4 Implications on pumping schemes

There exist two main ways to invert the 5 cm maser transition, (1) by radiative pumping of far infrared (FIR) photons or (2) by collisions with H₂ molecules (in combination with local and non-local line overlap). Chemical pumping does not seem applicable and interstellar UV photons are only responsible for the dissociation of H₂O molecules in the outer part of the circumstellar envelope to produce OH (the contribution of stellar UV photons is negligible except perhaps for the inner regions of the envelope).

4.1 About the ground-state masers

Theoretical studies of the pumping mechanism of the 18 cm OH lines are quite advanced. The absorption of FIR photons at 34.6 $\mu$m and 53.3 $\mu$m excites the OH from the ground state to the $^{ 2}{ \Pi }_{ 1/2}$ladder. Subsequent cascading of the populations through the J= 1/2 and J= 3/2 levels inverts the J= 3/2 ground state (Elitzur et al. 1976). This scheme explains rather well the strong 1612 MHz line and essentially avoids the $^{ 2}{ \Pi }_{ 3/2}$ ladder (Davis et al. 1979). Therefore, in circumstellar regions where 1612 MHz is strong and the above mechanism prevails one should not expect to detect 5 cm OH emission.

However, the pump cycle may differ in inner layers where highly excited lines are expected to be found. Moreover, Collison & Nedoluha (1994, 1995) argue that FIR line overlaps are also important to enhance the 1612 MHz line and may even be the primary inverting scheme for the 1612 MHz line in not too optically thick OH envelopes. In the latter case overlaps at 120 $\mu$m are important and populating the J= 5/2 level is essential. Collison & Nedoluha (1994) argue that FIR line overlaps alone cannot explain the main-line masers in stars (contrary to earlier models) and that near infrared (NIR) overlap effects (with OH or H₂O) are likely needed to explain the main line emission from thin circumstellar shells.

The most recent model published on OH masers in circumstellar envelopes (see Thai-Q-Tung et al. 1998) treats only the ground-state excitation and considers two models. The first one is with line overlapping limited by a Doppler shift of 2 km s^-1 and the other one with large overlapping (up to the expansion velocity). In both models the 1612 MHz appears much stronger than the other ground state lines by a factor 10² to 10³. They found that the pumping based on FIR hyperfine line overlapping is much smaller in the second case. They suggest that FIR line overlapping occurs inside clumps (small Doppler shift) of circumstellar envelopes (this idea was also invoked by Collison and Nedoluha), but no prediction is made about the excited state. Only the recent work of Pavlakis & Kylafis (1996, 2000) studied excited OH maser emission but only in the case of massive star forming regions.

Modeling the detected maser emission in Vy 2-2 at 1612 and 6035 MHz is a challenge. The particular nature of the source, a very young proto planetary nebulae, may certainly be a clue. The fact that two masers are observed at the same LSR velocity (-62 km s^-1) argues in favor of their spatial association. In such a case they would both originate from the thin ionization shell presenting similar conditions as in HII regions. Within this context, PPN shells may be characterized by particularly high densities and long path lengths for coherent amplification that have to be taken into account.

4.2 Infrared pumping?

Excitation of the OH radical results from complex competitive schemes involving both collisional and radiative pumping as well as line overlap effects that are correlated with the velocity field in the OH medium and local line broadening. The 5 cm OH lines arise from energy levels 84 cm ^-1 above the ground-state and we therefore expect that FIR photons around 100 $\mu$m are involved in the OH pumping cycle. To evaluate the possibility of a pumping scheme based only on IR photons, Fig. 5 compares the IRAS flux at 60 $\mu$m and the lower limits of OH emission at 6 GHz, assuming that the ratio between the radio solid angle and the IR solid angle is $\approx$1. Figure 5 shows that the number of FIR photons largely exceeds the emitted radio photons. Baudry et al. (1997) reached a similar conclusion for compact HII regions but in that case many 5 cm OH masers could be detected. From this we conclude that in stellar envelopes the OH pumping mechanism is different from that in massive star forming regions and that the available FIR radiation is unlikely to work as a pump for this maser. Moreover as the envelopes are dense, it is possible that even if the IR pumping were efficient, collisions could effectively quench the OH maser emission.