PIONIER Overview

The main features of PIONIER@VLTI for P96 are:

  • Unit Telescopes (UTs): UT adaptive optics guarantees diffraction-limited image quality on PIONIER for targets brighter than V=17.
  • Auxiliary telescopes (ATs): Tip-tilt correction for targets brighter than V=13.5.
  • The VLTI is considered non-operational at seeing values worse than 1.5". This is due to the operational limit for the adative optics above.
  • Off axis guiding is possible if the target does not meet the above limits. The presence of an appropriate guide star must then be checked and indicated in the proposal in the target notes.
  • Baselines: please look at the VLTI News webpage to get the list of the UT and AT baselines available.
  • Spectrograph optics: either dispersed (GRISM, 6 pixel over H-band), or undispersed (FREE, 1 pixel).
  • As a rule, Visitor Mode is discouraged and will only be accepted for observations that are well justified to require real-time decisions or non-standard operational procedures in order to achieve the proposed science goal.


Limiting Visibility and Magnitudes

PIONIER is tracking its own fringes. Due to the limited dynamic range of the detector, the fringe tracking requires a minimum visibility of V=0.05, or 5%. This requirement must, however, be met only on the short and intermediate baselines in a given quadruplet, since the fringe position on the longest (or most resolved) baselines can be calculated from the information obtained from the less resolved, fringe tracking baselines. Depending on the geometry of the baselines that can track fringes, two or three remaining baselines can be bootstrapped. If bootstrapping the fringe position is possible in this way, the limit at which fringes can be reduced and calibrated is much lower, namely V=0.01, or 1%.

The bright limit is given by the high detector sensitivity vs. its limited dynamic range. Targets cannot be observed in service mode if they are brighter than H=-1mag. Brighter targets can in principle be observed by inserting a vignetting optics into the beam or defocussing the telescopes (only possible with ATs), but this special operation is limited to visitor mode observations. In any case, the above visibility limits apply, which more typically will be the reason why these targets cannot be observed, rather than the brightness as such.

The faint limit depends on the observing conditions:

Dispersion Limit (H mag) Conditions to be used in phase 1
GRISM 7.1 to 8.0 0.8", CLR
 6.1 to 7.0 1.2", THN
GRISM  -1.0 to 6.0 1.5", THN


If the FREE mode is used (undispersed, 1 pixel) the limiting magnitude is somewhat fainter, by about 0.5 mag. This is recommended only for targets fainter than about H=8, brighter targets should be observed with GRISM and better seeing.

Calibrators should be similarly bright as the target. This is for anumber of reasons. First, large magnitude differences will lead to considerably lower signal in the fainter object, and hence to lower precision. For optimal precision of the calibration, the magnitude difference should not be more than 1 magnitude. Second, differences of more than 3 magnitudes become operationally difficult to secure with the same setup and thus not only diminish the chance of an observation to be carried out, but will introduce a slight and systematic mis-calibration of a few percent. In particular, for targets brighter than H=2.5 also the calibrator needs to be brighter than H=2.5, and fainter targets should have fainter calibrators, as around H=2.5 a different detector gain value will be set under normal conditions by the operator. Finally, there is some evidence that for large magnitude differences between SCI and CAL the calibration is systematically worse in the blueward channels.

It is of greatest importance that the coordinates of all objects in a concatenation are taken from a consistent source. Coordinate offsets lead to offsets in optical path difference, which may prevent the fringes to be found and cause the execution to fail.

Tests of PIONIER on the UTs have shown that, although fringes can be detected up to H~11, the data turned out to be very difficult to calibrate. In particular, no observation beyond H=8.5 (which is as well the AT limiting magnitude) could yet be verified as scientifically valid. The exact issues are possibly related to injection instabilities, to which PIONIER, designed for the ATs and never optimised for the UTs, is particularly sensitive. Therefore, UTs should only be used when other issues (such as availability of guide stars or need for small interferometric FoV) prohibit the use of ATs.

That said, the general issue of performance of VLTI with the UTs is of paramount importance for the second generation instruments and is currently actively investigated.

Typical Calibration Precision

Applying the normal observation and calibration scheme offered in service mode, the accuracy of PIONIER observables (squared visibility and closure phase) is limited by calibration systematics related to the separation between calibrators and science target. To minimize these systematics, calibrators should be selected within a few degrees from the science target. For a separation of 3 degree between the calibrators and the science target, the typical accuracy reached is 3% on the squared visibility and 2 degree on the closure phase measured. The use of different calibrators distributed around the science target is advisable. Further reducing this uncertainty is possible applying a more complex calibration scheme that can only be offered in visitor mode (see the PIONIER user manual for details).

Other Instrument and Detector Parameters

Other parameters, such as detector gain and integration times, will be chosen by the operator based on conditions vs. target parameters, as they will not affect the scientific results. In service mode we ensure that a given concatenation is observed with a single set of parameters only, except for very bright targets, where suitable calibrators may be to faint and the have to be observed with a higher detector gain. The calibratatablity of this procedure was tested and results are within the typical precision. In any case, we do not guarantee to use identical parameters across several concatenations, since they may be observed on different dates under different conditions.This is not waiverable in service mode, and users who can solidly justify the scientific need for consistent instrument and detector parameters across their program (such as the need to achieve exceptionally high calibration precision) should apply for visitor mode.

Execution times

In service mode, two operation sequences are offered, namely three OBs (in a CAL-SCI-CAL concatenation, suitable for snapshot observations) or five OBs (in a CAL-SCI-CAL-SCI-CAL concatenation, suitable for imaging programs). The execution times to be used for the phase 1 proposals are 30 minutes for a 3-OB concatenation and 45 minutes for a 5-OB concatenation, i.e. 10min and 9min per OB, respectively. The difference is due to the requirement to use the same target and the spatial proximity and the recommended similarity in magnitude of target and calibrators in the 5-OB concatenation, making the acquisition process slightly more efficient.

Visitor mode observers are advised to plan with about 10 to 12 minutes per OB, to allow some additional time for inter-OB decision making or special operation procedures, which typically will be the reason why VM would be accepted..

Data Reduction

Data reduction software has been made available to the public by IPAG. Additionally, all SM data will be routinely processed at IPAG and calibrated OIFITS data will be provided.