1. Congresso del Dipartimento di Fisica Highlights in Physics 2005 11–14 October 2005, Dipartimento di Fisica, Università di Milano Towards an optimized scanning strategy for the Planck satellite B. Cappellini 1, D. Maino 1, M. Bersanelli 1,2, A. Mennella 1,2, P. Platania 3, C. Burigana 4, M. Maris 5 and X. Dupac 5 1 Dipartimento di Fisica, Università di Milano, 2 INAF – IASF Sezione di Milano, 3 Istituto di Fisica dei Plasmi – CNR – Milano 4 INAF – IASF Sezione di Bologna, 5 INAF – Osservatorio Astronomico di Trieste Planck is an ESA mission planned to fully map the CMB anisotropy in a wide frequency range (30-857 GHz) with very high angular resolution (30'-5') and sensitivity (ΔT/T~10 -6 per resolution element). The satellite is planned to be launched in 2007. A key aspect of the mission design is the choice of the orbit and the scanning strategy for the satellite. These choices are driven by technical constraints, which involve the payload and spacecraft design and capabilities, and by scientific requirements, which involve for example the sky coverage, the noise distribution over the sky, and the data reduction pipeline. The main features for the orbit and scanning strategy for Planck are now well established and define the so-called ``Nominal Scanning Law''. Nevertheless, the details of this strategy are not frozen yet, and efforts are being done to identify the best choice which can ensure optimal scientific results. I SCANNING STRATEGY CONSTRAINTS The Planck scanning strategy has to be defined such that it satisfies well established attitude limits for the payload. The two driving constraints are thermal and communication constraints. 1. The payload always needs to be in the shadow of the Sun: the spin axis is required to be pointed within 10 o (this value is determined by solar panels dimension and payload geometry) of the Sun-satellite line. 2. Full-band download of data requires that the spin axis is oriented within 15 o with respect to the spacecraft (S/C) - Earth direction. We call “telemetry angle'‘ the angle between the spin axis and the Earth-S/C directions. In the figure on the left, we report results of the parametric analysis on the dipole pattern as a function of the spin-axis phase. The mean available ΔT for calibration in every scan circle for 1 year of the mission is plotted: this value gives a quantitative estimate of the dipole amplitude with respect to the achievable calibration accuracy. It can be seen that the situation changes mainly in the more critical periods for calibration, when the dipole pattern is lower (giving rise to situations which are either better or critical with respect to the nominal one), while it is essentially unchanged when the dipole offers high ΔTs. As one expects, the optimal choice of the phase is different for the first and the second half of the year, since the spin-axis motion period is equal to 6 months IV CALIBRATION ISSUE Changes in the geometry of the scanning strategy directly modify the dipole pattern as seen by instrument beams, and therefore also impact the achievable accuracy σ G of the in-flight calibration [2,3] which is based on the dipole pattern ΔT (σ G is proportional to 1/ΔT). The two lines (red and blue) refer to the two halves of the year. The horizontal line corresponds to the nominal scanning strategy. The best choice for half of the year just corresponds to the worst choice for the other half. This would suggest a re-adjustment of the phase after 6 months. This should be possible since a redefinition of the scanning strategy at the end of the first survey is planned. This analysis proposes an objective criterion for choosing the initial spin axis pointing direction, to allow the most accurate calibration on short and intermediate time scales (< 3-6 months). The figure below shows, as a function of the phase, the minimum value of ΔTs plotted in the figures on the left. The aim is to avoid phases for which calibration is highly degraded in the worst (also for the NSL) periods. - a complete coverage of the sky is not allowed: the polar caps, within a diameter of a few degrees, are not observed for each detector, given that the boresight angle of the telescope axis is 85 o ; - circle crossings are concentrated at high ecliptic latitudes. For systematic effects removal, it should be useful to have such crossing to be spread on a larger range of ecliptic latitudes. A displacement of the spin axis away from the Ecliptic plane can be a solution for both these limitations.It also allows the presence of connected deep fields (regions with a sensitivity much higher than the mean value) around the ecliptic polar caps. Moreover, it ensures a higher level of redundancy i.e. the repeated observation of a given pixel in the sky with different detectors, different satellite attitudes, and on different time scales. Redundancy is beneficial to the need of tight control on systematic effects. In the NSL (see the figure on the right ), the spin axis is always kept in the anti-Sun direction along the Ecliptic plane. The boresight angle between the spin-axis and the pointing direction is 85 o. The same ring in the sky is observed for an hour before repointing. This is the simples scanning strategy which satisfies main scientific requirements. Actually the NSL suffers from a few limitations: Planck orbit for a launch date on Agust 3, 2007. The orbital period around L2 is 6 months. On the other hand, such displacements of the spin-axis can also be a disadvantage. They may induce thermal and/or stray-light (mainly from the Sun) signal modulations, which may in turn introduce systematic effects in the data. Moreover, high sensitivity concentrated in a sky region free from foreground contributions is very useful for investigation of polarization anisotropy, expected at a level ~10% of temperature anisotropy, in sky patches, while a smoother noise distribution is not useful in this respect. The Baseline Scanning Strategy has to be chosen as the best trade-off among all of these factors compatibly with the constraints imposed by the mission plan. II THE NOMINAL SCANNING STRATEGY Planck will be placed in a Lissajous orbit (figure below) about the Earth-Sun Lagrangian point L2, at ~ 1.5 x10 6 Km from the Earth. 