Spectro-Polarimetric High-contrast Exoplanet REsearch - SPHERE

Principal Investigator P.I. : Jean-Luc Beuzit
CO-I: Markus Feldt

The S P H E R E Consortium:

Institut de Planétologie et d’Astrophysique de Grenoble
Max Planck Institut für Astronomie
Laboratoire d’Astrophysique de Marseille
Laboratoire d’Etudes Spatiales et d’Instrumentation en Astrophysique de Paris
Laboratoire Universitaire d’Astrophysique de Nice
Istituto Nazionale di Astrofisica
Eidgenössische Technische Hochschule Zürich
Observatoire de Genève
Universität van Amsterdam
Office National d’Etudes et de Recherches Aérospatiales
Stichting ASTRonomisch Onderzoek in Nederland

Instrument Scientist Ralf Siebenmorgen
WP Manager Markus Kasper
Instrument Science and Operation Team Ralf Siebenmorgen (chair), Dimitri Mawet, Markus Kasper, Mario van den Ancker, Wolfgang Hummel, Julien Girard, Lars Lundin, Jakob Vinther
Instrument Science Team Florian Kerber (chair), A. Collier Cameron, R. Rebolo Lopez, N. Santos, N.Thatte
Location VLT Nasmyth



  • Schedule: First light on telescope: goal 2014
  • Preliminary Acceptance Europe, Dec 2013
  • Final Design Review, 2008
  • Preliminary Design Review, 2007
  • Contract signature, 2007

Baseline Specification

Optical Throughput
40% (goal 45%) for each beam
60% (goal 70%)
25% (goal 40%)
Wavelength coverage
0.6 - 0.9µm
(goal: 0.5 - 0.9µm)
Spectral Resolution

DBI: R ~ 20-30
LS: R ~ 50 (Y-K), 500 (0.95-1.8µm)

R ~ 30
Field of View
>11" diameter
>1.35" square (goal 3" square)
>3" square
Spatial Sampling
12.25 mas ( l/2D at 0.95µm)
12.25 mas ( l/2D at 0.95µm)
<7.8 mas (l/2D at 0.6µm)
Contrast (5s)
at 0.1": <5e-5 (goal 1e-5)
at 0.5": <5e-6 (goal 5e-7)
at 0.5": <1e-6 (goal 1e-8 )
at 1": <1e-8 in 4hr (goal 3e-9 in 15 hr) for a 30% polarized planet
Observing modes
dual-band imaging (DBI),
dual-polarimetric imaging (DPI),
long-slit spectroscopy (LS)
Visible Imaging
Differential polarimetric Imaging

Scientific Objectives

The prime objective of the Spectro-Polarimetric High-contrast Exoplanet Research (SPHERE) instrument for the VLT is the discovery and study of new extra-solar giant planets orbiting nearby stars by direct imaging of their circumstellar environment. The challenge consists in the very large contrast between the host star and the planet, larger than 12.5 magnitudes (or 10 5 in flux ratio), at very small angular separations, typically inside the seeing halo. The whole design of such an instrument is therefore optimised towards reaching the highest contrast in a limited field of view and at short distances from the central star (Mouillet et al., 2001). Both evolved and young planetary systems will be detected, respectively through their reflected light (mostly by visible differential polarimetry) and through the intrinsic planet emission (using IR differential imaging and integral field spectroscopy). Both components of the near-infrared arm of SPHERE will provide complementary detection capacities and characterization potential, in terms of field of view, contrast, and spectral domain.

SPHERE will greatly contribute to the field of extra-solar planets studies, already very active, particularly by offering direct detections of planets more massive than Jupiter at various stages of their evolution, in the key separation regime 1 to 100 AUs. Migration mechanisms will then be better understood. The complementarities of direct imaging with other detection methods such as radial velocities and photometric transits, in terms of targets, detection biases and measured planetary parameters, and more specifically the combination with results from other projects like HARPS, COROT, VLTI/PRIMA, JWST and Kepler, will offer promising avenues. The present indications that massive distant planets could be numerous will be firmly confirmed or denied by SPHERE, if the number of observed targets with relevant detection limits is statistically acceptable, i.e. of the order of 300 to 400.

This would in particular fully justify a large effort in an extended observational survey of several hundred nights concentrating on the following classes of targets:

Nearby young associations will offer the best chance of detecting low mass planets, since they will have brighter sub-stellar companions, and therefore the greatest number of planets per star observed. Stars with known planets, especially any that exhibit long term residuals in their radial velocity curves, indicating the possible presence of a more distant planet. Nearest stars: measuring these targets will probe the smallest orbits and will thus the only opportunities for detecting planets by directly reflected light. Stars aged from 100 Myr to 1 Gyr: planets will still be over-luminous as compared to Solar System planets, so mass limit will be lower than for old systems. With such a prime objective, it is obvious that many other research fields will benefit from the large contrast performance of SPHERE: proto-planetary disks, brown dwarfs, evolved massive stars and marginally, Solar System and extragalactic science. These domains will nicely enrich the scientific impact of the instrument. Their instrumental needs should however not be in conflict with the high-contrast requirement.

Instrument Description

Instrument Design

Figure 1. Concept of SPHERE implementation on the VLT Nasmyth platform showing the common path with the XAO system and the three science instruments IRDIS, IFS and ZIMPOL.

SPHERE will be located at the Nasmyth focus of the VLT. The instrument as shown in the figure above is composed of 4 major subsystem: the common path including the powerfull AO system and the three science instruments IRDIS, IFS and ZIMPOL each fed by a sophisiticated pupil apodized Lyot, Lyot, or phasemask coronagraph. The conceptual design of these subsystems is summarized hereafter.

