The Solar Orbiter SPICE instrument -- An extreme UV imaging spectrometer
SPICE Consortium, M. Anderson, T. Appourchaux, F. Auchère, R. Aznar Cuadrado, J. Barbay, F. Baudin, S. Beardsley, K. Bocchialini, B. Borgo, D. Bruzzi, E. Buchlin, G. Burton, V. Blüchel, M. Caldwell, S. Caminade, M. Carlsson, W. Curdt, J. Davenne, J. Davila, C. E. DeForest, G. Del Zanna, D. Drummond, J. Dubau, C. Dumesnil, G. Dunn, P. Eccleston, A. Fludra, T. Fredvik, A. Gabriel, A. Giunta, A. Gottwald, D. Griffin, T. Grundy, S. Guest, M. Gyo, M. Haberreiter, V. Hansteen, R. Harrison, D. M. Hassler, S. V. H. Haugan, C. Howe, M. Janvier, R. Klein, S. Koller, T. A. Kucera, D. Kouliche, E. Marsch, A. Marshall, G. Marshall, S. A. Matthews, C. McQuirk, S. Meining, C. Mercier, N. Morris, T. Morse, G. Munro, S. Parenti, C. Pastor-Santos, H. Peter, D. Pfiffner, P. Phelan, A. Philippon, A. Richards, K. Rogers, C. Sawyer, P. Schlatter, W. Schmutz, U. Schühle, B. Shaughnessy, S. Sidher, S. K. Solanki, R. Speight, M. Spescha, N. Szwec, C. Tamiatto, L. Teriaca, W. Thompson, I. Tosh, S. Tustain, J.-C. Vial, B. Walls, N. Waltham, R. Wimmer-Schweingruber, S. Woodward, P. Young, A. De Groof, A. Pacros, D. Williams, D. Müller
AAstronomy & Astrophysics manuscript no. SO_Book_SPICE_paper © ESO 2020September 3, 2020
The Solar Orbiter SPICE instrument
An extreme UV imaging spectrometer
SPICE Consortium: M. Anderson , T. Appourchaux , F. Auchère , R. Aznar Cuadrado , J. Barbay , F. Baudin ,S. Beardsley , K. Bocchialini , B. Borgo , D. Bruzzi , E. Buchlin , G. Burton , V. Büchel , M. Caldwell ,S. Caminade , M. Carlsson , W. Curdt , J. Davenne , J. Davila , C. E. DeForest , G. Del Zanna , D. Drummond ,J. Dubau , C. Dumesnil , G. Dunn , P. Eccleston , A. Fludra , T. Fredvik , A. Gabriel , A. Giunta , A. Gottwald ,D. Gri ffi n , T. Grundy , S. Guest , M. Gyo , M. Haberreiter , V. Hansteen , R. Harrison , D. M. Hassler ,S. V. H. Haugan , C. Howe , M. Janvier , R. Klein , S. Koller , T. A. Kucera , D. Kouliche , , E. Marsch ,A. Marshall , G. Marshall , S. A. Matthews , C. McQuirk , S. Meining , C. Mercier , N. Morris , T. Morse ,G. Munro , S. Parenti , C. Pastor-Santos , H. Peter , D. Pfi ff ner , P. Phelan , A. Philippon , A. Richards , K. Rogers ,C. Sawyer , P. Schlatter , W. Schmutz , U. Schühle , B. Shaughnessy , S. Sidher , S. K. Solanki , , R. Speight ,M. Spescha , N. Szwec , C. Tamiatto , L. Teriaca , W. Thompson , I. Tosh , S. Tustain , J.-C. Vial , B. Walls ,N. Waltham , R. Wimmer-Schweingruber , S. Woodward , P. Young , , A. De Groof , A. Pacros , D. Williams ,D. Müller (cid:63) RAL Space, STFC Rutherford Appleton Laboratory, Harwell, Didcot, OX11 0QX, UK Institut d’Astrophysique Spatiale, 91405 Orsay Cedex, France Max-Planck-Institut für Sonnensystemforschung, Justus-von-Liebig-Weg 3, 37077 Göttingen, Germany PMOD / WRC, Dorfstrasse 33, 7260 Davos Dorf, Switzerland Institute of Theoretical Astrophysics, University of Oslo, P.O. Box 1029 Blindern, 0315 Oslo, Norway NASA Goddard Space Flight Center, Greenbelt, MD, USA Southwest Research Institute, 1050 Walnut Street, Boulder, CO, USA DAMTP, Centre for Mathematical Sciences, University of Cambridge Wilberforce Road Cambridge CB3 0WA, UK Southwest Research Institute, 6220 Culebra Rd, San Antonio, TX, USA Physikalisch-Technische Bundesanstalt, Abbestraße 2–12, 10587 Berlin, Germany University College London, Mullard Space Science Laboratory, Holmbury St. Mary, Dorking, Surrey, RH5 6NT, UK ESR Technology Ltd, 202 Cavendish Place, Birchwood Park, Warrington, Cheshire, WA3 6WU, UK Division for Extraterrestrial Physics, Institute for Experimental and Applied Physics (IEAP), Christian Albrechts University atKiel, Leibnizstr. 11, 24118 Kiel, Germany European Space Agency, ESAC, Camino Bajo del Castillo s / n, Urb. Villafranca del Castillo, 28692 Villanueva de la Cañada,Madrid, Spain European Space Agency, ESTEC, P.O. Box 299, 2200 AG Noordwijk, The Netherlands ADNET Systems, Inc., Lanham, MD, USA CESAM SEED, 52B Bd Saint-Jacques, 75014 Paris School of Space Research, Kyung Hee University, Yongin, Gyeonggi-Do, 446-701, Republic of Korea Northumbria University, Newcastle Upon Tyne, NE1 8ST, UKReceived 29 March 2019 / Accepted 19 August 2019
ABSTRACT
Aims.
The Spectral Imaging of the Coronal Environment (SPICE) instrument is a high-resolution imaging spectrometer operatingat extreme ultraviolet (EUV) wavelengths. In this paper, we present the concept, design, and pre-launch performance of this facilityinstrument on the ESA / NASA Solar Orbiter mission.
Methods.
The goal of this paper is to give prospective users a better understanding of the possible types of observations, the dataacquisition, and the sources that contribute to the instrument’s signal.
Results.
The paper discusses the science objectives, with a focus on the SPICE-specific aspects, before presenting the instrument’sdesign, including optical, mechanical, thermal, and electronics aspects. This is followed by a characterisation and calibration of theinstrument’s performance. The paper concludes with descriptions of the operations concept and data processing.
Conclusions.
The performance measurements of the various instrument parameters meet the requirements derived from the mission’sscience objectives. The SPICE instrument is ready to perform measurements that will provide vital contributions to the scientificsuccess of the Solar Orbiter mission.
Key words.
Sun: UV radiation – Sun: transition region – Sun: corona – Instrumentation: spectrographs – Techniques: imagingspectroscopy – Methods: observational Article number, page 1 of 26 a r X i v : . [ a s t r o - ph . I M ] S e p & A proofs: manuscript no. SO_Book_SPICE_paperArticle number, page 2 of 26PICE Consortium: The Solar Orbiter SPICE instrument
SPICE
Fig. 1.
Solar Orbiter spacecraft, with parts of the side panels removedto show the SPICE instrument.
1. Introduction
The Solar Orbiter mission (Müller et al. 2013, 2019), scheduledto launch in February 2020, will study the Sun and inner helio-sphere with a set of remote-sensing instruments observing theSun and solar corona and a set of in-situ instruments measur-ing the solar wind around the spacecraft. Together, the ten SolarOrbiter instruments will provide a complete description of theplasma making up the solar wind – its origin, transport and com-position – vastly improving on the Helios mission (Schwenn &Marsch 1990) launched in 1974. Solar Orbiter reaches a min-imum perihelion of 0.28 AU after a series of gravity assistsfrom Venus and Earth, which will also raise the inclination ofthe orbital plane to above 30 ◦ from the ecliptic plane (Garcia-Marirrodriga & et al. 2019). The Solar Orbiter minimum perihe-lion of 0.28 AU is very similar to the Helios perihelion of 0.3 AU,but combined with its unique out-of-ecliptic vantage point, SolarOrbiter will be able to address a fundamental question of solarphysics: How does the Sun create and control the heliosphere?Solar Orbiter will combine in-situ measurements with high-resolution remote-sensing observations of the Sun in a systemicapproach to resolve fundamental science problems needed toachieve this objective. These problems include the sources ofthe solar wind, the causes of eruptive releases of plasma andmagnetic field from the Sun known as coronal mass ejections(CMEs), the evolution of CMEs and their interaction with theambient solar wind flow, and the origins, acceleration mecha-nisms and transport of solar energetic particles that may be haz-ardous to both human explorers and robotic spacecraft that op-erate in the highly variable environment outside of Earth’s mag-netosphere.While essential to meeting Solar Orbiter’s scientific ob-jectives, the mission’s orbit also poses specific challenges tothe remote-sensing instruments. For example, the changing dis-tances to Sun and Earth result in large variations of the thermalconditions and telemetry rates along each orbit, respectively. Thestrategies devised jointly by the remote-sensing instruments toalleviate these constraints are described in Auchere et al. (2019).The SPICE instrument (Fig. 1) is a high-resolution imag-ing spectrometer operating at extreme ultraviolet (EUV) wave-lengths from 70.4 nm − − (cid:63) Corresponding author: D. Müller, e-mail: [email protected] cility instrument on the Solar Orbiter mission, funded by ESAmember states and ESA. SPICE is allocated 45.3 Gbits of dataper six-month orbit, to be acquired nominally during three ten-day remote-sensing windows, which corresponds to an averagedata rate of 17.5 kbit s − . Most scientific objectives do not re-quire downloading of full spectra but only selected windows cen-tred on typically ten spectral lines of interest. Further reductionof the data volume can be obtained either by data compressionor by computing on board the total intensity of the lines. Theallocated resources do not impose compressing the data beyonda ratio of 20:1 (Sect. 7.9). In Sect. 10, we provide examples ofobservations that illustrate the ability of SPICE to operate withinthe Solar Orbiter constraints.SPICE will address the key science goals of Solar Orbiter byproviding the quantitative knowledge of the physical state andcomposition of the plasma in the solar atmosphere, in partic-ular investigating the source regions of outflows and ejectionprocesses that link the solar surface and corona to the helio-sphere. SPICE is of particular importance for establishing thelink between remote-sensing and in-situ measurements as it isuniquely capable of remotely characterising the plasma prop-erties of source regions, which can directly be compared within-situ measurements taken by the Solar Wind Analyser (SWA)instrument suite (Owen & et al. 2019). In magnetically closedregions, SPICE will play an essential role in characterising theturbulent state of the plasma over a wide range of temperaturesfrom the chromosphere into the hottest parts of the corona. Thisis essential to understand which processes heat the plasma anddrive the dynamics we observe, be it through waves, field-linebraiding, or reconnection.
