CODEX

1
CODEX Opening a new parameter Space
2
Why a new window now

• High wavelength calibration accuracy
• Very accurate Doppler shift measurements
• Measurement of isotopic ratios
• High Precision Spectroscopy requires
• very high S/N ratios
• ? photon starved
• ? Very and Extremely Large Telescopes

3
ESO is operating the VLT

• Four 8 Meter telescopes
• 11 Instruments at Cass or Nasmyth Foci
• Combined coherent focusVLTI
• Combined Incoherent Focus (16m equivalent)

4
And ESO is planning the E-ELT!
42 Meter adaptive Telescope First light 2015
5
CODEX _at_ ELT

• ESO
• G. Avila, B. Delabre, H. Dekker, S. DOdorico,
• J. Liske, L. Pasquini, P. Shaver
• Observatoire Geneve
• M. Dessauges-Zavadsky, M. Fleury, C. Lovis, M.
  Mayor, F. Pepe,
• D. Queloz, S. Udry
• INAF-Trieste
• P. Bonifacio, S. Cristiani, V. DOdorico, P.
  Molaro, M. Nonino, E. Vanzella
• Institute of Astronomy Cambridge
• M. Haehnelt, M. Murphy, M. Viel
• Instituto de Astrofisica Canarias
• R. Garcia-Lopez, R. Rebolo, M.
  Zapatero
• Others
• Bouchy (Marseille), S. Borgani (Daut-Ts), A.
  Grazian (Roma), S. Levshakov
• (St-Petersburg), L. Moscardini (OABo-INAF), S.
  Zucker (Tel Aviv),
• T. Wilklind (ESA)

6
- Direct measurement of the dynamics of the
Universe - Cosmological variation of the
Fine-Structure Constant CODEX will aim at an
accuracy of ??/? 5x 10-9

• Four Outstanding Cases

- Terrestrial planets in extra-solar systems
RV of earth mass planets, spectroscopy of
transits
- Primordial nucleosynthesis probing SBB
nucleosynthesis primordial Li7, Li6/Li7
Many additional applications (as from this
conference....) Asteroseismology,Abundances in
Stars, Variability of µ, Temperature evolution
of CMB, Chemical evolution of IGM..
7
Dynamics of the Universe
Directly measure the expansion of the Universe by
observing the change of redshift with a time
interval of a few (10-20) years It should be
possible to choose between various models of the
expanding universe if the deceleration of a given
galaxy could be measured. Precise predictions of
the expected change in zdl/l0 for reasonable
observing times (say 100 years) is exceedingly
small. Nevertheless, the predictions are
interesting, since they form part of the
available theory for the evolution of the
universe Sandage 1962 ApJ 136,319
8
Cosmic Signal
t0actual epoch
In a homogeneous, isotropic Universe a FRW metric
teemission epoch
9
Direct measurement !

• Bias-free determination of cosmological
  parameters
• Different redshift (CMB, SNIa) measure dynamics
  and not geometry
• Not dependent from evolutionary effects of
  sources
• Legacy Mission to future astronomers
• (First epoch measurements)

10
High Redshift Supernovae..
SNae observe magnitudes at different Redshifts
Assume SNae are standard candles at all
z Require non-trivial K corrections
Require determination of Reddening
11
The measurement concept ..
But this is for 107 years Having much less
time at our disposal the shift is much smaller.
12
The expected acceleration Is SMALL! Note the
standard model was not always standard..
The change in sign is the signature of the non
zero cosmological constant
13
Results of Simulation (..)
Simulating the dependence on Z
Information saturates a) Too many
lines b) High Redshift makes them broader.
14
Results of simulations ()
Many simulations have been carried out
independently by 3 groups using observed and
fully simulated spectra. A very good agreement is
found, and we can produce a simple scaling law
?v 1.4(2350/(S/N))(30/NQSO)0.5 (5/(1Z))1.8
cm/sec Where ?v is the total uncertainty
(difference between 2 epochs) while the other
parameters refers to the characteristics of one
epoch observation Pixel size considered 0.0125.
About half of the signal is coming from the metal
lines associated to the Ly?
15
Result of simulations ()
The full experiment
NQSO 30 randomly distributed in the range 2 lt
zQSO lt4.5 S/N 3000 per 0.0125 Å pixel/epoch (no
metal lines used) ?t 20 years
Green points 0.1 z bins Blue 0.5 z bins Red
line Model with H070 Km/s/Mpc ?m0.3
??0.7 The cosmic signal is Detected at gt99
significance(!)
16
Scaling with telescope diameter
With a telescope in the range of the 30-40m the
experiment will be possible
20 yrs baseline

1. With a timeline of 20 yrs the accuracy scales to
   2 cm/sec
2. Use the full spectral range, including the Ly?
   region

