Minutes of the Leiden ALMA Science IPT meeting ============================================== December 18-20, 2002 Prepared by Bryan Butler and Ewine van Dishoeck Draft dd. February 9, 2003; send comments to Ewine by February 20 Wednesday afternoon; December 18 -------------------------------- Plenary meeting on HIFI and ALMA prepatory science --------------------------------------------------- - Xander Tielens (HIFI project scientist) - "European prepatory science" see www.astro.rug.nl/~european for information The FP6 ("Sixth Framework Program") was announced December 17. It will have lots of money for "Research Training Networks", for "Integrating Activities" and for "Development and Construction" of new facilities. Herschel will problably go for some of this money through different routes, but there is no clear view yet. Preparatory studies such as laboratory work may fit in well with the "Research Training Networks", whereas other initiatives such as the "Herschel Virtual Institute" may fit better through other budget lines. It is unclear how this ties in or competes with the "Astronomical Virtual Observatory" - the European equivalent of the NVO. - Pierre Cox (former ALMA ASAC chairman) - "ALMA in FP6" Possibilities that are being pursued: - build a "strong european network" (through networking activities) - train young radio/mm/submm astronomers (through an RTN) - develop software for science (in coordination with ALMA project) - develop ALMA advanced instrumentation (in coordination with ALMA project) Most of the networking activities of the millimeter community with the optical and radio communities will occur through the OPTICON and RADIONET "Integrating Activities". This will provide funds for organizing workshops on different topics. - Thomas Giesen - "Molecular Spectroscopy" (laboratory work) conclusions of preparatory science workshop in october '01: - get organized - find funding - name molecules of interest groups formed since (or highlighted since): - European Associated Laboratory for High Resolution Molecular Spectroscopy (HiRes) - Cologne Database for Molecular Spectroscopy (CDMS) molecules of interest: - light hydrides (e.g., CH2) - non-polar, low bending mode molecules - ions - water & relatives (is a light hydride, but important enough to separate out for Herschel) - Evelyne Roueff - "Molecular Excitation" effort: - electron impact studies - new results on H2O-H2 collisions - database for molecular excitation electron impact: "short range effects" important for deltaJ=1 transitions? have a look at: www.strw.leidenuniv.nl/~moldata also: basescol.obs-besancon.fr also: ICAMDATA - Adam Walter/Sven Thorwirth - "Databases/Webtools" AW portion: lots of different databases to handle: CDMS, NIST, JPL, HITRAN, etc.. no single one has everything that Herschel needs. how to solve? 1 - incorporate existing dBs into 1; 2 - just have pointers to the appropriates ones. unclear which way to proceed. measurements are (relatively) sparse from 500 GHz to 2 THz. ST portion: CDMS (www.cdms.de) - compare to NIST & JPL. currently 166 entries. same layout as JPL dB. - Ewine van Dishoeck - "Spectral Line Modeling" old way: - constant Tex - rotation diagrams - escape probability for constant T,n modern way: more complete treatment: iterate between physical model and observed spectrum, using a more complete description of the physics: __________________________ | _________ | | | | | | | \/\/ |\/ | n(r) -----> T(r) -----> x(r) -----> line spectra ^ ^ ^ | | | thermal chemistry rad balance xfer + telescope parms Many codes now available (at least in 1-D) based on Monte-Carlo or Lambda Iteration techniques. See: www.strw.leidenuniv.nl/~radtrans comparison of 7 1-D radiative transfer codes: van Zandelhoff et al. (2002, A&A). differences up to 20% or so seen for more complicated case, and and large differences in the structure as a function of radius. gridding is a big issue. This may well form a limitation for the accuracy with which Herschel and ALMA data can be interpreted. Also need good rate coefficients. - Maryvonne Gerin - "Ground-based Observations" "template sources" for Herschel? Some of this is being done already by individual groups and probably no need for coordinated effort. - Serena Viti - "Chemical Models" model characteristics: - time dependent - depth dependent - gas-grain/gas chemistry - degrees of freedom: - elemental abundances - T - rho - radiation field - cosmic ray ionization rate - dynamical 'switches' (simulate shocks, e.g.) - chemical reactions (number, type, & rates) - depletion (freezing out) need to benchmark current codes (compare results on a simple and more complicated model - similar to what the rad transfer group did). first comparison is underway for PDR models. See www.strw.leidenuniv.nl/~pdr for results and ongoing updates. "regions of interest" for Herschel: - ISM - hot cores - PDRs - circumstellar/protoplanetary disks - outflows/jets - shocked regions www.star.ucl.ac.uk/~sv/ChemMod/home.html ----- Bottom line conclusion: ALMA can benefit from these Herschel working groups, since many of the results are also relevant for interpretation of ALMA data. Make sure that there is at least one ALMA representative in each working group to make sure