DARK ENERGY SURVEY (DES)

1
DARK ENERGY SURVEY (DES)
Francisco Javier Castander Serentill
IEEC/CSIC
All material borrowed from DES collaboration
2
Announcement of Opportunity Blanco
Instrumentation Partnership

• Develop a major instrument for Blanco 4m CTIO
• Submit a science, technical management plan
• Community instrument
• Up to 30 of Blanco 4m for 5 years commencing in
  2007 or 2008
• Letter of intent March 15, 2004
• Proposals August 15 2004

3
The Science Case for the Dark Energy Survey

• James Annis
• For the DES Collaboration

4
The Dark Energy Survey

• We propose to make precision measurements of Dark
  Energy
• Cluster counting, weak lensing and supernovae
• Independent measurements
• by mapping the cosmological density field to z1
• Measuring 300 million galaxies
• Spread over 5000 sq-degrees
• using new instrumentation of our own design.
• 500 Megapixel camera
• 2.1 degree field of view corrector
• Install on the existing CTIO 4m

5
Cosmology in 2004
Sloan Digital Sky Survey measures the galaxy
density field at z lt 0.3
WMAP measures the CMB radiation density field at
z1000

• Combine to measure parameters of cosmology to
  10.
• We enter the era of precision cosmology.
• Confirms dark energy (!)

2003 Science breakthrough of the year
6
The Big Problems Dark Energy and
Dark Matter

The confirmation of Dark Energy points to major
holes in our understanding of fundamental physics

• Dark energy?
• Who ordered that? (said Rabi about muons)
• Dark energy is the dominant constituent of the
  Universe
• Dark matter is next

95 of the Universe is in forms unknown to us
1998 Science breakthrough of the year
7
Dark Energy

• The Cosmological Constant Problem
• Particle physics theory currently provides no
  understanding of why the vacuum energy density is
  so small ?DE (Theory) /?DE (obs) 10120
• The Cosmic Coincidence Problem
• Theory provides no understanding of why the Dark
  Energy density is just now comparable to the
  matter density.
• What is it?
• Is dark energy the vacuum energy? a new,
  ultra-light particle? a breakdown of General
  Relativity on large scales? Evidence for extra
  dimensions?

The nature of the Dark Energy is one of the
outstanding unsolved problems of fundamental
physics. Progress requires more precise probes
of Dark Energy.
8
Measuring Dark Energy

• One measures dark energy through how it affects
  the universe expansion rate, H(z)
• H2(z) H20 ?M (1z) 3 ?R (1z) 4 ?DE
  (1z) 3 (1w)
• matter
  radiation dark energy
• Note the parameter w, which describes the
  evolution of the density of dark energy with
  redshift. A cosmological constant has w ?1.
• w is currently constrained to 20 by WMAP,
  SDSS, and supernovae
• Measurements are usually integrals over H(z)
  r(z) ? dz/H(z)
• Standard Candles (e.g., supernova) measure
  dL(z) (1z) r(z)
• Standard Rulers measure
  da(z) (1z)?1 r(z)
• Volume Markers measure
  dV/dzd? r2(z)/H(z)
• The rate of growth of structure is a more
  complicated function of H(z)

9
DES Dark Energy Measurements

• New Probes of Dark Energy
• Galaxy Cluster counting
• 20,000 clusters to z1 with M gt 2x1014 M?
• Weak lensing
• 300 million galaxies with shape measurements
• Spatial clustering of galaxies
• 300 million galaxies
• Standard Probes of Dark Energy
• Type 1a Supernovae distances
• 2000 supernovae

10
Supernova

• Type 1a Supernovae magnitudes and redshifts
  provide a direct means to probe dark energy
• Standard candles
• DES will make the next logical step in this
  program
• Image 40 sq-degree repeatedly
• 2000 supernovae at z lt 0.8
• Well measured light curves

Current projects
Essence
CFHLS
SCP
LSST
SNAP
SDSS
PanStarrs
Proposed projects
DES
11
New Probes of Dark Energy

