Campuses:
| Name | Description | Effect | References | Mitigation | SRF | Simulation Capability |
|---|---|---|---|---|---|---|
| Pair Differencing | ||||||
| Bandpass mismatch | Edges and shapes of the the spectral filters vary from detector to detector. Can be thought of as spectral-source-dependent gain variation between detectors. | T → P, P → P leakage | HFI 2015 : Calibration & Maps, HT, GP, RB Bandpass paper, Planck low-ell, ECO-Systematics | a) Calibration b) The Planck papers give a prescription for correction and a map-making algorithm 'SRoll'. Ranajoy(RB) has developed a technique to filter out the spurious signal demonstrated in ECO paper and the independent paper will be out shortly. c) full I/Q/U maps for individual detectors d) polarization modulation | 4 | Recent CORE experience: based on different pre-integrated tophat bands, all tools are available. Current sets of all-sky simulations are based on top-hat bands. Andrea Zonca working on on-the-fly band integration for TOAST which should be done soon. |
| Beam mismatch | Beam shapes differ between two detectors which are pair-differences to measure P. | T→P, P→P | EH Quickpol, CW Beams, CW, Scan Strategy, ECO-Systematics Shimon et al, 2008 BICEP/Keck VII (Tolan), GR FEBeCoP, Planck LFI beams, Planck HFI beams | a) Optical modelling, Point Source Measurement b) Analytical techniques demonstrated by Eric Hivon(EH) using Quickpol and by Chris Wallis(CW). Real space convolution having knowledge of the beams and correction demonstrated in ECO-Systematics by RB(independent paper expected soon). c) purification methods for cleaning E leaked into B d) polarization modulation e) More exact techniques demonstrated by Sanjit Mitra, Graca Rocha (GR) and Kris Gorski with FEBeCoP, pixel space convolution used in Planck to generate large number of MCs and propagate beam effects from timelines to maps. Estimate the impact of neglecting cross-polar beam, etc.. | 4 | Quickpol was used for CORE study, that is easily available. Takes arbitrary scan information and creates leakage. This is a fast code that once the hits/pixel and oritentations/pixel are calculated from a timeline, other parameters like beam shapes can be done very fast (at least for Planck). Can be used to MC over beam shapes or relative gains, cross pol, etc. Not publicly available yet, but in principle fairly portable. Ranajoy has a real space pipeline available that might be more useful in some cases of combining multiple effects. TOAST uses libConviqt for on-the-fly beam convolution. |
| FEBeCop is a fast code that once the hits/pixel and orientations/pixel are calculated and optical beam maps are provided, cross-polar component etc, generates the polarised tensorial beam at each pixel, it contracts the beam tensors and I,Q,U maps in pixel space to generate beam convolved maps. It was used to generate thousands of Monte Carlos for planck. It also generates the TT, EE, and BB beam window functions. Quickpol, LevelS (a predecessor of TOAST also based on Conviqt) and FEBeCop were compared and cross-validated in Planck. Quickpol, Ranajoy’s code and TOAST also were compared and cross-validated during the CORe proposal study | ||||||
| Gain Mismatch | Gain mismatch between detectors in pair differencing creates T→P leakage | T→P | Rosset et al 2010 (Planck HFI) | (see Gain Stability below) | this effect is mostly degenerate with Gain Stability and other mismatch effects, so not rated. | |
| Optics | ||||||
| Instrumental Polarization: Sky T → P | T→P leakage by optical elements in the instrument (reflectors, lenses) | T→P | a) Physical Optics Simulations, Unpolarized point sources scans b) Didier thesis (2016) for mitigation technique | 3 | Only any time-varying effect would be problematic. The constant leakage will just be part of the calibration process, as well as identifying cross-pol and polarization efficiency. We should simulate it just to check. See equation 7.2 in CORE Mission paper | |
| Instrumented emitted T → P through IP | T→P leakage by optical elements in the instrument (reflectors, lenses) acting on instrument unpolarized emissions | 1/f in polarization | 4 | Variation in the instrument temperatures (possibly scan-synchronous) causes additive 1/f in polarization. This can be correlated amongst detectors. | ||
| Cross Polarization | Q↔U rotation by the optical elements in the instrument | E→B | Physical Optics Simulations, Polarized point source scans, EB power spectrum | 3 | See above, this is a calibration issue (need a known polarized source though) | |
| Sidelobes: Reflector Spillover | Rays / beam edges missing the primary or secondary mirrors. Can propagate to the sky before completing all of their intended “bounces” and/or terminate on non-reflective surfaces adding extra loading and picking up signals far off the main beam axis (i.e. the galaxy). Spillover can be highly polarized. | Extra loading and spurious polarization power | Tauber 2010 | Beam Maps, Physical Optics Simulations | 5 | All the beam effects will need to be accurately modeled, and will be a large effect. GRASP expert really needed. Beam Calibration on the ground impossible because no way of cooling the instrument. Details on the Inflation Probe Optics (talk with imager group) would be useful to further this conversation and assess effects (secondary optics in the instrument will make this worse) |
| Sidelobes: Stray reflections | Reflections (sometimes coupling to what are assumed to be absorptive elements) can send rays to off-axis angles | beam shape | QUIET instrument paper | Optical Modelling, Beam maps | 5 | |
| Sidelobes: Surface roughness | Random surfacer errors remove power from the main beam and redistribute it over large angles. The process results in a smooth beam shoulder often referred to as a Ruze envelope. | beam shape | Planck 2013 beam paper | Measured with interferometric/photogrammetry techniques | 5 | This can also encompass pollution during the mission |
