FLASHING: New high-velocity H 2 O masers in IRAS 18286 − 0959
Hiroshi Imai, Yuri Uno, Daichi Maeyama, Ryosuke Yamaguchi, Kei Amada, Yuhki Hamae, Gabor Orosz, José F. Gómez, Daniel Tafoya, Lucero Uscanga, Ross A. Burns
aa r X i v : . [ a s t r o - ph . S R ] S e p Publ. Astron. Soc. Japan (2018) 00(0), 1–9doi: 10.1093/pasj/xxx000 FLASHING: New high-velocity H O masers inIRAS 18286 − Hiroshi I
MAI , Yuri U NO , Daichi M AEYAMA , Ryosuke Y AMAGUCHI , KeiA MADA , Yuhki H AMAE , Gabor O ROSZ , Jos ´e F. G ´
OMEZ , DanielT AFOYA , Lucero U
SCANGA , and Ross A. B URNS Amanogawa Galaxy Astronomy Research Center, Graduate School of Science andEngineering, Kagoshima University,1-21-35 Korimoto, Kagoshima 890-0065 Center for General Education, Institute for Comprehensive Education, Kagoshima University,1-21-30 Korimoto, Kagoshima 890-0065 Department of Physics and Astronomy, Graduate School of Science and Engineering,Kagoshima University,1-21-35 Korimoto, Kagoshima 890-0065 Department of Physics and Astronomy, Faculty of Science, Kagoshima University,1-21-35 Korimoto, Kagoshima 890-0065 School of Natural Sciences, University of Tasmania, Private Bag 37, Hobart, Tasmania 7001,Australia Xinjiang Astronomical Observatory, Chinese Academy of Sciences, 150 Science 1-Street,Urumqi, Xinjiang 830011, China Instituto de Astrof´ısica de Andaluc´ıa, CSIC, Glorieta de la Astronom´ıa s/n, E-18008Granada, Spain National Astronomical Observatory of Japan, 2-21-1 Osawa, Mitaka, Tokyo, 181-8588, Japan Department of Space, Earth and Environment, Chalmers University of Technology, OnsalaSpace Observatory, 439 92 Onsala, Sweden Departamento de Astronom´ıa, Universidad de Guanajuato, A.P. 144, 36000 Guanajuato,Gto., Mexico Korea Astronomy and Space Science Institute, 776 Daedeokdae-ro, Yuseong-gu, Daejeon34055, Republic of Korea ∗ E-mail: [email protected]
Received 2019 October 13; Accepted 2020 April 16
Abstract
We discovered new high-velocity components of H O maser emission in one of the “waterfountain” sources, IRAS 18286 − O-maser Ignitionsby Nobeyama Generation) project since 2018 December. The maser spectra show new, ex-tremely high expansion velocities ( >
200 km s − projected in the line of sight) components,some of which are located symmetrically in the spectrum with respect to the systemic velocity.They were also mapped with KaVA (KVN and VERA Combined Array) in 2019 March. Welocated some of these maser components closer to the central stellar system than other highvelocity components (50–200 km s − ) that have been confirmed to be associated with theknown bipolar outflow. The new components would flash in the fast collimated jet at a speed c (cid:13) Publications of the Astronomical Society of Japan , (2018), Vol. 00, No. 0 over 300 km s − (soon) after 2011 when they had not been detected. The fastest of the newcomponents seem to indicate rapid deceleration in these spectra, however our present mon-itoring is still too sparse to unambiguously confirm it (up to 50 km s − yr − ) and too short toreveal their terminal expansion velocity, which will be equal to the expansion velocity that hasbeen observed ( v exp ∼
120 km s − ). Future occurrences of such extreme velocity componentsmay provide a good opportunity to investigate possible recurrent outflow ignitions. Thus sculp-ture of the parental envelope will be traced by the dense gas that is entrained by the fast jetand exhibits spectacular distributions of the relatively stable maser features. Key words: masers — stars: AGB and post-AGB — stars: individuals(IRAS 18286 − Our understanding of the final stellar evolution exhibitingcopious and inhomogeneous mass loss and the subsequentformation of planetary nebulae (PNe) with a wide varietyof their morphologies. It is considered that such transfor-mation will happen during a relatively short period, fromthe end of the asymptotic giant branch (AGB) phase ofstellar evolution to the beginning of the post-AGB phase,when an AGB stellar wind rapidly develops a circumstellarenvelope (CSE) (e.g. Lewis 2001). It has been confirmedthat highly collimated, fast jets will be newly ejected fromsome of those stars (e.g. Sahai & Trauger 1998). Thesejets shape the previously ejected CSE, so they may beultimately responsible for the final morphologies of PNe.However, two open questions remain: the launching mech-anism and the timescale of such jet ejections.“Water fountain” (WF) sources have been identified asAGB or post-AGB stars hosting fast bipolar jets tracedby H O maser emission (Imai 2007). The WF candidateswere first confirmed in the maser spectra with extreme ve-locity widths much larger than those of typical 1612 MHzOH maser emission exhibiting clear double-horned spec-tral profiles (10–20 km s − , e.g., Likkel et al. 1992, c.f.,Engels & Bunzel 2015). There exist only 15 WFs con-firmed to date (G´omez et al. 2017), implying an extremelyshort timescale of the phase when such high velocity cir-cumstellar H O masers are visible ( <
100 yr, Imai 2007). Interferometric observations of these H O masers showa wide variety of the spatio-kinematical structures of themasers, reflecting different types of the central stellar sys-tems driving these jets and/or different stages of the jetevolution. The diversity of the central star properties isindicated by the variety of the substructures in the maserspatio-kinematics such as multiple arcs, pairs of maserfeature groups distributed point symmetrically, groups ofmaser features containing a large internal velocity disper-sion ( >>
10 km s − , e.g., Imai et al. 2002; Claussen et al.2009; Imai et al 2013b; Orosz et al. 2019). The temporal evolution of each jet can be also moni-tored in the cases where its growth rate, namely the changerate of a total length of the maser jet on the sky, is con-sistent with the mean of the proper motion speeds of theindividual maser features (Imai et al. 2007; Chong et al.2015). Recent observations of thermal molecular line anddust emission with the Atacama Large Millimeter-and-submillimeter Array (ALMA) have elucidated the com-plete spatio-morphologies of WF sources (e.g., Sahai et al.2017; Tafoya et al. 2019; Tafoya et al. 2020). In someWFs, H O masers are suggested to trace the interactionof the fast jet with the parental CSE sculptured by the jetand resulting deceleration of the jet that leads entrainedmaterial along the jets. The jet deceleration is suggestedby the spatio-kinematics of the H O masers themselves(Orosz et al. 2019). However, other WFs show an oppositetrend, with accelerating masers (e.g., G´omez et al. 2015),which could indicate a different nature of these WF phe-nomenon, maybe related to the evolutionary phase of thesources.Taking into account the rarity of WFs, their possiblyrapid evolution, and the possibility of the appearance ofnew H O masers in the known WFs and WF candidates,monitoring of these H O masers in single-dish observationswith moderate cadence (1–8 weeks) is crucial for under-standing the phenomena of WFs in a unified scheme. Thepossible deceleration of the jets may be proven by detect-ing H O masers at higher velocities with decreasing veloc-ity drifts. A periodic behavior of the maser spectrum mayindicate either periodic ejections of jets or pulsation of thecentral evolved star. In the latter case, synergistic monitor-ing of 1612 MHz OH masers and SiO (in addition to H O)masers will provide more concrete evidence for the stellarpulsation (e.g. Herman & Habing 1985). The weaknessof some velocity components of H O masers in WFs re-quires highly sensitive telescopes for systematic monitoringobservations, and possibly an internationally coordinatedprogram that can manage regular monitoring activities forthe known WFs and WF candidates. The Nobeyama 45 ublications of the Astronomical Society of Japan , (2018), Vol. 00, No. 0 m telescope is one of the best-suited telescopes for suchmonitoring observations thanks to its dynamic schedul-ing , which provided good opportunities for high sensi-tivity H O maser observations. Moreover, the new quasi-optics enables us to simultaneously observe H O and SiOmasers (Okada et al. 2019). Therefore we have monitoredthe WFs in the FLASHING (Finest Legacy Acquisitions ofSiO and H O maser Ignitions by Nobeyama Generation)project since 2018 December.In this paper, we report the first scientific resultof FLASHING, yielded for IRAS 18286 − O masers in I18286 havebeen mapped in several very long baseline interferometry(VLBI) observations using the Very Long Baseline Array(VLBA) and the VLBI Exploration of Radio Astrometry(VERA) (Yung et al. 2011, Paper I; Imai et al. 2013a,Paper II; Imai et al 2013c, Paper III). Paper I modeledthe H O masers on a double helix pattern. Paper II de-termined the trigonometric parallax distance to I18286 tobe 3.61 +0 . − . kpc. Paper III determined the locations of1612 MHz OH masers and the low-velocity ( ∼
10 km s − )components of the H O masers in the H O maser maps,both of which may be associated with a relic CSE inI18286. Before 2010, the H O masers in I18286 covered avelocity range of − ≤ V LSR ≤
180 km s − with respect tothe local standard of rest (LSR). Yung et al. (2013) foundnew velocity components in 2011 around V LSR ∼ −
90 and230 km s − . This paper focuses on even newer, faster ve-locity components (surpassing the previously detected ve-locity range), found with our monitoring in FLASHING. While FLASHING is still active, in this paper we presentthe observed spectra towards I18286 until 2020 January 10.This dataset consist of eleven epochs: eight between 2018December and 2019 May, and three additional epochs after2019 November. Table 1 gives the summary of the single-dish observations. We employed the new quasi-optics thatenables us to observe simultaneously line emission at thefrequency bands around 22.2 and 43.0 GHz (Okada et al.2019), including the maser lines of H O J K − K + = 6 → at the rest frequency of 22.235080 GHz and SiO J = 1 → v = 3 , , , and 0) in 42.5–43.4 GHz. This paper showsonly the H O maser spectra because only this maser linewas detected toward I18286.The aperture efficiency of the telescope is 0.65, yielding The dynamic scheduling as Back-up Program had been conducted until2019 May. a flux-density conversion factor of 2.67 Jy K − , when em-ploying the new quasi-optics. The flux density scale wasobtained from the information of the system noise tem-perature measured with the chopper-wheel method. Themaser emission was received in both left- and right-handcircular polarization. Position switching was employed byalternating observations of a target maser and a blanksky position for an integration time of 30 s per scan.Three spectral windows were set to cover a bandwidth of31.25 MHz in each window and centered at the mean veloc-ity of the maser emission in the LSR frame ( ∼
65 km s − )and those higher and lower by 30 MHz than the first one.The SAM45 spectrometers were used to obtain 4096 spec-tral channels per spectral window, but the final spectrahad 2048 spectral channels after smoothing. The threespectral windows cover a velocity width of 1258 km s − intotal with a velocity resolution of 0.42 km s − .The data reduction was made using the JavaNEWSTAR software , in which integration and baselinesubtraction were performed in order to obtain the finalmaser spectra. Dependent on system noise temperatureand integration time, the 1- σ noise level of the spectraranged from 32 to 118 mJy. We identified detections ofspectral components if their flux density was higher thanthe 5- σ noise level (160–990 mJy). We then identified eachspectral peak if it was composed of three or more detectedcomponents in consecutive spectral channels. In the caseof multiple peaks, which are close in velocity and thusblended together, each peal was identified as an indepen-dent peak if its flux density was higher by 3- σ than that ofthe spectral channel of the tail of a contiguous intersectingpeak. The H O masers in I18286 were also observed on 2019March 4 for six hours with KaVA, an array composedof four telescopes of the Japanese VLBI Exploration ofRadio Astrometry (VERA) plus three of the Korean VLBINetwork (KVN). The fringe finder OT 081 and the de-lay calibrator J183005.9+061915 were also observed threetimes, every 25 min. These sources were observed in bothof left- and right-hand circular polarization with the KVNtelescopes, but only the former polarization data werevalid. With VERA, only left-hand circular polarizationwas observed, but the dual-beam system allowed simulta-neous observations of our target and the position referencesource J183220.8 − Publications of the Astronomical Society of Japan , (2018), Vol. 00, No. 0 quantization, yielding a total recording rate of 1024Mbps.The data correlation was made using DaejeonCorrelator in the Korea-Japan Correlation Center (KJCC).The recorded signals taken with both of the VERA’s dualbeams were correlated with those with the KVN tele-scopes. Two correlated data sets were provided from thetwo BBCs, including the maser source and J1832 respec-tively. 