The observations were made during July 1996 using the 13.7 m radome enclosed radio telescope of the Itapetinga Radio Observatory 1. The receiver front-end consisted of a circularly polarized corrugated horn connected to a cooled HEMPT. An acusto-optical spectrometer was used in the back-end, with spectral resolution and total bandwidth of 70 kHz and 41 MHz respectively. The total system temperature was about 200 K. The angular resolution of the radio telescope at the frequency of NH3(J,K) = (1,1) transition was 4.2´. The observations were made using the ON-OFF total-power technique, switching between positions every 20 s with amplitude of 20´ in azimuth, enough to guarantee that there was no source in the OFF position. The effective integration time of each observation was 210 s. A 15 K noise source and a room temperature load were used in the calibration to obtain the gain and to correct for atmospheric attenuation (Abraham & Kokubun 1992). The continuum point source Virgo A was used as a primary calibrator. The NGC 6334 region was mapped in the NH3(J,K) = (1,1) transition, covering 6´ in right ascension and 12´ in declination, with a spacing of 2´. The central position of the map, with equatorial coordinates and , was observed several times during the mapping period and was used as secondary calibrator. This position was also observed in the NH3(J,K) = (2,2) transition.
A polynomial baseline was subtracted from the observations and the five Gaussian functions were fitted to the hyperfine NH3(J,K) = (1,1) spectra. The velocity separation between the Gaussians was fixed and the line-width of the hyperfine satellites constrained to be equal to the main line. This last assumption is well justified, since the optical depth of the main line is small ().
The results from the observations are presented in Table 1. The first column shows the position in the map relative to the central coordinate, Columns 2-6 exhibit the antenna temperature of each transition corrected for atmospheric attenuation, Columns 7 and 8 the velocity and line-width of the main line and the last column the number of observations made in each position. The NH3(J,K) = (1,1) spectra of some points of special interest, as well as the NH3(J,K) = (2,2) spectrum of the central position are shown in Fig. 1.
Table 1. Parameters obtained from the Gaussian fitting for each position in the map.
The NH3(J,K) = (1,1) spectra in all the observed positions exhibit systematic differences between the line intensities in each pair of hyperfine transitions (F=10, F=01) and (F=12, F=21), which are, at least in the strongest points, much larger than the rms measured at the baseline. These intensity differences can be interpreted as the signature of non-LTE conditions. However, we should point out that this conclusion cannot be obtained by looking only at the residuals of the line fittings, which present larger rms than the baseline, both under LTE and non-LTE assumption. This is expected an expected result, since the superposition of sources and the existence of velocity gradients can make the line shapes differ from Gaussian functions.
The antenna temperature distribution and gradients of velocity and line-width are presented in Fig. 2. They agree with the results of Forster et al. (1987) and Kuiper et al. (1995), who mapped with high spatial resolution a small region (about 100" in diameter) around the central position in our map in the NH3(J,K) = (3,3) transition. However, the distribution of velocities differs from that found by Schwartz et al. (1978), since their map do not show an increase in the velocity's modulus towards negative declinations, possibly due to the low signal-to-noise ratio in their spectra.
© European Southern Observatory (ESO) 2000
Online publication: October 2, 2000