The GREGOR Fabry-Pérot Interferometer and its companion the Blue Imaging Solar Spectrometer
Klaus G. Puschmann, Carsten Denker, Horst Balthasar, Rohan E. Louis, Emil Popow, Manfred Woche, Christian Beck, Thomas Seelemann, Reiner Volkmer
aa r X i v : . [ a s t r o - ph . I M ] F e b Optical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8)
The GREGOR Fabry-P ´erot Interferometer and its com-panion the Blue Imaging Solar Spectrometer
Klaus G. PuschmannCarsten DenkerHorst BalthasarRohan E. LouisEmil PopowManfred Woche
Leibniz-Institut f ¨ur Astrophysik Potsdam14482 PotsdamGermanyE-mail: [email protected]
Christian Beck
Instituto de Astrof´ısica de Canarias38205 La Laguna, TenerifeSpain
Thomas Seelemann
LaVision37081 G ¨ottingenGermany
Reiner Volkmer
Kiepenheuer-Institut f ¨ur Sonnenphysik79104 FreiburgGermany
Abstract.
The GREGOR Fabry-P´erot Interferometer (GFPI) is one ofthree first-light instruments of the German 1.5-meter GREGOR solartelescope at the Observatorio del Teide, Tenerife, Spain. The GFPI al-lows fast narrow-band imaging and post-factum image restoration. Theretrieved physical parameters will be a fundamental building block forunderstanding the dynamic Sun and its magnetic field at spatial scalesdown to ∼ km on the solar surface. The GFPI is a tunable dual-etalon system in a collimated mounting. It is designed for spectrometricand spectropolarimetric observations between 530–860 nm and 580–660 nm, respectively, and possesses a theoretical spectral resolutionof R ≈ , . Large-format, high-cadence CCD detectors with so-phisticated computer hard- and software enable the scanning of spec-tral lines in time-spans equivalent to the evolution time of solar features.The field-of-view (FOV) of ′′ × ′′ covers a significant fraction of thetypical area of active regions in the spectroscopic mode. In case ofStokes-vector spectropolarimetry, the FOV reduces to ′′ × ′′ . Wepresent the main characteristics of the GFPI including advanced andautomated calibration and observing procedures. We discuss improve-ments in the optical design of the instrument and show first observa-tional results. Finally, we lay out first concrete ideas for the integrationof a second FPI, the Blue Imaging Solar Spectrometer (BLISS), whichwill explore the blue spectral region below 530 nm. c (cid:13) Paper 121432SSPR compiled June 5, 2018.
Solar physics has made tremendous progress during recentyears thanks to numerical simulations and high-resolutionspectropolarimetric observations with modern solar tele-scopes such as the Swedish Solar Telescope, the Solar Opt-ical Telescope on board the Japanese HINODE satellite, and the stratospheric Sunrise telescope. Taking the natureof sunspots as an example, many important new observa-tional results have been found, e.g., details about the bright-ness of penumbral filaments, the Evershed flow, the dark-cored penumbral filaments, the net circular polarization, andthe moving magnetic features in the sunspot moat. Tele-scopes with apertures of about 1.5 m such as the GRE-GOR solar telescope
4, 5 or the New Solar Telescope
6, 7 willhelp to discriminate among competing sunspot models andto explain the energy balance of sunspots. New results onthe emergence, evolution, and disappearance of magneticflux at smallest scales can also be expected. However, these1.5-meter telescopes are just the precursors of the next-generation solar telescopes, i.e., the Advanced TechnologySolar Telescope and the European Solar Telescope, whichwill finally be able to resolve the fundamental scales of thesolar photosphere, namely, the photon mean free path andthe pressure scale height.Fabry-P´erot interferometers (FPIs) have gained impor-tance in solar physics during the last decades because theydeliver high spatial and spectral resolution. A growing num-ber of such instruments is in operation at various telescopes. Although most of these instruments have been initially de-signed only for spectroscopy, most of them have now beenupgraded to provide full-Stokes polarimetry.
10, 11, 12
The Uni-versit¨ats-Sternwarte G¨ottingen developed an imaging spec-trometer for the German Vacuum Tower Telescope (VTT) inthe early 1990s. A fundamental renewal of the G¨ottingen FPI during thefirst half of 2005 was the starting point of the developmentof a new FPI for the 1.5-meter GREGOR telescope. Newnarrow-band etalons and new large-format, high-cadenceCCD detectors were integrated into the instrument, accom-panied by powerful computer hard- and software. From 2006to 2007, the optical design for the GREGOR Fabry-P´erotInterferometer (GFPI) was developed, the necessary opticalelements were purchased, and the opto-mechanical mountswere manufactured. An upgrade to full-Stokes spectropo-larimetry followed in 2008.
16, 17
In 2009, the Leibniz-Institutf¨ur Astrophysik Potsdam took over the scientific respon-sibility for the GFPI, and the instrument was finally in-stalled at the GREGOR telescope. During the commission-ing phase in 2011, the software was prepared for TCP/IPcommunication with external devices according to the De-vice Communication Protocol (DCP ) and three computer-controlled translation stages (two filter sliders and one mir-ror stage) were integrated into the GFPI. This permits au-tomated observing and calibration procedures and facilitateseasy operations during observing runs. The present article isa modified version of a contribution to the SPIE confer-1 ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8) ence ‘Ground-Based and Airborne Instrumentation for As-tronomy IV’, which took place in Amsterdam in June 2012.A complementary description of the GFPI has been pub-lished in the meanwhile in a special issue of Astronomis-che Nachrichten/AN (Volume 333, Issue 9) dedicated to theGREGOR solar telescope. The 1.5-meter GREGOR telescope is the largest Europeansolar telescope and is designed for high-precision measure-ments of dynamic photospheric and chromospheric struc-tures and their magnetic field. Some key scientific topics ofthe GREGOR telescope are: the interaction between convec-tion and magnetic fields in the photosphere, and the enig-matic heating mechanism of the chromosphere. The inclu-sion of a spectrograph for stellar activity studies and thesearch for solar twins expand the scientific usage of theGREGOR telescope to the nighttime domain. The GREGOR telescope replaced the 45-cm Gregory-Coud´e Telescope, which had been operated on Tenerife since1985. The construction of the new telescope was carriedout by a consortium of several German institutes, namely,the Kiepenheuer-Institut f¨ur Sonnenphysik Freiburg, theLeibniz-Institut f¨ur Astrophysik Potsdam (AIP), and the In-stitut f¨ur Astrophysik G¨ottingen. In 2009, the Max-Planck-Institut f¨ur Sonnensystemforschung in Katlenburg-Lindautook over the contingent of the latter institute. The consor-tium maintains partnerships with the Instituto de Astrof´ısicade Canarias in Spain and the Astronomical Institute of theAcademy of Sciences of the Czech Republic in Ondˇrejov.The GREGOR telescope is an alt-azimuthally mountedtelescope with an open structure and an actively cooledlight-weighted Zerodur primary mirror. The completely re-tractable dome enables wind flushing through the telescopeto facilitate cooling of telescope structure and optics.
26, 27
The water-cooled field stop at the primary focus providesa field-of-view (FOV) with a diameter of 150 ′′ . The lightis reflected via two elliptical mirrors and several flat mir-rors into the optical laboratory. Passing the GREGOR adap-tive optics system (GAOS), the light is finally distributedto the scientific instruments. The removable GREGOR Po-larimetric Unit (GPU), which was developed at the AIP, islocated near the secondary focus and ensures high-precisionpolarimetric calibration. In addition, a telescope polarizationmodel has been developed, since the instrumental polar-ization is time-dependent due to the alt-azimuthal mount.Three first-light instruments have been commissioned in2011 and passed through science verification in 2012: theGRating Infrared Spectrograph (GRIS), the Broad-BandImager (BBI), and the GFPI. GRIS and GFPI can be usedsimultaneously. A dichroic beamsplitter directs wavelengthsabove 650 nm to the spectrograph, whereas all shorter wave-lengths are reflected towards to the GFPI. In the near future,this beamsplitter can be exchanged with a different one witha cutoff above 860 nm.