4096 spectral channels were contained for the BBCobserving I18286, yielding a velocity channel spacing of0.42 km s − , while 128 spectral channels for other BBCobserving J1832. Note that the latter data were invalid inthe VERA–KVN baselines for the scans on J1832.The data reduction was made using NRAO AIPS, whichwas handled using Python scripts and the ParselTonguepackage . Standard a-priori calibration procedures wereperformed for calibrating visibility amplitudes using the in-formation of the antenna gains and system noise tempera-tures. The delay calibration was performed using the scanson J183005.9+061915. Fringe fitting and self-calibrationwere performed using the data in a reference spectral chan-nel at an LSR velocity of 139.5 km s − , which included abright and compact maser spot. Positional offsets of themaser spots given in this paper are referred to the po-sition of this reference spot. Image cubes of the maseremission were created using the CLEAN deconvolution al-gorithm with a Gaussian synthesized beam of 2.1 × − ◦ , yielding a1- σ noise level of ∼
20 mJy in spectral channels withoutbright maser spots. Maser spots brighter than a 7- σ noiselevel were identified in the individual spectral channelsand grouped into isolated maser features, each of whichincludes maser spots located within 1–2 mas in positionalong consecutive spectral channels. We assume that theindividual maser features correspond to physical featuresor gas clumps emitting the masers.The visibility data of J1832 observed only with VERAas well as the maser data were used for precise astrome-try of the maser source. These visibilities were calibratedusing a more precise delay-tracking model. The data ofthe calibrator scans were used for instrumental delay cal-ibration, including that for relative delay and phase off-sets between VERA’s dual beams. The fringe-fitting andself-calibration solutions obtained from the maser scans asmentioned above were applied to the J1832 data, yieldingthe relative position of J1832 with respect to that of thereference maser spot. However, the astrometric accuracy ∼ < less sensitivity for J1832 in VERA, while the latter a lim-ited velocity coverage of the maser source with the fullKaVA. The self-calibration solutions taken from the maserdata were also imperfect for astrometry due to insufficientperformance of KVN whose accurate information of thestation coordinates was still unavailable. Fig. 1 shows the spectra of H O masers in I18286 takenat the first eight epochs in FLASHING. They clearlyshow new components in V LSR < −
60 km s − and V LSR >
200 km s − , which had not been identified in the previousobservations during 2006 April–2010 April (Deguchi et al.2007, Paper I). In our observations, we also find that thevelocity components in the velocity range of V LSR =130–160 km s − are flaring compared to those seen in the pre-vious observations by a factor of up to four.The central velocity between the two extremelyhigh velocity components present in all observing ses-sions (found around V LSR ∼ −
150 and 240 km s − ) is V LSR (center) ≃
45 km s − , close to the velocity of the CO J = 3 → V LSR ≃
65 km s − , Imai et al. 2009) andthe velocity of the single 1612 MHz OH maser componentin I18286 ( V LSR (OH) ∼
40 km s − , Paper III). In evolvedstars, OH masers at 1612 MHz usually show a double-horned profile, with tracing an expansion of the CSE.Here we suggest that the central velocity of the CO emis-sion traces the systemic stellar velocity ( V ∗ ≃
65 km s − ),while the velocity difference between the CO and OHpeaks represents the expansion velocity of the relic enve-lope ( V expansion ≃
25 km s − ). Furthermore, we found evenhigher velocity red-shifted components at V LSR ≃
345 and336 km s − in the spectra on 2019 February 6 and May14, respectively. As discussed later, it is unclear whetherthese represent the same gas clump (undergoing rapid de-celeration, and a temporary maser quench between 2019March and April), or they are independent clumps. In thelatter case, each of the clumps would be short lived ( > − ) and the systemic veloc-ity is ≃
280 km s − ). This is much larger than the ex-pansion velocity of the relic envelope mentioned above.Instead, we consider it as the lower limit of the projectedjet velocity along the line of sight. This makes it one ofthe fastest water fountains, similar to IRAS 18113 − D =3.6 kpc) and the ublications of the Astronomical Society of Japan , (2018), Vol. 00, No. 0 maximum proper motion of the H O masers in the jet( ∼
10 mas yr − ), the deprojected velocity of the jet wouldbe ∼
330 km s − at an inclination angle of ∼ ◦ with re-spect to the line of sight.Fig. 2 shows the time series of velocity distributionsof the identified maser peaks. The time baseline of theFLASHING observations is still too short to trace any peri-odic variation in brightness or systematic drifts of the line-of-sight velocities of the maser peaks. It is also too shortto relate the maser flare around V LSR =130–160 km s − with the new extremely high velocity components. Onecan see their large velocity drifts over one year, however itis unclear whether the components at V LSR ≃
345 km s − on 2019 February 6, ≃
345 km s − on 2019 May 14, and ∼
300 km s − in 2019 December–2020 January were reallythe same feature. Similarly, one also can note a velocitydrift of the maser component at V LSR ∼ −
176 km s − on2019 February 6, ∼ −
173 km s − on 2019 May 29, and ∼ −
166 km s − on 2019 December 6. Table 2 gives the parameters of the I18286 H O maser fea-tures detected in the KaVA image cubes. The individualrows show, respectively, the LSR velocity at the intensitypeak, the Right Ascension offset and its uncertainty, thedeclination offset and its uncertainty, the peak intensity,and the full velocity width of the feature. Fig. 3 shows thelocation of the H O maser components in I18286. Thesecomponents are distributed in groups, tracing a point-symmetric structure with respect to the center, similar toW43A (Tafoya et al. 2020). In the figure, we label thesesymmetric groups with the same number, and with theletter a or b, depending whether they are located south ornorth of the center, respectively. The new extreme veloc-ity components ( V LSR ≈ −
