Water Fountain Sources

• Water fountain sources are evolved, oxygen-rich stars characterized by fast bipolar H2O maser jets juxtaposed with slower OH envelopes.
• Their identification relies on the H2O maser velocity range exceeding that of the OH masers, allowing clear kinematic separation of newly launched jets from remnant AGB winds.
• High-resolution imaging reveals diverse jet morphologies and shock interactions, providing insights into the transformation of circumstellar envelopes in late stellar evolution.

Searching arXiv for water-fountain source papers to ground the article in current literature. Searching arXiv for "water fountain sources evolved stars H2O masers". Water fountain sources are evolved, oxygen-rich stars—usually AGB or post-AGB objects, though some discussions also admit evolved massive red supergiants—that exhibit high-velocity molecular jets traced by the 22.235 GHz H$_2$O maser line. They are widely interpreted as systems in which fast, collimated bipolar outflows have recently been launched into a pre-existing, slower circumstellar envelope, so that masers provide a direct kinematic view of the transformation from largely spherical AGB mass loss to bipolar or multipolar structures associated with the planetary-nebula pathway (Ouyang et al., 2024). The class is rare and short-lived, with maser-traced dynamical timescales commonly of order years to decades or, more broadly, decades to a few centuries, and it is therefore a principal observational laboratory for studying how jet launching, envelope interaction, and equatorial density enhancement become imprinted on circumstellar morphology (Day et al., 2010).

1. Definition and astrophysical setting

The defining phenomenology of a water fountain is the coexistence of a slow OH-bearing circumstellar envelope and a faster H$_2$O-masing outflow. In the standard picture, the 1612 MHz OH maser line traces the remnant, expanding circumstellar envelope produced during the AGB phase, while the 22 GHz H$_2$O masers trace shocked, dense gas in newly launched bipolar jets (Fan et al., 2023). This contrast makes water fountains particularly valuable for separating the kinematics of the older spherical or quasi-spherical wind from those of the newer asymmetric flow.

The evolutionary context is brief and heterogeneous. Many studies place water fountains in the transition from the asymptotic giant branch to the post-AGB or pre-planetary-nebula phase, but the class is not restricted to a single instantaneous stage. Some members retain AGB diagnostics, some are post-AGB objects, and some may already be early planetary nebulae; accordingly, the presence of a water-fountain jet does not, by itself, fix a unique evolutionary status (Yung et al., 2016). This suggests that the water-fountain phenomenon is better understood as a maser-defined kinematic state than as a one-to-one evolutionary label.

Several properties recur across the class. H$_2$O maser spectra typically span much larger velocity intervals than OH 1612 MHz spectra, often from tens to hundreds of km s$^{-1}$ and in extreme cases up to about $500$–$540$ km s$^{-1}$, while the OH envelope expands at more modest AGB-like speeds of roughly $10$–$30$ km s$_2$0 (Uscanga et al., 2023). Spatially, very long baseline imaging commonly resolves compact bipolar structures on scales of hundreds to a few thousand AU, with bow shocks, arcs, precessing patterns, or paired red- and blue-shifted maser clusters that straddle the central star (Yung et al., 2011).

2. Maser diagnostics and selection criteria

The classical observational signature of a water fountain is that the radial-velocity range of the H$_2$1O masers exceeds that of the OH 1612 MHz masers. In kinematic terms, the maser velocity span is defined as $_2$2, and for a double-peaked OH shell one writes $_2$3 and $_2$4, so that $_2$5 (Fan et al., 2023). The working water-fountain selection criterion is then $_2$6, expressing the idea that the H$_2$7O-emitting flow outruns the older AGB shell.

This criterion is broader than the older practice of requiring very large H$_2$8O widths, often $_2$9 km s$_2$0. A systematic search using databases of circumstellar 1612 MHz OH and 22.235 GHz H$_2$1O maser sources explicitly adopted $_2$2 in order to identify water fountains whose H$_2$3O velocity ranges may be smaller than those of the previously known extreme cases (Fan et al., 2023). A related later formulation is the “velocity excess” diagnostic, in which OH 1612 MHz is treated as the terminal-velocity tracer of a spherical envelope and any maser feature outside that interval is assigned an excess $_2$4; this approach was developed specifically to identify dynamically young or unfavorably inclined systems that could be missed by the older width threshold (Xie et al., 18 Dec 2025).