Optical design
The GFPI is mounted on five optical tables and is protectedby an aluminum housing to prevent pollution of the optics by dust and to reduce stray-light. The optical layout of theinstrument is shown in Fig. 1. Behind the science focus F4 ,four achromatic lenses TL1 , TL2 , HL1 , &
HL2 create twomore foci F5 & F6 and two pupil images P1 & P2 in thenarrow-band channel NBC . The etalons
FPI1 & FPI2 areplaced in the vicinity of the secondary pupil in the collimatedlight beam. A neutral density filter
NDF between the twoetalons with a transmission of 63% reduces the inter-etalonreflexes. The beam in the
NBC is folded twice by M2 & M5 ata distance of 500 mm and 400 mm from F5 and HL2 , respec-tively, to minimize the instrument envelope. Two field stops
FS1 & FS2 prevent an over-illumination of the detectors,where the secondary adjustable field stop especially avoidsan overlap of the two images created by the removable dual-beam full-Stokes polarimeter. A beamsplitter cube
BS2 near F5 directs 5% of the light to the broad-band channel BBC .There, the two achromatic lenses
TL3 & TL4 are chosen suchthat the image scale of the detectors
CCD1 & CCD2 is ex-actly the same. The simultaneous recording of broad- andnarrow-band images is a prerequisite for post-factum imagereconstruction. Several image reconstruction techniques arealready implemented in a data pipeline presently being de-veloped at the AIP. A detailed description of these tech-niques and a comparison of related results has been recentlypublished. A dichroic beamsplitter
BS1 just behind thescience focus F4 sends the blue part of the spectrum (be-low 530 nm) to the Blue Imaging Channel BIC . One-to-oneimaging with the lenses
TL1a & TL2a provides the option ofrecording broad-band images in the wavelength range 380-530 nm with two pco.4000 cameras
CCD1a & CCD2a be-hind beamsplitter
BS2a in parallel to
NBC and
BBC or GRISobservations. This imaging channel will be replaced in thefuture by the Blue Imaging Solar Spectrometer (BLISS),which is described in Section 5.Three computer-controlled precision translation stagesfacilitate automated observing sequences. The two stages
FSL1 & FSL2 are used to switch between two sets of inter-ference filters. The filters restrict the bandpass of
BBC and
NBC to a full-width at half-maximum (FWHM) of 10 nmand 0.3–0.8 nm, respectively. The near-focus position of thesliders excluded the usage of a filterwheel because of therequired positioning accuracy in the order of micrometers.Only two different interference filters can be mounted oneach slider because of the limited space on the optical tables.The pre-filters of each channel can be tilted to optimize thewavelength of the transmission maximum. At present, theoptical elements of the GFPI are not super-achromatic. Thus,a chromatic focus shift restricts the maximum spectral dis-tance between two different pre-filters/spectral lines to about100 nm. A third stage inserts a deflection mirror M1 intothe light path to take calibration data with a continuum-lightsource for spectral calibration purposes (white-light chan-nel). A laser/photo-multiplier channel for finesse adjustmentof the etalons completes the optical setup. Cameras, etalons, and control software
The GFPI data acquisition system consists of two ImagerQE CCD cameras with Sony ICX285AL detectors, whichhave a full-well capacity of 18,000 e − and a read-out noiseof 4.5 e − . The detectors have a spectral response from 320–900 nm with a maximum quantum efficiency of ∼
60% at2 ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8)
Table 1Table 2 Table 3 Table 4 Table 5 F4FS1BS1TL1P1TL2F5FS2FSL1M2 HL1FPI1NDFP2 FPI2HL2M5POL CCD1/F6 BS2 TL3TL4FSL2 CCD2/F6 TL1aP1aTL2a BS2aIF1aCCD1a/F5aIF2aCCD2a/F5aM1 CL2FCL FSCLCL1WLSM4MP M3LFSLPFLaser LL1 LL2
Fig. 1
Up-to-scale drawing of the GFPI, including narrow-band channel (NBC, marked as red beam), broad-band channel (BBC, salmon beam), blueimaging channel (BIC, blue beam), and all auxiliary channels, i.e., white-light channel (off-wight beam) and laser/photo-multiplier channel (yellow beam).The light from the telescope enters the instrument at the science focus F4. Black arrows denote the light direction in the individual channels. The opticalelements of BIC are labeled with an extra “a”. CCD1, CCD2: Imager QE detectors; CCD1a, CCD2a: pco.4000 detectors; FPI1, FPI2: narrow-band etalons;NDF: neutral density filter; TL1 ( f = 600 mm, d = 63 mm), TL2 ( f = 250 mm, d = 40 mm), TL1a ( f = 500 mm, d = 63 mm), TL2a ( f = 500 mm, d = 80 mm), TL3 ( f = 400 mm, d = 63 mm), TL4 ( f = 600 mm, d = 63 mm), HL1 ( f = 1000 mm, d = 80 mm), HL2 ( f = 1500 mm, d = 100 mm):achromatic lenses; CL1 ( f = 300 mm, d = 63 mm), CL2 ( f = 150 mm, d = 40 mm): plano-convex lenses; M1, M3, M4: removable folding mirrors; M2,M5: fixed folding mirrors (60 mm ×
85 mm); F4, F5, F6, F5a,FCL: foci; P1, P1a, P2: pupil images; BS1, BS2, BS2a: beamsplitters (40 mm ×
40 mm); FS1,FS2, FSCL: field stops; WLS: white-light source (slide projector); FSL1, FSL2: filter sliders; IF1a, IF2a: interference filters; POL: full-Stokes polarimeter; LL1,LL2: laser lenses; LPF: laser polarization filter; LFS: laser field stop; and MP: photomultiplier.
550 nm. The chips have × pixels with a sizeof 6.45 µ m × µ m. The image scale at both camerasis 0.0361 ′′ pixel − , which leads to a FOV of ′′ × ′′ inthe spectroscopic mode. Due to the dual-beam configuration,the FOV shrinks to ′′ × ′′ when using the polarimet-ric mode of the instrument. The maximal blueshift due tothe collimated mounting of the etalons is about 4.3 pm at630 nm. The cameras are triggered by analog TTL signals ofa programmable timing unit (PTU). The analog-digital con-version of the CCD-images is carried out with 12-bit resolu-tion. The data recorded by the cameras are passed via digitalcoaxial cables to the GFPI control computer and are storedon a RAID 0 system. Two pco.4000 cameras in stock at the observatory can be used for imaging in the
BIC . The pco.4000 camera hasa full-well capacity of 60,000 e − and a read-out noise of11 e − . The detector has a spectral response from 320–900 nm with a quantum efficiency of ∼
32% ( ∼ × pixels, each with asize of 9 µ m × µ m. The image scale at both cameras is0.031 ′′ pixel − . Because of a vignetting of the beam by BS1 and
BS2a only × pixels can be used resulting ina FOV of ′′ × ′′ . Three interference filters for Ca II H λ . nm, Fraunhofer G-band λ . nm, and blue contin-uum λ . nm are available, which have a FWHM = 1 nmand a transmission better than 60%.The two GFPI etalons manufactured by IC Optical Sys-tems (ICOS) have a diameter ⊘ = 70 mm, a measured fi-3 ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8) Tab.
Channel λ [nm] ∆ t [ms] I [counts] PHA ( F4 ) PH ( F3 ) PH ( F2 ) TG QS SP PF
NBC ( × binning) 543.3 40 2400 x x – x x x xBBC ( × binning) 543.3 40 – x x – x x x –NBC ( × binning) 543.3 30 3500 – – – x x – xNBC ( × binning) 557.6 40 2100 – x – x x – xNBC ( × binning) 617.3 10 3200 – x – x x – xCa II H 396.8 15 9230 x x x x x x –G-band 430.7 6 10350 x x x x x x –Blue continuum 450.6 3 11300 x x x x x x –Note. — Central wavelength λ , exposure time ∆ t , and mean intensity I for all observations ( PHA : pinhole array, PH :pinhole, TG : target, QS : quiet Sun, SP : sunspot, and PF : pre-filter). nesse F ∼ , spacings d = 1 . and 1.4 mm, and a high-reflectivity coating ( R ∼ %) in the wavelength rangefrom 530–860 nm. The resulting FWHM of the instrumentis on the order of 1.9–5.6 pm and leads to a theoreticalspectral resolution of R ∼ , . However, the etalonspacings are not optimal because of historical reasons, thusoff-band contributions can become critical when using toobroad pre-filters. All etalons are operated by three-channelCS100 controllers manufactured by ICOS. The cavity spac-ings are digitally controlled by the GFPI control computervia RS-232 communication. A thermally insulated box pro-tects pupil and etalons from stray-light and air flows insidethe instrument.The communication between internal (cameras, etalons,and filter and mirror sliders) and peripheral devices (tele-scope, AO system, AO filter wheel, GPU, GRIS, etc.) iscontrolled by the software package DaVis from LaVision inG¨ottingen, which has been adapted to the needs of the spec-trometer.
14, 21
The modification of the software for TCP/IPcommunication with external devices using DCP allows aneasy implementation of automated observing procedures. All observing modes such as etalon adjustment, line finding,flat-fielding, recording of dark, pinhole, and target images,continuum scans, and recording of scientific data are nowautomated.