2. Scientific objectives and opportunities
The main science goals of SPICE are related to our understand-ing of the complex dynamic connection between the Sun and theinner heliosphere. In this sense, the scientific focus is on studiesthat combine the remote-sensing and the in-situ instruments onSolar Orbiter to work as one comprehensive suite. At the sametime, the unique instrumental capabilities of SPICE will also al-low stand-alone studies that will address other unsolved prob-lems in solar physics.By observing the intensities of selected spectral lines andtheir spectral profiles, SPICE will allow the temperature, den-sity, flow, elemental composition and the turbulent state of theplasma in the upper solar atmosphere to be characterised. Emis-sion lines originating between the top of the chromosphere andthe low corona cover the temperature range from 10,000 K to2 MK, augmented by two 10 MK lines seen in flaring plasma(see Table 1).Following a discussion of the observables provided bySPICE in Sect. 2.1, we present a selection of scientific topicsthat will be addressed by SPICE (Sect. 2.2). Naturally, this listwill be incomplete, but should give a flavour of the scientific op-portunities provided by SPICE.
SPICE is capable of measuring the full spectrum in its twowavelength bands. To optimise the science data return withinthe given telemetry budget, only the full profiles of the strongemission lines (marked in Table 1) will be measured routinely.This will provide the intensities, Doppler shifts and widths ofthe lines, from which the non-thermal broadening can be de-termined. The accuracy of the line shifts determined through
Article number, page 3 of 26 & A proofs: manuscript no. SO_Book_SPICE_paper
Table 1.
Selection of spectral lines covered by SPICE.
Ion λ [Å] log T [K] Intensity [ph.pix − .s − ]AR QS Ref. (cid:63) H i ii (cid:63) C iii v iv v (cid:63) O vi (cid:63) O vi vi vii (cid:63) Ne viii viii ix x xi (cid:63) Si xii † xviii (cid:63) Fe xx Notes.
For each line the rest wavelength, λ , the approximate logarithmicline formation temperature, T , and the number of photons detected bySPICE for active regions (AR) and quiet Sun (QS), are listed. The lastcolumn provides the reference for the solar flux used to simulate SPICEobservations. The intensities are for the 2 (cid:48)(cid:48) slit binned over two spatialpixels. Lines marked with a (cid:63) are strong lines for which full line profilescan routinely be returned. For the weaker lines, the intensity integratedacross the line will be computed on-board and sent down. The Fe xx intensity is from a M7.6 flare. Si xii , marked by † , will be observed inthe second spectral order. This line and Fe xviii intensities are for o ff -limb observations. [1] Curdt et al. (2001), [2] Curdt et al. (2004). centroiding will be of the order of 5 km s − at the longer wave-lengths (depending on the signal-to-noise ratio). There is an op-tion of on-board summing of the line profiles and, separately, ofthe adjacent background. This will be particularly useful for theweaker lines observed with shorter exposure times, where thelimited signal-to-noise ratio may prevent determining line shiftand width, but the intensity summed across the line can be ob-tained and downloaded, while using very little of the allocatedtelemetry. Maps of line intensities, shifts, and widths will be pro-vided as high-level data products.In addition to these products that are directly deducible fromthe line profiles, further higher-level data products can be de-rived. First and foremost is the possibility to investigate the el-emental abundances, and in particular the separation of the ele-ments according to the first ionisation potential (FIP). Throughthe FIP e ff ect, there is a preferential enhancement of elements oflow FIP compared to those with high FIP, often called the FIPbias (Fludra & Schmelz 1999; von Steiger et al. 2000). The en-hancement depends on the source region, see a recent review ofobservations in (Del Zanna & Mason 2018).The carefully selected lines from low-FIP elements (S, Si,Mg, Fe) and high-FIP elements (H, C, O, Ne) will allow mapsof the FIP bias to be produced from SPICE data. The broad tem-perature coverage from 10,000 K to 10 MK (see Table 1) will bewell suited to study the thermal structure of the solar atmosphere,all the way from the chromosphere to the corona, occasionallyeven including hot flare plasma around 10 MK. Spatial maps ofthe emission measure can be computed in this temperature range. The spatial resolution of SPICE (along the slit) will be about 4 (cid:48)(cid:48) .At perihelion (0.3 AU), this resolution corresponds to 1.2 (cid:48)(cid:48) foran instrument observing from Earth orbit. This is comparableto, or better than most previous EUV spectrographs. The spa-tial resolution of Hinode / EIS (Culhane et al. 2007) is about 2 (cid:48)(cid:48) ,and SOHO / SUMER (Wilhelm et al. 1995) provided a similar orslightly better resolution. Only the most recent IRIS spectrome-ter (De Pontieu et al. 2014) provides a resolution of about 0.4 (cid:48)(cid:48) that is significantly better than SPICE.In contrast to EIS and IRIS, SPICE provides a much morecomprehensive temperature coverage of the transition region,with lines spaced closely over a wide range of temperatures(see Table 1). EIS is mostly sensitive to hot plasma above 1 MK,while IRIS, designed to observe the Sun’s chromosphere andtransition region, is mostly blind to the temperature range from0.3 MK to 8 MK. While SUMER could observe a temperaturerange wider than SPICE, it had to step through wavelengths torecord spectral profiles, so that it needed considerable time tocover line profiles emitted over the full temperature range. Incontrast, SPICE records all the lines simultaneously. This is amajor advantage when studying the often very dynamic solar at-mosphere.In terms of temporal cadence, SPICE is comparable to pre-vious EUV spectrometers. The brightest lines can be observedwith exposure times of 1-5 s, while 30-60 s exposures are envis-aged for comprehensive coverage of weaker lines. This allowsdynamic phenomena to be followed in a fashion comparable toprevious instruments (although with restrictions set by the lim-ited telemetry). Considering its performance, SPICE will pro-vide unique sets of data. SPICE will be able to take full advan-tage of the special vantage points close to the Sun and from highlatitudes o ff ered by Solar Orbiter during the extended mission,and it will always operate in concert with other remote-sensingand in-situ instruments. Solar Orbiter’s vantage point of out-of-ecliptic latitudes will al-low an unprecedented view of the poles. SPICE will carry outthe first-ever out-of-ecliptic spectral observations of the solar po-lar regions. SPICE will provide maps of outflow velocities andidentify the sources of the fast solar wind inside the polar coro-nal holes, connecting them to solar wind structures observed byin-situ instruments. Joint observations with the SWA / HIS sensorwill allow the testing of models of the fast solar wind (Fludra &Landi 2018).The magnetic fields near the poles are poorly known, and So-lar Orbiter will provide major advances in this direction throughthe PHI instrument (Solanki et al. 2019). In this context, SPICEcan provide the response of the upper atmosphere to the surfacemagnetic field in a globally open magnetic environment. The in-vestigation of the emission from plasma over a wide range oftemperatures will show if the heating mechanisms in the mag-netically closed quiet Sun at lower latitudes are comparable tothose in the globally open coronal hole regions near the poles.Another important goal of SPICE, together with the otherremote-sensing instruments, is to understand small-scale heatingevents in the corona. Here the line profiles from SPICE will pro-vide the crucial information on the turbulent state of the plasmathrough the analysis of the non-thermal broadening. Likewise,propagating waves that transport energy through the atmospherereveal themselves through the spectral profiles. For example,a non-compressible wave like an Alfvén wave, will not be di-rectly visible in imaging observations, but will leave an imprint
Article number, page 4 of 26PICE Consortium: The Solar Orbiter SPICE instrument in spectral data. The very good coverage in temperature will al-low SPICE to determine the resulting thermal structure of thetransition region (through the emission measure) with consid-erably improved resolution compared to previous imaging andspectroscopic observations. In combination with the EUV imag-ing observations by EUI (Rochus & et al. 2019) that will providediagnostics of the spatial and temporal evolution, these SPICEobservations will o ff er a new and comprehensive picture of thestate of the plasma over a range of coronal structures. In partic-ular, SPICE can carry out further studies of the coronal heating,through correlations of the transition region emission with themagnetograms in active regions (Fludra & Warren 2010). Stud-ies of the ubiquitous magnetoacoustic waves, for example, abovesunspots (Fludra 2001) or in coronal holes (Banerjee et al. 2011)will also be possible.Solar Orbiter’s vantage point at high latitudes will also pro-vide a new view at structures near the equator. SPICE and EUIwill have a novel view of, for example, loops connecting theplage regions near the preceding and trailing sunspot in an ac-tive region. Such loops run mostly in the east-west direction andcan only be viewed edge-on from Earth.SPICE will play a crucial role in understanding the coupledsystem of the Sun and the inner heliosphere, the overarching sci-ence goal of Solar Orbiter. Only a spectrometer can provide re-liable tracers connecting measurements by in-situ instruments inthe inner heliosphere to the near-surface regions of the Sun, ob-served remotely.On the one hand, the Doppler maps provided by SPICEwill provide information on the source region of the solar windstreams, for example in the vicinity of active regions or withincoronal holes. On the other hand, SPICE will provide maps ofthe FIP bias that might be a helpful tracer to identify the sourceof the solar wind. In such investigations PHI will provide mea-surements of the underlying (changing) magnetic field and EUIwill provide images of the temporal and spatial structures ofthe chromospheric features through the Ly- α line. Together, thisopens new possibilities to study the acceleration and heating inthe actual source region of a solar wind stream that will then becaptured and characterised in terms of magnetic field, waves, andparticle properties by the in-situ instruments: MAG (Horbury &et al. 2019), RPW (Maksimovic et al. 2019), SWA (Owen &et al. 2019).Eruptive events, such as CMEs, will disrupt the coronalstructures and can be eventually detected by the in-situ instru-ment suite. The radiation from the associated flare will also bedetected by STIX (Krucker & et al. 2019). In those cases wherethe source region has been observed by SPICE, the spectrome-ter will provide crucial information on the initial stages of themagnetic disturbance. The shocks and increased turbulence as-sociated with such an event can be studied through line shifts,and widths and intensity enhancements of spectral lines. WithSPICE, these can be followed closely as a function of tempera-ture. Such observations are essential to understanding how (andwhere) shocks form in the corona. Ultimately, this is the key tounderstanding the generation of solar energetic particles which,again, can be directly measured by the in-situ instruments, in par-ticular EPD (Rodríguez-Pacheco et al. 2019), and which leavetheir trace through the FIP bias in the data acquired by SPICE.
