The EVLA and SKA pathfinder surveys

Jim CondonNRAO, Charlottesville

The EVLA recycles VLA infrastructure• The VLA is a 1960’s

design• 27 x 25-meter VLA

parabaloids – Off-axis Cassegrain

optics for > 1 GHz • Many km of railroad

tracks:– 4 scaled arrays, 1‒ 36

km– Selectable resolution

and brightness sensitivity

– Good imaging with nearly natural weighting

– Multiwavelength spectra at constant angular resolution

Leiden 2011 Feb 24

The EVLA recycles VLA infrastructure• VLA was designed in the

1960’s

• Advantages:– Already exists and

paid for! – Works well 1 to 50 GHz– Sensitive and flexible

• Disadvantages:– Small field of view– Many competing users

Leiden 2011 Feb 24

What’s new? 1. Receivers and feeds for continuous coverage from 1 to 50 GHzBand

(GHz)

Tsys/(best weather)

1-2 L 60 -- 80

2-4 S 55 -- 70

4-8 C 45 -- 60

8-12 X 45

12-18 Ku 50

18-26.5 K 70 -- 80

26.5-40 Ka 90 -- 130

40-50 Q 160 - 360

L

K

QKa

XC

S

Ku

FeedHeaters

What’s new? 2. The ‘WIDAR’ correlator

• The key element of the EVLA is its `WIDAR’ correlator – a 10 petaflop computer

• Major capabilities:– 8 GHz/polarization maximum instantaneous bandwidth– Spectral dynamic range up to 58 dB– Extensive special modes– 64 independently tunable full polarization ‘spectral

windows’, each of which effectively forms an independent ‘sub-correlator’

Leiden 2011 Feb 24

Summary of wideband coverage:

Freq. IF BW SW Width

# SWP

Freq. Res.

Vel. Res.

Nch per

Nch Data rate*

GHz GHz MHz kHz km/s spctrm total MB/s

L 1—2 1.024 16 64 15.6 3.1 1024 131072 50

S 2—4 2.048 32 64 62 6.3 512 65536 25

C 4—8 4.096 64 64 250 12.5 256 32767 12

X 8—12 4.096 64 64 250 7.5 256 32767 12

U 12—18 6.144 128 48 1000 20 128 12288 4.8

K 18—26.5 8.192 128 64 1000 13 128 16384 6.2

A 26.5—40 8.192 128 64 1000 9 128 16384 6.2

Q 40--50 8.192 128 64 1000 6.5 128 16384 6.2

• For dual (RR,LL) polarization, with recirculation• For full polarization, spectral channels are 2 x wider.

* With 10 second averaging

Leiden 2011 Feb 24

EVLA sensitivity• The table gives the 1-hour, 1- point-source

sensitivities, in continuum and line (1 km/sec)

Band Code EffectiveBW

SEFD (cont) (line)

GHz GHz Jy Jy mJy

1 – 2 L 0.75 400 5.5 2.2

2 – 4 S 1.75 350 3.9 1.7

4 – 8 C 3.5 300 2.4 1.0

8 – 12 X 4 250 1.8 0.65

12 – 18 Ku 6 280 1.7 0.61

18 – 27 K 8 450 2.3 0.77

27 – 40 Ka 8 620 3.2 0.90

40 -- 50 Q 8 1100 5.6 1.4

How can the EVLA contribute?

The EVLA already exists, so it can be used now to develop imaging techniques on real data:1.Wideband imaging2.High-fidelity imaging

The EVLA can support othersurveys with1.Reference fields2.High-frequency survey?3.CO spectroscopy4.Continuum spectral indices5.High angular resolution and position accuracy

Wideband imaging

• The EVLA will often be operating in ‘wideband’ mode, with large fractional bandwidths: ~ 1.

• Source structures change significantly over this range – cannot simply grid all channels on one transform plane.

• The primary telescope beam scales with -1 – introducing a ‘false’ spectral index.

• Deconvolving all background sources – even beyond the main beam – is necessary to achieve full sensitivity.

