atmospheric phase correction at the Plateau de Bure interferometer

1
atmospheric phase correction at the Plateau de
Bure interferometer

• IRAM interferometry school 2006
• Aris Karastergiou

2
No atmosphere

• Interferometers can image astronomical sources by
  measuring delays in arrival times of incoming
  wavefronts between pairs of antennas

3
Effects of the atmosphere

• Variations in the refractive index of the
  atmosphere above each antenna distort the
  wavefront and result in delays different to those
  predicted.

4
Water vapour

• At radio frequencies, the variations of the
  refractive index are almost entirely due to
  parcels of water vapour, which are poorly mixed
  with the atmosphere.
• Luckily, water vapour not only causes delays in
  propagation, it also emits radiation. Both
  effects are proportional to the water vapour
  present along the line of sight.

5

• therefore
• by monitoring the emission of water vapour in
  the line of sight of each antenna, we could
  compute the extra delay induced by the
  atmosphere, and correct the phase

Problem solved!
6
But how do we do it?
7
The atmosphere
Atmospheric transmission in the zenith direction
from the Plateau de Bure site. The six curves
correspond to different amounts of water vapour,
starting from 0 (light blue) to 10 mm (black).
Two blue arrows denote two water lines used for
atmospheric phase correction, at 22 and 183 GHz.
8
IRAM against atmospheric phase noise

• Until August 2004
• The amount of water was derived from the total
  power measurements of the astronomical receivers
  in the 1mm band. Good overall performance but
  some significant drawbacks. Clouds (high
  temperature, small path length effect) are not
  accounted for.
• After August 2004
• Dedicated receivers placed in each of the six
  antennas, monitor the 22 GHz emission line. New
  system is separate from the astronomical
  receivers.

9
The receivers
Uncooled, broadband (7 GHz). Extremely
stable Gain fluctuations of order 10-4
10

• Filters extract three 1 GHz bands from the total
  bandwidth.

Cloud contribution to Tsky has different
frequency dependence ( v2) to water vapour, so
three channel system allows to remove this
baseline, leaving only the water vapour
emission.
11

• Effect of clouds can be removed
• Tsky,H2O Tvapor Tcloud
• TAtm (1 - e-tv) TCloud (1 - e-tc) , tC n2
• linearize cloud exponential term, measure at two
  frequencies, build weighted mean
• DTdouble Tsky ,1 Tsky,2 (n1/n2)2 Tvapor,1
  Tvapor,2 (n1/n2)2
• Same for the other frequency pair, and then a
  subtraction of one double value from the other,
  to form DTtriple

12
An unexpected advantage of the triple channel
system

• A geostationary satellite, HOTBIRD 6 is emitting
  inside the first 22 GHz channel. Channels 2 and 3
  can still provide the information for correcting
  the phase.

13
The data
Channel 1
Channel 2
Channel 3
combination removes continuum baseline
14
The data
Antenna 1 Antenna 2
15
The data
Between 2 antennas, the difference in the
measured water vapour emission is clearly
correlated with the astronomical phase.
16
Converting water vapour emission to phase

• Atmospheric model
• Based on the total amount of water in the line of
  sight, parameters such as the ambient temperature
  and pressure, and a general description of the
  atmosphere at that moment.
• Empirical approach
• From observations of strong sources, where the
  phase is well determined, a conversion
  coefficient can be calculated. Not constant with
  time.

17
(No Transcript)
18
22 GHz calibration

• Each channel of each radiometer needs to be
  calibrated separately, although triple
  combination is useful in removing various harmful
  effects.
• The process involves observing a table at
  ambient temperature, to estimate the conversion
  factor between counts and Kelvin.

• Ambient temperature 300 K is far from sky
  temperature 25 K, so conversion factor is assumed
  linear over this large range.
• For absolute phase correction, the calibration
  must be extremely precise.

19
Relative Vs Absolute

• When moving between sources, the radiometers are
  pointing through different parts of the
  atmosphere. The phase may change by many full
  cycles! To track it, calibration of the 22 GHz
  receivers needs to be excellent!

20

• Relative phase correction
• Aim is to remove phase fluctuations within each
  on-source period.
• Reduced phase rms means higher amplitude of the
  visibilities.
• Required for longer baselines / higher
  frequencies.
• Phase is not tracked when the telescope is moving
  from one source to another. Less sensitive to
  calibration problems.

21

• Phase noise is larger at longer baselines. Power
  law dependence on baseline length (see talks by
  R. Lucas and J. M. Winters from yesterday).

Corrected phase does NOT depend on baseline
length.
22
(No Transcript)
23

• Absolute calibration
• Phase is constantly tracked, even when changing
  source.
• Calibration requirements much more stringent.
• Astronomical phase of point-like strong sources
  should be used in the 22 GHz calibration process.
• Has not been achieved yet.

24
Points to remember

• Water vapour in the atmosphere results in noisy
  phases. Effect is worst for longer baselines,
  higher frequencies.
• To attack the problem, the emission of a water
  line is monitored, in the line of sight of each
  antenna of the interferometer.
• 22 GHz line is better for conditions with high
  (gt4mm) water vapour. 183 GHz is more useful for
  dryer sites.

25
Points to remember

• The water vapour radiometers can be used for
  relative phase correction, which means the phase
  cannot be maintained after a source change.
• Calibration of the 22 GHz receivers is crucial,
  given the required precision.
• Conversion of calibrated signal into phase can be
  done either by atmospheric model, or empirically.

26

• The relative phase correction works well in
  reducing the phase noise and increasing the
  visibility amplitudes.
• No working absolute phase correction system
  exists today.
• Plans for arrays with extremely long baselines,
  observing at extremely high frequencies are well
  under way.

Problem should be solved soon...