Fast Solar Polarimeter Alex Feller Francisco Iglesias Nagaraju Krishnappa Sami K. Solanki 013-10-01 1
FSP in a nutshell Novel ground-based solar imaging polarimeter developed by MPS in collaboration with the MPG semiconductor lab and PNSensor corp. Funded by Max Planck society and European Commission (SOLARNET) Based on fast, low-noise pnccd sensor and ferro-electric liquid crystals (FLCs) for polarization modulation Polarimetric sensitivity: 0.01% Targeted mainly at statistical studies of weak photospheric and chromospheric polarization signals at high spatial or temporal resolution Development in phases: Phase I (011-013): proof of concept with small pnccd prototype (64 x 64 pixels), single-beam setup Phase II (014-016): Full-scale, science-ready instrument with two 1k x 1k pnccds 013-10-01
Main scientific focus of FSP Study of ubiquitous small-scale magnetic processes in the quiet Sun photosphere and chromosphere radiative processes (scattering polarization) on smaller spatial scales The limited photon flux suggests a statistical approach to reach an increased polarimetric sensitivity by Feature classification in high-resolution Stokes images, spatially binning into classes of pixels (e.g. granules / intergranular lanes, cf. Snik et al. 010) Feature tracking in time of highly dynamic structures, e.g. in the chromosphere 013-10-01 Predicted small-scale scattering polarization in Sr I 460.7 nm, based on MHD simulations (Trujillo Bueno & Shchukina 007) 3
Main scientific focus of FSP Study of ubiquitous small-scale magnetic processes in the quiet Sun photosphere and chromosphere radiative processes (scattering polarization) on smaller spatial scales The limited photon flux suggests a statistical approach to reach an increased polarimetric sensitivity by 013-10-01 Feature classification in high-resolution Stokes images, spatially binning into classes of pixels (e.g. granules / intergranular lanes, cf. Snik et al. 010) Feature tracking in time of highly dynamic structures, e.g. in the chromosphere Times series, Ca II H line core, SST 013 (credits: M. Van Noort) 4
Polarimetry basics Incoming pol. state Polarization is a phase property of light whereas detectors are sensitive to intensity only Modulator and analyzer transform the incoming polarization state into an intensity modulation Synchronous demodulation with the detector decodes the initial polarization state 013-10-01 Modulator (~100 Hz) Synchronisation Analyzer Detector Intensity = + = Polarization = - = 5
Why fast modulation? In order to detect a polarization signal of 0.01% we have to perform differential intensity measurements at the same level of precision Polarimetry is therefore very sensitive to instabilities during measurement, e.g.: Atmospheric turbulence Vibrations in the instrument 013-10-01 Top panels: Simulated measurement of Zeeman signals of the quiet Sun granulation in Fe I 630. nm with a random image jitter of 0.1 detector pixels rms. Bottom panels: clean reference measurement without jitter. Detector gain fluctuations See also Lites 1987, Judge et al. 004, Casini et al. 01. 6
Polarimetry basics Modulation matrix I = M.S S = (I, Q, U, V): incoming (solar) Stokes vector I = (I1,, In): measured intensities for the different modulator states Polarimetric accuracy Smeas = Mmeas-1.M.S Response matrix: R = Mmeas-1.M Crosstalk: non-diagonal elements of R Polarimetric sensitivity Smeas = Mmeas-1.I =: D.I n σi 1/,i= I, Q,U, V Efficiency: εi = σ =(n Dij ) i j=1 013-10-01 7
Modulation techniques Temporal modulation (single-beam setup) I(t) S Modulator I (t 1) = I (t ) = g ( I +Q) g (I Q) Polarizer I1 = g 1 1 I I 1 1 Q () ( )( ) Detector Qmeas I (t 1) I (t ) Q = = I meas I (t 1 )+ I (t ) I g : detector gain, optical transmission 013-10-01 8
Modulation techniques Temporal modulation (single-beam setup) I(t) S Modulator I (t 1) = I (t ) = g (I +δ I 1 +Q+ δ Q1) g (I +δ I Q δ Q ) Polarizer Qmeas I meas Detector = I 1 I I 1+ I = 1 1 Q+ (δ Q 1+ δ Q )+ ( δ I 1 δ I ) 1 1 I + (δ I 1 + δ I )+ (δ Q1 δ Q ) With disturbances δ Q, δ I, for example from jitter δ r : δi I δr, δq Q δr δ I δq 1 In practice, for slow modulation : 10... 10 I Q 013-10-01 9