Polarimetry with the GFPI
The science verification time in 2012 was mostly devotedto spectroscopy. Nevertheless, the GFPI is equipped with adual-beam full-Stokes polarimeter
16, 23 that can be insertedin front of the detector in the
NBC . The polarimeter consistsof two ferro-electric liquid crystal retarders (FLCRs) and amodified Savart-plate. The first liquid crystal acts as a half-wave plate and the second one as a quarter-wave plate at anominal wavelength of 630 nm. The modified Savart-plateconsists of two polarizing beamsplitters and an additionalhalf-wave plate, which exchanges the ordinary and the ex-traordinary beam. With this configuration, the separation ofthe two beams is optimized and the orientation of the astig-matism in both beams is the same so that it can be correctedby a cylindrical lens. For the present set of FLCRs their prop-erties (retardation and fast axis orientation) have been meas-ured so far only for the spectral range 580–660 nm. An au- tomated polarimetric calibration procedure will be integratedduring an upcoming technical campaign in April 2013. Atthat time the GREGOR polarimetric calibration unit will beavailable for the visible-light range. This will be the startingpoint for polarimetry with the GFPI at the GREGOR solartelescope. A future enlargement of the reduced FOV in thepolarimetric dual-beam mode would be possible by replac-ing the Imager QE cameras with modern and large-formatsCMOS cameras, which have also much higher frame rates(cf. Sect. 5). However, this would require an additional re-design of the polarimeter, because presently the FOV is alsolimited by the size of the calcites of the polarimeter.
We took several data sets in a technical campaign from15 May to 1 June 2012 for a first characterization of theGFPI performance. The AO system was not available be-cause of technical problems with the control computer. Thus,our efforts have been restricted to an optimization of the sys-tem and to observations of test data as measures for inten-sity levels, image quality, spectral resolution, stray-light, andother performance indicators of
NBC , BBC , and
BIC . Sev-eral problems related to the RS-232 communication with theFPI controllers were resolved so that a stable finesse is nowachieved over several days. In addition, the timing betweencameras and etalons has been optimized to ensure that im-ages are only taken when the etalon spacing has settled to itsnominal value.
Imaging spectrometric data
Three complete data sets with × binning ( ∼ ′′ pixel − ) were taken on 31 May and 1 June 2012, which in-cluded images of the target TG and pinhole PH mounted inthe AO filter wheel at F3 (telescope focal plane). The ob-servations covered neutral iron lines at 543.4 nm, 557.6 nm,and 617.3 nm. The Fe I λ . nm line had already beenscanned on 27 May at full spatial resolution (0.036 ′′ pixel − )including images of a pinhole array in F4 (science focus atthe entrance of the GFPI). Simultaneous BBC images wererecorded (Tab. 1). Line scans with the GFPI were usuallycarried out using a step width of eight digital-analog (DA)steps, whereas the pre-filter scans were performed with oneDA step. One DA step corresponds to 0.26–0.41 pm in the4 ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8)
Tab.
Filter λ [nm] FWHM [nm] T [%] Binning ∆ t [ms] I [counts] Frame rate [Hz]ANDV11436 543.4 0.4 38 - / - / ×
60 / 100 / 30 1200 / 2000 / 3500 7 / 5 / 111100 BARR9 543.4 0.6 70 - 40 / 60 1700 / 2700 7 / 6ANDV5288 557.6 0.3 40 ×
40 2100 11DV5289 AM-32389 569.1 0.3 45 - 60 / 100 1200 / 2000 6 / 5ANDV9330 617.3 0.7 80 - / - / ×
20 / 60 / 80 1200 / 3400 / 3200 9 / 6 / 16Note. — Continuum intensities I at wavelengths λ for filters with peak transmissions T and exposure times ∆ t . spectral range from 530–860 nm.Broad-band data were taken in the BIC on 26 and 27 May2012 just before the
NBC and
BBC measurements. The ob-serving scheme was the same for all available pre-filters(396.8 nm, 430.7 nm, 450.6 nm). In addition to a few ob-servations of the quiet Sun and a sunspot, images of the tar-get and pinhole in F3 , the pinhole-array in F4 , and the pin-hole mounted in the GPU at F2 are also included. All data inthe BIC were taken with a combination of two spare lenseswith f = 500 mm and 1250 mm because a second lens with f = 500 mm was not available yet. Thus, the image scaleof 0.0124 ′′ pixel − oversamples the diffraction limit at thesewavelengths by a factor of about three. The proper imagescale of 0.032 ′′ pixel − has been obtained in a followingcampaign by an one-to-one imaging. Intensity estimates for the narrow-band channel
The wavelength λ , FWHM, and transmission T of differ-ent NBC filters are summarized in Tab. 2 together with theselected binning, the counts at continuum wavelengths, andframe rates at a given exposure time ∆ T . The results revealthat at full spatial resolution very long exposure times of upto 100 ms are necessary to achieve at least 2000 counts in thecontinuum of most of the measured spectral lines for most ofthe filters with T ∼ % and a FWHM ∼ . –0.4 nm. Asa consequence, one can reach only very low frame rates. A × -pixel binning speeds up the observations and reducesthe exposure times. The situation changes when choosingfilters with higher transmission. The 617.3 nm filter with T ∼ % and a FWHM ∼ . nm yields reasonable framerates and exposure times even without binning. Estimates of the spatial point spread function
Knowledge of the instrumental point spread function (PSF)provides an estimate of both the spatial resolution and the spatial stray-light level.
32, 33
Using a reference such as a pin-hole or a blocking edge in the focal plane, the PSF of all op-tics downstream can be derived. The combined PSF of time-variable seeing, telescope, and post-focus instruments, canalso be derived from the observations of, e.g., sunspots withtheir steep spatial intensity gradients.The PSF of the optical train at the GREGOR telescope rel-evant for the different GFPI channels was calculated basedon reduced images of pinholes located in the focal planes F4 , F3 , F2 , and on sunspot images (see Tab. 1). All availablecalibration images, e.g., target or pinhole images, were aver-aged for each wavelength step or image burst. However, onlya single image was selected for solar observations because of −3 −2 −1 0 1 2 3arcsec0.00010.00100.01000.10001.0000 I/I m a x I/I c Fig. 2
Derivation of the PSF estimate at 430 nm for the pinhole in F3 ( top )and for the complete optical train from a sunspot observation ( bottom ),where the PSF estimates of the four different focal planes are displayed inthe lower left corner. the variable seeing. The pinhole images were normalized tothe maximum intensity I max inside the FOV, whereas thesunspot data were normalized to unity in a quiet Sun re-gion. All data were taken without real-time correction of theAO system and consequently correspond to the static perfor-mance of the optical system. Derivation of the point spread function estimates
To obtain an estimate of the PSF, we took cuts along the x -and y -axes of the CCD across the center of the pinhole ineach pinhole observation. We defined a step function that is5 ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8) E [ % ] diffraction limit Fig. 3
Energy enclosed within the radius r for the PSF estimates at430.7 nm of the four different focal planes. The dashed vertical lines de-note the radius where 90% of the energy is enclosed. The dotted verticalline denotes the diffraction limit at 430.7 nm. assumed to represent the physical extent of the true pinhole.The borders of the step function were set to intersect the ob-served intensity along the cuts at about the 50% level (toppanel of Fig. 2). We constructed a convolution kernel froma combination of a Gaussian (variance σ ) and a Lorentzianfunction (parameter a ) and convolved the step function withthe kernel. A modification of the parameters ( σ, a ) within aspecific range yielded finally the kernel that best matchedthe convolved step function and the observed intensity alongthe horizontal and vertical cuts. The same method was ap-plied to all pinhole observations ( F4 , F3 , and F2 ) to derivea PSF estimate for each focal plane and wavelength. Thesunspot observations were modeled by a similar step func-tion at the two transitions between quiet Sun, penumbra, andumbra (bottom panel of Fig. 2).The resulting PSF estimates of the four focal planes for the BIC data at 430.7 nm are displayed in the lower left corner ofthe bottom panel of Fig. 2. The width of the PSF estimatesis similar for F4 and F3 , where only static optical compo-nents contribute to the PSF, but increases significantly whenpassing to the focal plane F2 that experiences telescope see-ing and seeing fluctuations along the Coud´e train. The PSFderived from the sunspot observations includes all seeing ef-fects and roughly doubles its width relative to the PSF at F2 .PSF estimates for all four focal planes were obtained incase of the BIC data at 430.7 nm, but only for some of the fo-cal planes in case of the
BBC and
NBC data (see Tab. 3). Theoptical performance and the expected stray-light level canbe best quantified from the total energy enclosed in a givenradius, i.e., a radial integration of the PSF. Figure 3 showsthe enclosed energy E for the PSF estimates at 430.7 nm.The curves provide two characteristic values, namely, theenergy enclosed at the diffraction limit E ( r DL ) and the ra-dius r ( E = 90%) at which 90% of the energy is enclosed.The former value gives an estimate of the generic stray-lightwhen subtracted from 100%. The latter value can be usedas generic estimate of the spatial resolution. Table 3 lists thetwo values for all analyzed wavelengths and channels, i.e., BIC , BBC and
NBC at specific wavelengths.A comparison between the value of the diffraction limit
Channel BIC λ . nm BBC λ . nmSampling 0.0124 ′′ pixel − ′′ pixel − (0.0126 ′′ pixel − ) (0.036 ′′ pixel − )Diff. limit 0.072 ′′ ′′ Focal plane
F4 F3 F2
Solar obs.