3. Instrument overview
The SPICE instrument is an imaging spectrograph that recordshigh resolution EUV spectra of the Sun. The SPICE optical de-sign was first presented in Fludra et al. (2013). The instrument optics consists of a single-mirror telescope (o ff -axis paraboloidoperating at near-normal incidence), feeding an imaging spec-trometer. The spectrometer also uses just one optical element, aToroidal Variable Line Space (TVLS) grating (Thomas 2003),which images the entrance slit from the telescope focal planeonto a pair of detector arrays. Each detector consists of a pho-tocathode coated micro-channel plate (MCP) image intensifier,coupled to an active pixel sensor (APS). Particular features ofthe instrument needed due to the proximity to the Sun include:use of a dichroic coating on the telescope mirror to transmit andthus reject the majority of the solar spectrum (this overcomes thelarge heat load close to the Sun), a particle deflector to protectthe optics from the solar wind, and use of data compression dueto telemetry limitations. The mechanical design and layout of theSPICE Optics Unit (SOU) are shown in Fig. 2, and the opticalpath is plotted in Fig. 3.As shown in Fig. 3, the light enters the instrument throughthe entrance aperture. Then an image is formed at the slit bythe o ff -axis parabola mirror. The slit defines the portion of thesolar image that is allowed to pass onto a concave TVLS grat-ing, which disperses, magnifies, and re-images incident radia-tion onto two detectors. The two wavebands cover the same one-dimensional spatial field, and are recorded simultaneously. De-tails of the optical path are further described in Sect. 4. The in-strument contains four mechanisms: – The SPICE Door Mechanism (SDM), which can be actuatedto provide a contamination tight seal of the entrance apertureduring non-operational periods (both during ground handlingand non-operational periods in flight). – The telescope mirror is mounted onto a two-axis mechanism(tilt and focus), the Scan-Focus Mechanism, that is used todirect di ff erent portions of the solar image onto the selectedentrance slit and to focus the telescope relative to the en-trance slit. The image of the Sun is repeatedly scanned acrossthe entrance slit. During each scan the image of the Sun isstepped across the entrance slit in increments equal to the se-lected slit width, such that the region of interest is completelysampled. – A Slit Change Mechanism (SCM) provides four interchange-able slits of di ff erent widths, one of which can be selecteddepending upon the science activities to be conducted. Theseslits have a 2 (cid:48)(cid:48) , 4 (cid:48)(cid:48) , 6 (cid:48)(cid:48) , and 30 (cid:48)(cid:48) width on the external field ofview. They are interchangeable via a slit change mechanismand are arranged on the mechanism in this order. – A vacuum door mechanism on the Detector Assembly (DA).The MCP and image intensifier used to translate the inci-dent EUV photons into visible light photons that can be de-tected by the detectors must be maintained either at vacuumor in zero humidity during ground handling. Therefore thedetector assembly contains a door mechanism which is onlyopened during vacuum testing on ground, and opened finallyonce on-orbit.The instrument structure consists of a stable optics benchwith Aluminium honeycomb core and Titanium inserts. Base-plate facesheets, side walls and lids are made of Carbon FibreReinforced Plastic (CFRP). This is isostatically mounted to thespacecraft panel by means of Titanium flexures. The structureis designed to have approximately zero coe ffi cient of thermalexpansion (CTE), therefore maintaining instrument alignmentthroughout the wide operating temperature range.The instrument control function will be provided by a ded-icated electronics box, the SPICE Electronics Box (SEB). TheSEB drives and monitors all mechanisms, the acquisition and Article number, page 5 of 26 & A proofs: manuscript no. SO_Book_SPICE_paper processing of all housekeeping telemetry and the processing andpacketisation of science data. It controls and communicates withthe detector front-end electronics (FEE) via a SpaceWire link.The SEB also contains the SPICE flight software (FSW), whichis responsible for all control and monitoring of the instrument,plus the processing and compression of the science data to allowthe data rate and volume requirements to be achieved.
4. Optical design
SPICE is a grating-spectrometer (Figs. 2, 3) where a portionof the solar disc is imaged by the single-mirror telescope ontothe spectrometer entrance slit. The physical length of the slit is~2 mm, and for the mirror focal length (see Table 2) it gives anangular size of 11 (cid:48) (size of the along-slit instantaneous field ofview on the Sun, oriented solar north-south). The slit is imagedby the di ff raction grating on the two array detectors. The gratinghas a concave toroidal surface form, which images the slit withmagnification as given in the table; the physical size of the slitimage at the detectors is ∼
11 mm. This spectrometer magnifica-tion also scales the telescope focal length to give a system e ff ec-tive focal length of ~3.3 m. This is needed to give the requiredimaging-spectrometer spatial and spectral sampling (~1 (cid:48)(cid:48) / pixeland ~0.01 nm / pixel), with the given physical pixel-spacing of thearray detectors (see Table 2).The two detector arrays are separated in the focal planein the dispersion direction according to the two chosen bands,short wavelengths (SW, 70.4 nm–79.0 nm) and long wavelengths(LW, 97.3 nm–104.9 nm). They are each based on a format of1024 × (cid:48) . The detectors areoversized with respect to this image size such that they pro-vide images with spatial sampling along-slit of approximately14 (cid:48) /
800 pixels, or ~1 (cid:48)(cid:48) / pixel. In the spectral direction the sam-pling is ~9 nm / / pixel (cf. Table 2). Thedispersion relation of the spectrometer is given by the gratingequation:sin( θ m ) = m · λ d + sin( θ i ) , (1)where d is the ruling spacing, m is di ff raction order, θ i is theangle of incidence and θ m the angle of di ff raction, with values asgiven in Table 2.In order to create the 2D images, the 1D instantaneous-FOV(slit FOV) is scanned laterally, over a range of 16 (cid:48) , by rota-tion of the telescope mirror. The mirror is mounted on a flex-ure rotation mechanism, driven by a continuous range mecha-nism (piezo actuator), but with a functional minimum step sizeof 2 (cid:48)(cid:48) , due to the mechanism encoder and control system used.The full range of the spectral imaging is thus an area of sun of14 (cid:48) × (cid:48) times 2 × ×
480 pointsspatially, times 2048 spectral points (maximum size of x - y - λ datacube). The scan is controlled by closed loop, with a minimumtime per step of 0.25 s for a step size of less than 1 arcminute.This gives a best maximum frame rate of 3 Hz. For fixed-scanobservations, the overheads are reduced to 0.1s per frame, lead-ing to a maximum frame rate of 5 Hz (for the minimum exposuretime of 0.1s). The narrowest slit is 2 (cid:48)(cid:48) wide. Because the spectrometer magni-fication is the same in spatial and spectral directions, this corre-
Table 2.
Optical system parameters of the SPICE instrument.
Parameter Value
Telescope
Entrance aperture size 43.5 mm × µ m / (cid:48)(cid:48) Instantaneous FOV Slit length: 11 (cid:48) , plus ± (cid:48) ’dumbbell’ apertures,oriented solar north-southWidth, rastered FOV 16 (cid:48) Slits
Widths: 2 (cid:48)(cid:48) , 4 (cid:48)(cid:48) , 6 (cid:48)(cid:48) , and 30 (cid:48)(cid:48)
Spectrometer
Wavelength range, SW 70.387–79.019 nm (1st order)Wavelength range, LW 97.254–104.925 nm (1st order)48–53 nm (2nd order)Slit to grating distance 128 mmGrating groove densityand ± chirp 2400 ± − Grating image distance,SW mid-band 692.43 mmGrating image distance,LW mid-band 720 mmSpectrometermagnification(image-distance / slit-distance) ~5.5Grating ruling period d / ◦ Gratingangle-of-di ff raction (atSW, LW) + ◦ at 74.7 nm + ◦ at 101.12 nmDispersion, at imageplane 0.0095 at 74 nm 0.0083 at101 nm (per pixel spacing) Detector
Pixel spacing 0.018 mmAngle-of-incidence ondetector (SW, LW) 27.33 ◦ at 74.7 nm, 35.33 ◦ at101.12 nm System focal length (SW)
System spatial plate scale (cid:48)(cid:48) / px at 74 nm, 1.059 (cid:48)(cid:48) / pxat 101 nmsponds to 0.02 nm spectral width geometrically. However, whenthe optical-system and detector resolution e ff ects are added,the net spectral resolution is 0.04 nm (full-width half-maximum(FWHM) of line-spread function (LSF): ~4 pixels). The contrib-utors to imaging resolution are listed in Table 3, and each of thecontributions is explained in the relevant subsections below. SPICE has a single-mirror telescope. The mirror is an o ff -axisparaboloid made of UV-grade fused silica substrate with a clearaperture of 95 mm ×
95 mm and focal length of 622 mm. Thecentral area of 50 mm ×
50 mm on the substrate has a thin reflec-tive coating of boron carbide (B C). This single-layer coating isa novel design, which is a result of a compromise providing 30%
Article number, page 6 of 26PICE Consortium: The Solar Orbiter SPICE instrument
Fig. 2.