Modulation techniques Temporal modulation (single-beam setup) I(t) S Modulator I (t 1) = I (t ) = Polarizer g (I +δ I 1 +Q+ δ Q1) g (I +δ I Q δ Q ) Qmeas I meas Spatial smearing Detector = I 1 I I 1+ I = 1 1 Q+ (δ Q 1+ δ Q )+ ( δ I 1 δ I ) 1 1 I + (δ I 1 + δ I )+ (δ Q1 δ Q ) Polar. error With disturbances δ Q, δ I, for example from jitter δ r : δi I δr, δq Q δr δ I δq 1 In practice, for slow modulation : 10... 10 I Q 013-10-01 10
Modulation techniques Spatial modulation (dual-beam setup) Pol. beamsplitter I1 S Modulator Detector 1 I Detector I1 = I = g ( I + δ I +Q+ δ Q) g+δ g (I +δ I Q δ Q) Qmeas I meas = Differential detector gain or opt. transmission: I 1 I I 1+ I = ( ( δg δg (Q+ δ Q) ( I +δ I ) δg δg g+ (I +δ I ) (Q +δ Q) g+ ) ) g δ 10 3 g 013-10-01 11
Modulation techniques Spatial modulation (dual-beam setup) Pol. beamsplitter I1 S Modulator Detector 1 I Detector I1 = I = g ( I + δ I +Q+ δ Q) g+δ g (I +δ I Q δ Q) Qmeas I meas = Differential detector gain or opt. transmission: δ g 10 013-10-01 I 1 I I 1+ I = ( ( Polar. error δg δg (Q+ δ Q) ( I +δ I ) δg δg g+ (I +δ I ) (Q +δ Q) g+ ) ) 3 1
Modulation techniques Mixed spatial and temporal modulation (e.g. Semel et al. 1993) Pol. beamsplitter I1(t) S Modulator Detector 1 I(t) Detector I 1 (t 1 ) = I (t 1 ) = I 1 (t ) = I (t ) = 013-10-01 g ( I + δ I 1+ Q+δ Q 1) g+δ g (I +δ I 1 Q δ Q1) g ( I + δ I Q δ Q ) g+δ g (I +δ I +Q +δ Q ) Qmeas I meas = I 1 (t 1 ) I 1 (t ) I (t 1)+ I (t ) I 1 (t 1)+ I 1 (t )+ I (t 1 )+ I (t ) = ( )( ) δ I +δ I δg δg g + I + + ( )( ) 4 ( δq δ Q ) g+ δ Q 1+ δ Q δ g δg Q+ + (δ I δ I 1 ) 4 1 1 13
Modulation techniques Mixed spatial and temporal modulation (e.g. Semel et al. 1993) Pol. beamsplitter I1(t) S Modulator Detector 1 I(t) Detector I 1 (t 1 ) = I (t 1 ) = I 1 (t ) = I (t ) = 013-10-01 g ( I + δ I 1+ Q+δ Q 1) g+δ g (I +δ I 1 Q δ Q1) g ( I + δ I Q δ Q ) g+δ g (I +δ I +Q +δ Q ) Qmeas I meas I 1 (t 1 ) I 1 (t ) I (t 1)+ I (t ) I 1 (t 1)+ I 1 (t )+ I (t 1 )+ I (t ) = δ Q 1+ δ Q δ g δg Q+ + (δ I δ I 1 ) 4 ( ) ( ) = δ I +δ I δg δg g + I + + ( δq δ Q ) ) ( )( 4 Spatial smearing g+ 1 1 14
Why fast modulation? Simulated seeing induced crosstalk as a function of AO correction and modulation frequency for a single-beam setup (Nagaraju & Feller, Appl. Opt., 01). 013-10-01 15
Why fast modulation? Simulated seeing induced crosstalk as a function of AO correction and modulation frequency for a dual-beam setup (Nagaraju & Feller, Appl. Opt., 01). 013-10-01 16
Why fast modulation? Conclusions: For typical moderate seeing conditions (wind speed ~10 m/s, r0 ~ 10 cm), a polarization modulation frequency of order 100 Hz reduces seeing induced polarization crosstalk below 0.01%. N.B.: AO system does not relax the requirements on the modulation frequency. The seeing induced crosstalk is practically independent of the degree of AO correction. At slow modulation frequency, a slow dual-beam setup only suppresses I --> Q,U,V crosstalk and results in a significant spatial degradation of the images. 013-10-01 17
FSP first-light campaign First test of the prototype instrument at the German VTT on Tenerife in June 013 Spectrograph mode, low spatial resolution (0. - 0.4 arcsec/pixel) Focus on functional and performance tests at different modulation frequencies under different atmospheric seeing conditions Measurements Zeeman diagnostics in Fe I lines (55.0 nm, 630. nm), Sunrise II co-observations Scattering polarization at the solar limb (Ca I 4.7 nm, Sr I 460.7 nm) 013-10-01 18
FSP first-light campaign: setup FSP modulator, mounted on top of the spectrograph entrance slit 013-10-01 FSP camera with re-imaging optics, mounted at a spectrograph exit port 19
Some test results Scan of active region (NOAA 1176) in Fe I 55 nm Greyscales: 0.-1.3 <I>, ± 0. for Q/I, U/I, V/I 013-10-01 0
Some test results Scattering polarization in Ca I 4.7 nm at µ = 0.15 013-10-01 1
Some test results Scattering polarization in Ca I 4.7 nm at µ = 0.15 The second solar spectrum (Gandorfer 00) 013-10-01 FSP
Some test results Scattering polarization in Sr I 460.7 nm at µ = 0.15 013-10-01 3
Photon budget Assumptions: Diffraction limited critical sampling (~ 1 pixels for PSF core) Throughput: 10% Exposure time:.5 ms (400 fps) Spectral sampling: 15 må / pixel Polarimetric efficiency: 0.5 Example Ca II K 393 nm Fe I 55 nm Ca II 854. nm Intensity phot / (s m nm sterad) 1.5 101 10 1.3 10 Flux per pixel and frame 5 e- 300 e- 00 e- No. of pixels to average for 0.01% polarimetric sensitivity, after 1s integration 40'000 3300 (4% of detector area) 013-10-01 5000 4