Common path with eXtreme Adaptive Optics

The design of the SPHERE Adaptive optics for eXoplanet Observation (SAXO), resulting from a global trade-off combining optical design, technological aspects, cost and risk issues, leads to the use of a 41x41 actuator DM of 180 mm diameter with inter-actuator stroke >±1μm and maximum stroke >±3.5μm, and a 2-axis tip-tilt mirror (TTM) with ±0.5 mas resolution. The wavefront sensor is a 40x40 lenslets Shack-Hartmann sensor, with a spectral range between 0.45 and 0.95μm equipped with a focal plane filtering device of variable size (from λ/d to 3λ/d at 0.7μm, where d is the projected micro-lens diameter) for aliasing control. A temporal sampling frequency of 1.2kHz is achieved using a 240x240 pixel electron multiplying CCD detector (CCD220 from EEV) with a read-out-noise < 1 electron and a 1.4 excess photon noise factor. The global AO loop delay is maintained below 1 ms. SAXO will deliver corrected images with a Strehl ratio of close to 90% in H-band under 0.85" line of sight seeing.

Image and pupil stability are essential in high-contrast instruments. Differential image movements due to thermo-mechanical effects and ADC mechanism precision are therefore measured in real-time using an auxiliary NIR tip-tilt sensor located close to the coronagraphic focus and corrected via a differential tip-tilt mirror in the WFS arm. Likewise, pupil run-out is measured by analysis of the WFS sub-pupil intensity along the pupil edge and corrected by a pupil tip-tilt mirror close to the telescope focal plane at the entrance of the instrument. Non-common path aberrations are measured off-line using a phase diversity algorithm and compensated on-line by reference slope adjustments.

Infra-Red Dual-beam Imager and Spectrograph

The Infra-Red Dual-beam Imaging and Spectroscopy (IRDIS) sub system constitutes the main science module of SPHERE. The main specifications for IRDIS include a spectral range from 950 to 2320 nm and an image scale of 12.25 mas per pixel consistent with Nyquist sampling at 950 nm. A FOV greater than 11" diameter is required for both direct and dual imaging, leaving a slight margin for system optimization when using two "quadrants" of a 2kx2k detector. The main mode of IRDIS is the dual band imaging (DBI), providing images in two neighboring spectral channels with minimized (<10nm rms) differential aberrations. Ten different filter couples are defined corresponding to different spectral features in modeled exoplanet spectra. In the direct imaging mode, 12 broad, medium and narrow-band filters are defined. In addition to direct and dual imaging, long-slit spectroscopy (LS) at resolving powers of 50 and 500 is provided, as well as a dual polarimetric imaging mode (DPI). A pupil-imaging mode for system diagnosis is also implemented.
Dual imaging separation is done using a beam-splitter combined with a mirror, producing two beams in parallel. Each beam has its own camera doublet and band-limiting filter. The main challenge is to achieve the required 10nm differential aberrations requirement, but an error budget based on high-quality classical polishing technology is found to satisfy the requirement. This option has been favored over the alternative Wollaston-based option used for example in the NACO SDI camera because it eliminates spectral blurring problems, which would limit the useful FOV, and allows the use of high-quality materials with high homogeneity.

Infra-red Integral Field Spectrograph

While an integral field spectrograph (IFS) for planet imaging is conceptually challenging, it is widely recognized as a potentially extremely useful science module for a planet searching instrument. The reasons for this are two-fold: Firstly the IFS can be build with virtually zero differential aberrations, and secondly the multiple spectral channels allow for better correction of speckle chromaticity and even data analysis strategies that do not rely on the presence of a-priori assumed features in the planet’s spectrum.

For SPHERE we are pursuing a micro-lens based IFS concept BIGRE, that is an evolution of the classical TIGER concept modified for the case of high-contrast diffraction limited observations. The required 5-σ detectivity at 0.5" is 1e-7 with a goal of 1e-8 with respect to the un-occulted PSF peak, and the spectral range of the IFS is limited to the Y-H bands (0.95-1.7µm), allowing the use of a single detection channel and parallel operation of IRDIS (in H-band) and IFS (in J-band). A resolving power per pixel of 30 is maintained, with a minimum FOV of 1.35" square and a strong goal of 3" square. Nyquist-limited spatial sampling at 0.95µm is imposed as for IRDIS. Optimized commonality between IFS and IRDIS in terms of detector and associated equipment is seen as an important system goal. The same 2kx2k detector format is therefore adopted, and it is highly likely that the long-wavelength cut off defined for IRDIS will also be acceptable for IFS.

Visible Differential Imager

The Zurich Imaging Polarimeter (ZIMPOL) subsystem is a high-precision imaging polarimeter working in the visual range, covering at least from 600 to 900 nm. The instrument principle is based on fast modulation using a ferro-electric retarder, and demodulation of the polarization signal using a modified CCD array. Key advantages of this technique are the simultaneous detection of two perpendicular polarizations (the modulation is faster than seeing variations), and the recording of both images on the same pixel. Thanks to this approach, a polarimetric precision of 1e-5 or even better should be achieved. The CCD will cover a Nyquist sampled field of 3" x 3" square and it is foreseen that the FOV can be moved around the bright star so that a field with a radius of 4" can be covered. In addition to polarimetric imaging, ZIMPOL provides the possibility for high resolution imaging in the visual range using a set of broad and narrow band filters. This capacity will be unique in the post-HST era.