When low- to middleweight stars like our sun are near the end of their lives, they are red giant stars casting off their outer layers of gas and dust into space in a series of thermal pulses. Subsequently the central star will start to heat up and ionize the ejected material, creating beautiful planetary nebulae (PNe). The central star will continue to evolve until finally all nuclear reactions in its interior cease and the star starts to cool and fade. It becomes a white dwarf while the PN expands and dissolves into the interstellar medium.

This describes the usual evolutionary path, but in some cases the central star will experience one final thermal pulse after it has gone through the PN phase and has already started to cool. This is called a very late thermal pulse (VLTP), which dramatically changes the evolution of the star. The star will very quickly swell up to become a red giant again (a so-called born-again red giant) and will expel the remaining atmosphere in the process. This will form a new PN inside the old PN. Theory predicts that 10 – 20% of central stars should undergo a VLTP, but only a handful of objects have been observed. Of those, only two have been discovered shortly after the VLTP: V605 Aql in 1919 and V4334 Sgr (a.k.a. Sakurai’s Object, after the Japanese amateur astronomer who made the discovery) in 1996. V605 Aql was quickly lost after its discovery as it was not possible to continue observing it with contemporary instrumentation, leaving Sakurai’s Object as the only well-studied example of a VLTP.

Fig. 1: From left to right: Deconvolved Ks images taken in 2010 and 2013 from Hinkle & Joyce (2014), CN and CO emission detected by ALMA in 2015.

This describes the usual evolutionary path, but in some cases the central star will experience one final thermal pulse after it has gone through the PN phase and has already started to cool. This is called a very late thermal pulse (VLTP), which dramatically changes the evolution of the star. The star will very quickly swell up to become a red giant again (a so-called born-again red giant) and will expel the remaining atmosphere in the process. This will form a new PN inside the old PN. Theory predicts that 10 – 20% of central stars should undergo a VLTP, but only a handful of objects have been observed. Of those, only two have been discovered shortly after the VLTP: V605 Aql in 1919 and V4334 Sgr (a.k.a. Sakurai’s Object, after the Japanese amateur astronomer who made the discovery) in 1996. V605 Aql was quickly lost after its discovery as it was not possible to continue observing it with contemporary instrumentation, leaving Sakurai’s Object as the only well-studied example of a VLTP.

Fig. 2: In ALMA spectra we detect the presence of CO, CN, HC₃N, and ¹³C isotopologues.

Other groups also studied Sakurai’s Object. Chesneau et al. in 2009 detected the presence of a circumstellar disk using the Very Large Telescope Interferometer (VLTI), and Hinkle & Joyce in 2014 (hereafter HJ14) detected the presence of bipolar lobes using near infrared imaging (see the left panel in Fig. 1). In our ALMA spectrum we detected various molecules: CO, CN, and HC₃N (see Fig. 2). The CO and HC₃N emission was spatially unresolved, but surprisingly the CN emission showed a bipolar structure (see Fig. 1) that coincided with the lobes seen by HJ14. We also detected the dust continuum which was also spatially unresolved. In the optical spectra we see two sets of lines. The first is a set of nebular lines (recombination lines of hydrogen and helium, as well as forbidden lines of heavier elements) that was first detected in 1998 (helium) and 2001 (forbidden lines). These originate in the lobes detected by HJ14 (see Fig. 3). Since 2013 we see a second set of emission lines emerging. Not all lines are identified yet, but the strongest are electronically excited lines of CN (see Fig. 4). They originate much closer to the central star than the nebular lines.

Fig. 3: position-velocity diagram of the [N II] 658.3 nm line showing the velocity along the x-axis and spatial displacement along the y-axis. It is clear that the red-shifted emission comes from a different location than the blue-shifted emission. The spatial displacement shown in the right-hand side of the diagram agrees well with the physical dimensions of the bipolar lobes found by Hinkle & Joyce (2014).

Combining these results with our data allowed us to reach the following interpretation. Sakurai’s object ejected its remaining envelope in the 1990’s. Soon afterwards there was a brief but strong shock, forming the nebular lines detected from 1998 onwards. This shock likely was caused by faster ejecta hitting slower ejecta. This shocked material then started cooling. A disk (which is likely Keplerian) also formed very quickly. This disk contains all the dust and likely also the bulk of the molecules. From 2008 onwards we see the nebular lines steadily brightening. Using the data obtained with the X-Shooter instrument on the VLT, we could prove that these lines are excited in the bipolar lobes. Since 2008 there must be an increased mass loss, which could be funneled by the disk into a jet and is now hydrodynamically shaping the bipolar lobes. The nebular lines are formed in a strong J-type bow-shock at the end of the bipolar lobes. The optical CN emission is possibly formed where the molecular wind is funneled by the disk in a C-type shock, or alternatively is excited by ultraviolet radiation from the central star. The CN could be formed by dissociation of HCN in the shock and then carried into the lobes. This would agree with the ALMA observations that show that the CN emission coincides with the bipolar lobes seen by HJ14. This implies that we now have a unique and detailed data set following the very early stages of the formation of a bipolar nebula in time. Understanding this formation process has been the subject of intense study for many decades.

Fig. 4: the complex of electronically excited CN lines that has been emerging since 2013 in Sakurai’s object.