Optics Unit of the SPICE instrument. In this top view into the SPICE Optics Unit, its key components are identified, along with theproviding institutes and companies.
Fig. 3.
SPICE optical layout. The system parameters are listed in Ta-ble 2.
Table 3.
Imaging resolution contributions.
Contribution in pixels(spatial and spectraldirections)
Design, including nominalaberration, 2 (cid:48)(cid:48) slit-width × Total (RSS) 4.2
Fig. 4.
Field-of-view diagram for SPICE, showing detector size, slitsizes, and spectral ranges (SW: short wavelengths, LW: long wave-lengths, FOC: fibre optic coupler).
EUV reflectance while transmitting the rest of the solar spectrum(UV / VIS / IR) to space via a 45 ◦ fold mirror and an exit aperture.Such a dichroic design greatly reduces the heat load inside theinstrument. The thickness of 10 nm of the boron carbide coatingwas found to be advantageous for this purpose (Schühle et al.2007).The mirror substrate and the exit aperture are oversized withrespect to the reflective aperture surface. The size is requiredto pass the angular range of the whole solar disc, for Solar Or-biter pointing at any part of it. The rear side of the substrate hasan anti-reflective coating to maximise transmission of the solarspectrum passing through and beyond the reflective front coat-ing.For imaging in the EUV, the mirror surface quality must behigh. Thus, the figure error was specified as ~ λ /
20 RMS. Mea-sured at 633 nm the RMS figure deviation of the flight mirrorwas 0.028 waves. Also, the surface roughness must be low tolimit scattered light from the whole solar disc while only a verysmall part of it is passing on to the spectrograph. The micro-
Article number, page 7 of 26 & A proofs: manuscript no. SO_Book_SPICE_paper roughness was specified as < ◦ C in the mirror, as the silica has low thermal conduc-tivity. However it also has low thermal expansion, so the result-ing ‘swelling’ of the mirror’s front surface is small (predicted~0.04 µ m), which is not significant for aberrations, and is in-cluded in the term ’optical component tolerances’ in table 3. Thiswas also verified by tests, using a mirror illuminated with a solar-simulator UV beam, while monitoring its surface form using anoptical test interferometer.The mirror is mounted on a scan-focus mechanism (SFM,see Sect. 6.2), which performs the angle scan (see above) using apiezo drive, as well as a motor-driven focus adjustment (range of ± ffl e of the instrument structure. It consists of conduc-tive plates with an applied voltage of − . The four slits are arranged in-line on a single frame, and tochange slits this is raised and lowered within the telescope focalplane, on the linear slit change mechanism. Each slit is an aper-ture etched into a silicon slice of 0.5 mm thickness (vee-grooveetch, as shown in Fig. 5), and gold-coated. The three narrow slitsare 11 (cid:48) long (i.e. having 2 mm physical length), but also with asmall square aperture near each end (i.e. at a distance of ± (cid:48) fromcentre). These serve to image small regions of the sun (0.5 (cid:48) × (cid:48) ) to obtain pointing information during the observations (seeFig. 4). An electron-microscope test image showing the end ofone slit and its square aperture is shown in Fig. 5. The 30 (cid:48)(cid:48) slit(also known as a ’slot’) is 14 (cid:48) long, and has no additional squareapertures. This element allows pseudo-spatial images to be ac-quired from isolated spectral lines, which can be used for certainscience studies (e.g. movies), or for collecting instrument cali-bration data (e.g. detector flat-fielding). The di ff raction grating is of TVLS type, as developed for thistype of solar spectrographs. This enables slit-to-array detectorimaging to be performed directly with the grating, in order to beable to dispense with additional mirrors normally required foraberration-control (chromatic astigmatism), and thereby greatlyincreasing the EUV throughput (for near-normal incidence thereflectivity of coatings at these wavelengths is only ~30%). Theaberration control over the range of the two detectors requiresboth toroidal form and varying line-spacing (‘chirp’), that needto be precisely matched. The toroid radii of curvature have tomatch to ~1%. The grating is holographic, 2400 lines / mm (atgrating centre), and with linear ‘chirp’ variation of ~1% of this Fig. 5.
Electron-microscope image of the rear of a SPICE slit (etchedsilicon). The end of a slit and one of its square dumbbells are shown. across the used aperture. This level of linear variation is itselfcontrolled to ~5% to match the toroid radii. The surface opticalquality has to be similar to that of the telescope mirror, and theused aperture is much smaller (by the factor ~622 mm /
128 mm).The reflective coating of the grating is of the same material asthat of the mirror, boron carbide, but the thickness was increasedto 20 nm to increase reflectivity. This is possible due to the neg-ligible solar heat load on the grating.The di ff raction e ffi ciency of the grating is a critical parame-ter, and was measured (using synchrotron radiation) to be ~9%(absolute e ffi ciency). For the spectrometer build tolerances, thelarge magnification is a challenge for the grating focus-settingand the correct alignment of the spectrum on the detector. Fora de-focus blur radius equal to 1 pixel, the axial distance of thegrating from the slit has to be set to within ± µ m. This is apractical challenge in the planning. Since the grating only worksoptically in the vacuum ultraviolet (VUV), the set up in air hasto be done to this accuracy by dead-reckoning, meaning by me-chanical metrology using reference surfaces on the grating sub-strate and the slit mount. The alignment method was to then usea VUV test immediately after this set-up (i.e. before complet-ing the build), to confirm this critical focus and alignment of thegrating. The detector assembly is described in detail in Sect. 6.5. Theincident UV light is converted to visible light inside the assem-bly, for detection by two independent sensor arrays, each sizedat 1024 pixels square. The active area is large enough to recordimages of the full length of the slits plus dumbbells, with somemargin. However, the detector area limits the wavelength rangein each band (see Sect. 9). The pixel pitch of 18 µ m also setsthe spectral and spatial sampling, which in both cases is over-sampled relative to the instrument resolution (see Tables 2, 3). There are two main e ff ects, both due to non-ideal light-scatteringin the optical system: out-of-field light and out-of-band light.The out-of-field e ff ect is the light from the surrounding scene,meaning outside of the FOV, that is scattered into the FOV. It Article number, page 8 of 26PICE Consortium: The Solar Orbiter SPICE instrument
Fig. 6.
SPICE Optics Unit: Details of interfaces to spacecraft includingflux-exit aperture (upper right). has a worst-case for the viewing of the relatively faint corona(when spacecraft points at the limb, SPICE will view at up to~8 (cid:48) above the limb). The relatively bright light of the solar discscatters at instrument ba ffl es and at the optical surfaces. This ef-fect is mainly restricted to the telescope because most of the solardisc is blocked at the slit. The scatter from ba ffl es (vane edges) iskept low by designing the vanes as oversized from the used FOV(i.e. the beam envelope defined by combination of the entranceaperture and the slit) and giving them sharp edges. Also the baf-fle material is CFRP which is absorbing to VUV, and the vanesare designed to block any grazing-incidence light paths fromthe structure surrounding the optics. The mirror roughness andparticulate contamination are kept as low as possible ( < <
100 ppm surface area, respectively).The out-of-band light is the di ff use scatter within the spec-trometer that adds a constant background level to the measuredspectra, adding to the photon noise. This e ff ect is kept low,again by use of ba ffl e vanes, particularly around the slit mecha-nism where there are surfaces close to the beam, by the rough-ness quality of the grating surface and its lines ruling (gratinggrooves) plus its cleanliness, and by the visible-light blocking ofthe detector photo-cathode (so called solar-blindness). The grat-ing’s final roughness after the etching of its grooves (~40 nmdepth) is 0.8 nm RMS. In the final instrument testing, whenimaging spectral lines, the out-of-band level was found to be~0.1% relative to line peak, at ~0.1 nm from line centre.
5. Mechanical and thermal design
The SPICE Optics Unit (SOU) is primarily made of a CFRP-and-aluminium honeycomb optical bench structure, onto whichmost of the subsystems are mounted, along with CFRP panels toproduce a light-tight enclosure and stray-light ba ffl es. The rear ofthe SOU houses the heat rejection mirror and ba ffl e that allowsthe unwanted infrared (IR) radiation from the entrance apertureto pass out the rear of the instrument and out to space. These keyinterfaces can be seen in Fig. 6. The SOU has a total mass ofapproximately 13 kg and maximum dimensions of 1100 × ×
280 mm.The SOU interfaces to the spacecraft panel with three ‘quasi-kinematic’ mounts; one fixed foot at the front of the unit andtwo mounted on blades at the rear that allow for the di ff eren- Fig. 7.