that the applications are also relevant for ALMA. ***ACTION: make sure that one ALMA representative is on each Herschel preparatory science working group at any time (currently O.K.) -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- Thursday December 19 --------------------- Morning: Joint HIFI-ALMA calibration meeting ------------------------------------------- - Frank Helmich - "Herschel/HIFI Overview" HIFI launches in 2007/2008 HIFI: 22 institutes contributing ICC = "Instrument Control Center" (where all the calibration work is being done). PACS = 40 um - 200 um SPIRE = 200 - 600 um HIFI = two different receiver bands: 480-1250 GHz; 1410-1910 GHz. SIS mixers in 480-1250; HEB in 1410-1910. delta f = 0.14 MHz min; 1 MHz typical total bandwidth = 4 GHz single linear polarization 80 K ambient absolute calibration accuracy: 10% spec; 3% goal. 2-load calibration system. extensive documentation (2 years effort already, and well documented, including control),. "use cases" are critical. this is "ESA speak"... - Carsten Kramer - "Beam Pattern on Herschel" geometric obscuration = 8% primary accuracy = 3 um surface rms (6 um path-length error) OTFM? yes, but crude. beam-switching: 3' throw. - Volker Ossenkopf - "HIFI Observing Modes and Calibration Plan" 2.7 m primary diameter (effective) -> 13" to 40" resolution (FWHM) major problems: - sideband gain ratio - is it stable over time? - slew times are long: t = 10 + sqrt(delta theta) seconds where delta theta is in arcseconds - standing waves two-load chopper to calibrate Rx. 1 @ 15K, 1 @ 100 K. they will attempt to use OFF pos'n to "calibrate" the standing wave contribution. BUT, this won't work at the high band. ALSO - Pepe Cernicharo points out that the OFF calibration doesn't work for calibration standing wave problem, at least for Earth-based telescopes (they tried it at the 30-m?). Volker argues that this is an atmosphere problem; this needs to be further discussed. cal cycle example: 2,1,2,1,2,1,2,1,2,3,2,1,2,1,2,1,2,1,2,3,2,1,2,1,2,1,2,1,2,3,4,2... 1 = source 2 = "reference" 3 = hot-cold load cal 4 = standing wave cal 2,1,2,1... = "reference loop" 2,3,2 = "bandpass calibration loop" 2,3,4,2 = "standing wave calibration loop" loop times determined by combination of thermal rms and gain drift. current thinking is ~100 sec for bandpass cal loop; ~1000 sec for standing wave cal loop. Jesus-Martin Pintado points out problem in knowing or measuring the coupling coefficients. HIFI calibration URL for maillist: mail.ph1.uni-koeln.de/cgi-bin/mailman/listinfo - Maryvonne Gerin/David Teyssier - "Astronomical Calibration Sources for HIFI" (talk given by DT). distinguish between "photometric" vs "spectrometric" vs "instrumental" calibrators. visibility is a problem - Herschel is at L2 orbit point, and can't look "back" at the Earth/Sun direction. so, need good set of secondaries and monitoring. problem of variation timescale of secondaries though. Mars models - Griffin, Lellouch Uranus model - Moreno asteroids - Muller (but SPIRE driving this) evolved stars - Loup et al. 1993 need ground-based observations of secondaries (tough at these frequencies though). which telescopes? CSO currently used, but others may be added: APEX, ASTE, JCMT, etc... they can post-correct amplitude scale, if new absolute values are available. - Stephane Guilloteau - "ALMA Calibration Issues" see Stephane's presentation on ALMAEDM. - Bryan Butler - "Astronomical Calibration Sources for ALMA" see Bryan's presentation on ALMAEDM. ---- Bottom line conclusion: useful information was exchanged; main overlap in calibration is through the primary calibrators, the accuracy of which is limited by the models. ALMA may eventually help in mapping Herschel secondary calibrators at high spatial resolution to better understand their structure. Both teams should stay in contact by being on each other's e-mail distribution list. ***ACTION: a few members of the ALMA calibration group should subscribe to the Herschel calibration list (DONE?) --------------------------------------------------------- Thursday December 19, afternoon ------------------------------- - John Conway - "ALMA Configurations" summary of status. discussed moving 4 or 5 pad positions into broad quebradas which were covered by the mask, to minimize sidelobes for some intermediate configurations (note EvD: this plan was later discarded in january, since imaging qualities of original positions good enough) now working on the reconfiguration scheme. Stephane prefers a more dynamic reconfig scheme ("reaction to proposal pressure"). some disagreement on this amongst the audience. - All - science IPT milestone discussions problem with Juan-Ramon Pardo not being able to contribute effort in the next few months, with some milestones depending on the availability of ATM. lots of discussion on how to proceed. problem with polarization. seems to be the one place where we have expended little effort to this point, and we have no person assigned to this currently (Steve Myers was going to take it over, but with his new position as AIPS++ project scientist, this will likely not happen). can we find somebody else to take over this effort? ***ACTION: identify other people to work on polarization