Description of source "DR21(OH)" from Kang et al. (2016)

The 44 GHz masers of DR21(OH) have been imaged with the VLA (Kogan & Slysh 1998; Kurtz et al. 2004). The VLA results show that there are two groups of masers, one at vLSR 0 km s1 and the other at vLSR ~ -4 km s1, about 20 apart from each other. These two groups of masers are located at the end of approaching and receding outow in the DR21(OH) region(Zhang et al. 2014), supporting the theory that Class I methanol masers are stimulated by outows. Our observation was pointed to the brightest peak at vLSR 0 km s1 that contains 80% of the total ux and is located at the end of an receding outow lobe. This feature has an elongated, arc-shape morphology in the VLA map.

The linear polarization of DR21(OH) shows features indicating depolarization and a sudden ip of polarization angle in the 44 GHz transition line. The total intensity and the polarized emission of DR21(OH) at 44 GHz show two maser components, one at vLSR =0.5+0.2 km s1 and the other at vLSR =+0.2+0.8 km s1, which are indicated as A and B regions in panel (a) of Figure 10. Their polarization fractions are similar, 2%, while their polarization angles differ by ~90. At the velocity where the two components over-lapped, the fraction of linear polarization drops to minimum.

Taking into account the previous observations, elongated masers aligned along a curved receding shock front with the magnetic eld compressed along the shock can explain the observed proles. Such a ip of polarization angle is expected around q = 55 in maser theories (e.g., Goldreich et al. 1973), and has been observed previously in SiO masers by Kemball & Diamond (1997) and in H2O by Vlemmings & Diamond (2006). In these observations, the angle ip appeared in a single maser feature, which was interpreted as evidence of a curved magnetic eld, with being close to the van Vleck angle, i.e., q = 55. In our case, the ip happens in two different maser features separated in velocity by a single-dish spectrum, but the underlying physics could be the same. If the two maser features are located in a curved magnetic eld aligned along the shock front receding from us, each of their being q < 55 and q > 55, then, they would produce a 90 angle ip as well as a LOS velocity shift. One implication of this explanation is that the 44 GHz masers of DR21(OH) are unsaturated, or only moderately saturated according to the simulation of Nedoluha & Watson (1990). In the upper panels of Figure 3 in their paper, the polarization angle (f in their paper) ips by the change of only when log R G 1. When the maser is highly saturated, e.g., when log R G = 3, the polarization angle is not seriously affected by but rather constant.

The polarization angle can also differ by as much as 90 when the saturation levels of the two maser components are different. As seen in Figure 3 of the above paper, for example, if the saturation levels of the two components are log R G=1 and 3, respectively, the two components will show 90 of polarization angle difference even for the same eld orientation with q = 15. How much the saturation levels of the same 44 GHz transition line can differ is another issue. Assuming that the rare factor is similar, R could be determined by the physical size and ux of maser components. The observed uxes differ by less than a factor of two, while their physical sizes cannot be determined with the current observations.

The angle ip is not observed in the 95 GHz transition line prole (see Figure 4(n)). While the polarization angle of the 0.5+0.1 km s1 component at 95 GHz is almost perpend-icular to that of the 44 GHz maser, the angle of the +0.3+0.7 km s1 component is similar to that of the 44 GHz maser. If the angle ip in the 44 GHz maser is due to the van Vleck angle crossing, the polarization angles of the 95 GHz maser being relatively constant may imply that the 95 GHz transition is more highly saturated than the 44 GHz transition, a condition in which the PA is always perpendicular to the magnetic eld regardless of , as demonstrated by Nedoluha & Watson (1990). The saturation levels of masers are difcult to measure. Future high-resolution observations and theoretical studies dedicated to methanol maser lines will be able to reveal the morphology, eld orientation, and kinematics of the DR21(OH) region as well as the involved maser physics.