Global EoR Experoments

Ron Ekers, CSIRO

CAASTRO Global EoR WorkshopUluru, 17 July 2013

Summary

Strategy for technically difficult experiments– Either masochists or people who don’t know any better!– NB the Crick and Watson story on DNA structure

Global HI EoR v imaging HI EoR– Statistical v direct detection

CORE ZEBRA SARAS Other global HI experiments

– EDGES, BIGHORNS, COREII, DARE, Pulse calibration Epoch of (re)combinationJuly 2013 2

Global HI EoR prediction Pritchard et al, Nature 468, 772 (2010)

Z =

6.3

Peter Shaver

conjecture

(200/65)2.5=17

The 21cm EoR challenge

Global T ~30mK in few MHz– S/N easy – can reach a few mK in few hours

T/T < 10-4 to10-5

Calibrate the complex gain Minimize the number of unknowns that can couple to EoR Remove the forgrounds Remove the additive constant

– Correlation receiver» Eliminate LNA additive noise

– Position switching» T now very small so large antenna and long integration times » Correlation interferometer» Arrays

• Statistical detection• Direct detection

Zero spacing interferometerZero spacing

interferometer

From CoRE to ZEBRA

CoRE (Chippendale)

ZEBRA SARAS

ZEBRA II

Calibratable receiver

CSI

RO

RR

IM

RO

July 2013 5

COREfrequency independent antenna beam

Global EoR system RRI Bangalore

Ravi Subrahmanyan

Ron Ekers

Peter Shaver

A. Raghunathan

Zebra – fat dipole v1

ZEBRA Global EoR Experiment

ZEro-spacing measurement of the Background RAdio spectrum

Partially reflecting resistive screen

Virtual zero spacing interferometer

Removes all additive errors

Modulate screen ?

Subrahmanyan, EkersPatra

Partial reflector/transmitter

X

The space beam-splitter: a resistive wire mesh

Need a space beam-splitter before the antenna A lossless screen (e.g. a conducting grid)

– transmitted & reflected waves are orthogonal Resistive wire mesh

– Thickness of wire < skin depth– Frequency independent– Re-radiated fields no longer cancel the incident field on

the far side of the wire screen Lumped resistance on scale <<

– Practical solution instead of resistance wire

Building resistive screen

The Resistive Screen

copper wire + lumped resistors

resistor value = free space impedance/2

3x4 metres

holes to reduce wind loading

Roll up for transport

ZEBRA – interferometerfirst CMB correlation 20 Jan 2011

3.4m

1.5m separation Max sky coverage at zero spacing

26% Contributions to correlated output

– Global sky signal– Screen radiating– 1.5m interferometer sky

correlation» One path through screen» Both paths miss screen

Ferrite absorber

ZEBRA at Gauribidanur

Zebra correlated output

Baseline ripple – changes with LST– Repeats each day– Multipath scattering of galaxy foreground signal– Shifted location …….

SARAS receiver evolution

SARAS receiver

Patra & Subramanyan, EA (2013) 88-175MHZ Differential correlation spectrometer Digital correlator well separated from

receiver Minimize number of parameters in

solution (11) Solve for multipath propagation from

internal reflections Eg noise from receiver input

SARAS internal reflections

Short connections to keep broad bandwidth Long connections to decrease coupling

SARAS internal reflections

Short connections to keep broad bandwidth Long connections to decrease coupling

SARAS waterfall plot

Pulse calibration ?

Pulse injected at Parkes vertex

Pulse reflected from Parkes focus

Inject and integrate short (sec) pulses Calibrated noise spectrum Understand & calibrate reflections Nipanjana Patra, Paul Roberts

Pulse calibration

Band limited pulse with -20db reflection Pulse repetition rate 106 Hz Accuracy 0.05%

Other Global EoR Experiments

WSRT: – Lunar occultation

EDGES: Rogers & Bowman– Polynomial fits Δz > 0.06– Absolute calibration of components for wider bandwidths

BIGHORNS: Sokolowski, Tremblay, Wayth, Tingay ⑫– Low rfi site, high stability

Core II: Bannister, Chipendale, Dunning ①– Precision self calibrating receiver

DARE– Go to moon to avoid ionosphere and rfi

July 2013 23

Estimates of the sources of error and their magnitude expressed as the residuals to fits with increased numbers of parameters along with the bias in EOR estimation

Parameters of 10 parameter solution:1] EoR signature (30 mK, 50@145MHz)

2] scale (assumes spectral index of -2.5)

3] constant (ground emission)

4] frequency -2 (ionosphere emission)

5] frequency -4.5 (ionosphere absorption)

6] Magnitude of antenna S117] Magnitude of LNA S118] S11 phase error9] S11 delay error10] temperature scale

Estimate of errors using simulations – for more details see EDGES memo 99

Rogers & Bowman (EDGES memo #99)

BIGHORNS Sokolowski, Tremblay, Wayth,

Tingay ⑫– Low rfi site, high stability

Dynamic spectrum normalised by the median

Dynamic range 2%– Required 10-4

Day

200MHz

CORE2: A global EOR experiment with a self-calibrating receiver on two antennas

CORE2: A global EOR experiment with a self-calibrating receiver and two antennas| Keith Bannister | Page 26

CORE2-MONO60 Degree beamOptimised for frequency Independence and low RFI and

CORE2-DISH5 Degree beamOptimised for foreground removal

27

The Richness and Beauty of the Physics of Cosmological Recombination

well defined quasi-periodic spectral dependence

photons are coming from redshifts– z 1300−1400– i.e. before the time of

the formation of the CMB angular fluctuations

Chluba & Sunyaev A&A, 458, L29 (2006)

28

Observing

All sky so dish size is not relevant Needs a wideband spectrograph in 2-10 GHz range Can measure multiple independent patches of sky

– Many dishes/receivers Need lowest possible Tsys Can integrate over all oscillations

– Spectral dependence is accurately predicted

29

Sensitivity Required

Need ΔT/T = 10-8

Tsys = 25K Δν = 1010 Hz 2 pol 100 antennas Time = 1month (3.106 sec) ΔT/T = 25/(√(1010 . 2.100.3.106 .)) = 10-8 !