next previous
Up: Non-confirmation of reported H


2 Observations and data reduction

We made our observations with the refurbished 305 m Arecibo Gregorian radio telescope in May and June 2002. Data were taken with the L-Band Narrow receiver (circularly polarized at 1415 MHz), using nine-level sampling with two of the 2048 lag sub-correlators set to polarization A and two to polarization B. All observations were taken in the position-switching mode, with the off-source observation taken for the same length of time and over the same portion of the Arecibo dish as the on-source observations. Each on+off pair was followed by a 10 s on+off observation of a noise diode calibration source. In principle, each scan consisted of a 5 min on/off pair. The total net integration time (on+off) was on average 90 min per source, the maximum being 140 min for the faintest reported VLA source, sw-089, and the minimum 40 for sw-174, where a nearby strong continuum source made sensitive observations impossible. Each cloud was observed with each of the 4 sub-correlators centered at its redshifted H  I line frequency. Two of the sub-correlators, one per polarization, were set to a 12.5 MHz bandpass, resulting in a velocity coverage of about 2500 km s-1 and a velocity resolution of 1.3 km s-1, while the other two were set to a 3.25 MHz bandpass. The telescope's HPBW at 21 cm is $3'~\hspace{-1.7mm}.\hspace{.0mm}4$ $\times$  $3'~\hspace{-1.7mm}.\hspace{.0mm}6$. For the telescope's pointing positions the centre coordinates of the VLA H  I sources as given in D97 were used (see Table 1). For calibration purposes, a number of strong continuum sources as well as spiral galaxies with strong H  I lines from the catalogue of Lewis (1983) were observed throughout the run.

The observations were made at a mean frequency of 1368 MHz, in the 1350-1400 MHz frequency band allocated on a co-primary basis to the Radio Astronomy Service, where its protection from harmful interference is limited. Though care was taken to make the renovated Arecibo telescope more robust against radio frequency interference (RFI), and to coordinate its operation as well as possible with the frequency plan and emission periods of local radar installations, RFI signals with strengths that hamper the detection of faint H  I line signals were present during the first half of the observations. After the main terrestrial RFI source was identified halfway through the observing run, a blanker could be implemented that effectively removed these signals from our spectra.

The data were reduced using IDL routines developed at Arecibo Observatory. The two polarizations were averaged and corrections were applied for the variation in gain and system temperature of the telescope as function of azimuth and zenith angle, using the most recent calibration data available. A first-order baseline was then fitted to the data, excluding those velocity ranges with H  I line emission or RFI. Once the baselines were subtracted, the velocities were corrected to the heliocentric system, using the optical convention. All data were boxcar smoothed to a velocity resolution of 19.5 km s-1 for further analysis. If the average of all data on an object contained a signal with a peak level exceeding 1 mJy, it was checked if that feature corresponded to RFI signals that did not occur in all spectra - if so, the contaminated spectra were not used for the final analysis. The rms noise levels of the averaged spectra were determined in channels 300-1800, avoiding lines with a maximum exceeding about 1 mJy peak line flux density.

 \begin{figure}
\par\includegraphics[width=10.4cm]{aa3054fig1.eps}
\end{figure} Figure 1:I column density contours from the Dickey (1997) VLA survey superimposed on our V-band CCD images (from Iglesias-Páramo et al. 2002). The H  I clouds without optical counterparts in these images, reported as tentative detections in Dickey (1997), have been indicated.


next previous
Up: Non-confirmation of reported H

Copyright ESO 2003