovmassian G., G¨ansicke B., Zharikov S., Ramirez A., Diaz M.2011, IAUS, 275, 311Townsley, D. M., & Bildsten, L. 2003, ApJ, 596, 227Tulloch, S. M., Rodr´ıguez–Gil, P., & Dhillon, V. S. 2009, MNRAS,397, 82Uthas, H., Knigge, C., Long, K. S., Patterson, J., & Thorstensen,J. 2011, MNRAS, 414, 85Uthas, H., Patterson, J., Kemp, J., et al. 2012, MNRAS, 420, 379van Buren, D., & McCray, R. 1988, ApJ, 329, 93Vogt, N. 1982, ApJ, 252, 653Warner, B. 1995,
Cataclysmic Variable Stars , (Cambridge Uni-versity Press)Weaver, R., McCray, R., Castor, J., Shapiro, P., & Moore, R.1977, ApJ, 218, 377Wilkin, F. P. 1996, ApJ, 459, 31Yoon, D., & Heinz, S. 2017, MNRAS, 464, 3297Zharikov, S., Tovmassian, G., Napiwotzki, R., Michel, R. $Neustroev, V. V. 2008, A&A, 449, 645Zharikov, S., Tovmassian, G., Aviles, A., et al. 2013, A&A, 549,77Zharikov, S. & Tovmassian, G. 2015, Acta Polytechnica CTUProceedings 2, 41Zorotovic, M., Schreiber, M. R., & G¨ansicke, B. T. 2011, A&A,536, A42
APPENDIX A:
GAIA
POSTERIORDISTRIBUTIONS
We present the joint and marginal posterior distributions ofthe
Gaia distance and tangential velocity, shown in Fig. A1(see Sec. 3.1). We used corner.py (Foreman-Mackey 2017)to visualise the MCMC chains.
MNRAS , 1–14 (2018) Hern´andez Santisteban et al. d = 202.58 +7.646.74 v v = 123.20 +4.634.13
180 195 210 225 d . . .
360 112 120 128 136 v .
368 2 .
364 2 . = 2.36 +0.000.00 Figure A1.
Posterior probability distributions for the
Gaia distance and tangential velocity. Colour scale contours show the jointprobability for every combination of parameters. Units for each parameter are: distance in pc, velocity in km s − and angle in radians.Contours represent the 0.5 σ , 1 σ , 2 σ and 3 σ levels. Marginal posterior distributions are shown as histograms with the median and 1 σ marked as dashed lines.This paper has been typeset from a TEX/L A TEX file prepared bythe author. MNRAS000