calibration? -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- Friday December 20 ------------------- Science IPT: European tasks --------------------------- - Stephane Guilloteau - "Calibration Requirements with Science Examples" discussed science examples provided by ASAC; see ASAC www pages for details Tatematsu - need beam error pattern Wilson & Matthews - what is best freq. for thermal (& faster, since dust goes like nu^3) observations? .5-5% pol'n with SNR of 5-10. Stephane - does this provide a spec on differential gain stability on the 2 poln's of the Rx. of 5 X 10^-4 ? more demanding: SD inclusion. need to know pol'n angle to about 6 deg (value needs to be justified). French SIMS (even with ACA) indicate that this will be extremely difficult. van Dishoeck & Blake - only _relative_ amplitude calibration between receiver bands is needed. "few" %. (about 5%, after discussion). line survey - can we deconvolve sidebands? need a req. on sideband gain ratio stability for this. Peter Schilke will provide a number for this. Momose - really tough bandpass cal example. 1 part in 1000 or so, but absorption in front of a strong quasar (so no stronger astronomical bandpass sources). Gurwell - Mars atmosphere lines. similar problem to Momose. interesting that there were no examples for flux density ABSOLUTE scale. from discussion - 5% might be OK? for comparison with other wavelength instruments. also got no examples on, e.g., ant. beam props or pointing. Stephane considers these current specs to be OK and defended (historically), so specifically asked the ASAC for examples on pol'n and BP. how about baseline length calibration? science example on receiver stability comes from large scale mapping requirements. Anne noted that a science example on high angular resolution, low density contrast was missing. as a general comment, ALMA needs to explain calibration for general community in understandable terms ***ACTIONS: - decide on frequency for polarization observations (see Stephane's recent e-mail) - justify scientifically polarization angle accuracy of 6 degrees - check spec on differential gain stability on the 2 poln's of the Rx. of 5 X 10^-4 for polarization - provide req. on sideband gain ratio stability for sideband deconvolution in line surveys (action PS) - provide science example on high angular resolution, low density contrast, e.g. gap in disk - provide science example for ABSOLUTE calibration, e.g. planetary science - explain ALMA calibration to general astronomical community in understandable terms - Stephane Guilloteau - "ALMA Bandpass Calibration" presentation of idea on "bandpass normalization" - i.e., calibrating the bandpass with the autocorrelations. note that this will be in trouble if there is any strong line in the bandpass (including RFI). Bryan notes that this method gives only mixed results at the VLA, and although it's available, it's not often used. - Aurore Bacmann - "ALMA Bandpass Calibration, more..." this was originally planned to be a talk on bandpass cal. on astronomical sources but turned out to be a nice presentation on standing waves. v_ripple = c / 2D ~ 30 MHz for ALMA antennas amplitude is proportional to wavelength (worse at long wavelength). very nice work on simulating the standing wave problem in the ALMA antennas, full physical optics treatment. simulated/calculated results on standing wave ratio: - straight subreflector 0.7% - with absorbing disk 0.7% - with tangent cone 0.08% - with discontinuity (hole in subreflector) 1.3% - with scatter cone (even inside hole) 0.8% so, interestingly, cutting a hole in the subreflector is actually worse than just leaving it there (in terms of standing wave). Richard Hills says it agrees with his earlier analyses. problem with tangent cone is that optimal angle is such that some of the subreflector actually gets blocked off - solution is maybe to taper tangent cone. Kate Isaak points out that maybe one could do the single dish trick of modulating the subreflector in/out position by +- 1/8 wavelength to get around this. Stephane thinks that because of this study, there is no need to further simulate the layout, etc... of the dual-load in the subreflector cal. system. - Stephane Guilloteau - "Semi-transparent Vane" Richard Hills points out that the dual load *doesn't* suffer from the problem of needing to know the opacity (at least any more than the semi-transparent vane does), because one can think of it as a single load, which doesn't depend so strongly on opacity, *plus* additional info. Jesus-Martin Pintado points out that semi-transparent vane is *only* needed to avoid saturation (TBC with receiver IPT). but he also points out potential serious problems with polarization. also difficult to actually measure. JMP will attempt to measure it to 3% or so this spring (see next talk). Stephane and Richard both point out that if the spec is relaxed to 5% (or even maybe 3%), that there is no need to go with a fancy dual-load or semi-transparent vane calibration system, just use a simple ambient load. it was felt that we could slip the spec on the mechanical design of the dual-load in the subreflector, based on results already known now. - Jesus-Martin Pintado - "Semi-transparent Vane Test Plans" also involved: S. Navarro, M. Carter (IRAM); some others. to be done at the IRAM 30-m antenna. 