• Rely on mapping the cosmological density field
• Up to the decoupling of the radiation, the
  evolution depends on the interactions of the
  matter and radiation fields - CMB physics
• After decoupling, the evolution depends only on
  the cosmology - large-scale structure in the
  linear regime.
• Eventually the evolution becomes non-linear and
  complex structures like galaxies and clusters
  form - non-linear structure formation.

z 0
z 30
12
Spatial Clustering of Galaxies

• The distribution of galaxy positions on the sky
  reflects the initial positions of the mass
• Maps of galaxy positions are broken up in
  photometric redshift bins
• The spatial power spectrum is computed and
  compared with the CMB fiducial power spectrum.
• The peak and the baryon oscillations provide
  standard rulers.
• DES will
• Image 5000 sq-degrees
• Photo-z accuracy of ?z lt 0.1 to z 1
• 300 million galaxies

PanStarrs
Cooray, Hu, Huterer, Joffre 2001
13
Weak Lensing

• Weak lensing is the statistical measurement of
  shear due to foreground masses
• A shear map is a map of the shapes of background
  galaxies

14
Weak Lensing

• The strength of weak lensing by the same
  foreground galaxies varies with the distance to
  the background galaxies.
• Measure amplitude of shear vs. z
• shear-galaxy correlations
• shear-shear correlations
• DES will
• Image 5000 sq-degrees
• Photo-z accuracy of ?z lt 0.1 to z 1
• 10-20 galaxies/sq-arcminute

PanStarrs
DeepLens
CFHLS
15
Peaks in the Density Field

• Clusters of galaxies are peaks of the density
  field.
• Dark energy influences the number and
  distribution of clusters and how they evolve with
  time.

16 Mpc
2 Mpc
16
Cluster Masses
Optical

• Our mass estimators
• Galaxy count/luminosity
• Weak lensing
• Sunyaev-Zeldovich
• The South Pole Telescope project of J. Carlstrom
  et al.
• DES and SPT cover the same area of sky
• Self calibration
• Mass function shape allows independent checks
• Angular power spectrum of clusters
• Allows an approach at systematic error reduction

Lensing
Mass
SZ
X-ray
17
Cluster Counting

• Locate peaks in the density field using cluster
  finders
• Red sequence methods
• SZ peaks
• DES will
• Image 5000 sq-degrees
• Photo-z accuracy ?z 0.01 to z 1
• 20,000 massive clusters
• 200,000 groups and clusters

N
z
z 0
1
3
PanStarrs
18
We aim at 5 precision on Dark Energy
Weak Lensing
Cluster Counting
Supernova
w
w
w
?M
?DE
?M
? w 5 and ? ?DE 3
The Planck satellite will provide tighter input
CMB measurements, and the constraints will
improve slightly.
Joint constraints on w and wa are promising
initial results suggest ?wa 0.5.
19
The Dark Energy Survey

• We propose the Dark Energy Survey
• Construct a 500 Megapixel camera
• Use CTIO 4m to image 5000 sq-degrees
• Map the cosmological density field to z1
• Make precision measurements of the effects of
  Dark Energy on cosmological expansion
• Cluster counting
• Weak lensing
• Galaxy clustering
• Supernovae

5000 sq-degrees Overlapping SPT SZ survey 4
colors for photometric redshifts 300 million
galaxies
20
Design of the Dark Energy Survey

• James Annis

21
Science Goals to Science Objective

• To achieve our science goals
• Cluster counting to z gt 1
• Spatial angular power spectra of galaxies to z
  1
• Weak lensing, shear-galaxy and shear-shear
• 2000 zlt0.8 supernova light curves
• We have chosen our science objective
• 5000 sq-degree imaging survey
• Complete cluster catalog to z 1, photometric
  redshifts to z1.3
• Overlapping the South Pole Telescope SZ survey
• 30 telescope time over 5 years
• 40 sq-degree time domain survey
• 5 year, 6 months/year, 1 hour/night, 3 day
  cadence

