Talk g siringo_laboca_spie20080626_last

Giorgio Siringo

Bolometer DevelopmentMillimeter & Submillimeter Astronomy Group

Director: Karl M. MentenMax-Planck-Institut für Radioastronomie (MPIfR)

[email protected]://www.mpifr-bonn.mpg.de/staff/gsiringo

SPIE 2008 - June 26, 2008

LABOCALABOCALLarge arge AAPEX PEX BoBolometer lometer CaCameramera

Bolometer Development Group @ MPIfR

Walter EschHans-Peter GemündErnst Kreysa (group leader)

Gundula LundershausenGiorgio Siringo

StudentsNikhil Jethava (now at NIST)

Angel Colin (now Uni-Cantabria)

LABOCAcommissioning team

Giorgio SiringoErnst KreysaAxel WeißAttila KovacsFrederic Schuller

CollaborationE.Haller & J.Beeman (LBNL Berkeley)

Frank Bertoldi (Uni-Bonn)

Alexandre Beelen (Uni-Bonn)

Lars-Åke Nyman (ESO)


IntroductionIntroduction


LABOCA key facts – 1: the instrument

Bolometric continuum receiverfor operation in the 870 µm atmospheric window (345 GHz)

•Large array: 295 semiconducting composite bolometers

•Antenna-coupled

•DC-coupling: total power receiver

•AC-bias + real-time DSP: large, clean, post-detection frequency band

•Designed and assembled by the bolometer group of MPIfR, Bonn

G. Siringo, MPIfRG. Siringo, MPIfR LABOCA – LABOCA – The The LLarge arge AAPEX PEX BoBolometer lometer CaCameramera

LABOCA key facts – 2: on APEX

Commissioned in May 2007 as facility instrumenton the APEX telescope

•High-efficiency telescope: 12 m submillimeter telescope with 15 µm surface accuracy (~ /50)

•On Llano de Chajnantor: 5100 m, extremely good atmospheric conditions for most of the year

•Large field of view: diameter = 11’.4and high resolution: 1 beam = 19” FWHM

•Operated without chopping secondary mirror, therefore

•No limitations on the scan pattern

•Continuous scanning mode, scanning speed up to 4’/s

Largest/fastest camera for mapping sub-mm continuum


LABOCA timeline

•2003: initial design

•2004: development (first wafers, design of the electronics, …)

•2005: extensive testing on a pulse-tube cooling machine, butTest failed! High noise induced by the pulses

Moved to a different design based on a wet cryostat

•September 2006: installation on APEX and first light

•May 2007: successfully commissioned

•Since June 2007 - today: routinely operated as facility instrument

•First year of operation: •two observing semesters•about 2000 hours of observations•more than 6000 hours requested


Technical OverviewTechnical Overview


•Installed in the Cassegrain cabin of APEX

•Receiver and M4-M7 mounted on hexapod positioners

Tertiary Optics


M3M3

M5M5M7M7

M6M6

LABOCALABOCA

Tertiary Optics


The task of the optics is to transform the f-ratio from f/8 at the Cassegrain focus to f/1.75 at the horn array, while correcting the aberrations over the whole field (~12’) under the constraint of having parallel output beams

The final design is diffraction limited even for 350 m wavelength

side viewside view

M1M1

M2M2

f/8 at APEXf/8 at APEXfocal planefocal plane

M3M3

side viewside view

M1M1

M2M2

f/8 at APEXf/8 at APEXfocal planefocal plane

M3M3

Tertiary Optics

side viewside view

M3M3

M5M5

M7M7

M3M3

M4M4

M5M5

M6M6

M7M7

lenslens

view from aboveview from above

f/1.75 at f/1.75 at LABOCALABOCAfocal planefocal plane


M6M6

Tertiary Optics


spot diagram with Airy disk for = 350 mStrehl ratio > .995

.1442 degmax distortion:10%

The task is to transform the f-ratio from f/8 at the Cassegrain focus to f/1.75 at the horn array, while correcting the aberrations over the whole field (.2 deg) under the constraint of having parallel output beams