Selection effects are central. Surveys based on very large H$_2$5O velocity spreads are biased toward jets aligned close to the line of sight, whereas jets near the plane of the sky can show more modest line-of-sight spreads even when their true three-dimensional speeds are large (Vlemmings et al., 2014). This is one reason “low-velocity” water-fountain candidates are astrophysically important: they may represent either intrinsically slower jets at the onset of bipolarity or strongly projected fast jets (Yung et al., 2013).

The same velocity criterion can, however, recover contaminants or neighboring phenomena. The database search emphasized that the maser-velocity criterion can discover other astrophysically interesting objects than just water fountains, including peculiar planetary nebulae with maser emissions and stellar merger remnants (Fan et al., 2023). More generally, star-forming regions, maser-bearing planetary nebulae, and beam-confused single-dish spectra must be excluded through OH profile morphology, infrared colors, spatial association, and interferometric follow-up (Fan et al., 2023).

3. Spatial morphology and kinematic archetypes

Interferometric studies show that water fountains are not defined by spectroscopy alone; they are three-dimensional jet systems whose masers often trace bow shocks, jet heads, or cavity walls. In IRAS 19190+1102, very long baseline observations resolved two main arc-shaped H$_2$6O maser structures perpendicular to their NE–SW separation axis, expanding at $_2$7 mas yr$_2$8. The masers span $_2$9 to $_2$0 km s$_2$1, giving a three-dimensional outflow speed of about $_2$2 km s$_2$3, an inclination of about $_2$4 to the plane of the sky, and a dynamical age of about $_2$5 yr (Day et al., 2010). The same source also showed a variable off-axis blueshifted group whose velocity changed by more than $_2$6 km s$_2$7 between epochs, interpreted as possible evidence for the earliest stages of a new lobe (Day et al., 2010).

IRAS 18460$_2$80151 provides a more extreme configuration. Its H$_2$9O maser spectrum reached a line-of-sight velocity spread of about $^{-1}$0 km s$^{-1}$1 and resolved into a highly collimated bipolar jet roughly $^{-1}$2 mas in extent, corresponding to about $^{-1}$3 AU at the adopted distance, with a three-dimensional flow velocity of about $^{-1}$4 km s$^{-1}$5 and a dynamical age of only about $^{-1}$6 yr (Imai et al., 2013). At the same time, systemic-velocity H$^{-1}$7O masers formed a central ring-like distribution with radius about $^{-1}$8 AU and expansion speed about $^{-1}$9 km s$500$0, while the OH 1612 MHz masers displayed the classic double-peaked envelope profile with separation about $500$1 km s$500$2 (Imai et al., 2013). This combination makes the source a particularly clear demonstration that a very young jet can coexist with a slow, nearly spherical outflow in the same system.

Some water fountains appear even more extreme in speed and opening angle. In IRAS 18043$500$32116, new H$500$4O features were found at about $500$5 and $500$6 km s$500$7, yielding a total spread of about $500$8 km s$500$9, described as the largest yet measured in any water fountain (Uscanga et al., 2023). The highest-velocity components are offset from the main cluster axis, and together with compact ALMA CO emission they imply a wide outflow opening angle of about $540$0 and a dynamical timescale of about $540$1 yr (Uscanga et al., 2023). The same source also showed a rising 22 GHz continuum, consistent either with an ionized jet or with the onset of photoionization (Uscanga et al., 2023).

IRAS 18113$540$22503 and IRAS 18286$540$30959 illustrate two further dynamical modes. Multi-epoch VLBI of IRAS 18113$540$42503 resolved three nested bipolar bow-shock pairs with deprojected expansion speeds of about $540$5, $540$6, and $540$7 km s$540$8, interpreted as an episodic jet whose ejecta decelerate through quadratic drag in a dense circumstellar envelope; the inferred initial ejection velocity is about $540$9 km s$^{-1}$0 and the ejection period about $^{-1}$1 yr (Orosz et al., 2018). IRAS 18286$^{-1}$20959, by contrast, shows a double-helix pattern consistent with two bipolar precessing jets across about $^{-1}$3 mas, with fitted three-dimensional speeds of $^{-1}$4 and $^{-1}$5 km s$^{-1}$6 and a precession period of about $^{-1}$7 yr for the principal jet (Yung et al., 2011). Together these systems establish that water fountains include highly collimated jets, wide-angle fast winds, episodic bow shocks, and precessing helical outflows rather than a single universal geometry.