22
DARK ENERGY SURVEY (DES)
Science Goal measure wp/?, the dark energy
equation of state, to a precision of dw 5, with

• Cluster Survey
• Weak Lensing
• Galaxy Angular Power Spectrum
• Supernovae

23
DES Requirements
Science Goals
Science Requirements

• Cluster Survey
• Weak Lensing
• Galaxy Angular Power Spectrum
• Supernovae

redshifts, area, filters, limit mag, red
image quality, area
photometry, area, limit mag
repeat, area, filters, red
24
Science Requirements

• 5000 sq-degrees
• Significantly overlapping the SPT SZ survey area
• To be completed in 5 years with a 30 duty cycle
• 4 bandpasses covering 390 to 1100 nm
• SDSS g,r,i,z
• z modified with Y cutoff
• Limiting magnitudes
• g,r,i,z 24,24,24,23.6
• 10s for small galaxies
• Photometric calibration to 2
• 1 enhanced goal

• Astrometric calibration to 0.1
• Point spread function
• Seeing lt 1.1 FWHM
• Median seeing lt 0.9
• g-band PSF can be 10 worse
• Stable to 0.1 over 9 sq-arcminute scales
• From chapter 3 of NOAO proposal version 3 of
  requirements.
• Version 4, under review, will be a formal science
  requirements document.

25
Limiting Magnitude
Red Galaxy

• Limiting magnitude (10s for small galaxies) was
  set by flow down of science goals
• ½ L cluster galaxies at redshift 4000A break
  leaving blue filter
• g,r,i,z 22.8,23.4,24.0,23.3
• Complete cluster catalog
• Galaxy catalog completeness
• g,r,i,z 22.8,23.4,24.0,23.6
• Simple selection function
• Blue galaxy photo-z at faint mags
• g,r,i,z 24.0,24.0,24.0,23.6
• Photo-z for angular power spectra and weak
  lensing

Mag of ½ L galaxy
0 redshift 1.5
i 23-24
photo-z spectro-z
0 redshift 1.5
26
Galaxy Cluster Redshifts
four filters (griz) track 4000 Å break. Need z
band filter to get out to redshift gt 1
Theory

• DES data will enable cluster photometric
  redshifts with dz0.02 for clusters out to z1.3

for M gt 2x1014 M?

• the distribution of the number of clusters as a
  function of redshift is sensitive to the dark
  energy equation of state parameter, w.

27
Photometric Redshifts
Red galaxies

• Resulting limiting magnitudes give very good
  photometric redshifts
• Monte Carlo simulations of photometric redshift
  precision
• Evolving old stellar pop. SED
• Redshifted and convolved with filter curves.
  Noise added.
• Polynomial fit to photo-z
• For clusters, averaging all galaxies in the
  cluster above limiting magnitude.
• Template fit for photo-z
• These are sufficient to achieve our science goals.

½ L
2 L
Clusters
1.0x1014 M0
28
The Footprint

• Requirements
• Overlap with SPT SZ survey
• Redshift survey overlap
• Footprint
• -60 lt Dec lt -30
• SDSS Stripe 82 VLT surveys

DIRBE dust map, galactic coordinates
29
Survey Strategy I

• Design decision 1 area is more important than
  depth
• Image the entire survey area multiple times
• Design decision 2 tilings are important for
  calibration
• An imaging of the entire area is a tiling
• Multiple tilings are a core means of meeting the
  photometric calibration requirement offset
  tilings, not dithers
• Design decision 3 substantial science with year
  2 data
• We will aim for substantial science publications
  jointly with the public release of the year 2
  data.

30
Survey Strategy II

• Year 2
• g,r,i,z 100 sec exposures
• g,r,i,z 24.6, 24.1, 23.6, 23.0
• Calibration abs2.5 rel1.2
• Clusters to z0.8
• Weak lensing at 12 gals/sq-arcmin
• Year 5
• z 400 sec exposures
• g,r,i,z 24.6, 24.1, 24.3, 23.9
• Calibration abslt2 rellt1
• Clusters to z1.3
• Weak lensing at 28 gals/sq-arcmin

• Two tilings/year/bandpass
• In year 1-2, 100 sec/exp
• In year 3, drop g,r and devote time to i,z 200
  sec/exp
• In year 5, drop i and devote time to z 400
  sec/exp
• If year 1 or 2 include an El Nino event, we lose
  1 tiling, leaving three tilings at the end of
  year 2. This is sufficient to produce substantial
  key project science.

