Demonstration Observations: W3OH

In preparation for this large project we conducted trigonometric parallax observations on the 12 GHz methanol masers in W3OH. We observed in the same manner that we propose here (see below), including the measurement of and correction for the atmospheric/ionospheric delay errors in the VLBA correlator (typically 10 cm of vertical path), which greatly increases the accuracy of the observations. We used the masers as the phase-reference source and switched rapidly among 3 QSOs. The angular separations of the QSOs from W3OH ranged from 0.7 to 1.5 degrees. The results were outstanding!

We switched rapidly among W3OH and 3 background sources, repeating the following pattern: W3OH, J0235+6216, W3OH, J0231+6250, W3OH, J0230+6209. Sources were changed every 40 s, typically achieving 30~s of on-source data. We used a methanol maser as the phase-reference source, because it is considerably stronger than the background sources and could be detected on individual baselines with signal-to-noise ratios typically more than 100 in the available on-source time.

The data were reduced using NRAO Astronomical Image Processing System (AIPS). The calibration sequence included: 1) parallactic angle, reference source position, and atmospheric delay error correction; 2) correlator bias and system temperature and gain curve corrections; 3) electronic phase-calibration, phase referencing and self-calibration. After applying these corrections and calibrations, we made an image of a strong maser (ie, the reference channel) using a 1.2 mas (CLEAN) restoring beam. We show the first and last epoch images of the reference spectral channel at V(LSR) = -44.2 km/s in Fig.~1. One can see that there is little change in the masers over a year time scale.
Fig. 1: The first and last epoch maps of the spectral channel at V(LSR) = -44.2 km/s containing the maser reference spot. Note that the maser emission structures change little over the time range of our parallax measurements. The contour levels are multiples of 2 Jy, with the zero contour surpressed. The origin of the maps is at R.A.(J2000) = 02 27 03.8192 and Dec.(J2000) = 61 52 25.230.
In order to provide the data needed to measure the parallax and proper motion, we fitted Gaussian brightness distributions to the nine brightest maser spots and the three background radio sources for all five epochs. In Fig. 3 we plot the positions of one maser spot in W3OH relative to the three background radio sources. The change in position of W3OH relative to a background radio source was then modeled by the parallax sinusoid in both coordinates, completely determined by one parameter (the parallax), and a secular proper motion in each coordinate.
Fig. 2: Position versus time for a maser spot at V(LSR) = -44.2 km/s relative to three background radio sources. The large difference in position between W3OH and each background source has been removed and the data for the different background sources have been offset for clarity. The top, middle, and bottom plots are for the background sources 0230+621, 0231+625, and 0235+621, respectively. Also plotted are the best fitting models, specified by five parameters: one parameter for the parallax and two parameters for the proper motion in each of the coordinates.
An unweighted average parallax for nine maser spots of W3OH measured against each of the background sources yields 0.502 +/- 0.011 mas using J0230+621, 0.526 +/- 0.014 mas using J0231+625, and 0.515 +/- 0.015 mas using J0235+621. These results are consistent within their formal errors, and a weighted average of these three parallaxes yields 0.512 +/- 0.007 mas (corresponding to 1.953 +/- 0.03 kpc). The parallaxes from the three calibrators increase slightly with increasing maser--calibrator separation, possibly suggesting that atmospheric systematics may not have been entirely removed, since the three calibrators are all offset roughly north-east of W3OH. However, with only three calibrators it is difficult to assess the significance of this effect. Still, allowing for the possibility of some uncompensated atmospheric systematics, we adopt a parallax uncertainty of 0.010 mas (0.04 kpc), corresponding to a distance uncertainty of 2%.

The proper motions in the RA and Dec coordinates were -1.14 +/- 0.05 and -0.26 +/- 0.05 mas/yr, respectively. Thus, the uncertainties in proper motions translate to an impressive 0.5 km/s (for the 1.93 kpc distance). As a check on this accuracy, we computed the motion of W3OH perpendicular to the Galactic plane to be -0.8 +/- 0.5 km/s (after correction for the 7 km/s motion of the Sun in this direction). This ``null result'' is consistent with what one might expect for the motion of a massive forming star.

Note that the Perseus arm sources, including W3OH, have kinematic distances of about 4 kpc, while the Georgelin & Georgelin model used a luminosity distance of 2.2 kpc for this arm (Humphreys 1978). We have now resolved this significant discrepancy; the luminosity distance is approximately correct and the Perseus arm has strong kinematic anomalies.


Humphreys, R. M. 1978, ApJS, 38, 309