4. Circumstellar envelopes, chemistry, and launching environments

Although H$^{-1}$8O masers define the class observationally, water fountains are embedded in more extended dusty and molecular structures. A sensitive CO and $^{-1}$9CO survey of ten water fountain stars found wide velocity components of about $10$0–$10$1 km s$10$2 centered at the stellar velocities in IRAS 18460$10$30151 and IRAS 18596+0315, interpreted as the former AGB envelopes of the progenitor stars (Rizzo et al., 2013). For these envelopes, large-velocity-gradient modeling yielded molecular masses close to $10$4 solar masses, mean densities of order $10$5 cm$10$6, mass-loss rates of order $10$7, and kinetic temperatures between $10$8 and $10$9 K, implying that the thermal CO emission arises from cool outer regions into which the maser jets are being launched (Rizzo et al., 2013).

Detailed work on IRAS 16342$30$03814 shows that the circumstellar chemistry can carry direct nucleosynthetic information. ASTE observations detected CO only in this source among four surveyed water fountains and confirmed an exceptionally low $30$1CO/$30$2CO intensity ratio of about $30$3 at the systemic velocity, which the authors interpreted as evidence for a very low $30$4C/$30$5C ratio and hence hot-bottom burning in an oxygen-rich evolved star (Imai et al., 2012). In the same source, two-dimensional radiative-transfer modeling of the dust shell required not only an optically thick torus but also an additional optically and geometrically thick inner disk; the estimated masses were about $30$6 for the disk and about $30$7 for the torus (Murakawa et al., 2012). This geometry was proposed to arise through shielding, equatorial concentration, and possible binary interaction, linking the dust distribution directly to the conditions under which jets are launched and collimated (Murakawa et al., 2012).

Water fountains also probe unusually hot shocked gas. The first detections of the 321.226 GHz ortho-H$30$8O maser transition toward water-fountain nebulae showed that in IRAS 18043$30$92116 and IRAS 18286$_2$000959 the 321 GHz velocity range coincides with that of the 22 GHz masers, implying that the two transitions likely coexist (Tafoya et al., 2013). Because the 321 GHz and 22 GHz intensities are comparable in these sources, the inferred kinetic temperature of the masing region is $_2$01 K, and a two-shock scenario was proposed in which a fast J-type shock is followed by a slower C-type shock through already accelerated gas (Tafoya et al., 2013).

The OH environment can be equally informative. In IRAS 18460$_2$020151, excited-state OH masers at 4660 and 6031 MHz were discovered together with ground-state OH masers at 1612 and 1665 MHz; this was reported as the first detection of 4660 MHz excited-state OH maser emission in an evolved-star circumstellar environment and the first excited-state OH masers reported toward a water-fountain source (Ouyang et al., 2024). The slight extension of the 1665 MHz velocity span beyond that of the 1612 MHz line was suggested to mean that inner OH regions may have been accelerated through interaction with the water-fountain jets within the OH shell (Ouyang et al., 2024).

On the theoretical side, a radial self-similar non-relativistic MHD model was applied to the archetypal water fountain W43A. Using observational constraints on magnetic field, density, and velocity at the H$_2$03O maser regions, the study concluded that a toroidal field $_2$04 mG at the H$_2$05O maser location is required for the jet momentum rate to match or exceed the observed value, and that radiative driving alone is insufficient (Ceccobello et al., 2020). This places water fountains within the broader astrophysical class of magnetized disk-launched outflows, while preserving the specific circumstellar-envelope physics that makes their masers observable (Ceccobello et al., 2020).

5. Evolutionary interpretation and major controversies

Water fountains are often described as marking the onset of the morphological metamorphosis of circumstellar envelopes, but this interpretation is not universally valid. A dedicated study of infrared SEDs argued that water fountains can have different envelope morphologies: some possess spherical envelopes resembling usual AGB stars, whereas others have aspherical envelopes more common to post-AGB stars (Yung et al., 2016). On that basis, the authors concluded that water fountains may not represent the earliest stage of the morphological metamorphosis in every case (Yung et al., 2016).