31
DES Time Allocation Model
Time to the Community and to
the Dark Energy Survey

• September 4 bright 4 dark nights
  22 nights
• October 4 bright 5 dark nights
  22 nights
• November 4 bright 4 dark nights
  22 nights
• December 4 bright 4 dark nights
  21 nights
• Telescope shut down Dec 25,
  31
• January 4 bright 5 dark nights
  11 nights
• and the 2nd half of all
  nights
• February 3 bright 3 dark nights
  11 nights
• and the 2nd half of all
  nights
• March August all
  none
• Total 257 nights
  108 nights

32
Time Allocation
CTIO mean weather year

• Analytic calculation of time available
• 30 year CTIO weather statistics
• 5 year moving averages
• Calculate photometric time
• Can complete imaging survey and time domain
  survey with 3 sq-degree field of view camera
• Simulations of observing process
• Use mean weather year
• Survey geometry
• Observing overhead
• NOAO time allocation model
• High probability of completing core survey area
  in time allocated

gt DES time allocation model just sufficient to
achieve science objective.
Probability of obtaining 8 tilings per year over
survey area. Dark is 100, light yellow 50
33
Photometric Calibration Strategy

• Calibrate system response
• Convolve calibrated spectrum with system response
  curves to predict colors to 2
• Dedicated measurement response system integrated
  into instrument
• Absolute calibration
• Absolute calibration should be good to 0.5
• Per bandpass magnitudes, not colors
• Given flat map, the problem reduces to
  judiciously spaced standard stars

• Relative calibration
• Photometry good to 2
• Per bandpass mags, not colors
• Use offset tilings to do relative photometry
• Multiple observations of same stars through
  different parts of the camera allow reduction of
  systematic errors
• Hexagon tiling
• 3 tilings at 3x30 overlap
• 3 more at 2x40 overlap
• Aim is to produce rigid flat map of single
  bandpass
• Check using colors
• Stellar locus principal colors

34
Survey Simulation

• We plan a full scale simulation effort
• Led by Huan Lin
• Centered at Fermilab and Chicago
• Using analytic, catalog and full image simulation
  techniques
• Over 4 years
• Underway, starting with photometric redshift
  simulations
• Use the simulations in 3 ways
• Check reduction code
• Mock data reduction challenge
• Chris Stoughton
• Prepare analysis codes
• Mock data analysis challenge
• Josh Frieman
• Prepare for science

• Survey simulations
• Jim Annis
• Catalog level simulations
• Lin, Frieman, students for photo-z and galaxy
  distributions
• Risa Weschlers Hubble Volume n-body
• Albert Stebbinss multi-gaussian approximation
• Mike Gladders empirical halo model
• Image level simulations
• Erin Sheldon for weak lensing
• Doug Tucker and Chris Stoughton
• Terapix skyMaker
• Masseys Shapelets code

35
Survey Planning Summary

• We have well defined science goals and a well
  defined science objective
• A 5000 sq-degree survey substantially overlapping
  the SPT survey
• A time domain survey using 10 of time
• The science requirements are achievable.
• A good seeing, 4 bandpass, 2 calibration, i 24
  survey
• Multiple tilings of the survey area the core of
  the survey strategy and photometric calibration.
• The survey can be completed using
• 22 nights a month between September and October
• 21 nights in December
• 22 half nights a month in January and February

36
DES Instrument Project

• OUTLINE
• Science and Technical Requirements
• Instrument Description
• Cost and Schedule
• Prime Focus Cage of the Blanco Telescope
• We plan to replace this and everything inside it