The final design is diffraction limited even for 350 m wavelength

Bolometer Array

295 semiconducting composite bolometer foroperation at 300 mK and 870 µm wavelength (345 GHz)

Photons

An absorber is kept at low temperatureby a weak thermal link to a heat sink

Absorption of photons produces a temporary increase in the temperature of the absorber

A ultra-sensitive thermometer (thermistor) transforms the temperature variations of the absorber in electric signals

The electric signals are consequently amplified and processed by electronic devices


Thermal+mechanical support:unstructured silicon nitride membrane (400 nm)

Absorber:titanium layer

Thermistor:neutron-transmutation doped (NTD)germanium semiconducting chips(from E. E. Haller and J. W. Beeman)

Electrical connection:gold and niobium wires

Bolometer Array

295 semiconducting composite bolometer foroperation at 300 mK and 870 µm wavelength (345 GHz)

Photons

NTD-Ge

300 mK

870 µm

Ti layer

0.4 µm Si3N4 membrane


Bolometer Array

A naked array:

•wiring side

•4” Si wafer

•295 bolometers

Manufactured byE. Kreysa at theBerkeley Microlab


Bolometer Array

A naked array:

•wiring side

•4” Si wafer

•295 bolometers

1 central beam9 concentric hexagons+other 6x5 bolometers


Bolometer Array

opposite side: bolometer cellsNTDs are attached on the wiring side(only manual step in the manufacture)


Pictures of the two sides of the array

Bolometer Array

wiring sidebonding wires

(backreflector at /4)

other side: array of conical horn antennasRF filters

bias circuitry load resistors


Bolometer Array

295 conical horns machinedin a single aluminum block(manufactured at MPIfR)

Bolometer Array

cutfront

NTDniobium

wiresbackreflector

cut

Si wafer

Cryogenics

•The cryostat of LABOCA incorporates•a 3-liter reservoir of liquid nitrogen•a 5-liter reservoir of liquid helium

•At APEX (5107 m above the sea level) the air pressure(~540 mbar) is almost half of the standard one:

•liquid nitrogen provides thermal shielding at ~73 K•liquid helium provides thermal shielding at ~3.7 K

•The cryostat must be refilled once per day

•The operation temperature of 290 mK is provided bya two-stage closed-cycle sorption cooler

•The system works without pumping on the helium bath

sorp

tion

coo

lerso

rptio

n co

oler

liquidliquidheliumhelium

tanktank

liquidliquidnitrogennitrogen

tanktank


Cold Optics

•Filters designed and assembled at MPIfR

•Theoretical support and electromagnetic simulations by V. Hansen (University of Wuppertal, Germany)

•Band-pass formed by an interference filter made of inductive and capacitive meshes embedded in polypropylene

•The low frequency edge of the band is defined by the cut-off of the cylindrical waveguide embedded in the horn antennas.

•A freestanding inductive mesh behind the window provides shielding against radio frequency interference


low-pass filterlow-pass filterat nitrogen shieldat nitrogen shield

band-pass filterband-pass filterat helium shieldat helium shield

Cold Optics



Cold Electronics

sorp

tion

coo

lerso

rptio

n co

oler


tanktank


tanktank

focal plane

312 channels:

12 printed circuit boards 26 bias resistors each(30 MOhm nichrome/Si, MSI)

RF filters


Cold Electronics

arrayarray

12 boards12 boards26 bias resistors 26 bias resistors

per boardper board

12 flat cables12 flat cables1 cable = 26 manganin1 cable = 26 manganin

wires embedded in Kaptonwires embedded in Kapton

liquid helium liquid helium cold platecold plate

sorp

tion

coo

lerso

rptio

n co

oler


tanktank


tanktank


Cold Electronics

sorp

tion

coo

lerso

rptio

n co

oler


tanktank


tanktank

• 312 JFET impedence adapters in 4 boxes thermally shunted to the liquid nitrogen bath