F4 F3
Solar obs. r ( E = 90%) [ ′′ ] E ( r DL )
90% 73% 58% 55% 85% 74% 52%Channel NBC NBC NBC λ . nm λ . nm λ . nmSampling 0.034 ′′ pixel − ′′ pixel − ′′ pixel − (0.036 ′′ pixel − ) (0.072 ′′ pixel − ) (0.072 ′′ pixel − )Diff. limit 0.091 ′′ ′′ ′′ Focal plane
F4 F3
Solar obs.
F3 F3 r ( E = 90%) [ ′′ ] E ( r DL )
85% 71% 51% 75% 76%
Tab. and the radius where 90% of the energy is enclosed showsthat the optics behind F4 performs close to the diffrac-tion limit. The optics downstream of F3 performs slightlyworse with a total enclosed energy of about 70–75% ofthe diffraction-limited case. This implies a spatial stray-lightlevel of about 25% created by the optics downstream of F3 .The corresponding value at F4 is about 10–15%. The valuesof both the stray-light level and r ( E = 90%) experience aprofound jump when passing to F2 and beyond. Howeverwe note, that all these data were taken at mediocre seeingconditions and without AO correction. Spectral resolution, spectral stray-light, and blue-shift
The spectral resolution and the spectral stray-light insidethe
NBC was estimated by a convolution of Fourier Trans-form Spectrograph (FTS) atlas spectra
34, 35 with a Gaus-sian of width σ and a subsequent addition of a constantwavelength-independent stray-light offset β .
36, 37
This com-ponent of stray-light corresponds to light scattered onto theCCD detector without being spectrally resolved. Therefore,it changes the line depth of observed spectral lines. The con-volved FTS spectra in each wavelength range were com-pared with sets of spatially averaged observational profilesthat either covered the full pre-filter transmission curve oronly the line inside the same range that is usually recordedin science observations (543.4 nm, 557.6 nm, 617.3 nm). Theupper panel of Fig. 4 shows the average observed spectrumat Fe I λ . nm, the original FTS spectrum, and the FTSspectrum after the convolution with the best-fit Gaussian ker-nel and the addition of the stray-light offset. The method hassome ambiguity between σ and β , which can be modified inopposite directions over some range near the best-fit valueswithout significantly degrading the reproduction of the ob-served spectra. Therefore, the values listed in Tab. 4 have anerror of about ± . pm in σ and ± % in β . The stray-lightlevel β inside the NBC is below 10% and the spectral reso-lution is
R ∼ , , which is significantly below the the-oretically expected value of R ∼ , . The dispersionvalues derived from the observed spectra are listed in Tab. 4and correspond to eight DA steps. In addition, we also con-6 ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8) Tab.
Wavelength 543.4 557.6 617.3 630.25 [nm] [nm] [nm]Dispersion (8 DA) [pm] 2.08 2.15 2.36 2.31(2.09) (2.15) (2.36) (2.41)Dispersion (1 DA) [pm] – – – –(0.261) (0.268) (0.296) σ [pm] 2.62 2.07 2.81 1.65 β [%] 6.6 6.5 7.8 14 λ/σ – 207200 269500 219700 381800 λ/ ∆ λ – 88200 114700 93500 162500Max. blueshift [pm] 3.91 3.96 4.49 –(3.72) (3.82) (4.23) (4.32)Note. — Theoretical values are indicated in parentheses. firmed the maximal blueshift induced in the spectra, whichis caused by the etalon mounting inside a collimated beam.The blueshift was determined from a set of observationalflat-field data. All resulting numbers for both dispersion andblueshift are close to the theoretically expected values (seeTab. 4).Because of the surprisingly low spectral resolution, wespent substantial effort to clarify the source of this prob-lem. We tried to match FTS spectra convolved with the Airy-function, resulting from a theoretical simulation based on aleast-squares fit with the effective finesse and an additionalparasitic light contribution as free parameters, with observedprofiles from campaigns performed in July and August 2012.It turned out that a finesse of ∼
16 instead of ∼
46 and an ad-ditional parasitic light contribution of ∼
7% are needed to fitthe observed profile of, e.g., the 543.4 nm line (lower panelof Fig. 4).A finesse of 16 is in clear contradiction to the results ob-tained from the finesse-adjustment by laser light that is per-formed every day at the beginning of each observational run.These adjustments revealed always a finesse of 40-50 forboth etalons. Furthermore, a better spectral resolution usingthe same etalons was achieved at the VTT (
R ∼ , ). Several possibilities which might explain these observationswere explored, e.g., the variation of the finesse between thecenter and periphery of the etalon plates. However, noneof the additional tests were conclusive.A precise re-alignment of the GFPI in 2013 will show, if astricter requirement for the collimation of the beam throughthe etalons together with an adjustment of the parallelismof the etalons with a wider laser beam will result in a higherfinesse. Additionally we will carry out tests in an optical lab-oratory to quantify cavity errors caused by, e.g., the micro-roughness of the etalon surfaces.
39, 40, 41
First observations supported by adaptive optics
First observations with the GFPI at GREGOR were obtainedbetween 28 July and 7 August 2012 with real-time correc-tion of wavefront deformations provided by GAOS. Severaldata sets covering different spectral lines (e.g., 543.4 nm,617.3 nm, 630.2 nm), and different objects and disk positionshave been reduced and analyzed during the last months.Figure 5 and Movie 1 show a detailed example of two- λ [nm]0.40.60.81.0 I n t e n s it y I / I Wavelength λ [nm]0.00.20.40.60.81.0 I n t e n s it y I / I Fig. 4
Top: close-up of the Fe I λ . nm line – observed spectrum ( blackpluses ), FTS atlas spectrum ( red solid line ), and FTS convolved with thebest-fit Gaussian kernel ( blue dashed line ). An appropriate stray-light off-set ( β ) was added after the convolution. The Gaussian kernel in the leftcorner is displayed in arbitrary units. Bottom: comparison of averaged ob-served quiet Sun spectrum ( dotted orange line ) of the Fe I λ . nm line,FTS atlas spectrum ( solid red line ), and FTS-atlas spectrum convolved withan Airy-function ( dash-dot turquoise line ) obtained from a least-squares fitwith the effective finesse and an additional parasitic light contribution asfree parameters ( dashed green line ). dimensional (2D) spectral line scans with the GFPI that wereobserved on 7 August 2012. The narrow-band continuumimage around 630.2 nm is displayed in the left panel ofFig. 5. The spectral scan covered the Fe I λ . nm lineand the nearby teluric O line with 70 steps in total and a stepsize of ∆ λ = 1 . pm (right panel of Fig. 5). Movie 1 showsall narrow-band images (left panel) at the consecutive spec-tral positions denoted by stars (right panel). The penumbralfinestructure and individual umbral dots of the active regionNOAA 11538 located at S22 W24 can be clearly identifiedin the spectral data. While throughout the solar line a signif-icant change of structure and intensity contrast with heightin the solar atmosphere is visible, the narrow-band imagesof the O line just reflect the continuum intensity because ofthe teluric origin of the latter.On 31 July, in total a 22-min time-series of the active re-gion NOAA 11530 located at S24 W18 (heliocentric angle θ = 30 ◦ ) was obtained between 8:04 and 8:26 UT. Because7 ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8) I / I C Average Spectra of Umbra, "QS", FTS @ 630.2 nm[nm]0 5 10 15 20 25 30Mm05101520 M m Narrow Band Image at 48.51 pm from Line Center
Fig. 5
Left panel: Deconvolved continuum narrow-band image of the active region NOAA 11538 at 48.5 pm distance from the line center of the Fe I λ . nm line. Right panel: measured average profiles of quiet sun (red), umbra (blue), and the FTS-atlas profile for comparison. The stars indicate thesampled spectral positions. Movie 1 shows the entire line-scan, i.e., the deconvolved narrow-band images (left panel) at the consecutive spectral positions(right panel) in steps of 1.20 pm. M m −1.5 −1.0 −0.5 0.0 0.5 1.0 1.5LOS velocity [km s −1 ] 11 12 13 14 15 16 17FWHM [pm] 5 6 7 8 9Eqivalent width [pm] I / I I / I I / I Fig. 6
Snapshot of a 22-min time-series of the active region NOAA 11530. Upper panels: speckle-reconstructed broad-band image at 620 nm (left), speckle-deconvolved continuum narrow-band image (middle), and line-core narrow-band image (right) of the Fe I λ . nm line. Bottom panels: line-core velocity(left), full-width at half-maximum (middle), and equivalent width (right). Movie 2 shows the temporal evolution of the individual parameters. the angle between solar south and the center of symmetrywas 33.8 ◦ , the direction from center to limb corresponds tothe x -axis of the displayed images. The spectral scan cov-ered the Fe I λ . nm line with 86 steps of ∆ λ = 1 . pmstep size.Figure 6 shows 2D maps of individual line parame-ters, i.e., normalized continuum intensity, line-core inten-sity, line-of-sight velocity, full-width at half-maximum, andequivalent width together with the speckle-reconstructed broad-band image at 620 nm for one spectral scan. The tem-poral evolution of these line parameters over the entire time-series is presented in Movie 2. In the LOS velocity map,the thinnest radial filamentary structures correspond to ar-eas with almost zero velocity (spines), whereas the flow fil-aments on both center and limb sides are significantly wider(intraspines). Even with some fluctuations in the final imagequality, the evolution of small-scale structures can be wellfollowed, e.g., the inward motion of penumbral grains. The8 ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8) Table 1Table 2 Table 3 Table 4 Table 5 F4FS1BS1TL1 TL1aP1 P1aTL2 TL2aF5FS2 F5a/FS2aFSL1 FSL1aM2 M2aHL1 HL1aFPI1 FPI1aNDF NDFaP2 P2aFPI2 FPI2aHL2 HL2aM5M5aPOL CCD1/F6 POLa CCD1a/F6aBS2 BS2aTL3 TL3aTL4 TL4aFSL2FSL2a CCD2/F6CCD2a/F6aM4MP M3LFSLPFLaserLL1 LL2 M4aMPa M3a LFSaLPFa Laser_aLL1aLL2a
