A high-resolution study of episodic mass loss from the carbon star TT Cygni

CO radio line observations with the IRAM Plateau de Bure interferometer sho w that the carbon star TT Cyg is surrounded by a large (radius similar to 3 5 " or 2.7 x 10(17) cm), geometrically thin (average width similar to 2." 5 or 1.9 x 10(16) cm) shell of gas, which has a remarkable overall spherical symmetry (e.g., its radius varies by less than +/- 3%). It expands with a velocity of similar to 12.6 km s(-1). The emitting gas is very evenly distr ibuted in the shell when averaged over a solid angle of about 0.2 steradian s. We estimate a molecular hydrogen density of similar to 250 cm(-3), a gas kinetic temperature of similar to 100 K, and a mass of similar to 0.007 M. for the shell if the medium is homogeneous. There is no evidence for matte r immediately inside or outside the shell, nor is there any evidence for st ructure in the radial direction of the shell brightness distribution (it is essentialy perfectly fitted with Gaussians). The shell centre is displaced similar to 1." 7 (position angle similar to-20 degrees) with respect to th e star. We favour an interpretation of this displacement in terms of TT Cyg being a member of a binary system. We put forward several arguments for a shell medium that consists almost entirely of a large number of small (less than or similar to 1 ") clumps (in which case the density required to fit the observational data is much higher, similar to 10(4) cm-3, and the kinet ic temperature is considerably lower, less than or similar to 20 K). TT Cyg is presently losing mass at a modest rate, similar to 3 x 10(-8) M. yr(-1) , and with a low expansion velocity, similar to 3.8 km s(-1). This is infer red from CO line emission from a region centred on the present position of the star. The systemic velocity is estimated, from both the centre and the shell emission, to be -27.3 +/- 0.1 km s(-1) in the LSR system. All quantit ative results are obtained assuming the Hipparcos distance of 510 pc. These data strongly support that TT Cyg has recently (similar to 7 x 10(3) yr ago) gone through a period of drastically varying mass loss properties. We discuss briefly two scenarios: a short period (a few hundred years) of v ery intense mass loss (a rate in excess of 10(-5) M. yr(-1)), and a related scenario with a more modest mass ejection and where most of the shell gas is swept-up from a previous, slower stellar wind. It is presently not possi ble to favour any of these two scenarios, but we suggest that in either cas e it is a coordinated mass ejection that caused the shell formation. The He -shell flash phenomenon in AGE-stars can provide this coordination, and it also fits the time scales involved.