150 and 240 km s − ) correspondto Group 2a and 2b. Group 2a is located close the thebrightest and flared maser features (Group 4a).Fig. 4 shows the comparison among the maser featuredistributions in 2006–2007, 2009, and 2019, and clarifiesthis secular variation. It is difficult to superimpose themaser maps in the common coordinate system preciselyfor comparison, because of the the uncertainty in the sec-ular proper motion of the whole system. The astromet-ric information obtained in Paper II was composed of theGalactic rotation and the intrinsic motions of the individ-ual maser features selected as position reference in the as-trometry. Alternatively, map registration shown in Figure4 has been produced by roughly referencing the possiblelocation of the dynamical center of the outflow.Although stream lines of the gas hosting these masers might be bent by a magnetic force as suggested elsewhere(e.g. Vlemmings et al. 2006), we here suppose that the in-dividual maser features are moving in straight lines alongthe bipolar outflow or jet driven by the stellar system. Inthis case, the stellar system may be located at the mid-dle point between the new extreme velocity components,( X,Y ) ≈ ( − , V LSR ≈ −
50 and160 km s − , Paper I, II, III) or even slightly closer to thestellar system.We also note that the alignment axis of the highest ve-locity components seems to rotate clockwise in this obser-vation with respect to those in the previous epochs. Thisclockwise rotation of the maser alignment does not agreewith the expectation from the previous kinematic model oftwo highly collimated precessing jets (Paper I), in whichthe precession is counter-clockwise (with a precessing angleand period of ∼ ◦ and ∼
56 yr, respectively). The spatio-kinematics of the outflow now seems to be better explainedby a model of wider-angle bipolar lobes from a single driv-ing source as proposed for W 43A (Chong et al. 2015), inwhich H O masers are located along a bipolar cavity evac-uated by a series of precessing collimated outflows. Thisscenario was also considered in Fig. 13 of Paper I, but wasnot favored there due to its inconsistency with the appar-ent systematic variation of the LSR velocities of the maserfeatures along the outflow. However, Tafoya et al. (2019)and Tafoya et al. (2020) proposed a more plausible model ofa bipolar outflow, which is composed of a faster collimatedjet and slower lobes surrounding the jet. They suggest thatH O masers are excited in the lobes formed in the materialentrained by the jet. Although we see a very rapid expan-sion of the whole maser distribution by ∼
140 mas in 13yr, the expansion rate is consistent with the mean of themaser feature proper motions (8–10 mas yr − , Paper I, II,and III). Thus we are likely watching the evolution of theI18286 jet traced by the H O masers (Section 4).Fig. 5 shows the distribution of the maser features ina velocity–distance diagram, suggesting episodic gas ejec-tions traced by the H O masers. Taking into account thespacing between the groups of maser components (20–30mas) and the mean maser proper motions (8–10 mas yr − ),the time difference between the eruptions is roughly esti-mated to be 2–4 yr. The mean LSR velocity of the maserfeature group pairs (1a–b, 3a–b, 5a–b) is derived to be ∼
75 km s − . Although the deviation of the mean velocity Publications of the Astronomical Society of Japan , (2018), Vol. 00, No. 0 from the systemic velocity (∆ V ∼
10 km s − ) will providea clue of the kinematics of the central stellar system, itshould be examined in future thermal line mapping obser-vations such as those made with ALMA.Unfortunately, the highest velocity components seenin our single-dish monitoring campaign ( V LSR ≈
336 and345 km s − ) could not be detected in the KaVA ses-sion. On the other hand, thanks to the precise astrom-etry of KaVA, we determined the absolute coordinates ofthe maser feature containing the position-reference maserspot at V LSR =139.5 km s − : R.A.(J2000)= 18 h m s .93440 ± s .00002, decl.(J2000)= − ◦ ′ ′′ .6654 ± ′′ .0003.This is useful for future precise registration between themaser emission and other spectral line maps. The previous single dish observations in 2006 April–2010April (Paper I) did not detect the new extremely highvelocity components ( V LSR < −
60 km s − and V LSR >
200 km s − ) reported in this paper, with a 5- σ upper limitto 0.57 Jy, which suggests that these components wouldappear after 2010 April. We note that the velocity cover-age of our VERA and VLBA observations (22 epochs intotal during 2006 September–2009 September, Paper I, IIand III) did not cover the velocity ranges including thenew components, therefore we cannot ascertain whethersuch extreme components existed during those observa-tions. One may wonder whether the material now showingextreme-velocity maser components was ejected much be-fore their maser emission could be excited. However, thisis unlikely. The material ejected at the expected initialspeed ( >
300 km s − ) would interact within 1–2 yr of theejection with the ambient gas that excites the maser fea-tures located within a few ten AU ( ∼
10 mas) of the stellarsystem (Paper III). The travel time of the ejected materialto arrive at the present positions would be only 5 yr orshorter.The absence of the extreme components before 2010(Paper I) should indicate the absence of the gas cor-responding to these maser components at that time.Yung et al. (2013) found new velocity components in 2011around V LSR ∼ −
90 and 230 km s − . However, the ab-sence of information of their locations makes it difficult forus to suggest that they were also associated with the samegas clumps as those hosting our discovered maser compo-nents. In fact, there is a relatively large shift ( >
10 km s − )between the extreme velocity components in Paper I andthose in this paper.Previous observations could have missed the high-velocity components ( < −
60 and >
200 km s − ) if their lifetime was short, as seen in the case of the componentsat 336 to 345 km s − and −
178 to −
144 km s − in thispaper. However, the components around −
90 to −