3mm and 1.3mm Rx's simultaneously. the material is a 1cm thick foam. transmission = 90%, but they will measure it (with 0.5% accuracy), as a function of frequency. will attempt to measure polarization effects (to 0.8%). will attempt to measure Tvane, to 3%. Richard points out to vary the physical temp. of the vane, to get a good test also, will attempt to measure standing waves - this could be a serious problem will compare results to what is already used on the 30-m: dual-load and single-load chopper wheel methods. will attempt to do this during the spring, but time is not formally allocated, so at the whims of the IRAM scheduling. Jesus thinks that they might do the observations as early as Feb., but this is optimistic. they might attempt to do measurements with a lab mock-up of the Rx system, at least as much as they can. (update EvD: some tests can be done in february, but time not allocated before late april) - Juan-Ramon Pardo - "ATM" did one WVR test in 2002 at CSO/JCMT (JRP and Martina Wiedner). coupling of the WVR to the atmosphere is only ~ 63%! new measurements go up to ~1.6 THz. new modelling of the continuum to go > 1 THz. nu^2 dependence @ nu<900 GHz starts to break down @ 1.5 THz. phase correction tests on SMA in March 2002. observed Uranus. 230 GHz. results not so good (in terms of WVR variation matching observed interferometric phase)! want to do more tests, but a problem - can't get time on telescopes, and Martina is no longer at SMA. ATM tasks in 2002: - code cleaning - separating code levels - improved speed - ATM library for ALMA - interfacing to other software done under the auspices of ADACE - "ALMA Data Analysis Center in Europe", which is coordinating ALMA software efforts software includes 183 GHz WVR fitting. Bryan: this should be used to look at site 183 GHz radiometer data some discussion on the profile of O3. has been modelling hydrometeors - both liquid and solid. has an array simulator (see talk by Francois later). can have full 2-D phase screens (not fully 3-D, except that you specify whether they are thin or thick) which move over the array. ***ACTION: Bryan to supply Juan-Ramon with more realistic site data - Kate Isaak - "WVR" spec = 10 (1 + PWVmm) um in 1 second (1-sigma) 1 deg in elevation two options: - Dicke switch (DSB) - correlation receiver (SSB) 4 IF chains: IF nu-nu0 delta nu 1 0.96 0.16 GHz 2 1.94 0.75 3 3.175 1.25 3 5.20 2.50 giving up on correlating backend (a'la Andy Harris) [ also, a group at Cambridge has a wider bandwidth version - Richard had a work package request to look at using it for the WVR but it wasn't funded] - Richard Hills - "WVR, more" what is the test plan? Richard thinks that testing it at the ATF is best. requires an interferometer, of course. he would like to have it available to have WVRs on it by early 2005. two options: - analytical (e.g., ATM) fitting - emperical model & adjust over time (e.g., neural nets) possibility of doing tip-tilt correction. rotating mirror or prism, or chopping to antenna quadrants (@ ~ 10 Hz). but, should this be pursued? should he try to get a student to look into it/work on it? mixed reactions to it from science IPT. Effects only important for pointing and probably average out randomly. Needs 1 month of Jerome Pety/Fredric Gueth to simulate the importance ***ACTION: decision on whether to pursue "tip-tilt" correction and associated simulations - Francois Viallefond - "ALMA Simulations" 2 pieces: - Obs Tool (which is also a Prop Tool). relatively simple, with uv coverage + model + simulate); now called "Sim-Lite" - full-blown stand-alone simulator GIPSY & GILDAS versions SIMULATM (in GIPSY). combination of simulation + ATM (Juan-Ramon and Francois). done under auspices of ADACE. full-blown simulator with turbulent screen. simulates visibilities & WVR observations, and the correction of the vis with the WVR observations (WVR fitting, etc...). ============================================================================ Summary of ACTION ITEMS: ======================== ***ACTION: make sure that one ALMA representative is on each Herschel preparatory science working group at any time (currently O.K.) ***ACTION: a few members of the ALMA calibration group should subscribe to the Herschel calibration list (DONE?) ***ACTION: identify other people to work on polarization calibration ***ACTIONS Science examples for calibration: - decide on frequency for polarization observations (see also Stephane's recent e-mail) - justify scientifically polarization angle accuracy of 6 degrees - check spec on differential gain stability on the 2 poln's of the Rx. of 5 X 10^-4 for polarization - provide req. on sideband gain ratio stability for sideband deconvolution in line surveys (action Peter Schilke) - provide science example on high angular resolution, low density contrast, e.g. gap in disk - provide science example for ABSOLUTE calibration, e.g. planetary science - explain ALMA calibration to general astronomical community in understandable terms ***ACTION: Bryan to supply Juan-Ramon with more realistic site data for ATM simulations ***ACTION: decision on whether to pursue WVR "tip-tilt" correction and associated simulations