3 mm spectralline observing with Mopra Tony Wong

3 mm spectral-line observing with Mopra Tony Wong ATNF & UNSW Mopra Induction Weekend - 28 May 2005

Outline 1. How bright is your source? – Estimate expected brightness temperature 2. What is the required integration time? – Depends on Tmb, Tsys, bandwidth 3. What are the required calibrations? – OFF scans, Tsys, pointing, standards 4. How do I process the data? – Emphasis on OTF; see also tutorials 2

Flux density units: Janskys Flux density is the power received per unit collecting area per unit frequency interval, and depends on luminosity and distance. W m-2 Hz-1 = 1026 Jy Generally used for discrete objects of definite size like stars, galaxies, and quasars. For spectral line sources, can integrate in frequency (velocity) to get units of Jy km s-1. 3

Brightness units: Kelvins Brightness (intensity) is the flux density per solid angle, and thus depends also on the beam size of the telescope. Generally used for extended objects such as clouds, but may be used in all situations since single-dish telescopes are calibrated in K. If using Tmb from another telescope for a compact source, must rescale to Mopra’s Tmb using ratio of beam areas! 4

Getting K from Jy At 3 mm wavelength, a source of 1 Jy will produce Tmb of 0. 1 K in the =35” Mopra beam. Applying a beam efficiency for Mopra of b=0. 4, this gives TA* = b. Tmb = 0. 04 K = 40 m. K. Note this is only valid when the source size < beam. 5

Sensitivity Radiometer equation: • ton is integration time in seconds • ∆ is bandwidth per channel in Hz (s-1). • Tsys is the “effective” (corrected for atmospheric absorption) system temperature. • Additional factors due to correlator efficiency etc. 6

Position switching Strong and variable atmosphere at 3 mm requires frequent bandpass calibration. 7

Position switching Strong and variable atmosphere at 3 mm requires frequent bandpass calibration. 8

Position switching The “quotient” spectrum is in units of TA*. 9

Position switching Sensitivity equation for position switching: • Usual practice is to set ton = toff ~ 1 minute then repeat. • For Tsys=300 K, ∆ =1 MHz, need ton=30 minutes to get rms down to 10 m. K (TA*). • In general would like to detect a signal at the 5 level. • Extra 2 improvement by averaging 2 polarisations. 10

Position switching schedule $ unit closefile observer = Ned project = standardspec telescope = Mopra 22 m receiver = SIS corrmode = NORMAL nfreq = 1 freq 1 = 110201. 393 config = ac_64_1024_2 bandw = 64 chans = 1024 source = Orion. KL raj = 05: 35: 14. 5 decj = -05: 22: 29. 56 vel = 9 obsunit = cycles obsval = 36 average = 12 srcoff = Reference lngref = -00: 08: 00. 0 latref = -01: 00. 0 obstype = TRACK go $ unit srcoff = signal lngsig = 00: 00. 0 latsig = 00: 00. 0 obstype = TRACK go $ unit srcoff = signal obstype = TRACK go $ unit srcoff = Reference obstype = TRACK go 11

Multiple ONs per OFF If spectra are taken quickly, possible to use the same OFF spectrum for several ON spectra. This strategy is employed in OTF mapping, where a single OFF spectrum is followed by ~45 ON spectra. One can show (http: //kp 12 m. as. arizona. edu/12_obs_manual/appendix_F. htm) that for N ONs per OFF, the optimal ton/toff =1/ N. This assumes that (Nton+toff) < a few minutes, the time over which the atmosphere can change significantly. 12

OTF (Raster) mapping one scan 13

Scanning in alternate directions 14

Advantages of OTF 1. Less time spent off source (assuming whole map is interesting!) 2. Averaging multiple maps can help smooth out systematics related to pointing and weather. 3. Antenna position recorded continuously during scanning, so tracking errors not a problem. Also reduces overheads involved in acquiring a demanded position. 15