SPICE Optics Unit: Interface Mounts (left: fixed mount, right:bladed flexible mount). tial CTE between the optics bench and the spacecraft panel –these are shown in Fig. 7. The mounts (manufactured from ti-tanium alloy Ti-6Al-4V) are optimised in order to be compliantenough to compensate for the ~1 mm in-plane di ff erence in CTEbetween the spacecraft and optics bench (without distorting thebench), while also being sti ff enough for the SOU to meet theminimum resonance frequency requirement of 140 Hz and sur-vive the launch loads (the actual resonance frequency is 224 Hz).The blade mounts were manufactured in the RAL Space Preci-sion Development Facility and heat treated and surface treatedin order to maximise their performance.The optical bench prior to installation of the subsystems canbe seen in the upper panel of Fig. 8 and the single-skin honey-comb panels used as stray-light control and sti ff ening ribs can beseen in the lower panel. The design is based on the use of a verylow-CTE CFRP in order to maximise the thermo-mechanical sta-bility of the instrument across a wide range of temperatures. Ini-tial coupon testing using interferometry demonstrated CTE of0.2–0.7 ppm / °C for a representative sample, although the finaloptical bench measurement showed a higher (2 . ± . / °C)figure which could be accepted by the use of margin within thealignment budget. This increase is thought to be due to strongerinteraction between the CFRP face sheet, the adhesive film andthe honeycomb core than had been assumed in the theoreticalmodel. The main requirements of the thermal design of SPICE are: – To manage the solar load and maintain all instrument com-ponents to within their operational and non-operational tem-perature limits – To control the detectors to a stable temperature of less than −
20 °C during all operational periods – To minimise the heat flow that is rejected (either radiativelyor conductively) to the spacecraft thermal interfaces – To ensure that the primary mirror is warmer than its sur-roundings during the cold early phases of the mission, toavoid contamination of its surfaceThe primary thermal challenge for the SOU is managing theextreme heat input (~17 kW / m ) during operation at perihelion.The thermal control system must also be compliant during peri-ods with little solar loading, with the conditions during the Earthand Venus gravity assist manoeuvres required for orbit adjust-ment (with additional planetary IR and albedo thermal loads) and Article number, page 9 of 26 & A proofs: manuscript no. SO_Book_SPICE_paper
Fig. 8.
SPICE optical bench structure without (top) and with (bottom)internal stray-light ba ffl ing. with conditions with the instrument o ff -pointed from its centralaxis.The SOU is accommodated within the spacecraft, behind itsheat shield. A feedthrough in the heat shield provides a view forthe instrument. The heat shield includes a door that can be usedto prevent direct sunlight entering the instrument feedthroughduring non-operational periods.The SOU is, with the exception of two designated radiativeand conductive heat rejection interfaces, thermally decoupledfrom the spacecraft. Conductive decoupling is achieved throughthe use of low thermal conductivity titanium for the kinematicmounts, and the natural isolation required by the quasi-kinematicmount design. Radiative decoupling is achieved through the ap-plication of a low-emissivity aluminised coating to the externalsurfaces of the SOU structure.The thermal design utilises a synthetic quartz (Suprasil ® C coating that re-flects the EUV radiation of interest for science but transmitsthe visible and near-infrared solar radiation with little absorp-tion. Consequently, much of the high-flux solar radiation enter-ing through the aperture passes through the instrument and isthen reflected to space by the heat-rejection mirror (HRM) at-tached to the rear of the instrument. The HRM assembly is aCFRP structure mounted to the rear of the instrument that housesa highly reflective diamond turned aluminium fold mirror.The majority of the solar radiation that is reflected withinthe SOU by the primary mirror is intercepted by three pre-slitheat rejection mirrors (mounted before the slit) and reflected to asingle high-absorptivity heat dump radiator. As these heat loads
Fig. 9.
Predicted thermal energy balance for the SPICE Optics Unitduring perihelion operation, at end-of-life. are relatively low, this re-radiates to the internal surfaces of thespacecraft. The pre-slit mirrors are configured so that just the re-quired science beam is passed through to the slit (anything thatreflects on to them via the primary mirror is not part of the sci-ence beam). Ba ffl es also intercept radiation that either divergesas it comes into the instrument or is o ff -axis due to the spacecraftpointing away from the Sun centre.At perihelion, the instrument can survive thermally when thespacecraft is o ff -pointed by up to 3.5 ◦ in any axis in steady-state,and by 6.5 ◦ for a period of up to 50 seconds. This gives su ffi -cient margin for nominal spacecraft operations which will neverpoint beyond the solar limb, maximising at 0.94 ◦ at perihelion.At the distance of 0.7 AU, the steady-state o ff -pointing limit in-creases to 4.5 ◦ , and at 0.95 AU the spacecraft can be oriented atany attitude without thermally a ff ecting SPICE. The above ther-mal analysis shows how far of an o ff -axis pointing SPICE cansurvive (without necessarily being required or able to do sci-ence).The primary thermal design driver is to manage, at perihe-lion, the solar load incident through the 52 ×
52 mm aperture inthe spacecraft heat shield. Fig. 6 illustrates the thermal modelpredictions for the nominal perihelion case, with end-of-life(EOL) thermo-optical properties. About 66% of the 31.7 W en-tering the SOU cavity is transmitted through the primary mir-ror and reflected directly to space by the HRM. The remaining10.7 W is absorbed within the SOU structure. Of this, 2.5 W isdirected to the heat dump radiator, where it is radiated to thespacecraft. In the perihelion case, the surrounding spacecrafttemperature is specified as 50°C. The instrument structure aver-ages about 55°C due to the addition of absorbed solar loads andinternal dissipation. It is noted that the view of the HRM struc-ture to deep space provides radiative cooling, which reduces theheating e ff ect of the absorbed solar loads. About half of the inter-nal absorption of solar flux occurs at the primary mirror (at thecoating and within the silica substrate). The primary mirror istherefore warmer than its surroundings, operating at about 70°C.The detector assembly is conductively and radiatively iso-lated from the instrument surroundings at the mounting interfaceusing polyether ether ketone (PEEK) stando ff s. Heat generatedinternally and the low levels of parasitic heat to the assembly arerejected to the spacecraft-provided cold element interface. Thisallows the active pixel sensors within the DA to be passivelycooled. They are then individually PID (Proportional-Integral-Di ff erential) controlled by SEB-powered heaters to a set-pointof − Article number, page 10 of 26PICE Consortium: The Solar Orbiter SPICE instrument
6. Mechanisms and detector assembly design
The SPICE Door Mechanism (SDM) provides a contaminationseal at the entrance aperture of the instrument. It protects thehighly sensitive internal optics during non-operational periodsduring the ground integration and test phase, and during thecruise and non-operational phases in flight. The mechanism con-sists of the door itself (with a highly reflective finish and spher-ical shape to reject the incoming high intensity solar flux dur-ing some flight phases), which is articulated on linear bearingsand driven by a stepper motor (with reduction gear-box) and ballscrew. The door maximum temperature is ~125 ◦ C when closedat perihelion. However, the door and mechanism design is notthermally qualified for the transient case of being opened orclosed while sun-illuminated, and this means that for these oper-ations, the outer heat-shield door must be closed.The SDM provides the defining aperture for the instrumentoptical design, including a knife edge to control the stray-lightimpact of the aperture and reject the oversized beam passingthrough the heat shield feedthroughs. The door forms a labyrinthseal which is contamination-tight, but allows purging of the op-tical cavity during Assembly, Integration and Verification (AIV)up to launch. The SDM is designed and qualified for up to 100open-and-close cycles so that it can be used repeatedly both dur-ing AIV and during flight between the remote-sensing windows(to limit contamination from the spacecraft entering the instru-ment). The door opening and closing operations involve drivingthrough motor steps, between the end positions. These are de-tected by position switches at each end of the range; in addition,the steps are counted. At the step rate used, the time required toopen and close is approximately 50 s. The component parts ofthe SDM design are illustrated in the left panel of Fig. 10 andthe flight model mechanism (integrated to the front panel of theSOU) is shown in the right panel.
The required range of motion for the mirror is 0 − (cid:48) in rotation(about an axis parallel to the slit-direction), and ± . ff ness in the other degrees of freedomto maintain alignment and stability. This is achieved by using aflexure-based design, with multiple blade flexures for strength.It is a two-stage design comprising (1) a linear stage for focus,driven by a roller-screw mechanism with stepper motor, and (2)a rotation (scan) stage. This is mounted on the linear stage, andis driven by a piezo-electric actuator via a lever arm. The mirrorassembly itself is mounted on this rotation stage, and the adjust-ments for optical alignment during the build are made at thisinterface.The motions of both stages are sensed by linear variable dif-ferential transducer (LVDT) sensors. The sensor for rotation isconnected to a lever arm, which amplifies its displacement to al-low a suitable accuracy of the sensor for the control system. Thisrotation (scan) has closed-loop control to give the required sta-bility and step resolution ( ≈ (cid:48)(cid:48) ). This mirror scanning is usedregularly during many of SPICE’s observing sequences, and theresponse time for scan stepping with a small range is typically0.25 secs. The linear (focus) stage has step size 0.4 mm, and po-sition accuracy is < µ m at a given temperature. During obser-vations the focus setting is changed only infrequently, for exam- ple during calibrations or after change in instrument temperature(depending on mission phases). The slit change mechanism has the function of positioning anyone of the four slits into the active slit position, to the re-quired absolute and stability tolerances (in particular as regardsthe spectrometer focus and spectral calibration). The slits aremounted in-line in a single carrier mounted to the mechanismstage. The physical length of each slit is approximately 2.5 mm(slit length including dumbbells, for 14 (cid:48) angular size on the Sun),and the spacings between adjacent slits are 5, 6, and 5 mm, so thetotal range of motion needed is 16 mm plus margin. To changefrom one slit to any other, is a single linear movement. Due to thefixed mechanism speed the time taken is a minimum for movebetween adjacent slits, to a maximum for a move between theend slitsFor the mechanism design the slits carrier is mounted be-tween two large leaf-spring flexure blades (titanium), in orderto provide the required motion range (along-slit direction) whilemaintaining sti ff ness in the across-slit (spectral) and focus axes.It is driven by a stepper motor, which drives a Rollvis satellitescrew, which translates the rotary output of the stepper motorinto linear motion. The step size of the slit motion is 0.02 mm.The assembly includes mu-metal magnetic shielding that greatlyreduces the magnetic signature of the Sagem stepper motor. Alimit switch is implemented at one end of the range of travel,0.25 mm from the 6 (cid:48)(cid:48) slit. All operations of the mechanism areperformed by counting the number of motor steps moved rela-tive to this limit switch, which defines the ‘zero’ position. Stan-dard operations will include driving the mechanism back to thelimit switch on a regular basis (after every few slit changes), toensure that the positional reference is maintained. The standardtravel speed of 0.5 mm / s allows an adjacent slit to be selectedin 10 −
12 s. A slower drive speed is used for finding the limitswitch, but the reset operation can always be completed in lessthan 1 minute, from any given starting position.