pnccd Main specifications Sensor size (pixels) Phase I Phase II 64 x 64 104 x 104 Pixel size [µm] 48 36 Max. framerate 850 fps 400 fps QE > 90% 550 800 nm 380 nm 650 nm (tbc) Readout noise 3 e- ENC Non-linearity < 0.1% after calibration Duty cycle 0.95 013-10-01 0.90 Schematic layout of the pnccd sensor and readout electronics (Hartmann et al. 006) 6
pnccd Main specifications Sensor size (pixels) Phase I Phase II 64 x 64 104 x 104 Pixel size [µm] 48 36 Max. framerate 850 fps 400 fps QE > 90% 550 800 nm 380 nm 650 nm (tbc) Readout noise 3 e- ENC Non-linearity < 0.1% after calibration Duty cycle 0.95 013-10-01 Schematic cross-section through the pnccd sensor along one transfer channel (Hartmann et al. 006) 0.90 7
pnccd Quantum efficiency Measured QE of the pnccd prototype, and comparison with theoretical models. 013-10-01 8
pnccd - frame transfer correction Iglesias et al. 014, in prep. Shuttering is difficult at frames rates of 400 fps or higher. We work without shutter! Numerical correction of artifacts due to finite frame transfer time (34 µs) finite modulator transition times (~50 µs) 013-10-01 Uncorrected modulation states recorded of a high-contrast target; 100% linear polarization, 700 fps 9
pnccd - frame transfer correction Iglesias et al. 014, in prep. Shuttering is difficult at frames rates of 400 fps or higher. We work without shutter! Numerical correction of artifacts due to finite frame transfer time (34 µs) finite modulator transition times (~50 µs) 013-10-01 Same measurement after frame transfer correction. 30
pnccd - frame transfer correction Iglesias et al. 014, in prep. where: 013-10-01 31
pnccd - frame transfer correction Iglesias et al. 014, in prep. Normalized residual error vs. number of frames where: 013-10-01 3
Modulator SOLIS / ZIMPOL design based on FLCs static zero-order retarders Pol. Beamsplitter Optimization process (Gisler 006): Temperature controlled (± 0.1 C) Static retardances specified, following an optimization step based on measured FLC retardances Angle optimization of all 4 components using a merit function based on wavelength-dependent pol. efficiencies FLCs (variable retarders) Static retarders Polarizing beamsplitter 013-10-01 33
Modulator SOLIS / ZIMPOL design based on FLCs static zero-order retarders Pol. Beamsplitter Optimization process (Gisler 006): Static retardances specified, following an optimization step based on measured FLC retardances Angle optimization of all 4 components using a merit function based on wavelength-dependent pol. efficiencies Temperature controlled (± 0.1 C) 013-10-01 Component Birefringence Δn d [nm] Opt. axis pos. angle [deg.] FLC 1 10 nm -71.8 Static retarder 1 60 nm 6.7 FLC 50 nm -41.5 Static retarder 19 nm 64.8 34
FSP performance Pol. efficiency at 630 nm vs. modulation frequency Mod. frequency [Hz] Pol. efficiency at 5 Hz vs. wavelength Wavelength [nm] 013-10-01 35
FSP performance Nagaraju et al. 014, in prep. Analysis of pol. efficiency with modulator switched off (Fe I 630. nm) Mod. freq. [Hz] Quiet Sun Pore region 5 50 100 013-10-01 37
Lessons learned so far... The small FSP prototype has performed reliably at the VTT during its first-light campaign in June. Shutterless operation with post-facto frame transfer correction works sufficiently well. Room for improvement taking into account the finite FLC response. Polarimetric efficiency close to theoretical expectations. Stable response in time thus requiring less frequent calibration. A frame rate of order 400 fps (modulation frequency 100 Hz) is crucial for observing high contrast targets on the Sun at 0.01% noise level. Polarimetric sensitivity of 0.01% - 0.0% is currently reached. However, at a noise level below 0.1% we see some artifacts, related to modulator and camera, which need further analysis. Telescope polarization compensation needed! 013-10-01 39
What's next? Phase I: continued work with small prototype Second VTT observing campaign in November, using the TESOS filtergraph instrument GREGOR spectrograph campaign in 014 (tbc) Phase II: development of full-scale, science-ready instrument 013-014 015 Early 016 013-10-01 MPG semicond. lab Development of 1k x 1k pnccd sensors MPS Camera housing, dual-beam setup MPS System integration and verification First light at telescope 40