37
DES Instrument Reference Design
Instrument Construction Organization
1.2.1 CCDs 1.2.2 CCD Packaging 1.2.3 Front End
Electronics 1.2.4 CCD Testing 1.2.5 Data
Aquisition 1.2.6 Camera Vessel 1.2.7
Cooling 1.2.8 Optics 1.2.9 Prime Focus
Cage 1.2.10 Auxiliary Components 1.2.11 Assembly
and Testing
3556 mm
Camera
Scroll Shutter
1575 mm
Filters
The Reference Design represents our current
design choices and may change with more analysis
Optical Lenses
38
Optics Design

• 2.2 deg. FOV Corrector
• 5 powered elements (Fused Silica)
• one aspheric surface (C4)
• four filters griz needed for DES
• others can be used
• More details of the design in the next talk
  (Steve Kent)
• Cost for the glass 660k
• Cost for figuring 1M
• 1.5 yr delivery

Corrector
39
Dark Energy SurveyOptical Design and Issues

• 2.2 Deg. Field of View Corrector
• Requirements
• Performance
• Issues

Steve Kent, Fermilab, for the DES
Collaboration Dark Energy Survey BIRP, Aug 12,
2004
40
2.2 Deg. Field of View Corrector

• 14 requirements total
• 0.39 to 1.1 µ (SDSS filter bandpasses)
• Scale 17.7 arcsec/mm
• Field size 450 mm diameter (2.2 degrees)
• D80 lt 0.64 arcsec everywhere (FWHM lt .4 arcsec)
• No ADC (Atmospheric Dispersion Corrector)
• Minimize ghosting
• Space for filter, shutter
• Design choices should minimize procurement,
  fabrication schedule.

41
Gladders may11 design

• Features
• Flat focal plane
• Five lenses Filter
• (including dewar window)
• All fused silica
• One aspheric surface
• Largest diameter 1.1 meters
• Flexibility spacing elements
• Low distortion (lt1)
• Good ghosting properties
• star halos
• exit pupil image

C5
Shutter
C4
Filter
Filter
C3
C2
C1
42
CCDs

• Reference Design LBNL CCDs
• QEgt 50 at 1000 nm
• 2k x 4k
• 15 micron pixels
• 250 microns thick
• fully depleted (high resistivity)
• back illuminated
• 4 side buttable
• readout 250 kpix/sec
• 2 RO channels/device
• readout time 17sec
• fringing eliminated
• PSF controlled by bias voltage

RD on LBNL CCDs nearly finished. LBNL CCDs
have been used at LICK and on the WIYN
Telescope and on the Mayall
43
CCD QE and Read noise
To get redshifts of 1 we spend 50 of survey
time in z-band. LBNL CCDs are much more
efficient in the z band than the current devices
in Mosaic II
Read noise for a recently finished DALSA 2k x 4k
250 kHz ? 7e-
44
CCD Acquisition Model

• Reference Design Acquisition Model
• Order CCDs through LBNL good relationship with
  commercial foundry
• Foundry delivers wafers to LBNL (650 microns
  thick)
• LBNL
• applies backside coatings for back illuminated
  operation
• oversees thinning ( 250 microns thick) and
  dicing
• tests all devices on cold probe station
• LBNL delivers all tested, unpackaged devices to
  FNAL
• FNAL packages and tests CCDs
• Prepared to package 160 CCDs (spares, yield)

• CCD Wafers
• Existing masks have 2/wafer
• to be cost efficient we will make new masks with
  4/wafer

45
Packaging

• CCD Packaging will be done at Fermilab
• LICK and LBNL have already successfully packaged
  small quantities.
• We are developing a working relationship with R.
  Stover at LICK (we visited in July) to learn
  packaging techniques

Invar Foot
CCD packaged at LBNL
AlN circuit board
Wirebonds to CCD
CCD Packaging is very similar to building the
components of silicon vertex detectors. Fermilab
has built many vertex detectors for CDF and D0,
and is contributing to CMS
46
Packaging and Testing Process

• Packaging and testing keep up with anticipated
  CCD delivery rate of 20/month (5 wafers).
• Packaging
• one CCD takes 1 week to complete
• Plan to have capabilities to start 2/day
• Testing
• estimate 2 days/CCD
• 3 identical test stands needed to keep up with 5
  CCDs/week
• LBNL cold probe test results will guide which
  CCDs to package 1st
• Assume 60 good devices from production run and up
  to 18 good devices from preproduction run

