1 Introduction

The present study about the origin of OH maser radiation near new born stars was motivated by a recent study by [Baudry & Diamond] (1998) on the $^2\Pi_{3/2}(j=7/2)$ 13.44 GHz OH maser emission in W3(OH). This study was done with high spatial resolution and revealed a pronounced spatial structure with filaments containing single small and very intense maser spots. The compact area in which all of the j=7/2 masers are observed is very close to the maximum of 23.8 GHz continuum emission, which is suspected to be the position of a massive new born star ([Baudry & Diamond] 1998). There is one spot that extends over only 1 astronomical unit (AU) and yields a microwave flux of more than 20 Jy on earth. Because the bandwidth is $\Delta \nu \approx 10\,{\rm kHz}$ the observed line integrated flux density from this spot is $j_{\rm obs}$ $\approx 2\ 10^{- 21}\,{\rm W/m}^2$. A general background emission of 20 mJy is observed from the whole area where OH j=7/2 masers are found. The 13.44 GHz masers operate between the two $\Lambda$-doublet states of the j=7/2 rotational level in the $^2\Pi_{3/2}$ multiplet state of OH between the F=4 hyperfine states.

Because the maser emission originates from a rotationally excited state in OH with a high IR relaxation loss rate of $\approx 0.5\,{\rm s}^{-1}$ a very efficient mechanism must pump OH molecules into the j=7/2 state to account for the high loss rate. Despite this high loss rate the maser emits a large photon flux. [Baudry & Diamond] (1998) discuss the observations on the currently available maser models which assume collisional and/or IR pumping schemes (e.g. [Elitzur] 1992) including radiation transport. For the excitation of the high lying maser in the j=7/2 state the observations are extremely difficult to explain by these models. [Baudry & Diamond] (1998) speculate that a shock might excite the maser. However, although the shock should show up in the data, it is not seen. Here we try to explain this maser radiation by a completely different mechanism, based on grains that evaporate water in the presence of VUV which photolyzes water in the first absorption band creating inversion in the nascent OH photoproduct. This fundamental mechanism was not even mentioned by [Baudry & Diamond] (1998) although it has been suggested almost 15 years ago ([Andresen] 1985).

In this paper we will work out a quantitative OH maser model including evaporation of water from grains, generation of OH by the photodissociation of water, subsequent IR relaxation and destruction of OH by photolysis. We will show that this model yields population inversion and gain for all masers observed in star forming regions. In this still crude model we neglect collisions and FIR-pumping, which will affect the population to some extent, in particular at low photodissociation rates. We give quantitative numbers for the gain for the different maser transitions as a function of photodissociation and evaporation rates. It is interesting to note that all masers can operate over a wide range of photodissociation rates and that the output is determined to a larger extent by the evaporation rate rather than by the photodissociation rate. Nevertheless there are some OH maser features which cannot be explained with our model up to now, e.g. the absorption line at 13434 MHz or the 1667 MHz/1665 MHz emission ratio. Therefore we are working on various aspects to improve the present model. One aspect is the inclusion of FIR pumping and collisions, others are related to steric effects. We will show that collisions can be neglected at high photodissociation and IR relaxation rates.

We also elaborate on the astrophysical conditions required for the application of the present maser model. In contrast to the maser models discussed by [Baudry & Diamond] (1998), which do not even explain the origin of the high OH abundance, we use the well accepted picture of star formation in which the star is surrounded by a large cloud of grains. We will show that grains at the border between the HII region and the cold cloud are exposed to the heat- and VUV- flux from the central star and that in this situation masing should occur. We discuss the role of different grain sizes on the evaporation. The present model predicts that OH maser radiation should occur whenever grains are exposed to the heat flux and the VUV flux from the central star.

According to the present model ground state OH masers should be located directly at the border between the HII region and the grains. We will also discuss under which conditions the present maser model might explain the very high flux density obtained from the small diameter spot observed by [Baudry & Diamond] (1998) for the 13.44 GHz maser.