70 and230 to 250 km s − are actually relatively stable as seen inour monitoring observations, so we consider these compo-nents as a product of a new ejection after 2010 April. Herewe propose that the stable high-velocity components seenat present would have formed soon after 2010 April withmuch higher velocities, like those seen in the unstable ve-locity components observed in 2019 February and May (orthe 336 and 345 km s − components). If this is the case,they could have already been quenched during the rapiddeceleration of the jet, while the most extreme componentsin this paper would be excited by a new, more recent jeteruption.Here we note the colocation of the extreme velocitycomponents in the southern side of the maser pair 2a and2b with the components that have much lower velocitiesand flared flux densities in Group 4a (Fig. 3). This isalso consistent with the scenario suggested by Tafoya et al.2020, in which a newly jet is interacting with the ambient,slower material that is then accelerated as entrained ma-terial along the jet. However, any physical relationshipbetween the extreme velocity and flared components, in-cluding the information of their relative locations along theline of sight, should be examined with further kinematicinformation.On the other hand, as shown in Fig. 3, we can seea point-symmetric pattern in the distributions of maserfeature groups with respect to the central stellar system.This implies recurrent maser activity with a period of2–4 yr (Section 3.2) or longer together with recurrentgas eruptions. Taking into account the present size ofthe maser distribution ( ∼
260 mas in the largest distancefrom the central stellar system) and its growth rate (8–10 mas yr − ), the launching of the jet traced by H O isestimated to have happened in the year 1990 ± < V LSR <
105 km s − ) that are located closer to thecentral system than the new components, such as a pairof maser features (1a and b in Fig. 3) and others possi-bly associated with the relic AGB envelope (Paper III). Itis possible to consider that a pair of bipolar gas ejectionsmay host not only the pair of the maser features mapped inthis paper such as 2a and 2b, but also the fastest velocitycomponents identified in single-dish spectra, but that aremissing in this map. Similarly to the estimate describedin Sect. 3.2 and as supposed above, one can consider thatgas clumps hosting these most extreme velocity, short-livedcomponents were ejected within 2–4 years of the presentday. ublications of the Astronomical Society of Japan , (2018), Vol. 00, No. 0 Here we also consider another scenario for the I18286jet, where a jet has been launched at a projected expansionvelocity of ∼
280 km s − as seen in the short-lived mostextreme velocity components, which then decelerated to ∼
110 km s − as seen in the stable components detectedover the last 10 years due to drag by the ambient materialin the CSE. This model is similar to that presented forI18113 by Orosz et al. (2019). Adopting the inclinationangle of ∼ ◦ these projected expansion velocities corre-spond to ∼
330 km s − and ∼
130 km s − , respectively. Thejet velocity v is now expressed as v ( t ) = v (1 + v kt ) − ,where v is the initial velocity of the jet, k the coeffi-cient of the exponential decay of the expansion velocityas a function of the mass density of the CSE, the massof the individual jet clumps, the drag coefficient of thejet clumps, and the cross-sectional area of the jet clumps.We adopt v =330 km s − , v ( t ) =130 km s − . Wheninserting t ∼
30 yr as the dynamical age of I18286 (totallength divided by a growth rate of the outflow estimatedfrom maser locations and proper motions), k is derived tobe ∼ . × − cm − , which is well consistent with thevalue derived for IRAS 18113 − k ∼ . × − cm − ,Orosz et al. 2019).However, one should remember the absence of thefastest maser features 10 years ago and take into ac-count the possibility that such components had alreadybeen decelerated to the velocity of the stable components( ∼
110 km s − ) within 10 years. In this case, we get a muchlarger k value, and a velocity drift over 10 km s − yr − should be observed and easily confirmed by weekly moni-toring observations of the maser spectrum with a velocityresolution better than 1 km s − . In our present work,we could not confirm such extremely fast velocity driftsdue to the sparseness of the monitoring observations. Onthe other hand, we can see a general trend in which thewhole maser distribution seems to exhibit accelerationsfrom Group 1a to 6a and from 1b to 6b (Fig. 4). Notethat the distant maser features have an angular offset inthe radial ejection direction from that of the closer andhigher velocity components. A possible explanation forthe distant features is that they are moving more slowly ina less dense region of the outer CSE and therefore, with-out the strong deceleration suggested for the high-velocitycomponents. This is possible if we consider the precessionof the jet. The first ejections have cleared up the CSEalong the jet major axis, so that later ejections along thatdirection would travel more smoothly and with less de-celeration than the highest-velocity ejections we see nowalong a different direction.Paper I suggested the existence of twin jets, although itis unclear whether they are driven by a single stellar com- ponent or individual components in a stellar binary. Thepresent results do not support that the twin jets are pre-cessing anti-clockwise. This is evident when comparing thewhole maser feature distributions over a decade as shownin Fig. 4. Even with this rough superimposition of themaser maps as described in Sect. 3.2, it is clear that someof the representative pattern of the maser features such asarcs have been moving radially from the dynamical centerat a rate of ∼
10 mas yr − , as expected from the speed ofthe jet projected on the sky ( > ∼
150 km s − , Paper I, II, II).The maser distribution seems to have been growing in thewestern and eastern sides in the north-west and south-eastlobes, respectively, suggesting clockwise jet precession andlobe development.Nevertheless, it is noteworthy that the dynamical agesof the twin jets modeled in Paper I had a difference by ∼
10 yr. Interestingly, the value of this time lag roughlycorresponds to 2–3 cycles of the possibly repeating out-bursts. Thus it is possible that some of the outbursts of asingle jet could be further enhanced with a longer periodas mentioned, so as to form arc-shaped structures of maserdistribution as seen in I18113 (Orosz et al. 2019).In I18286, even with a single jet, enhanced periodic out-burst, if they exist, may form maser distributions that ap-parently look like twin jets as illustrated in Fig. 14 of PaperI. The model simulations by Vel´azquez et al. (2012) alsoreproduce multiple bipolar lobes produced by a precessingbipolar jet from a binary system. Although those modelsfocused on planetary nebulae with nebular lobe lengths of ∼ AU and binary orbital periods of ∼
10 yr, one canconsider a closer binary system for the case of I18286. Wehope that morphological evolution of the H O masers willbe further elucidated in the next decade, enabling us tocompare it with that predicted by the models, in order toderive the physical parameters of the jet and the possiblebinary system.
