Transcript Document

Cosmology with the GMRT
Jayaram N Chengalur
NCRA/TIFR
Outline
• Brief introduction to the Giant Metrewave
Radio Telescope (GMRT)
• Constraining the variation in Fundamental
Constants
• Near field cosmology from Dwarf Galaxy
observations
The Giant Metrewave Radio
Telescope
•
The Giant Metre-wave Radio Telescope
(GMRT) is a large aperture synthesis
radio telescope optimized for operation
at low frequencies
– Wavelengths of 21cm and longer
•
Designed and built (near Pune)
primarily by NCRA, a national centre of
TIFR.
•
Array telescope consisting of 30
antennas, each 45m across
– Novel ‘SMART’ antenna design
– The most sensitive synthesis radio
telescope in the world at most of its
frequencies of operation,
GMRT Antenna Layout
25
Unique hybrid configuration
with mix of long and short
baselines – allows
simultaneous imaging of
extended as well as compact
emission
Low and high resolution
images of CH3CHO emission
from SgrB2 made from a
single GMRT observation
1km
Using the GMRT
• Time is allocated to
proposals by an
independent time
allocation committee
• Two calls per proposals
per year
• At present time allocation
is roughly evenly split
between Indian and
Foreign PI proposals
Fundamental
Constants
Nissim Kanekar
Tapasi Ghosh
Introduction
• Low energy fundamental constants (α≡e2/ħc, μ≡mp/me)
are expected to show spatio-temporal evolution
[Uzan, 2003, Rev. of Modern Physics]
• Timescales of these changes are poorly constrained
– Need to search for changes over as wide a range of timescales
as possible
• Terrestrial methods place extremely tight limits, but over
very short timescales
– e.g.  /  < 10-17/yr (over 1 year)
[Rosenband et al. Science 319, 1808, (2008)]
• Astrophysical techniques offer less precision, but probe much
longer timescales
Astrophysical methods
•
Precise spectral line frequency
depends on the values of various
fundamental constants
•
Comparing the line frequency in a
distant source to that observed on
earth will allow one to measure
variations in the values of the
fundamental constants
•
The redshift of the distant source is
unknown a priori
– one needs at least two lines (with
different dependence on the
fundamental constants) to measure
any possible change.
– Narrow absorption lines from cold gas
are best suited for precise frequency
(redshift) measurements.
Optical Spectral Lines
• The fractional separation between the alkali doublet (e.g. Si IV, MgII)
lines Δλ/λ~ α2
– Current limits Δα/α < 1-2 x 10-5 (2 < z < 3 )
[Murphy et al. MNRAS, 327, 1237, 2001; Chand et al. A&A, 430, 47, 2005]
• (MM) Relativistic first order corrections lead to different fine structure
transitions in different species having different dependencies on α
– Average over a large number of transitions to reduce statistical errors
– Current limits Δα/α ~ few x 10-6, but with conflict between groups
[Murphy et al. MNRAS 345,609,2003;Chand et al. A&A,417, 853, 2004;
Levshakov et al. A&A, 466. 1077, 2007]
• The MM method gives lower statistical errors, but larger systematic
ones, e.g.
– calibration errors on different echelle orders,
– kinematical velocity shifts between species,
– Isotopic abundance variations
•
Comparisons of the HI 21cm hyperfine line frequency
with fine structure (e.g. MgII) line frequency constrains
X ≡ gpμα2
–
–
•
HI 21cm absorption studies were one of the first major
projects at the GMRT (Kanekar, PhD thesis)
–
–
But early, commissioning phase observations had small,
systematic errors because of imprecise correction for the
earth’s motion
Preliminary results from fresh, high precision, have
substantially improved precision
are in the process of obtaining more precise optical
redshifts
Large number of absorbers need to be averaged
–
–
Because of the possibility of kinematical shifts between
the HI 21cm absorbing gas and the MgII absorbing gas.
HI 21cm absorption comes from cold (~ 80K) gas, while
MgII absorbing gas is often warm (~ 5000-10,000K)
Kanekar, Chengalur & Lane 2007
–
•
Radio spectral line frequencies can be easily measured to
high precision.
Current published constraints ΔX/X < 2 x 10-5 (0.23 <z <
2.35)
Kanekar et al. 2008
HI 21cm vs fine structure lines
OH 18cm radio lines
Selection rule
ΔF=0,±1
•
The OH molecule has 4 spectral lines with wavelengths ~ 18cm
– These lines arise from a combination of Λ doubling and hyperfine
interaction
•
Observations of redshifted OH 18cm line absorption from GMRT was also
part of Nissim Kanekar’s PhD thesis
– Some of the earliest observations of OH at cosmological redshifts.
Constraints from OH
measurements
•
Comparison of the redshifts of the OH
main (ΔF=0) lines with redshifts of the HI
21cm (hyperfine) transition and CO mm
(rotational) transitions allows one to
simultaneously constrain Δα/α and Δμ/μ
and Δgp/gp
•
Constraints are relatively weak unless
one assumes Δgp/gp is small
•
Results are subject to possibility of
kinematical shifts between HI, OH and
CO absorbing gas.