When to OTF Map? • If covering a region of 3’ x 3’ or larger. • If the line is bright (~0. 5 -1 K Tmb). • For smaller regions, a large fraction of time is spent turning the telescope around. • For weak lines, a large number of maps must be taken and averaged (32 MB each!). • Alternative to OTF: do position switched observations on a grid (“grid mapping”). Can step outwards in spiral pattern. 16

Standard OTF parameters • Map size: 5’ x 5’, 31 rows, 45 spectra per row. • Scanning rate 3. 5” per second (14” Nyquist cell). • ton=4 sec, toff = 19 sec => rms = 0. 57 K (Tsys=400 K, ∆ = 75 k. Hz). • After gridding, rms reduced by about half (0. 28 K), since several spectra contribute to each cell. • Total time spent mapping: about 75 minutes! • Same sensitivity could be achieved at a single position in 1 minute (ton+toff)! 17

otfsched program • Interactive perl script to generate Mopra schedule files. • Scan in single or alternate directions, in RA or DEC, with 1 or 2 rows per OFF integration. • OFF position can be relative to map centre or given as a fixed RA and DEC. • Default is Nyquist sampling along row with 10” spacing between rows. • With 2 s cycle time, a 5’ square map takes about 80 minutes (~1500 spectra). 18

otfsched. pl <input-file> project = m 143 config = ac_64_1024_2 nfreq = 1 freq 1 = 115271. 202 bandw = 64 chans = 1024 obsfreq = 110000 observer = tw epoch = j 2000 (equatorial; use “gal” for GLON/GLAT) centre = 11: 06: 45. 0 -77: 22: 51 pmotion = 0 0 width = 5 (size in arcminutes) rowspoff = 1 scandir = 1 (DEC scanning) istart = 1 (starting corner, 1=NW and going cclockwise) absref = 11: 26: 00 -77: 20: 00 (absolute ref specified) calint = 30 source = cha 1 a-codc velocity = 3. 0 frame = lsr sched = cha 1 new 2 19

Si. O maser beam mapping 2003 Oct Beff ≈ 0. 4 20

Si. O maser beam mapping 2004 Jun Beff ≈ 0. 5 21

Data processing • The problem: coordinate information only recorded for the beginning of each scan. • The solution: query ACC, record RA, DEC, and a timestamp in (u, v, w) variables of RPFITS. • Interpolate to timestamps of spectra in aips++ reader (or better yet, using mapread/mapfix programs). • Livedata package does bandpass calibration (using previous OFF scan) and baseline subtraction. • Gridzilla package takes spectra from SDFITS files, grids them into FITS cubes 22

Useful software • rpflog: summarise an RPFITS file • rpfread: detailed information about each scan in an RPFITS file. • Otflook. csh: quick-look OTF processing, produces plots of spectra and “dot plot”. • Otfmap. csh: much the same as Otflook. csh, but also runs mapfix to produce a new RPFITS file with “corrected” position stamps. • Filewatch. csh: runs rpflog on files as they complete and appends to a daily observation log. • Comment. csh: add comments to the daily obslog. 23

Horsehead Nebula in Orion 12 CO 13 CO 6’ x 6’ fields 24

Horsehead Nebula in Orion 12 CO 13 CO 6’ x 6’ fields 25

Cha I region - N 2 H+ Cha-MMS 1 26

Cha I region - N 2 H+ 27

Cha I region - N 2 H+ 28

Other calibrations • System temperature: every half hour should provide ~5% relative accuracy in good conditions. Best to do off source. Can be included in schedule file. • Pointing: every 45 -90 minutes, on an Si. O maser <30˚ from source. • Tsys (and probably pointing) change most as a function of elevation. • Standard spectra: once a day, to check antenna gain and sideband rejection. Orion KL has lots of lines. • Si. O beam map: once a week, to check for deformations. Standard OTF schedules should be available; consult ATNF staff. 29

Final remarks • Always start with Si. O pointing to test the overall system. • When possible, observe your standard first to test the tuning. • Calibrate more often in poorer conditions. • When OTF mapping, process the first map immediately through Livedata/Gridzilla. • Keep good logs (preferably electronic) of everything you do, esp. pointing & tuning. 30