Leiden 2011 Feb 24

High-fidelity imaging

• The super-linear design of the EVLA should enable much higher DR imaging than before, so DR will be limited at virtually all bands by various primary-beam issues:– Systematic pointing offsets– Unsteady tracking– Noncircular beams (alt-az mount telescopes)– Polarization beams (special case of above).

Leiden 2011 Feb 24

3C147 deep field @ 1440 MHz

• Detailed testing underway with 12-station, full polarization, 4 subband initial configuration.

• Most demanding testing is at L-band, short spacings.

• 12 antennas, 110 MHz, 6 hours integration

• Best image so far: 400,000:1 DR for 3C147. (850,000:1 to noise in corners).

• Some artifacts visible – due to noncircular primary beams.

• RMS noise in corners ~ 1 millionth of the peak.

First Null

50% power

Leiden 2011 Feb 24

3C147: Residual errors in the full field

Smearing and W-Term errors

Errors due toPB pointing

Errors due toPB sidelobes

Leiden 2011 Feb 24

High-Fidelity Imaging – Cygnus A

• Cygnus A, observed with WIDAR last year at X band (3.6 cm), self-calibrated and imaged with the best modern methods, provided an image:– No better than that which Rick Perley made 25 years ago with the VLA– A noise level 1000 times the thermal noise!

Leiden 2011 Feb 24

What Went Wrong?

• What is wrong with Cyg A? • Clearly associated with

imaging/self-calibration. • Many ideas – little progress…• Important problem that must be

solved.

• Image of the calibrator for Cygnus A

• DR is ~650,000:1• Noise is 6 Jy (thermal)

0.004% of peak!

Leiden 2011 Feb 24

CO in submm galaxies

Leiden 2011 Feb 24

Low-excitation molecular gas in active star forming galaxies at z~2, during the ‘epoch of galaxy assembly.’ These observations show more total gas, by a factor 2, than previously derived from higher order transitions and indicate that the gas is extended on scales ~ 16 kpc. They challenge many preconceptions on massive galaxy formation at high z. (Ivison et al 2011, MNRAS, in

press).

Molecular gas in lensed LBGs

• The Ka band has performed the first imaging of molecular gas in normal star forming galaxies at high redshift (LBGs)

• Two strongly lensed were detected in CO1-0.• The gas masses, gas fractions, moderate CO excitation, and

star formation efficiencies are comparable to low-redshift star-forming galaxies

16

Riechers et al., 2011, ApJ,724, L153

CO in the most distant starburst galaxy at z = 5.3

The EVLA and the PdBI found CO emission from the most distant known submm galaxy at z=5.3. These observations show the cold gas that fuels the star formation, implying extreme amounts (>1010 M) of dense gas in this forming elliptical galaxy. (Riechers et al. 2010, ApJ, 720, L131)

4.051

z=4.055

4.056

0.7mJyCO2-1 46GHz

0.4mJy

1000 km/s

0.3mJy

• SFR ~ 103 Mo/year

• Mgas ~ 1011 Mo

• Early, clustered massive galaxy formation

GN20 molecule-rich proto-cluster at z=4 (Daddi)CO 2-1 in 3 submm galaxies, all in 256 MHz band

Evolving Universe

Leiden 2011 Feb 24

EVLA as SKA pathfinder: counts of very faint sourcesLocal visibility function

ϕ(L) = L5/2ρ(L) yieldscounts normalized by static Euclidean counts

∆ log(L) ~ 1.2 evolutionfor all radio sources <z> ~ 0.8, ∆ log(z) ~ 0.5over wide flux range(more sensitive survey ≠“deeper” survey)

stars

AGN

AGN

stars

Simple model counts ≈ Wilner et al. (2008)

The EVLA as an SKA pathfinder:deeper knowledge through confusion

Ultimate limits for radio continuum surveys

• confusion: c~ .03 K at 1.4

GHz) instrumental natural• field of view Ω• dynamic range• noise

Summary: Examples of possible EVLA contributions• CO J=1-0 spectra of high-z galaxies, clusters, and quasars• Continuum radio spectra of individual sources and fields• All-sky 6 cm NVSS for spectra of stronger (more luminous)

AGN• Sky density of faint sources• “Reference” field with higher resolution and sensitivity than

EMU/WODAN surveys

Leiden 2011 Feb 24