VLT and Keck have devoted already gt 1000 h each
to QSOs..
17
CODEX design parameters
Location Underground in
nested stabilized environment Telescope diameter
42 m Feed Coude Fibre or Fibre
only Overall DQE 20 Coude
Fibre , 13 Fibre only Entrance aperture
1 arcsec Wavelength range 380 680
nm Spectral Resolution 150 000 Number of
Spectrographs 3 (for 1 arcsec aperture at 42
m) Main disperser 3 x R4 echelle 42
l/mm 160 x 20 cm Crossdisperser 6 x VPHG 1500
l/mm 20 x 10 cm Camera 6 x F/1.4-2.8 CCD 6 x
8K x 8K (15 um pixels)
18
How to improve accuracy and stability

• Scramblers to reduce effect of guiding errors
• Image dissector, multiple instruments
• Simultaneous wavelength calibration
• Use of wavelength calibration based on laser
  comb
• Fully passive instrument, ultra-high temperature
  stability
• Instrument in vacuum tank
• High precision control of detector temperature
• Underground facility, zero human access
• Blaze Correction and Flat Fielding
• System PRECURSOR _at_ VLT

19
CODEX laboratory floor plan
Underground hall 20 x 30 m height 8 m 1 K
Instrument room 10 x 20 m height 5 m 0.1 K
Instrument tanks, dia. 2.5 x 4 m, 0.01 K
Optical bench and detector 0.001 K
Control room and aux. equipment (laser) 1 K
20
Telescope Link Transmission
The case was studied for the combined focus of
the 4 VLT telescopes Coude train is the most
efficient solution and the only one which
preserves the BLUE
21
CODEX Unit Spectrograph
B. Delabre ESO
Light from fibres enters here
22
Calibration System Laser Frequency Comb

• Metrology labs recently revolutionized by
  introduction
• of femtosecond-pulsed, self-referenced lasers
  driven by
• atomic clock standards
• Cesium atomic clock
  n 2n
• (or even GPS signal!)
• Result is a reproducible, stable comb of evenly
  spaced lines whos frequencies are known a priori
  to better than 1 in 1015

23
Comb spectra simulations R100k, Dn15GHz,
l5000Å
Total velocity precision better than 1cms-1 in a
single shot for SNR 1000 (4500-6500Å)
Running study with MPQ (Nobel Prize 2005)
24
Precision Spectroscopy _at_ ELT
Extremely Large Telescopes will provide for the
first time enough photons to guarantee a quantum
jump in high resolution spectroscopy
To open these new windows, High Resolution must
be coupled to extremely high accuracy and
stability
We have developed a design for such an
instrument, and we will propose ESPRESSO, the
CODEX precursor _at_ the VLT
25
Laser Comb Advantages and Challenges

• Advantages
• Absolute calibration
• Long term frequency stability
• Evenly spaced and highly precise frequencies
  allow
• mapping of distortions, drifts and
  intra-pixel sensitivity
• variations of CCD
• Naturally fibre feed system. HARPS prototype
  possible.
• Challenges
• Line-spacing currently limited to 1GHz. We need
  1015 GHz...
• Large spectral range required (380-680 nm)
  Existing technology can possibly be extended.

26
Peculiar motions at the Earth
Parameter Induced error on the correction cm s-1 Comment
Earth orbital velocity - Solar system ephemerides lt 0.1 JPL DE405
Earth rotation - Geoid shape - Observatory coordinates - Observatory altitude - Precession/nutation corrections 0.5 lt 0.1 lt 0.1 lt 0.1 Any location in atm. along photon path may be chosen
Target coordinates - RA and DEC - Proper motion - Parallax ? 0 0 70 mas ? 1 cm s-1 negligible negligible
Relativistic corrections - Local gravitational potential lt 0.1
Timing - Flux-weighted date of observation ? 0.6 s ? 1 cm s-1
The solar acceleration in the Galaxy will be
measured with an accuracy of 0.5 mm/sec/yr by
GAIA
27
Why measure dynamics?
All the results so far obtained in the
concordance model assumes that GR in the FRW
formulation is the correct theory. ??0.7 but
we do not know what ?? is and how it evolves.
If GR holds, geometry and dynamics are related,
matter and energy content of the Universe
determine both. Dynamics has, however, never
been measured. All other experiments, extremely
successful such as High Z SNae search and WMAP
measure geometry dimming of magnitudes and
scattering at the recombination surface.
28
Challenges Feedback

• Calibration source, stable, reproducible, equally
  spaced..
• LASER COMB
• CCD control (thermal )
• High throughput of the spectrograph and
  injection system
• EFFICIENT COUDE FOCUS
• Light scrambling capabilities (1 cm/sec ?
  0.0000003
• arcsec centering accuracy on a slit)
• TELESCOPE POINTING AND CENTERING 0.02
  arcsec
• System aspects (from pointing to calibration to
  data reduction)
• ALL aspects strongly suggest extensive prototyping