Fig. 7
Up-to-scale drawing for the integration of BLISS into the GFPI system. The design is based on the study presented in Fig. 8. The optical elements ofBLISS are labeled with an extra ‘a’. Apart from TL2a, HL2a, and TL4a, they are identical to those of the GFPI. The NBC of GFPI and BLISS is denoted by ared and blue beam, respectively. For the BBC of both instruments, a salmon (GFPI) and turquoise (BLISS) beam distinguish between the two instruments.Laser/Photomultiplier channels of both instruments are highlighted by a yellow beam. The light from the telescope enters at the science focus F4. Blackarrows denote the light direction inside both instruments. cadence of the observations is high enough to trace nearlycontinuously the temporal evolution.The restored broad- and narrow-band images presentedhere have been obtained by a speckle reconstruction
42, 43 ofthe broad-band raw data and a subsequent speckle deconvo-lution
45, 44 of the narrow-band raw data. In the case of timeseries, the temporal evolution of line parameters in 2D-mapshas been obtained after additional data reduction, e.g., im-age de-rotation/correlation and sub-sonic filtering. In sum-mary, these first observations with the GFPI and GAOS arevery promising and suggest that high-cadence 2D spectro-scopic observations near the diffraction limit are very likelyto be obtained in coming observing campaigns at GREGOR.
The spatial resolution of a telescope scales inversely pro-portional with the observed wavelength. Therefore, observa-tions at short wavelengths (below 530 nm) offer the oppor- tunity to obtain data with higher spatial resolution. There arerelatively few ground-based instruments for spectral obser-vations in the blue spectral region.
46, 47, 48, 49
All of these ob-servations were carried out with slit-spectrographs that per-mit only limited improvements by post-factum restorationtechniques, which are however applicable to Fabry-P´erot-based imaging spectrometers. This motivated the design ofBLISS, which will supplant the blue imaging channel of theGFPI in the near future. Spectral lines in the blue part of the visible spectrum
In the wavelength range covered by BLISS, there are severalspectral lines and two molecular bands (G-band λ nmand CN band-head λ nm) of high scientific interest.All Balmer-lines except H α are at shorter wavelengths than530 nm, and the Ca II H and K lines are the strongest linesin the visible part of the solar spectrum probing the chromo-sphere. Several photospheric resonance lines such as Mg I ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8)
606 P1
500 HL1
400 CCD1F6 BS2
400 50TL4
600 CCD2F6 FS1 F4 30BS1 33 TL1a
350 500 500P1a TL2a
250 500BS2a CCD2aF5a 30 CCD1aF5a (a)(b) (c)
150 TL1a
606 P1a
800 HL1a
700 CCD1aF6a BS2a
400 350 TL4a
900 CCD2aF6a (d)(e) Fig. 8
Schematical optical design of the current GFPI narrow-band channel NBC (a) , broad-band channel BBC (b) , and blue imaging channel BIC (c) . In thefuture, NBC (d) and BBC (e) of BLISS will replace the BIC of the GFPI. The diameter of the beam at all optical surfaces, the focal length of the lenses, andthe positions of other optical elements are given in millimeters. The optical elements of BIC and BLISS are labeled with an extra ‘a’. The achromatic lensesTL2a, HL2a, and TL4a of BLISS have a focal length of 220 mm, 2250 mm, and 900 mm, respectively. The mirrors M2 and M5 of the GFPI will be moved forthe integration of BLISS by 200 mm and 50 mm, respectively. λ nm, Sr II λ nm, and Ba II λ nm are found inblue part of the visible spectrum, which also contains severalmagnetically insensitive lines with g eff = 0 and lines ex-hibiting a Zeeman-triplet with splitting factors of g eff = 2 . and higher. A special case is the pair Fe I λ . nm( g eff = 0 ) and Cr I λ . nm ( g eff = 2 . ) that can berecorded at the same time. Optical design
The geometrical design of BLISS and its integration into theGFPI is depicted in Fig. 7. Most of the optical componentsand their labels are identical to those in Fig. 1. The calcula-tions for this initial optical design revealed that a pupil P2 of about 70 mm diameter will be sufficient for an accept-able maximal blue-shift over the entire wavelength range ata given image scale and FOV. Thus, the design of BLISS only requires different camera lenses HL2a and
TL4a in the
NBC and
BBC to obtain an adequate image scale. The reduc-tion of the diameter of P2 , caused by several modificationsof the telescope since 2007, is compensated in the design ofBLISS. The schematic optical designs of both instrumentsare compared in Fig. 8. The diameter of the light beam atthe different foci, pupils, and on the relevant optical surfaceshas been calculated by means of geometrical optics and con-firmed by a ZEMAX ray-tracing as for the GFPI. The ZE-MAX ray-tracing is presented in Fig. 9.A new type of cameras has been envisaged for BLISS,namely, two sCMOS-cameras from PCO (pco.edge). Thesecameras have × pixels with a pixel size of 6.5 µ m × µ m similar to the GFPI Imager QE cameras. Using afocal length of f = 2250 mm for HL2 and f = 900 mmfor TL4 yields an image scale of 0.028 ′′ pixel − and a FOV10 ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8) Fig. 9
Arrangement of GFPI and BLISS in a ZEMAX multi-configuration file in shaded modeling for the respective central wave length of each instrumentand their maximal field dimension. The design confirms the calculations in geometrical optics presented in Figs. 7 and 8. The optical elements of BLISS arelabeled with an extra ‘a’. The light coming from the telescope enters at the science focus F4. of ′′ × ′′ on both cameras. With this configuration, themaximal blue-shift across the FOV amounts to 4.4 pm and6.2 pm at 380 nm and 530 nm, respectively.In addition, we checked the possibility of interchangingthe sCMOS and Imager QE cameras between the GFPIand BLISS. The almost identical pixel sizes of both cam-eras facilitates this task. For the integration of the sCMOScameras into the GFPI, a circular FOV with a diameter of d = 79 ′′ and an image scale of 0.036 ′′ pixel − for spec-troscopy would result in a maximal blueshift of 6.9 pm at630 nm. However, the FOV for polarimetry is limited by thefull-Stokes polarimeter and would remain at its previous sizeof ′′ × ′′ unless an upgrade of the polarimeter will be un-dertaken in addition (see Section 3.3). The frame rate at fullresolution would increase from 7 to 40 Hz, which perfectlysuits this observing mode.The integration of the Imager QE cameras into BLISSwould result in an image scale of 0.027 ′′ pixel − and aFOV of ′′ × ′′ with a maximum blueshift of 1.1 pm and1.6 pm at 380 nm and 530 nm, respectively. The lower framerates would match the longer exposure times at shorter wave-length. A super-achromatic optical setup for BLISS – as alsoforeseen for the GFPI – would be preferable but otherwise alllenses except TL4a of BLISS could be purchased as off-the-shelf achromats.The distribution of the two instruments on the five opticaltables in the GREGOR observing room is shown in Fig. 7.All optical elements of BLISS have been labeled with anadditional “a” to distinguish them from those of the GFPI.Behind the common dichroic beamsplitter cube
BS1 , eachinstrument has its own
NBC and
BBC . A displacement ofthe folding mirrors M2 and M5 of the GFPI to a distance of700 mm and 350 mm from F5 and HL2 , respectively, yields some free space on optical table 3 that can be used for the
NBC of BLISS. The ZEMAX ray-tracing revealed changesin the optical path when considering the etalons plates in thedesign. Thus, the GFPI
NBC has to be shortened further byreducing the distance between P2 and HL2 from 500 mm to350 mm. This detail is not considered in Figs. 7 and 8.The beam in the
NBC of BLISS is also folded twice by
M2a and
M5a at a distance of 800 mm and 700 mm from
F5a and
HL2a , respectively. In the
BBC of BLISS,
TL3a and
TL4a are separated by 350 mm, in contrast to the 50 mmbetween
TL3 and
TL4 in case of the GFPI. Two computer-controlled filter sliders
FSL1a and
FSL2a will again switchbetween two sets of interference filters. BLISS is mainly de-signed for spectroscopy because of the expected low pho-ton numbers in the blue spectral region. Nevertheless, a full-Stokes polarimeter can easily be integrated into the system.As in case of the GFPI, a laser/photo-multiplier channel forfinesse adjustment of the etalons will be implemented. Thewhite-light channel of the GFPI will be removed and be re-placed by an external white-light source shared by all post-focus instruments.