The primary purpose of the detector assembly door mechanismis to control contamination, mainly molecular, as well as humid-ity, which can both be detrimental to the accuracy of the detectoras well as the functionality of the MCP. It is critical that the doornever be opened when not at vacuum or in <
30% humidity envi-ronment, or the intensifiers will be damaged.The actuator for the door mechanism is a controllable driveactuator. This motor was selected as a smaller mass option to theother motors on the SPICE project while still meeting require-ments. The motor is able to achieve 2.3 Nm of torque across theoperating temperature range of the SPICE DA. Conversion ofthe rotary stepper motor motion into linear displacement will beperformed using a worm gear, while a pin-and-slot feature on thedoor part will allow the door to open and close.Due to the contamination requirements of the instrument andthe relatively undemanding life requirements, dry lubrication isused on the motor mechanism. The door actuation cycles will below for this mechanism. The qualified number of cycles for thismechanism is one in flight and 20 during ground testing.
Article number, page 11 of 26 & A proofs: manuscript no. SO_Book_SPICE_paper
Fig. 10.
Door Mechanism. The drawing on the left highlights its components, the picture on the right shows the assembled flight model.
Fig. 11.
Scan-Focus Mechanism. The picture on the left highlights focus and scan directions, the drawing on the right shows how the instrument’sprimary mirror is mounted on the mechanism.
The SPICE DA (Fig. 13) consist of two independent, identical,intensified APS camera systems mounted in a common sealedhousing. Each camera consists of a HAS2 (High Accuracy Star-tracker 2) 1024 × µ m MCP pores and are accelerated by a2800 V potential across a sealed proximity gap (0.5 mm) onto analuminised phosphor screen deposited onto a fibre-optic outputwindow. Electrons are converted into a visible-light image at thephosphor screen. The resulting image is transferred through a fi-bre optic coupler to the APS sensor. A direct bond between thefibre optic and the APS seals the APS active area and eliminatesenvironmental contamination on the APS. The fibre optic cou-pler is similarly bonded to the MCP fibre-optic output window,again eliminating environmental contamination. The MCPs arescrubbed to stabilise the MCP gain against localised charge de-pletion. After scrubbing, the MCP housing is maintained at lowhumidity levels until launch to maintain sensitivity. Article number, page 12 of 26PICE Consortium: The Solar Orbiter SPICE instrument
Fig. 12.
Slit Change Mechanism.
The APS detectors are cooled (by conductive link to thespacecraft cold-element) to minimise the dark-current, and thetemperature is stabilised using heaters on the detector thermalstraps to achieve − ◦ C. This approach allows the spacecraftcold element temperature to vary during the orbit, while stillachieving stable detector performance (which is sensitive to tem-perature).The front-end electronics (Sect. 7) are located close to thefocal plane to maintain the integrity of the clocking and ana-logue output signals. Within the front-end electronics, a Field-Programmable Gate Array (FPGA) will accept configurationcommands and generate the timing signals needed to operate theHAS2 sensor as configured. The analogue video signal from thesensor will be digitised to 14 bits precision.
7. Electronics (including FPA hardware and flightsoftware)
The SPICE instrument has two electronic units: the SPICE Elec-tronics Box (SEB), which contains the control and data process-ing, and the Front-End Electronics (FEE), which form part of thedetector assembly within the SOU.
The SEB circuit boards are:(a) High Voltage Power Supply (HVPS) for the MCP, Gap andParticle Deflector(b) Data Processing Module (DPM) incorporating the spacecraftinterface, signal processing and instrument control(c) Two Mechanism Interface Modules (MIM) driving the mo-tors, scan mirror, position sensors and SOU heaters(d) Low Voltage Power Supply (LVPS) which converts thespacecraft 28V to the SEB internal supplies(e) Backplane which is used to transfer the various signals be-tween the other 4 circuit boards.
The SEB utilises a single chassis with 4 mm wall thickness thathouses the six circuit assemblies. The five plug-in cards are de-signed with wedge type card retainers that interface to machined card slots, providing both structural support and a conductivethermal path to the chassis. After all modules have been insertedinto the chassis, a single front panel is installed on the front,which along with rabbet joints at all panel to panel mating, min-imises electromagnetic radiation exiting the chassis, and cos-mic radiation entering the chassis. The chassis uses vent holessized such that there is adequate ascent depressurisation capa-bility while maintaining a small aspect ratio to give good EMC(electromagnetic compatibility) performance.
The DPM provides essential data processing, instrument com-manding and science operations management for the SPICE in-strument. The DPM is the central part of the SEB and performsthe following functions: – Hosts 8051 microcontroller and flight software (FSW) – Command & telemetry interface to the spacecraft – Control of instrument mechanisms – Command & control of FEE – Image processing and compression – Control of the high and low voltage power suppliesControl and management of the mechanism operation is pro-vided by the DPM through a combination of software functionsoperating in the 8051 and hardware resources within the com-mand and control FPGA. The DPM mechanism control inter-faces include: – SPICE Door Mechanism – Stepper motor for opening and closing SPICE door – Thermistor for monitoring mechanism temperature – Microswitches for detecting when door position – Slit Change Mechanism – Stepper motor for moving the slits carriage, movable inhalf or full steps – Microswitch for detecting when slit is in home position – Thermistors for monitoring the temperature of the slitmechanism – Scan Mechanism – Lead zirconate titanate (PZT) actuator for controllingscan mechanism
Article number, page 13 of 26 & A proofs: manuscript no. SO_Book_SPICE_paper
Fig. 13.
Left: Detector Assembly design concept (with door open), Right: Detector Assembly within the SPICE SOU (door closed to protect theMCPs).
Fig. 14.
Detector Assembly, exploded view. – Linear variable di ff erential transducer (LVDT) sensor formonitoring the position of the scan mechanism – Closed loop control algorithm – Thermistor for monitoring temperature of mirror mount – Thermistor for monitoring temperature of scan stage – Focus Mechanism – Stepper motor for controlling focus mechanism – LVDT for monitoring focus mechanism position – Thermistor for monitoring temperature on focus stage – Detector Assembly – Stepper motor for controlling detector assembly door. – Thermistors for monitoring the temperature of the detec-tor assembly – Microswitches for monitoring detector door position – SPICE Optical Unit – Thermistors for monitoring the temperature of the SPICEoptical unit – HeatersThe DPM controls the high voltage power supply outputsthrough a set of digital to analogue converters located on the DPM and routed via the backplane to the HVPS together with anumber of discrete digital signals for on / o ff control etc.The DPM uses a single SpaceWire interface for FEE con-figuration and control. The FEE and science acquisition settings(e.g. exposure duration, pixel selection for readout, gain) are pro-vided by the DPM resident FSW utilising a simple set of regis-ters within the DPM image processing FPGA. These registersare used to pass configuration information via the SpaceWire in-terface to the FEE.The LVPS module provides the power interface to the space-craft. It accepts the 28V input, provides noise filtering, and con-verts it to the required voltages to run the other boards and com-ponents inside the SEB (1.5, 2.5, 3.3, 5.0, 5.6, and ±
12 V). Afiltered 28 V supply is also provided for the heaters and step-per motor drivers. The design considers the need for good EMCperformance, in order to minimise any conducted emissions thatcould a ff ect other instruments on the spacecraft. The HVPS for the SPICE instrument provides the high voltage(HV) required to run two MCP and intensifier pairs as well as aparticle deflector. The HVPS consists of five individual suppliessharing a common board, each with its own set of control signalswithin a common low interface, and each responsible for a HVoutput.The high voltages are: – Two MCP supplies: 0 V to 1275 V (nominal 850 V) at up to50 µ A for each supply. – Two gap supplies: 0 V to 3570 V above MCP voltage at upto 10 µ A for each supply. – Particle deflector supply: -2500 V at 10 µ A. The SPICE FSW is written primarily in the C programming lan-guage, with a small amount of assembly language for initial bootup at power on. The Keil Integrated Development Environment(IDE) was used for creating, editing, compiling, linking and test-ing (through 8051 emulation) the FSW.
Article number, page 14 of 26PICE Consortium: The Solar Orbiter SPICE instrument
Fig. 15.
Software Modes.
The SEB FSW runs on a Microsemi Core8051 IP core thatis part of the DPM CPU FPGA. The primary function of theFSW is to act as the data manager for the SPICE instrument bysending and receiving messages over the SpaceWire commandand telemetry interface to the spacecraft, controlling the mecha-nisms, heaters and HVPS outputs, managing the acquisition andprocessing of science data, monitoring and reporting instrumenthealth, and helping manage the safety of SPICE.The FSW resides in two non-volatile memories of the DPM:The boot code image is stored in the 32 KB PROM, and the sci-ence code image is stored in the 256 KB EEPROM device. TheEEPROM contains two science code images (prime and redun-dant) that can be up to 64 KB and the lookup tables (LUTs), sim-ilarly stored as primary and redundant images, each up to 64 KBin size.The modes for SPICE are Startup, Standby, Engineering andOperate as shown in Fig. 15. When SPICE is switched on theFSW immediately enters STARTUP mode and runs the bootcode contained in the PROM to execute initialisation and self-test routines. The boot image uses the SpaceWire interface tothe spacecraft to implement functions such as memory load anddump so that the LUTs or the science code images can be up-dated or checked. Prior to transition to the STANDBY mode,a check is performed on the science and LUT images stored inEEPROM to ensure that they are uncorrupted. If the images aresound, they are copied to SRAM and run automatically.When the boot to science image transfer occurs, the FSWwill be in STANDBY mode which is a stable and safe configu-ration for the SPICE instrument. Transition to the ENGINEER-ING mode is accomplished with a command. This mode allowsfor the ramping of the HVPS outputs connected to the micro-channel plates (MCPs), gaps and particle deflector, prior to tran-
Fig. 16.
