Detecting HII bubbles in 21 cm HI maps

The reionization of the Universe, it is believed, occurred by the growth of ionized regions (bubbles) in the neutral intergalactic medium (IGM). We are studying  the possibility of detecting these bubbles in radio-interferometric observations of redshifted neutral hydrogen (HI) 21 cm radiation.

The signal (<1 mJy) will be buried in noise and foregrounds, the latter being at least a few orders of magnitude stronger than the signal. We have developed  a visibility based formalism that uses a filter to optimally combine the entire signal from a bubble while minimizing the noise and foreground contributions. This formalism makes definite predictions on the ability to detect an ionized bubble or conclusively rule out its presence in a radio-interferometric observation.

We make predictions for the currently functioning GMRT and a forthcoming instrument, the MWA at a frequency of 150 MHz (corresponding to a redshift of 8.5). For both instruments, we show that it will be possible to detect a bubble of comoving radius R_b > 40 Mpc (assuming them to be spherical) in 100 hrs of observation and R_b > 22 Mpc in 1000 hrs of observation, provided the bubble is at the center of the field of view. In both these cases the filter effectively removes the expected foreground contribution so that it is below the signal, and the system noise is the deciding criteria.

We find that there is a fundamental limitation on the smallest bubble that can be detected arising from the statistical fluctuations in the HI distribution. Assuming that the HI traces the dark matter we find that it will not be possible to detect bubbles with R_b < 8 Mpc using the GMRT and R_b < 16 Mpc using the MWA, however large be the integration time


Detecting ionized bubbles in redshifted 21 cm maps,  Datta, Kanan K.; Bharadwaj, Somnath; Choudhury, T. Roy
Mon. Not. R. Astron. Soc. 382 809 (2007)

Simulating the impact of HI fluctuations on matched filter search for ionized bubbles in redshifted 21-cm maps,  Datta, Kanan K.; Majumdar  Suman; Bharadwaj, Somnath; Choudhury, T. Roy
Mon. Not. R. Astron. Soc. 391 1900 (2008)

The impact of anisotropy from finite light traveltime on detecting ionized bubbles in redshifted 21-cm maps ,  Majumdar  Suman; Bharadwaj, Somnath; Datta, Kanan K.; Choudhury, T. Roy
Mon. Not. R. Astron. Soc. 413 1409 (2011)

Constraing quasar parameters through bubble detection in redshifted 21-cm maps,   Majumdar  Suman; Bharadwaj, Somnath; Choudhury, T. Roy
To be submitted in MNRAS


Ongoing and Future Research

1. Ooptimal strategy, blind search or targeted search around known QSOs?  Optimal z?

2. Can we combine other observations like S-Z effect to increase the possibility of a detection?

3. Simulations: 

A. We are currently using dark matter N-body simulations with the assumption that the HI outside  the bubble traces the dark matter. A bubble is put in by hand.  The simulations have 256^3 particles on a 256^3 grid and grid spacing of 2 Mpc. The simulations are evolved to a redshift z=6.

HII bubble
This shows a single channel image of 
256 Mpc x 256 Mpc area with a
bubble 20 Mpc radius  at center  

2 Mpc corresponds to 0.9' on the sky and 150 kHz in frequency

There are three major challenges for detecting HII bubbles.

 a. Foregrounds. Current models propose these have a very smooth frequency dependence. If so, our earlier work shows that foregrounds do not pose a big problem for bubble detection.

b. HI fluctuations.  HI fluctuations outside the bubble we wish to detect will create confusion. This is a big problem.

c. System Noise This decides the observation time needed for a detection.

We are currently focusing on the two latter effects only.

How does noise affect the image shown above? Can we still make out the bubble?

How do the HI fluctuations affect the visibilities?

Simulate  HI fluctuation contribution to estimator, and compare with earlier predictions.