The spectroscopic monitoring in FLASHING and theVLBI follow-up imaging with KaVA of the H O masersin I18286 have revealed newly excited, extremely high ve-locity components in the H O masers and a correspondingaxisymmetric ejections of gas in the water fountain out-flow. In this paper, we discussed the possibility that thestellar system driving the outflow has developed its out-flow lobes with a wide opening angle through periodic gasejections with a clockwise precession of the outflow axis,rather than through highly collimated twin jets with ananti-clockwise axis precession. The new model explains theobserved maser feature distributions biased to the down-
Publications of the Astronomical Society of Japan , (2018), Vol. 00, No. 0
Table 1.
Summary of the FLASHINGobservations toward IRAS 18286 − Epoch T sys ∗ t int † Noise ‡ (yy/mm/dd) (K) (m) (mJy)18/12/23 ∼
150 18 11819/01/05 ∼
110 106 3219/02/06 ∼
150 47 7719/03/06 ∼
130 35 7819/04/15 ∼
190 58 11419/04/28 ∼
130 39 5619/05/14 ∼
170 37 10819/05/29 ∼
230 152 8119/12/06 ∼
120 28 7719/12/20 ∼
100 49 5320/01/10 ∼
110 23 74 ∗ System noise temperature. † On-source integration time. ‡ σ noise level of the spectrum. stream sides of the bipolar lobes, namely the western andeastern sides of the northern and southern lobes, respec-tively.Further VLBI monitoring observations of the new com-ponents of H O masers will enable us to find their deceler-ations, suggested by Tafoya et al. (2019) and Tafoya et al.(2020), as a result of interactions between the ambient ma-terial and the entrained material around the high-speed jet.Thus the combination of single-dish maser monitoring pro-grams with targeted follow-up VLBI mapping observationsprovide an effective approach to investigating the ejectioncharacteristics of water fountain jets/outflows. If the indi-vidual features in a variety of PN shapes are related to theindividual events of such ejection behaviors seen in maserobservations, this kind of approach will provide essentialinsights into the final stages of stellar evolution.New ejections of the jet in I18286 should again be de-tected within several years if they are recurrent with timeseparations of 2–4 yr, as predicted by the spacing betweenmaser groups. Since the size of the central binary sys-tem is expected be only a few AU (e.g., Vel´azquez et al.2012), which may correspond to a few mas or smaller, itis challenging to spatially resolve the system into a ther-mal collimated jet from the driving star and an equato-rial disk/torus driving the jet, even with state-of-the-artinstruments such as ALMA. Therefore, the H O masersmapped with VLBI provide unique probes to elucidate thespatio-kinematical information and a possible periodic be-havior of the central stellar system and the jet in I18286.
Acknowledgments
The Nobeyama 45-m radio telescope and VERA are oper-ated by Nobeyama Radio Observatory and Mizusawa VLBIObservatory, respectively, branches of National AstronomicalObservatory of Japan (NAOJ), National Institutes of NaturalSciences. KVN in KaVA and the Daejeon Correlator in KJCCare operated by Korea Astronomy and Space Science Institute(KASI). Our data analysis was in part carried out on the com-mon use data analysis computer system at the Astronomy DataCenter, ADC, of NAOJ. HI and GO are supported by the MEXTKAKENHI program (16H02167). JFG is partially supportedby MINECO (Spain) grant AYA2017-84390-C2-R (co-funded byFEDER) and by the State Agency for Research of the SpanishMCIU through the “Center of Excellence Severo Ochoa” awardfor the Instituto de Astrof´ısica de Andaluc´ıa (SEV-2017-0709).HI and JFG were supported by the Invitation Program forForeign Researchers of the Japan Society for Promotion ofScience (JSPS grant S14128) and i-LINK+2019 Programmein IAA/CSIC. GO was supported by the Australian ResearchCouncil Discovery project DP180101061 of the Australian gov-ernment, and the grants of CAS LCWR 2018-XBQNXZ-B-021and National Key R&D Program of China 2018YFA0404602.DT was supported by the ERC consolidator grant 614264. RBacknowledges support through the EACOA Fellowship fromthe East Asian Core Observatories Association. LU acknowl-edges support from Convocatoria de Apoyo a la Investigaci´onCient´ıf——´ıca 2020 of the University of Guanajuato.