Camera system
Modern sCMOS cameras are capable of delivering highframe rates using large-format sensors with low readoutnoise, which makes them ideally suited for an applicationin solar physics. The pco.edge is a potential candidate forBLISS. The camera has a full well capacity of 30,000 e − anduses 16-bit digitization. The camera has a maximum framerate of 40 Hz when operated in the global shutter mode. Thecameras are running in a “fast scan mode” (286 MHz) con-verting the 16-bit signal internally to 12-bit because of speed11 ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8) limitations of the camera link interface. The signal losses dueto the compression are roughly a factor of ten smaller thanthe shot noise of the camera signal. The readout noise is inthe order of 2.3 e − . The dark current consists of a part re-lated to the exposure time, i.e., 2–6 e − pixel − s − and apart related to the sensor readout time, which is constant fora given pixel clock, i.e., 0.6 e − pixel − . Peltier cooling of thesensor ensures an operating temperature of +5 ◦ C. The cam-era has a quantum efficiency of ∼
30% and ∼
54% at 380 nmand 530 nm, respectively, similar to the Imager QE camerascurrently used in the GFPI ( ∼
36% and ∼ − at frame rates of 40 Hz, each cam-era will be controlled by a separate PC, in which the datawill be stored on local RAID 0 systems. The integration ofthe cameras in the control software is straightforward be-cause the DaVis software will be operational on the newsystem with minor changes only. We will perform first in-situ test-measurements with these cameras during an upcom-ing GFPI-technical campaign in April 2013 at the GREGORtelescope, whose results will undergo subsequently a carefuland detailed analysis.However, the feasibility of using the pco.edge for BLISShas still to be demonstrated. A detailed evaluation of thephoton statistic in the wavelength range 380–530 nm willhelp with the final decision. Rather long exposure times canbe expected for the blue wavelength range and the exposuretime of the cameras currently has an upper limit of 100 msdue to a relatively high dark current in the global shuttermode. On the other hand, high frame rates would be ex-tremely beneficial when operating the GFPI in the vector po-larimetric mode, because at present this instrument is limitedto just 5–7 Hz at full resolution. The almost identical pixelsize facilitates interchanging the cameras between the twoinstruments. Etalons
The design of a dual-etalon system requires the consider-ation of several etalon parameters, i.e., absorption, reflec-tivity, and resulting peak transmission and finesse of bothetalons, as well as the correct plate-spacing in order to obtainthe appropriate spectral resolution for a given wavelengthrange together with a low parasitic-light fraction. The lat-ter is defined as the ratio of the intensity transmitted out-offband to that of the central transmission peak of the result-ing Airy-function of the instrument and depends also on theused pre-filter. The reflectivity of etalons for imaging in theblue spectral region has to be lower to increase the FWHMand to accommodate the smaller number of available pho-tons. Therefore, the rejection of parasitic light further awayfrom the central wavelength λ becomes more importantand narrower interference filters are required. Interferencefilters with a FWHM of 0.3–0.8 nm provide a good com-promise between parasitic light, transmission, and signal-to-noise. Figure 10 shows the results of such a parameter studyfor BLISS together with the results obtained for the currentetalon properties of the GFPI. In both panels, the transmis-sion profiles of the individual etalons and the order-sortinginterference filters ( FWHM = 0 . nm, 40% transmission)are presented for center wavelengths of 600 nm (GFPI, up-per panel) and 400 nm (BLISS, lower panel), respectively. λ [nm]10 −8 −6 −4 −2 I n t e n s it y I / I λ [nm]10 −8 −6 −4 −2 I n t e n s it y I / I Fig. 10
Comparison of GFPI ( top ) and BLISS ( bottom ) transmissionprofiles. The curves correspond to a narrow-band interference filter(
FWHM = 0 . nm, thin solid line ), etalon 1 ( R = 0 . / . , d =1 . / . mm, FWHM = 2 . / . pm, thin dashed line ), etalon 2 ( R =0 . / . , d = 1 . / . mm, FWHM = 2 . / . pm, thin dash-dottedline ), and all transmission curves combined ( FWHM = 1 . / . pm,parasitic-light fraction 0.8/0.4%, thick solid line ). The transmission profilesare normalized to unity at the central wavelength λ = 600 . / . nm. The final BLISS and GFPI transmission profiles (thin solidcurve) are multiplied by the transmission curve of the pre-filters (thick solid curve) and have a central peak with aFWHM of 1.5 pm and 4.2 pm, respectively. The parasitic-light fraction has been calculated according to the wave-length range of the assumed pre-filter, i.e., ± . nm, andamounts to 0.8% and 0.4% for GFPI and BLISS, respec-tively. The lower parasitic-light fraction of BLISS is a di-rect result of optimizing the ratio of the plate spacings whichcould be done only partly in case of the GFPI. In princi-ple, a larger parasitic light fraction should be expected in theblue. Our results also show that only 0.3 nm interference fil-ters will work with BLISS. In contrast to the GFPI etalonswith plate spacings of d = 1 . and . mm ( FWHM = 2 . and . pm) and reflectivities of 95%, for BLISS the combi-nation of etalons with narrower plate spacings of d = 0 . and . mm ( FWHM = 8 . and . pm) and lower re-flectivities of 92% will increase the light level and result ina theoretical spectral resolution of R ∼ , . A proper12 ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8) wavelength-dependent parameter study like in the case of theGFPI requires the knowledge of the etalon characteristicsprovided by the manufacturer, thus the present calculationsshould be extended as soon as this information is available. Much effort has been spent to provide with the GFPI a mod-ern first-light post-focus instrument for the GREGOR solartelescope. New software and hardware provide now an easyhandling of the instrument with automated procedures forboth observation and calibration. The GFPI can be controlledwith a well-structured Graphical User Interface (GUI) andunderwent an extended comissioning in 2011 and a carefulscience verification throughout 2012. The latter revealed thatmost of the characteristic parameters of the instrument com-ply with the theoretical expectations. However, we are awareof a problem concerning the measured effective spectral res-olution of the instrument, which warrants further investiga-tion.The results from first 2D spectroscopic observations sup-ported by adaptive optics at the end of July 2012 are verypromising and suggest that high-cadence 2D spectroscopicobservations near the diffraction limit are very likely tobe obtained in coming observing campaigns at the GRE-GOR telescope as soon the apparent stray-light problem issolved. This will also increase significantly the performanceof the adaptive optics system and will finally permit a res-olution of about 50 km on the solar surface in combina-tion with post-factum image restoration techniques. Vector-polarimetric observations will follow in the first half of 2013.Possibilities for the extension of polarimetric observationsto the entire wavelength range covered by the instrument(530–860 nm) are under investigation. We also discussed afuture upgrade with new cameras, which would both yieldhigher frame rates and a larger field-of-view. This wouldbe crucial especially for the dual-beam spectropolarimetricmode of the GFPI, although the enlargement of the field-of-view would require also a re-design of the polarimeter.The future combination of the GFPI with its companion,the BLue Imaging Solar Spectrometer (BLISS), whose de-sign was presented here in detail, and the GRIS will permitmulti-wavelength observations over the entire optical spec-tral range 380–1600 nm. This will allow one to study theconnection of different physical processes from the lower-most photosphere, i.e., the continuum forming layers, up tothe upper chromosphere, e.g., the line core of H α , at highestspatial resolution in unprecedented intrinsic detail. Acknowledgments
The 1.5-meter GREGOR solar telescope was build by a Ger-man consortium under the leadership of the Kiepenheuer-Institut f¨ur Sonnenphysik in Freiburg with the Leibniz-Institut f¨ur Astrophysik Potsdam, the Institut f¨ur AstrophysikG¨ottingen, and the Max-Planck-Institut f¨ur Sonnensystem-forschung in Katlenburg-Lindau as partners, and with con-tributions by the Instituto de Astrof´ısica de Canarias and theAstronomical Institute of the Academy of Sciences of theCzech Republic. CD was supported by grant DE 787/3-1of the Deutsche Forschungsgemeinschaft (DFG). We wouldlike to thank Robert Geissler for his competent and extended help during the science verification.