29
BASIC SPECTROGRAPH REQUIREMENTS
Once QSOs are established targets, through basic
knowledge and simulations we determine the
spectrograph parameters. Spectral range QSO in
the range Z1.5 - 5. At higher Z too many
absorbers wipe out information, Ly? is visible
from earth at Zgt1.8 For lower Z, metal lines
only can be used. But to span a large Z range is
important also to link it to lower Z
experiments ideal range 300-680 nm. UV
trade-off (less sensitivity, few Ly? absorbers..)
Resolving Power Ly? line have a typical width
of 20-30 Km/sec and R50000 would suffice. Higher
spectral resolution is required by metallic lines
(blt3 Km/sec) and by calibration accuracy
requirement R150000 Such a resolution is
challenging for any seeing limited ELT the final
number will be a trade off between size, sky
aperture, detector noise
30
CODEX Image and Pupil Evolution
Several new features
Anamorphic collimator
Pupil Slicer
Anamorphic Crossidsperser (Slanted VPH)
31
Standard Cosmological Model
With the assumptions of homogeneity and
isotropy, the concordance model finds a FRW
metric with a non zero cosmological constant
32
How to Measure this signal?
Masers in principle very good candidates lines
are very narrow and measurements accurate
however they sit at the center of huge potential
wells large peculiar motions, larger than the
Cosmic Signal are expected Radio Galaxies with
ALMA The CODEX aim has been independently
studied for ALMA as for Masers, local motions
of the emitters are real killers. Few radio
galaxies so far observed show variability at
a level much higher than the signal we should
detect Ly? forest Absorption from the many
intervening lines in front of high Z QSOs are the
most promising candidates. Simulations,
observations and analysis all concur in
indicating that Ly? forest and associated metal
lines are produced by systems sitting in a warm
IGM following beautifully the Hubble flow !
33
Results of simulations (1)
Simulating the dependence on resolution ( Ly?
forest only)
Above a R50000 there is no more gain for the
Ly? Forest. Higher Resolution is required by
metal lines and Calibration accuracy.
34
Results of simulation (2) real spectrum
Dependence on cumulative S/N/pixel (0.015 A)
35
Can we do it ?
Telescope Instrument Efficiency required to
complete the experiment under the following
assumptions V16.5 QSO 36 QSO each S/N 2000
(0.015 pixel), for a total of 2000 obs.
hs/epoch ( 1cm/sec/yr) Red line VLTUVES peak
efficiency In 1st approx. precision scales as D
for a given efficiency.
36
PECULIAR MOTIONS
Ly? ?vacc100 Km/sec Tacc109 yr..
negligible Metal systems associated to damped
Ly??vacc3-400 Km/sec Tacc108 BUT hundreds
systems-statistically level out Different from
the maser and radio-galaxy case Calls for many
line of sight
37
PECULIAR MOTIONS Simulations
Detailed, state of the art hydrodynamical
simulations confirm that the effects are
negligible
38
Whats new

• VLT-UVES Keck HIRES observed hundreds of QSOs
  at High Res (Rgt40000), z between 2 and 5,
  V16-18. Ly? clouds have been extensively
  simulated their hot gas belongs to the IGM and
  they trace the Hubble flow
• Exoplanets (HARPS) long term accuracy 1m/s, short
  term (hours) 0.1m/s (and largely understood)
• ELT !! LOT OF PHOTONS (we need them!!)

39
Verification of targets and reference mission
QSO have been selected from existing catalogues
and compilations (Veron, SDSS ) Selection
criteria magnitude and z Magnitudes redwards of
Ly?, selected band depends on z
In this Figure only the 5 brightest QSOs of each
0.25 z bin are shown. Hypothesis e.g. 2000 h
observations with an 80 m telescope and 14
efficiency all QSO brighter or around the
iso-accuracy lines are suitable.
40
Fiber feed Pre-slit
Entrance aperture 1 on 60 m or 0.65 on 100 m 37
Fibres array, 8 fibres/spectrograph
OWL
Lightpipe forming an homogeneously illuminated
Slit scrambling
41
CODEX _at_ ELT
More info discussion at the Aveiro
Conference on Precision Spectroscopy
in Astrophysics 11-15 September 2006,
Aveiro, Portugal http//www.oal.ul.pt/psa2006
42
Results of simulations (.)
The full experiment
DT10 yrs 1500 Hours Metals V16.5 Eff. 15
43
QSO absorption lines
Quasar
To Earth
Lyaem
CIV
SiII
SiIV
CII
SiII
Lyman limit
Lya
Lyb
Lybem
NVem
Lya forest
CIVem
SiIVem