•3 structured printed circuit boards per box

•26 JFET per board self heated to ~120 K

Outside the cryostat

320 channels:

295 bolometers

+

25 extra lines

320 amplifiers in 4 identical units:

•16 printed circuit boards per unit

•5 low-noise, high-gain amplifiers per board

for a total of 80 channels per unit

The amplification units also provide the AC bias circuitry and perform real-time demodulation

Data Readout


top viewtop view

side viewside view

4 data acquisition (DAQ) boards

in the backend computer

80 channels per board, in a modular 1-to-1 scheme:

1 DAQ board = 1 amplif. unit

via a direct (private) network, the data are sent to the bridge computer

Data Readout


backend backend computercomputer

4x(20x4)4x(20x4)

Data Readout


•The data acquisition hardware provides a reference frequency for the AC biasing:

the AC bias is thus synchronized to the data sampling

•AC biasing: low 1/f onset, clean post-detection bandwidth

down to 0.1 Hz crucial for the detection of large structures in fast mapping mode

•Use of a bridge computer: data sampled at high rate (1 kHz) by the DAQ hardware are low-pass filtered and downsampled in real-time

•Data stored in MB FITS format (Multi Beam FITS) by the data writer embedded in the APEX Control Software (APECS)

•Use of a frontend computer: allows remote monitor and control of the electronics (gain setting, DC offset removal) and of part of the cryogenics (temperature monitor, operation of the sorption cooler, automatic recycling)

Data Readout

~ 0.1 Hz

noise floor

1/f


•AC biasing, DC coupling: low 1/f onset, clean post-detection bandwidth down to 0.1 Hz

Data Reduction


•A new software package has been specifically developed to reduce LABOCA data: the Bolometer array data Analysis software (BoA)

•Mostly written in Python, except for the most demanding tasks, written in Fortran90

•BoA was first installed and integrated in APECS in early 2006 (extensive description in Schuller et al. in prep.)

•The BoA software can be used to process any kind of bolometer data acquired at APEX

•During the observations, the online-BoA (as part of APECS) performs a quick data reduction of each scan, to provide the observer with a quick preview of the maps being observed. In particular, for the basic pointing and focus scans, it computes and sends back to the observer the pointing offsets or focus corrections to be applied

LABOCALABOCA on Sky on Sky


LABOCA On Sky


Performance

• 248/295 useful channels (84%)•2 blinded to monitor temperature and noise•18 show high noise •29 dead

•Positions and relative gains of each bolometer are derived from fully sampledmaps (beam-maps) on the planetsMars and Saturn

The circles represent the FWHM shape of the 248 beams on sky by 2-dimensional Gaussian fit to the single-channel map of each bolometer

Only bolometers with useful signal-to-noise are shown in this map.

Footprint of LABOCA on sky

from a beam-map on Mars

LABOCA On Sky


Performance

•The beam shape was derived for individual bolometers from

•beam-maps on Mars•pointing scans on Uranus and Neptune

Both methods lead to comparable results

•The main beam is well described by aGaussian with a FWHM of 19”.2 ± 0”.7

•The beam starts to deviate at ~ -20 dB

Radial profile of the LABOCA beam derived averaging the beams of all functional bolometers (248) from fully sampled maps on Mars

The error bars show the standard deviation

LABOCA On Sky


PerformancePerformance

• Mean point-source sensitivity of the array: 53 mJy·sqrt(s)(NEFD per channel, low frequencies filtering applied)

• Extended emission sensitivity: 95 mJy · sqrt(s)(without low frequencies filtering)

Number of bolometers per sensitivity interval

LABOCA On Sky


Observing Modes

•The frequencies of the signal produced by scanning across the source must fall into the white noise part of the post-detection frequency band (0.1 - 20 Hz)

•Min scanning speed: 30"/sminimum required for sufficient source modulation, depends on the atmospheric stability and source shape

•Max scanning speed: 4'/slimited by the time resolution in the coordinates given by the telescope’s control software

•Possible scan patterns limited by the telescope’s control software, which allows only constant speed, in rectangular or polar coordinates: spiral patterns are possible