•
If one assumes that Δgp/gp is small (e.g.
Langacker et al. 2002) then one gets
– Δα/α = -5 ± 1.5 x10-6
– Δμ/μ = -7.8 ± 2.4 x 10-6
Chengalur & Kanekar PRL 91, 241302 (2003)
Conjugate OH lines
OH 18cm lines from
Centaurus A (z ~ 0)
[van Langevelde et al. (1995)]
•
The OH ‘satellite’ (ΔF=±1) lines are often “conjugate” i.e. have same
spectral shape,but opposite signs
– Consequence of selection rule driven “competitive pumping”
•
Since line shape is the same, non parametric, cross correlation techniques
can be used to determine spectral shifts
– High level of independence from systematic effects (kinematic doppler shifts,
isotopic variations, calibration errors…)
– Can be applied to a single object
– Cross correlation of Centaurs A (z~0) lines gives ΔV = 0.05 ± 0.11 km/s
Conjugate OH lines at
cosmological distances
• First detection of conjugate
lines at cosmological
distances was for
PKS1413+135, (z= 0.247)
• Data constrain G≡gp[α2μ]1.849
• Original data leads to
ΔG/G = 2.2 ± 3.8 x 10-5
Kanekar, Chengalur & Ghosh PRL 93,
051302, (2004)
Current constraints from
PKS1413+134
• Limits from the new high sensitivity data are
ΔG/G < 1.4 x 10-6
– Δα/α = 3.1 x 10 -6 (2σ, if Δμ/μ is constant)
– Δμ/μ = 3.1 x 10 -6 (2σ, if Δα/α is constant)
• There have been theoretical suggestions that
changes in Δμ/μ are correlated with changes in
Δα/α with Δμ/μ ~ 50 Δα/α
[Calmet & Frisch Eur. Phy. J. C, 24, 639, 2002]
• In such a model, our data constrains Δα/α < ~
10-7
– One of the most sensitive existing limits
Comparison of OH based and
optical spectra based constraints
• OH constraints are offer similar precision, but:
– Apply to a single object (optical results are averages over large redshift
range)
– Not subject to the same systematics
– Currently probe a complementary redshift range
Near field Cosmology
Dwarf Galaxies as Cosmological Probes
Ayesha Begum
Sambit Roychowdhury
I. D. Karachentsev
S. Kaisin
M. Sharina
Cosmic Evolution: The quick tour
•
The universe starts in a hot big bang
and expands and cools steadily.
•
Inflation makes the density distribution
very (but not perfectly) smooth
•
Perturbations observed to be ~ 10-5 at
the epoch when protons and electrons
combine
•
Small perturbations collapse to form
the first stars and blackholes
• Energy release from these
objects reionizes the universe
•
Perturbations continue to grow to form
galaxies and clusters of galaxies
Hierarchical Galaxy Formation
The smallest
objects collapse
first, bigger objects
form by the merger
of smaller ones
Kauffman & White 1993
The growth of galaxies by mergers is driven by the gravity of
the non baryonic dark matter – the baryonic matter (stars,
gas) occupy a small region in the center of a much larger dark
matter “halo”
Near field cosmology from Dwarf
Galaxies
The process of galaxy merger is highly
inefficient
 Every large galaxy should be
surrounded by dozens of left over
“dwarf galaxies” which are remnants of
the primordial galaxy population
 As the earliest formed systems, with
relatively simple internal structure,
properties of dwarfs are sensitive to
cosmology.
3D map of the local group:
Two large galaxies (Milkyway,Andromeda)
surrounded by several small dwarf galaxies
Numerical
simulation: Each
large Dark Matter halo
is surrounded by
several, as yet
unmerged, smaller
halos.
Dwarf Galaxies as cosmological
probes
 Since dark matter is typically dominant even in
the central regions, the dark matter density
distribution in dwarfs should reflect that
predicted by numerical simulations
 Details of ‘baryon physics’, e.g.
 the mass to light ratio of the stellar population,
 feedback from baryonic cooling and collapse on the
structure of the Dark Matter Halo
make it difficult to accurately determine the dark
matter density profile in big galaxies.
 Baryons are easily lost from the shallow dark matter
potential wells of small galaxies
 Reheating during the epoch of reionization, as well as
from feedback from star formation should lead to dwarf
galaxies having baryon fractions smaller than the cosmic
mean.
The Faint Irregular Galaxy GMRT
Survey: FIGGS
A survey of the neutral hydrogen (HI)
emission in a large, systematically
selected, sample of dwarf galaxies.
Faintest sample galaxies are ~ 104 times less
luminous than the Milkyway
HI 21cm observations are preferred
because:
Accurately trace the dark matter potential
because the gas is “cold” compared to the stars
Dark matter potential can be traced to large
galacto-centric distances because the gas disk is
extended compared to the stellar disk
Doppler shifts can be easily measured to high
accuracy.