References
Amiri, N., Vlemmings, W. H. T., & van Langevelde, H. J. 2010,A&A, 509, A26Amiri, N., Vlemmings, W. H. T., & van Langevelde, H. J. 2011,A&A, 532, A149Beasley, J. A., & Conway, J. E. 1995, in ASP Conf. Ser.82, Very Long Baseline Interferometry and the VLBA, ed.J. A. Zensus, P. J. Diamond, & P. J. Napier (San Francisco,CA: ASP), 328Boboltz, D. A., & Marvel, K. B. 2005, ApJL, 627, L45Chong, S. N., Imai, H., & Diamond, P. J. 2015, ApJ, 805, 53Claussen, M., Sahai, R., & Morris, M. R. 2009, ApJ, 691, 219Day, F. M., Pihlstr¨om, Y. M., Claussen, M. J., & Sahai, R. 2010,ApJ, 713, 986Deguchi, S., Nakashima, J, Kwok, S., & Koning, N. 2007, ApJ,664, 1130de Gregorio-Monsalvo, I., G´omez, Y., Angalada, G., et al.2004, ApJ, 601, 921 & W. H. T. Vlemmings (Cambridge:Cambridge Univ. Press), 217Elitzur, M. 1992, Astronomical Masers (Dordrecht: Kluwer)Engels, D., & Bunzel, F. 2015, A&A, 582, A68G´omez, J. F. et al. 2017, MNRAS, 468, 2081G´omez, J. F. et al. 2015, ApJ, 799, 186G´omez, J. F. et al. 2011, ApJL, 739, L14Herman, J. & Habing, H. J. 1985, A&A, 59, 523Huggins, P. J., 2007, ApJ, 663, 342Imai, H. 2007, in IAU Symp. 242, Astrophysical Masersand their Environments, ed. W. Baan & J. Chapman(Cambridge: Cambridge Univ. Press), 279 ublications of the Astronomical Society of Japan , (2018), Vol. 00, No. 0 Imai, H., Deguchi, S., & Sasao, T. 2002, ApJ, 567, 971Imai, H., Diamond, P. J, Nakashima, J., Kwok, S., &Deguchi, S. 2008, in Proc. 9th European VLBI NetworkSymposium on the Role of VLBI in the Golden Age forRadio Astronomy and EVN Users Meeting, PoS (Trieste:SISSA), 60Imai, H., He, J.-H., Nakashima, J., Ukita, N., Deguchi, S., &Koning, N. 2009, PASJ, 61, 1365Imai, H., Kurayama, T., Honma, M., & Miyaji, T. 2013a, PASJ,65, 28 (Paper II)Imai, H., Nakashima, J., Deguchi, S., Kwok, S., &Diamond, P. J. 2013b, ApJ, 773, 182Imai, H., Nakashima, J., Yung, B. H.K., Deguchi, S., Kwok, S.,& Diamond, P. J. 2013c, ApJ, 771, 47 (Paper III)Imai, H., Nakashima, J., Diamond, P. J., Miyazaki, A., &Deguchi, S. 2005, ApJL, 622, L125Imai, H. 2007, in: IAU Symposium 242, Astrophysical Masersand their Environments, Baan, W., & Chapman, J.(Cambridge University Press: Cambridge), p279Imai, H., Sahai, R., & Morris, M. 2007, ApJ, 669, 424Imai, H., Obara, K., Diamond, P. J., Omodaka, T., & Sasao, T.2002, Nature, 417, 829Likkel, L., Morris, M., & Maddalena, R. J. 1992, A&A, 256, 581Okada, N., et al. 2019, PASJ, advanced issue (psz126)Orosz, G., et al. 2019, MNRAS, 482, L40Lewis, B. M. 2001, Nature, 560, 400te Lintel Hekkert, P., Versteege-Hensel, H. A., Habing, H. J., &Wiertz, M. 1989, A&AS, 78, 399Miranda, L. F., G´omez, Y., Anglada, G., & Torrelles, J. M.2001, Nature, 414, 284Sahai, R. 2004, in ASP Conf. Ser. 313, Asymmetrical PlanetaryNebulae III: Winds, Structure and the Thunderbird, ed.M. Meixner et al. (San Francisco, CA: ASP), 141Sahai, R., Morris, M., S´anchez Contreras, C., & Claussen, M.2007, AJ, 134, 2200Sahai, R., & Trauger, J. 1998, AJ, 116, 1357Sahai, R., et al. 2017, ApJ, 835, L13Sevenster, M. N., van Langevelde, H. J., Moody, R. A., et al.2001, A&A, 366, 481Su´arez, O., G´omez, J. F., Bendjoya, P.˜et al. 2012, in IAU Symp.287, Cosmic Masers — from OH to H , ed. R. S. Booth,E. M. L. Humphreys, & W. H. T. Vlemmings (Cambridge:Cambridge Univ. Press), 230Su´arez, O., G´omez, J. F., Miranda, L. F., et al. 2009, A&A,505, 217Tafoya, D., Orosz, G., Vlemmings, W. H. T., Sahai, R., P´erez-S´anchez, A. F. 2019, A&A, 629, A8Tafoya, D., et al. 2020, ApJL, in pressTafoya, D., et al. 2011, PASJ, 63, 71Vel´azquez, P., et al. 2012, MNRAS, 419, 3529Vinkovi´c, D., Bl¨ocker, T., Hofmann, K.-H., & Elitzur, M. 2004,MNRAS, 352, 852Vlemmings, W. H. T., Diamond, P. J., & Imai, H. 2006, Nature,440, 58Vlemmings, W. H. T., & van Langevelde, H. J. 2007, A&A, 472,547Yung, B. H. K., et al. 2013, ApJ, 769, 20 Yung, B. H. K., Nakashima, J., Imai, H., et al. 2011, ApJ, 741,94 (Paper I) Publications of the Astronomical Society of Japan , (2018), Vol. 00, No. 0
Peak: 70.7 JyPeak: 77.7 JyPeak: 67.4 JyPeak: 83.5 JyPeak: 72.1 Jy
Peak: 97.0 Jy
Peak: 80.4 Jy
Peak: 69.6 Jy
310 330 350 0.6 0.4 0.2 0.0-0.2 0.6 0.4 0.2 0.0-0.2320
Peak: 17.3 Jy(cross-power) C r o ss - po w e r f l u x d e n s i t y ( Jy ) (Yung et al. 2011) Fig. 1.
Time series of the H O maser spectra taken towardIRAS 18286 − V LSR ≈
336 and 345 km s − . For comparison,the spectra taken on 2010 April 2 cited from Paper I is displayed in the rightbottom panel with the same vertical scale as those in other 45 m spectra.The scalar-averaging cross-power spectrum taken in the KaVA observationis displayed in the left bottom panel. The two vertical dotted lines indicatethe LSR velocities of −
60 and 200 km s − , over which the new high-velocitymaser components appeared in the present observations. Log ( F ν [Jy]) L S R v e l o c i t y [ k m s - ] Fig. 2.
Velocity distributions of the spectral peaks ( > σ ) of theIRAS 18286 − O masers at the eleven epochs. A logarithm of theflux density of each spectral peak is indicated by the right-side color-scalebar. ublications of the Astronomical Society of Japan , (2018), Vol. 00, No. 0 - - - L S R v e l o c i t y ( k m s - ) R.A. offset (mas) D e c l . o ff s e t ( m a s ) - -
0 -100 +
1a 1b2b2a3a 3b4b4a 5a5b 6b6a
200 AUat 3.6 kpc
Fig. 3.
Distribution of the H O maser features in IRAS 18286 −
0 -100 D ec l. o ff se t ( m as ) R.A. offset (mas) +
200 AUat 3.6 kpc (cid:39918)
Fig. 4.