Front End Electronics: Block diagram. sitioning to OPERATE mode. A command is required to tran-sition to OPERATE, which is the mode used for science imageacquisition, processing and telemetering.SPICE uses a bespoke, Consultative Committee for SpaceData System (CCSDS)-based science packet format to organ-ise the data e ffi ciently, while still including enough metadatato make it self-describing. This is required due to the complexstructure of SPICE science observations, which are organisedby windows (wavelength ranges) and compressed in one of sev-eral formats (see Sect. 7.9). Detailed information is presented inthe SPICE Data Interface Control Document (ICD). The sciencepacket contains image header information in the first packet, fol-lowed by image data in the remaining packets. A checksum ofthe image header and data is appended and stored in the lastpacket.If an anomaly occurs during Engineering or Operate mode,the fault detection, isolation and recovery system generates anevent message and then has a subsequent action of either to con-tinue operation if the anomaly is benign or to perform recoveryactivities. The Front End Electronics (FEE) boards are contained withinthe Detector Assembly (see Sect. 6.5) and due to space con-straints are a single assembly comprising three circuit boardswith flexi-rigid interconnects. This approach uses less volumeand improves the reliability of the FEE due to reduction in num-ber of individual connectors. The flexi-rigid approach also al-lows the boards to be ‘folded’ up so that that HAS2 active pixelsensors can be mounted in the correct position within the detec-tor assembly. The FEE is used to convert optical signals to elec-trical signals and transmit these to the SPICE Electronics Box.The assembly contains two HAS2 active pixel sensors (board 1),an analogue to digital converter (board 2) and a control FPGA(board 3), see Fig. 17.As shown in Fig. 16, the FEE provides the control and clocksignals needed by the two HAS2 detectors, along with the read-out links to a single ADC chip. The HAS2 detectors are config-ured and operated in such a way as they appear to be one largerdetector of 1024 × Article number, page 15 of 26 & A proofs: manuscript no. SO_Book_SPICE_paper
Fig. 17.
Front End Electronics board before assembly into the DetectorAssembly (left to right: boards 1 to 3). (windows). Windowed readout reduces the images size and thusbandwidth required for each image.The FEE is programmed by sending it two-byte SpaceWirepackets which write data to the 8-bit registers in the FEE FPGA.In configuring the FEE for a particular mode of image acquisi-tion, the SEB must: – Load the required image readout waveforms and tables intothe FEE waveform generator (WG) RAM – Load the window registers in the FEE to select the requiredpixel columns (wavelength ranges) – Configure the analogue front end, including the gains (coarse& fine) and o ff set for each sensor – Load configuration registers in the sensor.The above RAM and registers are loaded by writing ASCII char-acters to a FPGA register. These characters represent the I2Cdata link ‘start’ and ‘stop’ signals as well as hex digits whichconvey addresses or data.The FEE supports both destructive and non-destructive readout modes. Destructive readout is when the pixel is reset imme-diately after the signal level has been sampled. This black levelreset signal is then subtracted from the signal level during pixelreadout, thus eliminating any static pixel-to-pixel o ff sets of thesensor. Non-destructive readout mode is when the sensor is re-set and then immediately readout. This black level image is thenstored by the SEB. After the exposure time has elapsed anotherimage is readout and again stored by the SEB. The correspond-ing black level image is then subtracted from the signal image.This is correlated double sampling, which eliminates static o ff -sets as well as thermal (‘kTC’) noise.The RAL SpaceWire Interface FPGA IP core provides theFEE’s communications with the control and data acquisition in-terface. The communications interface is a SpaceWire adaptationof the IEEE1355 serial interface standard, featuring an LVDS-driven SpaceWire serial data link running at 100 Mbits / s. TheSpaceWire link can support a range of communication speeds,which are programmed by writing to registers. At reset, the trans-mit and receive links are configured to run at the default speedof 10 Mbits / s. A science ‘observation’ campaign is expected to be a series ofso-called ‘studies’, which are scheduled in the mission timeline.A study consists of a series of spectra (or sets of windowed spec-tral lines) acquired at successive positions of the primary mirrorscanning mechanism. For the end user, a study can be thought of as one λ, y , x raster of a ( λ, y , x , t ) data set (for this purpose, sit-and-stare observations can be thought of as having an X step sizeof 0). A raster cube is commanded by sending a STUDY_RUNtelecommand with the number of rasters (study repetitions) tobe performed as sole parameter. Studies are defined in on-boardlookup tables (LUT), which are loaded with the required defini-tions for the operational period (see Sect. 10.1). Some complexoperations will also use macros, which are internal commandsequences to be run on an execution engine within the FSW.Macros are also defined in the onboard LUT, and consist of a se-ries of SPICE commands with relative time tags that determinewhen each command is executed. These can (and often will) in-clude commands to run a study.When a study starts, the FSW reads the corresponding sci-ence LUTs which hold parameters to configure the SPICE in-strument for science acquisition. The LUTs contain: slit choice;exposure time, scan mirror start, stop and increment (step size)information; the number of scan repetition ‘loops’; and win-dow position, size, binning and processing information such aswhether to compress the data (and the method of compression)or to keep the data in a raw state prior to packetisation. The FSWthen commands the mechanisms, communicates with the FrontEnd Electronics (FEE), sets up the Image Processing FPGA(IPF) for acquisition, configures the Discrete Wavelet Trans-form (DWT) Bit Plane Encoder (BPE) Application-Specific In-tegrated Circuits (ASICs) for compression (if desired), and coor-dinates the science data product packetisation process and trans-mission to the spacecraft mass memory unit.The FEE must initially be programmed by the SEB to per-form the desired acquisition of spectral or image data and thenthe Waveform Generator (WG) to execute the appropriate tableto perform readout of the detectors. The resulting pixel valuesrepresenting an image frame are then received by the SEB inthe form of a SpaceWire packet from the FEE. The selection ofpixels contained in the packet depends on how the WG and win-dow registers have been programmed. The WG table that readsout the detector also determines the exposure by controlling thetime between reset and readout of pixel rows.The WG table used for readout, and its component wave-forms, also determines where SpaceWire end-of-packets (EOP)are inserted; the EOP is inserted after each complete (i.e. of allactive windows) readout of the detector. For SPICE, the read-out table is used to generate only a single acquisition (image) ata time. The timing of the start of each acquisition is managedby the FSW, while the readout table controls the timing of allFEE operations within each acquisition. This includes the nec-essary overheads for resetting the active pixels (i.e. the windows)within the image, which requires 0.42 s for reading a whole ar-ray (i.e. 2048 × The SPICE DPM IPF is a single FPGA containing the followingfunctions: – One FEE SpaceWire Interface – Black Level and Dark Current Corrections
Article number, page 16 of 26PICE Consortium: The Solar Orbiter SPICE instrument – Binning in Y (spatial) and λ (wavelength) – Bu ff er Synchronous Dynamic Random Access Memory(SDRAM) Interfaces – Flash Interface – Fast Fourier Transform (FFT) Processing – Interfaces to DWT and BPE ASICs Central to the sciencedata acquisition and processing is a pipelined structure in-volving the following pipeline stages: – Non-destructive and destructive pixel readout – Selectable black level correction – Selectable binning – Selectable dumbbell extraction for alignment tracking – Frame accumulation for SHC (Spectral Hybrid Compres-sion, see Sect. 7.9).The first processing blocks in the IPF after data is read fromthe FEE via the SpaceWire interface are the Corrections and YBinning blocks. In the Corrections and Y Binning blocks, pos-itive o ff set, black level subtraction, dark current correction, andY binning can be applied to the incoming image data. Thesecorrections are enabled based on FSW configuration data fromLUTs loaded into the IPF. The processing steps are conducted inthe following order with the result being stored in the windowbu ff er:(a) Black level subtraction(b) Positive o ff set(c) Dark current correction(d) Binning in the Y (spatial) direction (2, 4, 8, 16, 32, and64 pixels sizes).FSW manages the contents of this dark current correctionmap by initialising its contents with data. The FSW also copiesthe data from the non-volatile storage for its calculations, scalingthe pixel amplitude for the appropriate exposure time and writingthe result in the window bu ff er for dark map subtraction use. Thisscaling module provides arbitrary multiplication of a 14-bit pixelvalue and 13-bit exposure value as well as left shift by 6, 9, 11,or 12 based on an input shift code. The exposure value and theshift code are provided to the IPF at run time by flight software.If Y binning is enabled and if any one (or more) of the pix-els being binned are saturated above a specified 14-bit number( < = ff er block as an intermediate storage location.The window bu ff er comprises a single SDRAM memory device.From the window bu ff er block, data is routed to the Lambda( λ ) Binning block. Binning in the wavelength direction, if en-abled, can be selected in 2, 4, 8, and 16 pixel sizes. Like Y bin-ning, if any one (or more) of the pixels being binned are saturatedabove a specified 14-bit number ( ≤ λ window. This is a32-point FFT with 18 bits of precision. Data is sent to the FFTusing either the lower 12 or 14 bits of unsigned data, pre-scaledto occupy the most-significant bits of the input. Data internal tothe FFT is scaled at each of the internal FFT stages to preventdata overflow. Data retrieved from the FFT consists of the low16 bits of output. There are 2 real and 15 complex coe ffi cients per FFT. This amounts to 32 scalars, each being assigned to aparticular ‘coe ffi cient plane.’ Note, if SHC is enabled, the FFTfunction is executed automatically as part of that process andprecludes any use of lambda binning.After the FFT block, data enters the scan bu ff er block, whichis a second SDRAM intermediate storage location. Additionally,alignment windows (also known as dumbbells) are filtered di-rectly to the scan bu ff er after being received from the FEE if anobservation was configured to collect them. Alignment windowsare not corrected, binned, or FFT-ed. The scan bu ff er block is capable of organising multiple exposureframes of data accumulating a ‘cube’ of X by Y by λ images.From the scan bu ff er, data is transferred to the compression In-terface block.In order to meet the downlink telemetry volume constraints,the DPM incorporates a novel compression algorithm known asSpectral Hybrid Compression (SHC). SHC performs compres-sion on the Fourier coe ffi cients to achieve an up to 20:1 com-pression ratio. The data flow associated with the SHC algorithmincludes compression in the X by Y plane for each FFT coe ffi -cient plane with the following options: – A menu of 8 possible configuration ‘recipes’ that can be cho-sen from for best SHC compression are made available. – Each recipe has di ff erent BPE settings. – Compression recipes designate values for the configurationsettings made available by the ASIC interface module, de-scribed below. – Compression algorithm is CCSDS 122.0-B-1.Compression is performed by two specially programmed ASICswhich operate at a maximum rate of 5 MPixels / sec to reduceoverall power consumption. Each ASIC is connected to two ded-icated SRAMs for use during compression. The compressionhardware works on individual slices of the ‘cube’ consisting ofseveral FEE images, corresponding to di ff erent scan mirror (i.e.X) positions. The ‘data cube’ thus has two spatial axes and onewavelength axis ( x , y and λ ). It is also possible to disable thecompression methods such that raw pixel data is transmitted tothe ground.The ASIC interface module within the IPF provides a mecha-nism to configure each ASIC independently with the image com-pression parameters required to achieve the desired compressionfunction for the current image data. The DWT ASIC must be re-set prior to reconfiguration (e.g. a change to the image height orwidth), and as a consequence all ongoing data must be flushedfrom the chip prior to the change. The BPE ASIC can be recon-figured while compression processing is ongoing and providessynchronisation mechanisms so new parameters are applied onlyon an image boundary.The ASIC interface module provides the following configu-ration settings: – Height : Height of the image in pixels (32 to 1024). – Width : Width of the image in pixels (32 or 64). – SegByteLimit : This is the number of 16-bit compressedwords used to represent the image. – S : The number of 8 × – SignedPixels : Whether the input image data is signed orunsigned.