| Beam ghosting | Off-axis beam response caused by internal reflections. Often associated with non-idealities in for example anti-reflection coatings or filters. | beam shape, T→P | Bicep systematics, SPIDER systematics, Sean Bryan thesis | Far field beam maps | 5 | |
| Scattering | Diffuse scattering of light on various optical elements, in some cases due to surface impurities. This is an issue for ACTPol | beam shape, T→P | Chiang thesis | Detailed modelling and testing of each cold optical element | 5 | |
| Sidelobes: diffraction | Light diffracts on edges that vary on scales that are comparable or smaller than the wavelength of light in question. Diffraction often results in a polarized signal. | T→P | Fraisse et al SPIDER, Essinger thesis ABS | Can be mitigated by adding smoothly varying transitions. | 5 | |
| Pointing systematic error & distortion | Detector pointing at position different from flight telemetry data. | Level of leakage demonstrated in the ECO paper for different cases. | Technique similar to bandpass leakage correction by filtering out spurious signal developed by RB, to be published soon. | 4 | This needs to be measured during flight, to assess offset between star trackers and detectors over time / in multiple pointing configuration. We should model before flight what level is ok | |
| Pointing Jitter | Random Pointing Error mixes T, E and B at high \ell | T→P, P→P | Miller et al 08 for impact on lensing, Didier thesis for E->B | 3 | This will translate on a requirement on star trackers and gyroscopes. Can get analytic requirement from E→B leakage and lensing (see quoted papers). With 1.4m mirror, Inflation Probe should behave roughly like Planck, for which this was not a problem | |
| Chromatic beam shape | Beam shape is a function of source SED: measured using a planet, used to build a window function to correct CMB power spectrum. | T→P, P→P, miscalibration of window function | Planck 2013 VII | Physical optics modeling | Note: this was a most important effect for Planck. Jacques says he needs to re-visit this, he doesn't remember this being fundamental. | |
| Detectors | ||||||
| Time Constant Accuracy & Stability | Uncertainty in time constants of the detectors (either from measurements or variation with time) creates uncertainty in polarization angle, pointing and beam size. Incoming power variation creates loop gain variation hence time constant variation. For a constant spin-rate spacecraft this effect is exactly degenerate with the beam shape. | P→P, beam size error, pointing error | Irwing & hilton, | Stimulator flash, Lab measurement, must be measured on-orbit because its difficult to create ground-based simulations of on-orbit. Time response measurement is degenerate with many other parameters. | 4 | |
| Readout Cross Talk | Power in one detector leaks into other detectors | T→P, P→P, T supression, high-\ell beam distorsion, frequency mixing | Dfmux: Dobbs et al 2011, Darcy Barron thesis | Electrical model, Measure mixing matrix with point source observation, cosmic ray events. Likely to be larger for massively multiplexed readouts than it was for Planck (for example, Polarbear sees this at the percent level). Has not been limiting for any experiment. | 4 - large effect that we think we understand well | make sure we have a number of blind detectors to make this measureable. |
| Polarization Angle | Uncertainty in polarization calibration creates E→B leakage | E→B | Sacrifice cosmic birefringence limits to zero TB,TE,EB Improvement of calibration sources: difficult across broad bands and large. Onboard calibrator? | 5 | Compact polarized source measurements should be revisitied? Hardware solution potentially? | |
| Gain Stability | Time-variation of gain (dT/dI) or sensitivity (dP/dI) coming from various sources (bath temperature variations, optical power drifts | Miscalibration, T→P, P→P | Irwing & hilton, der Kuur et al | Repeated calibration measurements with point sources, measure loop gain variations (IV curves), estimate optical power drifts through temperature monitoring, bias steps dipole provided best on-orbit gain measurement | 5 | This was simulated in the COR paper, but not propagated through to errors in science. The answer is dependent on the scan precession angle. Is this in the litebird white paper? All implemented in TOAST so could be adapted to probe. |
| Detector Non-Linearity | Detector has limited range to convert incoming power to current (i.e. gain change with large power variations) | Miscalibration | Irwing & hilton, Rostem et al, Salatino et al, Satoru et al PB | Detector design with high loop gain | 3 | Possibly simulate the detector physics as part of TOAST. Polarbear/Ebex had a waveplate-synchronous load modulation that created nonlinearity. Is dipole vs. DC load on-orbit a similar order of magnitude? to be checked. Polarbear code could potentially give order of magnitude effects for some range of detector parameters. |
| Residual correlated cosmic-ray hits | Detectors used to form a difference have correlated cosmic ray hits below detection threshold | Correlated noise / additive polarization power | ] | Thermal design of detectors to minimize effect | 3 | Noise needs to be measured and simulated no matter the level of correlation-detector to detector; this is just an additional component. Not much to do beyond what we already need to do for an analysis, and not a big worry. |
| ADC non-linearity | ] | 2 | This should not be a problem at all. we just need to make sure the readout design and the ground test plan incorporates lessons learned from Planck HFI and LFI. | |||
The importance of a good scanning strategy and focal plane design is to be emphasised and their design will be influenced by their effect on systematics.