A related issue concerns dynamical age. Because many published ages are derived from maser proper motions and compact angular extents, water-fountain jets are often said to be only a few to a few hundred years old. However, the same SED-based study argued that the maser dynamical age may not be the real age of the jet, because H$_2$06O masers may trace only the innermost portion of a larger, already developed outflow (Yung et al., 2016). This suggests that a small kinematic age measured from masers should be treated as the age of the currently masing segment, not necessarily the age of the full jet.

The diversity of source classes reinforces this caution. W43A shows AGB indicators such as OH maser flux variations with a $_2$07-day period and SiO maser emission, yet it hosts a water-fountain jet; conversely, IRAS 15103$_2$085754 has been identified as a planetary-nebula candidate with water-fountain-like H$_2$09O masers (Yung et al., 2016). Thus, water-fountain jets can occur during late AGB evolution, across post-AGB evolution, and into early planetary-nebula conditions (Yung et al., 2016). A plausible implication is that the water-fountain phenomenon records the presence of high-velocity, collimated molecular outflow more reliably than it records a unique stellar-evolutionary timestamp.

6. Searches, candidate classes, and future directions

Systematic searches have increasingly shifted from a focus on only the most extreme maser spectra to a broader search for velocity excess relative to OH 1612 MHz. A database study examined 8,474 OH observations from 2,195 sources and 6,085 H$_2$10O observations from 3,642 sources, applied the criterion that the H$_2$11O velocity range exceed the OH velocity range, and identified 11 sources that meet this criterion (Fan et al., 2023). Inspection of IRAS colors showed that two of these, IRAS 19069+0916 and IRAS 19319+2214, lie in the color region for post-AGB stars (Fan et al., 2023). The explicit conclusion was that the maser-velocity criterion broadens the discovery space beyond the previously known extreme water fountains (Fan et al., 2023).

A complementary Effelsberg 22 GHz survey of 204 AKARI- and IRAS-selected evolved-star candidates was aimed specifically at “low-velocity” water fountains. It obtained 63 detections, including 36 new ones, and proposed four candidates—IRAS 15193+3132, IRAS 18056$_2$121514, OH 16.3$_2$133.0, and IRAS 18455+0448—with H$_2$14O velocity coverages smaller than in classical water fountains but still larger than their OH 1612 MHz ranges (Yung et al., 2013). The study suggested that such objects may be oxygen-rich late AGB or early post-AGB stars in a stage of evolution immediately after spherically symmetric AGB mass loss has ceased (Yung et al., 2013).

IRAS 18455+0448 later became an instructive case of this strategy’s success. JVLA mapping confirmed an extended bipolar H$_2$15O maser outflow with a red/blue centroid separation of $_2$16 mas, $_2$17 km s$_2$18, and a kinematic age of about $_2$19 yr for the far distance and an assumed $_2$20 inclination (Vlemmings et al., 2014). Historically, the source had been classified as a “dying” OH/IR star because of the exponential decrease of its 1612 MHz OH maser and the earlier absence of H$_2$21O masers, so its later water-fountain behavior was described as a rebirth (Vlemmings et al., 2014).

The most recent extension of this search logic is the “velocity excess” approach applied to 17 sources with maser components outside the OH 1612 MHz terminal-velocity interval. One-dimensional DUSTY modeling showed that seven of these are well reproduced by spherical outer-envelope models, including five with single-peaked AGB-like SEDs and two with double-peaked post-AGB-like SEDs (Xie et al., 18 Dec 2025). Two objects, IRAS 19229+1708 and IRAS 19052+0922, were highlighted as possible incipient water-fountain candidates, because they combine measurable maser velocity excess with outer circumstellar envelopes that remain nearly spherical (Xie et al., 18 Dec 2025). This supports a morphological sequence in which bipolar jets and tori arise first in the central regions while the outer circumstellar envelope remains spherical (Xie et al., 18 Dec 2025).

Confirmation still depends on high-resolution follow-up. Across the literature, the recommended program is consistent: VLBI imaging of 22 GHz H$_2$22O masers to trace spot distributions and proper motions; OH interferometry to map the slower shell and constrain $_2$23; CO rotational-line imaging to measure the molecular envelope and its mass-loss history; and radio or infrared continuum imaging to locate the central source and exclude star-forming contaminants (Fan et al., 2023). In practice, water-fountain research now proceeds on two linked fronts: expanding the census through velocity-based searches, and converting spectroscopic candidates into physically secure jet systems through astrometry, imaging, and multi-transition maser diagnostics.