6 months III CYCLOIDAL OPTIMIZED PERIOD AND PHASE Different scenarios have been taken into account for chosing the definitive scanning strategy, according to the Planck Scanning Strategy reference document [1]. The choice is going towards a cycloidal pattern for the spin axis, characterized by a constant Sun aspect angle and hence free from dangerous thermal fluctuations. We consider this option with amplitude between 5 o and 10 o. Fast scanning strategies are not a viable option because of the dynamical design of the satellite. The period of the spin axis pattern is in principle a free parameter, but its value is constrained by the telemetry requirement. This requirement translates in limitation on the choice of the initial position of the spin axis along the traced pattern, which is determined by a “phase'' parameter, given as a value in [0,1]. A phase value is allowed if the telemetry limit is always satisfied, during a survey. A period is not allowed if no allowed phase can be found. We compute the telemetry angle during the mission as a function of the phase, of the spin axis period and of the motion direction (in principle the cycloidal pattern can be traced by the spin axis both in the prograde and in the retrograde direction). As an example the figure on the right shows this angle in the optimized (as derived through the analysis) configuration. In every configuration we derive the width of the allowed phases (blanck regions in the lower figure) bin. Larger is this bin and higher is the degree of freedom for choosing the optimal phase. As expected, the best choice for the period value is about 6 months, synchronized to the orbital period around L2, independently of the cycloidal amplitude. On the other hand, higher is this amplitude and stronger is the limitation on the allowed periods and phases value. The cycloidal pattern needs to be traced in the clockwise – as seen from the Earth- direction, at least for amplitudes higher than 6.7 o, to follow the direction of the Lissajous orbit around L2. Chosen the favourite configuration, we can optimize the phase value from the “telemetry point of view”. This can be done minimizing the mean or the maximum telemetry angle during the survey. These two criteria are well correlated, as shown in the last figure (symbols set the best choice). The analysis also pointed out that allowed phases and period values highly depend on the considered Lissajous orbit. V POLAR HOLES AND DEEP FIELDS One of the main disadvantage of the Nominal Scanning Strategy is the presence of unobserved regions around the ecliptic poles. This is due to the value of boresight angle, i.e. the angle between the spin axis and the line-of-sight of the telescope. This means that no complete coverage of the sky is achievable, and furthermore, this causes deep fields to be not connected regions. Both these properties are not optimal with respect to the scientific results one aims to obtain with Planck data. Deep fields in fact promise to be used for CMB polarization anisotropy (expected at 10% value of the temperature) detection and imaging. A cycloidal scenario helps in resolving these problems. The figures compares the distributions of sensitivities [4] for two different scanning strategies: Nominal scanning strategy 7° precession with ~6 months period. Precession better fills the polar holes and redistributes the sensitivity in the sky. The higher sensitivity at poles is due to the pile-up of the scan circles near the poles. This cause Planck to have deep fields near the poles, with integration times of up to about 50 min / pixel / feed-horn. Once a scenario (cycloidal amplitude and period) is set, the phase value does not affect, at first order, the senstivity distribution over the sky, while it determines the orientation of the deep field with respect to the observed sky. This can be used to choose the phase value to optimize the deep filed properties,e.g. minimizing foregrounds contamination. No Precession Precession Polar Hole Deep Field Regions VI STRAYLIGHT AND THERMAL EFFECTS Planck requires high sensitivity and angular resolution, together with strong rejection of systematic effects at the μK level and highly stringent thermal requirements (both absolute levels and stability), which drive the design of the instruments and payload. Also the choice of a far-Earth orbit and of the simplest Nominal scanning strategy (orientation of the spin axis in the anti-Sun direction) are mainly driven by the requirements of thermal stability and minimization of environmental disturbances (e.g. from the Sun, Earth and the atmosphere). Neverthless, stray-light rejection is an assett for each scanning strategy scenario analysed to derive the Baseline choice. Comparing scanning strategies it has been found [5] that a cycloidal precession: - improves slightly Earth and Moon stray-light rejection - worsens substantially the Solar stray-light rejection. However in all cases for the Moon ad the Earth straylights are largely below the signal and the noise levels. The same for the Sun for no precession, while for the case of precessions, this is true provided that the beam response in the relevant pattern region never exceed ~100 dB (value at 70 GHz). A cycloidal pattern of the spin axis has been preferred mainly beacuse it ensures a constant aspect angle of the spacecraft with respect to the Sun. This leads to minimization of potential seasonal variations in the heat balance of the spacecraft Service Module and of changes in the dissipation efficiency of radiators. Amplitudes higher than 10 o are ruled out by the thermal constraint to prevent direct solar irradiation of thermal radiators or any optical part, while it has been verified that a minimum angle of ~7.4 o is needed to allow full-sky coverage for every detector. With these prescriptions long term thermal instabilities due to the scanning strategy shall be smaller than those induced by the 7% variation of the radiation flux due to spacecraft – Sun time dependence due to Earth orbit ellipticity. References [1] Tauber, Planck Tec. Note, Planck/PSO/2003-001 [2] Bersanelli et al., 1997, A&AS, 121, 393 [3] Cappellini et al., 2003, A&A, 409, 375[4] Dupac & Tauber, 2005, A&A, 430, 363 [5] Burigana et al., 2004, A&A, 428, 311http://www.rssd.esa.int/Planck