Accurate distances are known for a large
fraction of the sample
complimentary multi-wavelength data is also
available with our collaborators or in the public
domain.
By far the largest such study of
dwarf galaxies, possible due to
high sensitivity of the GMRT
Dark matter in faint dwarf galaxies
DDO 210 (MB -10.6 mag)
Need to observe the circular velocity in order to reconstruct the underlying density
distribution
Earlier, less sensitive, observations indicated that gas in the faintest dwarf galaxies
has chaotic velocity fields
Fresh, high sensitivity GMRT observations established that even the faintest dwarfs
have well defined coherent large scale velocity fields
DV ~ 6.5 km/s (VLA)
Lo et al. 1993 AJ, 106, 507
DV ~ 1.6 km/s (GMRT)
Begum & Chengalur 2004 A&A,
413, 525
Dark matter density profiles
Traditionally used (phenomenological) dark halo models have constant
density cores (‘psuedo isothermal’ halos)
ρ(r)=ρ0/[1+(r/rc)2]
Numerical simulations of hierarchical CDM models predict cusped
density core (“NFW”) dark matter halos
ρNFW (r)=ρi / [(r/rs)(1+r/rs)2]
(Navarro et al. 1997 ApJ 490 493)
From measurements of the circular velocity as a function of galactocentric radius (“rotation curve”) one can reconstruct the underlying
mass distribution
Rotation curves of FIGGS galaxies can be used to check if dark
matter density distribution matches numerical predictions
Velocity (km/s)
GMRT Observations of Camelopardalis B
(MB ~ -12.3)
Galactocentric distance (arcsec)
a
a
One of the faintest galaxies with a well measured rotation curve
Begum, Chengalur & Hopp New Ast, 2003, 8, 267
Dark matter in Camelopardalis B
Begum et al. New Ast, 2003, 8, 267
Halos with constant density cores
provide a good fit, but cuspy halos
do not
» “NFW” halos in general do
not provide a good fit to our
sample galaxies.
Rotation curve derived at a
range of spatial resolutions
» results are not a consequence
of limited angular resolution
Tension between the predictions of CDM heirarchichal galaxy formation
numerical simulations and observations is probably indicative of baryonic
processes (e.g. cooling and collapse) shaping the centers of the dark
matter halos even in dwarf galaxies. Alternatively it has been taken as
evidence for WDM
Serendipitous discoveries of extremely gas rich galaxies
(The baryon fraction in the faintest dwarfs)
NGC 3741: A dwarf galaxy with a
giant HI disk
Rotation curve measured to a record 38 optical disk scalelengths
Mass/Luminosity ~ 107 – one of the “darkest” galaxies
known.
HI in Andromeda IV
HI disk extends out to more than 6 Holmberg radii
Mass/Luminosity ~ 237 !
Do “dark” galaxies also have anomalously low baryon
fractions?
Baryon fraction in dwarf galaxies
• Small halos are less
efficient at capturing
baryons
– hot baryons escape
during the epoch of
reionization
– Feed back from star
formation drives baryons
out of shallow dwarf
galaxy potential wells
• Baryon fraction
expected to vary
inversely with galaxy
mass
Gnedin ApJ 542, 535, (2000)
Baryon fraction: Theory vs
Observation
• Since baryons cool and collect
at the center of the halo, the
baryon fraction increases with
decreasing radius
• Simulations give baryon fraction
as measured at the virial radius
• Observations determine the
baryon fraction up to the last
measured point of the rotation
curve
• Simulations suggest that the
baryon fraction within the last
measured point of the rotation
curve should vary inversely with
halo mass
Baryon fraction in gas rich galaxies
Large scatter in baryon fraction for all
galaxies
Dwarf galaxies don’t have systematically
smaller baryon fractions
AndIV and N3741 are not particularly
baryon deficient – but for some reason
they have been unable to convert gas into
stars
(See Roychowdhury et al 08 for star formation
recipes in dwarfs)
Baryon fraction in galaxies with well
measured HI rotation curves
Discrepancy between predicted and observed baryon fractions
is probably again indicative of our lack of understanding of the
detailed processes involved in baryon capture and cooling, star
formation etc.
Summary
•
Radio spectral lines from redshifted absorbers provide very competitive
constraints on the cosmic variation of fundamental constants
– Can be applied to a single object
– Are not subject to the same systematics as optical lines
– Probe a complementary redshift range
•
Detailed observations of nearby, extremely faint dwarf galaxies allow one to
do “near field” cosmology
– Dark matter distribution in these galaxies does not conform to predictions of
CDM numerical simulations
– Baryon content also does not decrease with halo mass as expected
•
While these discrepancies could be interpreted as being problems related to
the CDM model, it is more likely that they are a consequence of our poor
understanding of baryonic processes
– Cooling and collapse of gas into stars, feed back from star formation etc.
Thank you