References
1. Scharmer, G. B., Bjelksjo, K., Korhonen, T. K., Lindberg, B., and Pet-terson, B., “The 1-meter Swedish solar telescope”,
Proc. SPIE ,341 (2003).2. Kosugi, T., Matsuzaki, K., Sakao, T., Shimizu, T., Sone, Y.,Tachikawa, S., et al., “The Hinode (Solar-B) mission: an overview,”
Sol. Phys. , 3 (2007).3. Solanki, S. K., Barthol, P., Danilovic, S., Feller, A., Gandorfer, A.,Hirzberger, J., et al., “SUNRISE: instrument, mission, data, and firstresults”,
ApJL , L127 (2010).4. Schmidt, W., von der L¨uhe, O., Volkmer, R., Denker, C., Solanki,S., Balthasar, H., et al., “The GREGOR solar telescope on Tenerife”,
ASP Conf. Ser. , 365 (2012).5. Schmidt, W., von der L¨uhe, O., Volkmer, R., Denker, C., Solanki,S., Balthasar, H., et al., “The 1.5 meter solar telescope GREGOR”, AN , 796 (2012).6. Denker, C., Goode, P. R., Ren, D., Saadeghvaziri, M. A., Verdoni,A. P., Wang, H., et al., “Progress on the 1.6-meter New Solar Tele-scope at Big Bear Solar Observatory”, Proc. SPIE , 62670A(2006).7. Cao, W., Gorceix, N., Coulter, R., Coulter, A., and Goode, P. R., “Firstlight of the 1.6 meter off-axis New Solar Telescope at Big Bear SolarObservatory”,
Proc. SPIE , 773330 (2010).8. Wagner, J., Rimmele, T. R., Keil, S., Hubbard, R., Hansen, E., Phelps,L., et al., “Advanced Technology Solar Telescope: a progress report”,
Proc. SPIE , 70120I (2008).9. Collados, M., Bettonvil, F., Cavaller, L., Ermolli, I., Gelly, B., Grivel-Gelly, C., et al., “European Solar Telescope: project status”,
Proc.SPIE , 77330H (2010).10. Puschmann, K. G. and Beck, C., “Application of speckle and (multi-object) multi-frame blind deconvolution techniques on imaging andimaging spectropolarimetric data”,
A&A , A21 (2011).11. Mart´ınez Pillet, V., Del Toro Iniesta, J. C., ´Alvarez-Herrero, A.,Domingo, V., Bonet, J. A., Gonz´alez Fern´andez, L., et al., “The Imag-ing Magnetograph eXperiment (IMaX) for the Sunrise balloon-bornesolar observatory,”
Sol. Phys. , 57 (2011).12. Rimmele, T. R., Wagner, J., Keil, S., Elmore, D., Hubbard, R.,Hansen, E., et al., “The Advanced Technology Solar Telescope: be-ginning construction of the world’s largest solar telescope”,
Proc.SPIE , 77330G (2010).13. Bendlin, C., Volkmer, R., and Kneer, F., “A new instrument forhigh resolution, two-dimensional solar spectroscopy,”
A&A , 817(1992).14. Puschmann, K. G., Kneer, F., Seelemann, T., and Wittmann, A. D.,“The new G¨ottingen Fabry-P´erot spectrometer for two-dimensionalobservations of the Sun,”
A&A , 1151 (2006).15. Puschmann, K. G., Kneer, F., Nicklas, H., and Wittmann, A. D.,“From the ”G¨ottingen” Fabry-P´erot interferometer to the GREGORFPI”, 2007masfa.conf, 45 (2007).16. Bello Gonz´alez, N. and Kneer, F., “Narrow-band full Stokes po-larimetry of small structures on the Sun with speckle methods,”
A&A , 265 (2008).17. Balthasar, H., Bello Gonz´alez, N., Collados, M., Denker, C., Hof-mann, A., Kneer, F., and Puschmann, K. G., “A full-Stokes polarime-ter for the GREGOR Fabry-P´erot interferometer”,
IAU Symp. ,665 (2009).18. Balthasar, H., Bello Gonz´alez, N., Collados, M., Denker, C., Feller,A., Hofmann, A., et al., “Polarimetry with GREGOR,”
ASP Conf.Ser. , 351 (2011).19. Denker, C., Balthasar, H., Hofmann, A., Bello Gonz´alez, N., andVolkmer, R., “The GREGOR Fabry-P´erot interferometer: a new in-strument for high-resolution solar observations”,
Proc. SPIE ,77356M (2010).20. Halbgewachs, C., Caligari, P., Glogowski, K., Heidecke, F.,Knobloch, M., Mustedanagic, M., et al., “ Telescope Control System,” AN , 840 (2012).21. Puschmann, K. G., Balthasar, H., Bauer, S. M., Hahn, T., Popow, E.,Seelemann, T., et al., “The GREGOR Fabry-P´erot Interferometer – anew instrument for high-resolution spectropolarimetric solar observa-tions”, ASP Conf. Ser. , 423 (2012).22. Puschmann, K. G., Balthasar, H., Beck, C., Louis, R. E., Popow, E.,Seelemann, T., et al., “The GREGOR Fabry-P´erot Interferometer – ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8) status report and prospects”, Proc. SPIE , 844679 (2012).23. Puschmann, K. G., Denker, C., Kneer, F., Al Erdogan, N., Balthasar,H., Beck, C., et al., “The GREGOR Fabry-P´erot Interferometer”, AN , 880 (2012).24. Strassmeier, K. G., Ilyin, V., Woche, M. Granzer, T., Weber, M., We-ingrill, J., et al., “Gregor@night: the future high-resolution stellarspectrograph for the GREGOR solar telescope”, AN , 901 (2012).25. Hammerschlag, R.H. Kommers, J. N., Visser, S., Bettonvil, F. C. M.,van Schie, S. J. van Leverink, G., “Open-foldable domes with high-tension textile membranes – The GREGOR dome,” AN , 830(2012).26. Soltau, D. Volkmer, R. von der L¨uhe, O and Berkefeld, T.,“Opticaldesign of the new solar telescope GREGOR”, AN , 847 (2012).27. Volkmer, R., Eisentr¨ager, P., Emde, P. Fischer, A. von der L¨uhe, O.,Nicklas, H., et al., “ Mechanical design of the solar telescope GRE-GOR,” AN , 816 (2012).28. Berkefeld, T., Schmidt, D., Soltau, D., von der L¨uhe, O., and Hei-decke, F., “The GREGOR Adaptive Optics System ”, AN , 863(2012).29. Hofmann, A., Arlt, K., Balthasar, H., Bauer, S.M., Bittner, W.,Paschke, J., et al., “The GREGOR polarimetric calibration unit”, AN , 854 (2012).30. Collados, M., L´opez, R., P´aez, E., Hern´andez, E., Reyes, E., Calcines,A., et al., “GRIS – The GRating Infrared Spectrograph for GRE-GOR”, AN , 872 (2012).31. von der L¨uhe, O., Volkmer, R., Kentischer, T., and R. Geißler, “TheGregor Broad Band Imager”, AN , 894 (2012).32. Beck, C., Rezaei, R., and Fabbian, D., “Stray-light contamination andspatial deconvolution of slit-spectrograph observations”, A&A ,A129 (2011).33. L¨ofdahl, M. G. and Scharmer, G. B., “Sources of straylight in thepost-focus imaging instrumentation of the Swedish 1-m Solar Tele-scope”,
A&A , A80 (2012).34. Kurucz, R. L., Furenlid, I., Brault, J. and Testerman, L., “Solar fluxatlas from 296 to 1300 nm”, National Solar Observatory, Sunspot,NM (1984)35. Neckel, H.,“Announcement: spectral atlas of solar absolute disk-averaged and disk-center intensity from 3290–12510 ˚A ”,
Sol.Phys. , 421 (1999).36. Allende Prieto, C., Asplund, M., and Fabiani Bendicho, P., “Center-to-limb variation of solar line profiles as a test of NLTE line formationcalculations”,
A&A , 1109 (2004).37. Cabrera Solana, D., Bellot Rubio, L. R., Beck, C., and Del Toro Ini-esta, J. C., “Temporal evolution of the Evershed flow in sunspots. I.Observational characterization of Evershed clouds”,
A&A , 1067(2007).38. Denker, C. and Tritschler, A., “Measuring and Maintaining the PlateParallelism of Fabry-P´erot Etalons”,