Comparison of the maser feature distributions in IRAS 18286 − Publications of the Astronomical Society of Japan , (2018), Vol. 00, No. 0
Fig. 5.
Distribution of the maser features in IRAS 18286 − ( X, Y ) = ( −
43, 66)[mas], is displayed foreach maser feature. A horizontal dashed line shows the assumed systemicvelocity of the central stellar system, V LSR =
65 km s − . The labels of maserfeature groups (from 1a-b to 6a-b) are the same as those in Fig. 3. ublications of the Astronomical Society of Japan , (2018), Vol. 00, No. 0 Table 2. H O maser features detected with KaVA V LSR
X σ X Y σ Y I peak ∆ V [km s − ] [mas] [mas] [mas] [mas] [Jy beam − ] [km s − ] − . − .
64 0.25 117 .
44 0.50 0.12 0.84 − . − .
16 0.09 121 .
15 0.10 0.72 3.16 − . − .
59 0.13 113 .
39 0.11 0.17 0.63 − . − .
99 0.06 109 .
08 0.09 0.55 1.26 − . − .
06 0.05 109 .
63 0.13 0.35 0.84 − . − .
09 0.06 179 .
46 0.09 1.18 2.32 − . − .
67 0.02 182 .
60 0.02 3.49 2.32 − . − .
18 0.18 256 .
19 0.11 0.80 2.95 − . − .
94 0.19 256 .
16 0.12 1.33 6.11 − . − .
23 0.20 251 .
61 0.10 0.26 1.47 − . − .
69 0.07 256 .
88 0.12 0.61 2.11 − . − .
45 0.07 251 .
87 0.09 0.56 0.63 − . − .
15 0.05 179 .
59 0.09 1.34 1.05 − . − .
74 0.12 252 .
00 0.10 1.11 1.69 − . − .
54 0.11 243 .
10 0.11 0.56 2.53 − . − .
82 0.10 179 .
80 0.05 0.36 0.633 . − .
07 0.08 156 .
41 0.06 0.45 1.906 . − .
95 0.08 156 .
23 0.14 0.24 1.2617 . − .
67 1.21 178 .
60 0.35 0.37 2.7419 . − .
75 0.09 205 .
44 0.11 0.23 0.6327 . − .
41 0.08 150 .
62 0.18 0.21 0.6328 . − .
87 0.06 88 .
11 0.36 0.99 4.0034 . − .
64 0.47 90 .
45 0.66 0.80 1.2637 . − .
59 0.12 91 .
21 0.23 0.43 1.4737 . − .
05 0.05 231 .
91 0.08 1.35 0.8439 . − .
44 0.13 90 .
83 0.09 0.32 0.6342 . − .
94 0.05 92 .
29 0.10 0.79 1.6950 . − .
43 0.02 10 .
23 0.04 2.83 2.1160 . − .
33 0.05 68 .
19 0.11 0.54 0.8462 . − .
43 0.09 58 .
00 0.17 1.14 1.2662 . − .
24 0.31 − .
19 0.18 0.59 0.8464 . − .
45 0.11 55 .
00 0.21 0.48 2.1170 . − .
38 0.30 − .
79 0.10 0.37 0.6376 .
65 11 .
50 0.11 − .
09 0.15 0.70 2.5378 . − .
78 0.10 52 .
43 0.22 0.56 1.2787 . − .
51 0.20 − .
24 0.06 0.35 1.0590 .
09 13 .
58 1.15 − .
22 0.16 0.41 0.8491 .
52 2 .
93 0.13 7 .
71 0.05 1.03 1.6891 . − .
28 0.42 7 .
94 0.26 0.79 1.6895 .
11 16 .
40 0.11 − .
72 0.11 0.49 1.0596 . − .
82 0.13 37 .
17 0.06 0.55 0.8497 .
78 16 .
29 0.05 − .
13 0.07 0.77 2.3299 . − .
51 0.31 − .
97 0.09 0.59 1.0599 . − .
84 0.37 − .
79 0.15 0.60 1.2699 . − .
18 0.16 − .
37 0.06 1.34 2.9599 .
47 18 .
93 0.61 − .
49 0.21 0.55 2.53103 . − .
26 0.51 − .
82 0.31 0.42 2.74103 . − .
62 0.31 − .
11 0.49 1.40 2.53 Publications of the Astronomical Society of Japan , (2018), Vol. 00, No. 0
Table 2. (Continued) . − .
66 0.07 35 .
81 0.07 0.73 2.32106 . − .
42 0.05 − .
59 0.04 0.86 1.27107 .
31 2 .
58 0.15 7 .
48 0.14 0.35 2.32109 . − .
68 0.10 − .
39 0.15 0.24 0.63112 .
58 2 .
45 0.08 7 .
32 0.12 0.52 1.47116 .
56 15 .
86 0.16 − .
99 0.11 0.51 4.64117 . − .
86 0.08 − .
40 0.06 0.59 1.05122 .
20 5 .
33 0.08 − .
43 0.11 0.72 2.11123 .
68 14 .
24 0.10 − .
73 0.15 0.27 1.26125 .
38 14 .
18 0.12 − .
51 0.11 0.36 0.84125 . − .
66 0.08 4 .
29 0.17 0.24 1.47127 .
19 14 .
84 0.12 − .
61 0.09 0.33 1.05136 .
59 0 .
06 0.02 0 .
04 0.09 11.96 6.32139 .
05 0 .
01 0.02 0 .
00 0.02 28.00 3.58147 .
08 1 .
74 0.10 − .
64 0.25 5.59 6.74151 .
10 9 .
03 0.19 − .
21 0.30 23.69 24.02171 .
52 7 .
26 0.06 − .
01 0.09 0.47 0.84177 .
03 6 .
37 0.11 − .
14 0.05 0.88 1.68236 . − .
37 0.08 15 .
23 0.04 1.28 7.16241 . − .
43 0.11 15 .
11 0.06 1.38 6.32247 . − .
84 0.19 15 ..