47
CCD Test Stand and Acceptance Criteria

• Testing
• linearity, full well depth, QE, CTE, readnoise,
  dark current
• Testing and acceptance criteria will be defined
  as we gain experience with LBNL CCDs
• Will also consider impact of acceptance criteria
  on community
• Multiple tilings reduces impact of bad regions
• Study with 100 consecutive bad columns found
  1.5 of tiling area was imaged less than 3 times
  after 5 complete tilings

48
Camera Reference Design
Focal Plane
Camera Design
feed through board
62 2k x 4k CCDs for main image, 4-side
buttable, 15 micron pixels 8 1k x 1k CCDs
for guiding and focus
Frontend electronics
Focal Plane
49
Camera Vessel
Cooling/ Vacuum spool piece
Camera is separated into two spool pieces one
for signal feed throughs one for cooling and
vacuum services Removal of cooling spool piece
allows access to back of focal plane and cables

• Vacuum feed through board brings signals out of
  cryostat

50
Cooling and Integration

• Reference Design has LN2 reservoir inside
  cryostat
• Fill from recondensing dewars on floor
• investigating alternative Gifford-McMahon cryo
  coolers on cooling spool piece which condense N2
  directly into reservoir

Will fully assemble prime focus cage at FNAL and
test all systems together (corrector, focal
plane, cooling, data acquisition, data
management....) before shipping to Chile
51
Front End Electronics and DAQ
off the cage

• Large focal plane implies long cables between CCD
  and electronics crates
• Reference design has clock drivers and preamps as
  part of the cable assembly
• Goals are noise lt 5 e-, linearity lt0.25, support
  a readout rate of 250 kpix/sec
• Reliable operation requires careful consideration
  of internal and external components
• Minimize heat generated in the PF cage by
  locating DAQ off telescope

52
Data Acquisition
DES data rates are relatively high by astronomy
standards, but not for particle physics.

• We will use the Monsoon data acquisition system,
  developed by NOAO.
• We will modify it to separate digital and analog
  functionality.

Using Monsoon shortens development time and
enables collaboration with NOAO and other Monsoon
users.
53
Data Acquisition
Monsoon architecture

• DES Modifications
• ADCs will reside on the telescope.The rest of
  the electronics will be off the telescope.
• Save space and power on the telescope.
• Reduce noise (ADCs are closer to the CCDs).
• Save money.

54
We Can Do This!
The DES collaboration has assembled a team of
experienced scientists, engineers, designers and
technicians

• The Silicon facility at Fermilab has experience
  building the Run 0, I, II silicon vertex
  detectors Micron precision assembly
  Wirebonding
• Thermal Management Cleanrooms
• Building a CCD focal plane uses many of the same
  skills, but has many fewer devices.
• LBNL has extensive experience with CCD
  development and packaging for SNAP/JDEM

UIUC has experience building large, high rate
data acquisition systems at SLAC, Fermilab, and
Cornell. U Chicago has experience with optical
design and optical systems on SDSS DES does
not depend on pioneering development work. The
main issues are cost, schedule, and integration.
55
Schedule Milestones

• Optics and CCDs are the most Challenging tasks
• CCDs Preproduction run FY05, Production run
  FY06 and FY07
• Optics Order glass in FY06, Figuring/polishing
  in FY07

Fully Commissioned by June 2009!
56
Total Cost profile in Then Yr
(excluding institutional overhead)
The Reference Design represents our current
choices for meeting the science goals Total cost
for the Instrument project is 18.4 M excluding
institutional overheads and 22.5M with overhead
in then year . We will be ready for
observations by June 2009. This schedule is
funding limited.
57
Instrument Project Organization
58
Conclusions

• We have a strong collaboration with a wide
  variety of skills that cover all aspects of this
  project
• With this collaboration we can complete the
  instrument and start survey operations on the
  telescope in 2009