PASP , 1435 (2005).39. Scharmer, G. B. “Comments on the optimization of high resolutionFabry-P´erot filtergraphs”,
A&A , 1111 (2006).40. Righini, A., Cavallini, F., and Reardon, K. P., “Imaging performanceof multi-etalon bidimensional spectrometers”,
A&A , 7 (2010).41. Schnerr, R. S., de La Cruz Rodr´ıguez, J., van Noort, M. S., “Stokesimaging polarimetry using image restoration: a calibration strategyfor Fabry-P´erot based instruments ”,
A&A , 45 (2011).42. de Boer, C. R., “
Speckle-Interferometrie und ihre Anwendungenauf die Sonnenbeobachtung ”, PhD thesis, Universit¨ats-SternwarteG¨ottingen, Germany (1993).43. Puschmann, K. G. and Sailer, M., “Speckle reconstruction of photo-metric data observed with adaptive optics”,
A&A , 1011 (2006).44. Janssen, K., “Structure and dynamics of small scale magnetic fields inthe solar atmosphere Results of high resolution polarimetry and imagereconstruction”, PhD Thesis, Georg-August Universit¨at G¨ottingen,Germany (2003)45. Keller, C. U. and von der L¨uhe, O., “Solar speckle polarimetry”,
A&A , 321 (1992)46. Lites, B. W., Rutten, R. J., and Kalkofen, W., “Dynamics of the solarchromosphere. I - Long-period network oscillations”,
ApJ , 345(1993).47. Rouppe van der Voort, L. H. M., “Penumbral structure and kinematicsfrom high-spatial-resolution observations of Ca II K”,
A&A , 1020(2002).48. Beck, C., Schmidt, W., Kentischer, T., and Elmore, D., “Polarimet-ric Littrow Spectrograph - instrument calibration and first measure-ments”,
A&A , 1159 (2005). 49. Beck, C., Schmidt, W., Rezaei, R., and Rammacher, W., “The sig-nature of chromospheric heating in Ca II H spectra”,
A&A , 213(2008).50. Sistla, G. and Harvey, J. W., “Fraunhofer lines without Zeeman split-ting”,
Sol. Phys. , 66 (1970).51. Harvey, J. W., “Fraunhofer lines with large Zeeman splitting”, Sol.Phys. , 9 (1973). ptical Engineering 52(8),121432SSPR (June 5, 2018)Optical Engineering June 5, 2018/Vol. 52(8) Klaus G. Puschmann performed his thesis at theInstituto de Astrof´ısica de Canarias (IAC) and re-ceived his doctoral degree from the Karl-FranzensUniversit¨at Graz in 2002. After post-doctoral posi-tions at the Institut f¨ur Astrophysik G¨ottingen andthe IAC, he is presently employed at the Leibniz-Institut f¨ur Astrophysik Potsdam. His research in-terests include image reconstruction, spectropo-larimetry and Stokes inversion of sunspot spectra.He is member of the GREGOR project with focuson the development of the GREGOR Fabry-P´erotInterferometer.
Carsten Denker received his doctoral degreefrom the Georg-August-Universit¨at G¨ottingen in1996. At the Leibniz-Institut f¨ur Astrophysik Pots-dam, he is the head of the Optical Solar PhysicsGroup and the Solar Observatory Einstein Tower.He is teaching astronomy, astrophysics, and so-lar physics at the Universit¨at Potsdam and theHumboldt-Universit¨at zu Berlin. His main researchinterests are in the areas of solar physics, instru-mentation for high-resolution solar observations,and image restoration.
Horst Balthasar studied physics at the Georg-August-Universit¨at G¨ottingen where he got hisdoctoral degree in 1984. After scientific stationsat the Instituto de Astrof´ısica de Canarias, theUniversit¨at G¨ottingen, the Fachhochschule Wies-baden and the Kiepenheuer Institut f¨ur Sonnen-physik Freiburg, he works since 1997 as a scien-tist at the Leibniz-Institut f¨ur Astrophysik Potsdam.His main scientific interest is the investigation ofthe dynamics and magnetic field in sunspots.
Rohan E. Louis completed his PhD at theUdaipur Solar Observatory, India in 2010. Heis currently working as a post-doctoral fellowat the Leibniz-Institut f¨ur Astrophysik Potsdamwhere he is involved in scientific investigationsof high-resolution imaging-spectropolarimetricobservations from the GREGOR Fabry-P´erotInterferometer. His research interests includesunspot fine structure, photospheric magneticfields, solar adaptive optics, and Stokes inversioncodes.
Emil Popow studied physics at the Friedrich-Schiller-Universit¨at Jena and the Humboldt-Universit¨at zu Berlin. He is employee of theLeibniz-Institut f¨ur Astrophysik Potsdam, wherehe worked on the development of scientificinstruments and astronomical instrumentation.In 2003, he became the head of AIP’s technicaldivision.
Manfred Woche obtained his diploma in physicsat the Friedrich-Schiller-Universit¨at Jena in1976. Until 1996 he was employed at theKarl-Schwarzschild-Observatorium Tautenburgwith core responsibilities in spectroscopy andinstrumentation. ¿From 1996 until 2001 heworked in optical design and astronomicalinstrumentation for different institutes, especiallyfor the Max-Planck-Institut f¨ur extraterrestrischePhysik Garching and the Skinakas ObservatoryCrete. Since 2001 he is responsible for theoptical design of astronomical instrumentation at the Leibniz-Institut f¨urAstrophysik Potsdam.
Christian Beck obtained his PhD at the Albert-Ludwigs-Universit¨at Freiburg in 2006. He workedas a post-doctoral researcher at the Instituto deAstrof´ısica de Canarias from 2006 to 2012 witha research focus on instrumentation, the struc-ture of sunspots, and the solar chromosphere. Herecently joined the National Solar Observatory inSunspot, New Mexico, as an instrument scientist.
Thomas Seelemann studied physics atthe Georg-August-Universit¨at G¨ottingen. Forthe diploma thesis and doctoral dissertationhe worked at the Max-Planck Institut f¨urStr¨omungsforschung in G¨ottingen in the fieldof molecular physics and received his doctoraldegree in 1980. Since then, he is employed atLaVision GmbH, where he works on camerasand their integration into dedicated experimentalapplications.
Reiner Volkmer received his doctoral degreein physics from the Georg-August-Universit¨atG¨ottingen in 1995. After his PhD, he was projectmanager of the data handling system for IBISon board of the INTEGRAL satellite at theUniversit¨at T¨ubingen. Later on, he was managerof the GREGOR project and is now leading thedevelopment of the image stabilization system forthe PHI instrument on board the Solar Orbiter.received his doctoral degreein physics from the Georg-August-Universit¨atG¨ottingen in 1995. After his PhD, he was projectmanager of the data handling system for IBISon board of the INTEGRAL satellite at theUniversit¨at T¨ubingen. Later on, he was managerof the GREGOR project and is now leading thedevelopment of the image stabilization system forthe PHI instrument on board the Solar Orbiter.