Max-Planck-Institut für extraterrestrische Physik

- Infrared/Submillimeter Astronomy -



 .MPE . Infrared Astronomy . Projects . GRAVITY . Instrument







Jobs / Thesis




Instrument Concept

Following the requirements from the reference science missions, GRAVITY is designed for:
  • infrared wavefront sensing down to mK > 10;
  • internal fringe tracking down to mK > 10;
  • multiple baseline narrow angle astrometry with 10 microarcsec accuracy for UT operations;
  • interferometric imaging of faint objects with mK > 19 in 1 hour observing time.
The following figure gives an overview how GRAVITY will look like:

Observation scheme

The following figure shows how a typical observation with GRAVITY will be done:

Single Key Components

Adaptive Optics (AO)

The GRAVITY adaptive optics system will be made of four individual adaptive optics systems to correct the atmospheric turbulence of every UT. Hence the GRAVITY AO system will be the first to control simultaneously four individual AO systems. The challenge is to make the AO system transparent for the global instrument and the observer, who has to focus on the interferometric observations.

GRAVITY will take advantage of as much of the VLTI adaptive optics infrastructure as possible:
  • in its baseline design, the MACAO 60-actuator bimorph deformable mirrors can be used to compensate atmospheric turbulence and tunnel seeing for GRAVITY observations with the UTs;
  • in order to sense the tunnel seeing as well as static optical aberrations, the wavefront sensors should be placed in the VLTI lab close to the GRAVITY beam combiner;
  • the real time computing needs for the analysis of the wavefront sensor signals and the control of the DMs could possibly be handled by ESO's adaptive optics real-time computer platform SPARTA (Standard Platform for Adaptive Optics Real Time Applications), if it meets the requirements specific to GRAVITY.

Beam Combiner

Two beam combiner options are considered for GRAVITY: integrated-optics and fiber X-couplers. The baseline is integrated optics. Widely used in the telecommunication NIR bands (e.g. for beam switching), the technique has already been applied in astronomical interferometry, and has also been demonstrated to work in the K-band. The second option would be fiber X-couplers. The usage of fluoride-glass fiber combiners for the K-band has been demonstrated. This is not the preferred solution, since the space envelope would be much larger than that of integrated optics. In addition, integrated optics beam combiners are more versatile and all-guided optics solutions can be designed for both co-axial and multi-axial schemes, therefore allowing the best optimization of the instrument sensitivity.
The following image shows a external link near infrared three-telescope integrated optics beam combiner made by LETI:


Field selector, path length compensation, and metrology

In order to not move the fibers during observation, we foresee a K-mirror for de-rotation, a tip-tilt mirror for laterally moving the field, and a device to adjust the projected separation of the two fibers. A fiber stretching unit will introduce the differential OPD to one of the objects. All motions may be implemented for cryogenic operation, for example using stepper motors and inductive resolvers for the K-mirror, piezo stacks with capacity sensors for the tip-tilt mirror and OPD control, and piezo-electric translation stages for fiber coupling. A preliminary optical and mechanical layout of the field selection and fiber coupling unit is shown in the following fgure:
We intend to use a pure reflective design of the optical components, in order to minimize dispersion that might complicate the OPD calibration. After the K-mirror, an off-axis parabola re-images the pupil onto the tip-tilt mirror, which is equipped with a cold stop. An image plane is created with a second parabola. In this plane, a roof-prism splits the field for the selection of the two objects within the 2'' field of the VLTI beam. The light from the two objects is then re-imaged onto the single mode fibers. The fibers are mounted on piezo driven stages for field selection, for optimization of the coupling, and for the zero-OPD calibration using the metrology system.

Since we are picking two objects within the FoV of the VLTI, a path length measurement and compensation system will be required for GRAVITY. The path length difference originates out of the angular separation of the objects on the sky. Measuring this angular separation with an accuracy of 10 microarcsec sets high demands on the accuracy to which the instrumental OPD has to be known. With the largest baselines available within the VLTI, typical errors allowed are on the order of 5 nm.

For a two-baseline interferometric setup such as in PRIMA, the standard solution for measuring the differential OPD is a dual wavelength super-heterodyne laser metrology. The metrology beams will be launched directly at the beam combiner. The extension of this technology to the multi-baseline measurement and four telescopes is demanding.

Therefore we are also exploring another concept that uses backward propagation of laser beams through the entire VLTI. The concept is illustrated in th following figure:
Injecting the laser light into the science fibers will take place in the temperature-stabilized part of the instrument and before any active phase control. This can either be done by wavelength dispersive elements (as dichroic fiber splitters) or the injection takes place within the beam combination unit. When overlapping in a pupil plane, the two laser wavefronts create an interference pattern. Using phase-shifting technologies, or adequate Fourier filtering methods, the relative phase difference between the beams can be measured with high accuracy. For detecting the fringe pattern we intend to use commercially available cameras. With the high number of photons within the laser beams we expect that even the scattered light from the VLT secondary mirror can be used for detection. The laser fringe spacing and phase movement can be measured at high frequency, delivering the OPD difference, i.e. the projected angular separation of the fibers on the sky. The laser wavelength should be as close as possible to the observing wavelength, in order to minimize dispersion effects and incomplete mode matching in the fibers. Best candidates for the lasers are Tm-doped fiber lasers, emitting tunable single mode radiation at 1.8 - 2 micron. Wavelength stabilization can be performed using Argon or Iodine atomic lines.

Fringe Tracker

GRAVITY will track on fringes on a reference source within the 2'' field of view. The instrument will therefore have two beam combiners whose designs may be different: the fringe tracker beam combiner will be optimized for short exposures and high phase accuracy, and the science beam combiner will be optimized for long exposures on stabilized fringes. The fringe tracker will lock on the white light fringe. The interferograms of the fringe tracker will be dispersed to ensure white light fringe identification and to avoid fringe jumping. The quality of the lock will be monitored by the real-time computer. Integration in the science beam combiner will be paused whenever the quality of the lock is poor or when the lock is lost. Integration will resume as soon as the lock is restored. The spectral resolution in the fringe tracker will be a compromise between sensitivity and fringe tracking stability. In practice, the signal is dispersed over a few spectral channels to allow disentangling between the white light fringe and a side fringe. The actuator of the fringe tracker should be the VLTI delay line unless the band-pass is too low, in which case a GRAVITY internal device will be used as actuator.
Valid HTML 4.01! © Infrared and Submillimeter Astronomy Group at MPE
last update: 27/02/2006, editor of this page: Stefan Gillessen
up © Max-Planck-Institut für extraterrestrische Physik