LABOCA On Sky


Observing Modes

Mapping modes

•On-the-fly maps (otf): rectangular scanning patterns with a constant scanning linear speed, in horizontal or equatorial coordinates. For maps on the scale of the FoV up to a few degrees

•Spirals: done with a constant angular speed in polar coordinates, the linear scanning velocity is not constant. In ~30 s can produce a fully sampled map for the FoV with linear scanning velocities limited between 1’/s and 4’/s

•Rasters of spirals: for fainter sources, the basic spiral pattern can be combined with a raster grid of pointing positions resulting in an denser sampling and longer integration time. This mode gives excellent results for sources smaller than the FoVand is suitable for integrations on faint sources

Other modes

•Pointing: short spiral, offsets determined with a 2-dimensional Gaussian fit•Focus: 5 s integration on a bright source at 5 different subreflector positions•Skydip: a continuous tip scan in elevation, to measure the power of the atmospheric emission as function of the airmass

LABOCA On Sky


Observing Modes

Example of a raster of spirals:spiral mapping modecombined with a rastermap of 25 positions

pattern in Az/El of the central beamof the array

complete scan = 12 minutesfinal map = 2000x2000 arcsecwith uniform rms residual noise

LABOCA On Sky


Calibrations

•Calibration accuracy, based on planets: ~5-10%(depending on weather stability)

• Sky opacity is determined with skydips

•Secondary calibrators list available(based on MAMBO and SCUBA lists)

Example of calibration on planets: 5 observing runs on consecutive days

LABOCA On Sky


ATLASGAL: APEX Telescope Large Area Survey of the GalaxyMPIfR Bonn (Schuller et al., in prep.) + MPIA Heidelberg + ESO + Uni. de Chile

Unbiased survey of the inner Galactic Plane at 870 µmGoal: mapping |l| ≤ 60°, |b| ≤ 1.5°, down to 50 mJy/beam2007: covered 95 deg2 in ~75 hours of observing time2008: covered another 150 deg2 , project ongoing…

4x2 4x2 degdeg22 map of the Galactic Center map of the Galactic Center (courtesy of F. Schuller)(courtesy of F. Schuller)

Integration time: Integration time: 6 hours only6 hours only - rms: 50 mJy/beam- rms: 50 mJy/beam

FoVFoV

NGC 253: 8 h - rms: 3.5 mJy/bCen A: 5 h - rms: 4.0 mJy/bNGC 4945: 3 h - rms: 5.0 mJy/b

(courtesy of A. Weiß, submitted to A&A)

LABOCA On Sky


Nearby Galaxies

First detection at 870 µm

NGC 253

NGC 4945

Cen A

LABOCA On Sky


S/N pixel histogram (1 pixel=1/3 of beam)

rms vs time

30’ x 30’ deep field200 hour of integrationrms: 1.5 mJy/beam63 sources detected

Weiss, Smail, Walter et al.

Chandra Deep Field – South (CDFS)

LABOCALarge APEX Bolometer Camera

The future

In collaboration with the Institute for Photonics Technology (IPHT) of Jena we are already working on LABOCA-2 :

a new bolometer array using superconducting technology• TES (transition edge sensor) thermistors• dipole absorbers• SQUID multiplexing and amplification

A system already using the same technology, SABOCAthe Submillimeter APEX Bolometer Camera (870 GHz)has been succesfully tested on APEX last May and will be commissioned in October 2008 as facility instrument(see also N. Jethava’s talk about TES in the afternoon)

Given the low sensitivity of superconducting bolometers to microphonics, it will be possible to move from the wet cryostat to a pulse-tube cooler: already succesfully tested in our labs in Bonn

350 µm map of the NGC6334 molecular cloud – SABOCA on APEX - May 16, 2008total observing time: 1h 40m, including observations of calibrators; units: Jy/beam

LABOCA map

SABOCA = Submillimeter APEX Bolometer Camera (350 µm)

37-elements TES + SQUIDs multiplexing and amplification