Article number, page 17 of 26 & A proofs: manuscript no. SO_Book_SPICE_paper
The other ASIC configuration settings have been left at their de-fault values so that the compression conforms to the CCSDS122.0-B-1 specification. After compression, data is transferredto the output bu ff er block where it is stored until transmission tothe spacecraft using the appropriate science packet structure (seeSect. 7.5). This bu ff er is used for both compressed and uncom-pressed science data.
8. SPICE testing
In this section we describe the testing methods used to verifythe key performance and engineering requirements of the instru-ment. The molecular and particulates cleanliness of the instru-ment is a key requirement, and so in the instrument build andthroughout all of the testing the cleanliness control is an impor-tant aspect. The overall flow of the tests described in this sectionare summarised in the flow diagram in Fig. 18.
In common with all space instruments that observe in the UVspectrum, and especially in the EUV, SPICE is highly suscepti-ble to degradation e ff ects of contaminants: – Darkening of optical surfaces, meaning loss of reflectivity,due to build-up of deposited molecular contamination – Increase in scattered light (stray-light e ff ects) due to particu-late contamination on optical surfaces.For both of these, the e ff ect is so great that very stringentcontrol of these contaminants is needed throughout the flight-hardware program. The parts in the optics unit must be manufac-tured as initially very clean or ‘pristine’, and then protected fromcontamination (and verified as clean), throughout the groundtesting. For the molecular e ff ect, the main source is hydrocar-bons and silicones such as from wire insulation and epoxies inthe assemblies. The total allowance is 200 ng / cm per opticalsurface (~50 nm film thickness), for which the light-throughputloss at the SPICE bands is ~20%. For the scattered light e ff ect,the allowance is 100 ppm, for which the e ff ect is then as low asthe scatter from optical surface roughness ( < – In the design, selection of approved materials with correctmechanical and low-outgassing properties, and preparationprocedures; – In the fabrication of all parts, use of precision cleaning, fol-lowed by pre-bake at the highest possible temperature beforeassembly; – For use of epoxies, use of low-outgassing types, verified bybatch-testing, de-gassing before use, and curing using bu ff ergas flushing to prevent re-deposition of outgassing products.For the complete parts, further features needed were: – Protection of the parts from lab air, or during transport, bystorage in purged double-bags or containers. – For the instrument, continuous gas-purge of the unit withclean gas (dry N ), to prevent ingress of contaminated labair (this purge is maintained until launch). – A separate continuous purge of the detector assembly withdry high-purity nitrogen gas, to protect the KBr photocath-ode coatings from exposure to humidity. – Precautions in vacuum-tests, to prevent out-gassed productsfrom re-deposition on optics, especially during the processof warm-up and venting of the chamber at the end of a test. – Verification of the cleanliness levels throughout the buildand test phase, by monitoring the environment using wit-ness samples. For molecular contaminants, these were op-tical plates that were measured using the ESA FTIR spec-trum method (with sensitivity to key contaminant species of~50 ng / cm ). For particulates, particle-fall-out plates wereused. These were used for regular monitoring of the cleanroom, and for each specific test. In addition, the vacuum testfacilities and all bake-outs had to be verified using quartzcrystal microbalance (QCM) measurements.These cleanliness precautions are at similar stringency to thosein ultra-high-vacuum experiments in other science areas usingUV (e.g. synchrotron facilities). For several of the subsystems, performance and calibration testswere made by the supplier, at unit-level. This was in part to al-low the tests to be made early to reduce risk, and also in somecases because more detailed characterisation could be done atthis level. The main examples are: – Mechanisms: for scan-focus, and slit change, detailedmetrology of the mechanism responses, such as stability, re-peatability and thermo-mechanical e ff ect – Optics (mirror, grating and photo-cathode) spectral re-sponses – Detector assembly, functional and UV tests, and environ-ment test, especially thermal as detector cooling is a criticalrequirement
The di ff erent stages of instrument-level testing are given in theflow diagram in Fig. 18. The main features are:(a) The first instrument-level test was the ‘first VUV test’ forwhich the instrument SOU was only part-built. This test wasneeded immediately after the optics build (and is described inSect. 9), and before completion of the rest of the SOU parts.This is because it was anticipated that alignment-adjustmentmight be needed after these first VUV images were obtained.This test is considered as instrument-level as it uses the samesystem set-up as the subsequent VUV tests.(b) The SOU environmental tests (vibration and thermal-vacuum), are made on the final build of the unit. A VUVreference test, meaning a set of tests to verify all the sensorperformances, with the instrument in the vacuum-test cham-ber, was made before the vibration test, in between vibrationtest and thermal-vacuum and thermal-balance (TV / TB) test,and again after TV / TB.(c) The instrument has extensive bake-out (out-gassing) test re-quirements, and the set-up required for this is similar to thatfor TV-test. Likewise the final calibration tests require thefull temperature range of the instrument. For this reason, andto save time as well as minimise handling, all four tests (post-vibe reference test, TV / TB, bake-out, and calibration) wereplanned into the same vacuum-test procedure. The approxi-mate durations of each phase were: TV / TB: 2 weeks, bake-out: 1 week, calibration: 1 week.
Article number, page 18 of 26PICE Consortium: The Solar Orbiter SPICE instrument
Fig. 18.
Flow diagram of instrument-level tests.
Fig. 19.
Instrument calibration: Test set-up with hollow-cathode (HC)source at the Physikalisch-Technische Bundesanstalt, Berlin. (d) After completion of final calibration, the SOU was removedfrom the test-chamber, and the remaining tests done werebench tests, for functional, EMC and magnetics testing.With this scheme, there were three test-campaigns in the vacuumchamber facility in total (see Fig. 18).
A schematic of the facility is shown in Fig. 19. The hollow-cathode (HC) based calibration source was originally developedfor use on the SOHO / CDS instrument (Lang et al. 2000), incollaboration with the Physikalisch-Technische Bundesanstalt(PTB, Berlin), and later used for the Hinode / EIS instrument,where a detailed description of its use was given (Lang et al.2006).The solar instrument test chamber is a 1-metre diametercylinder tank, in which the SPICE optics unit is positioned asshown, with the optical axis along the chamber axis. The flightharnesses are included on the unit, but for reasons of molecu-lar cleanliness (out-gassing) the Electronics Box is positionedoutside the chamber, using harness extensions through chamberfeedthroughs.The VUV test source used has been described previously byLang et al. (2000). It is a collimated test beam produced by thehollow-cathode plasma discharge lamp (Hollandt et al. 1993), which has an exit aperture of 0.6 mm diameter placed at thefocal-point of a 2-mirror grazing-incidence telescope acting asa collimator. For use in the test, the source vacuum-system hasto be coupled to the instrument chamber via a beam-pipe; this istemporarily installed for these tests and so is isolated from theinstrument and source chambers via gate-valves. In this config-uration it is not possible to view directly either the instrumentaperture, nor its alignment-cube, and this is an obstacle to set-ting up the source, and to determining the instrument pointing(detector line-of-sight relative to the cube), which is an impor-tant part of the test aims. To overcome this and align the system,the chamber uses a viewport which is o ff -axis (adjacent to thebeam pipe), to allow an optical alignment reference attached tothe instrument bench to be viewed. The alignment of this refer-ence relative to SPICE then has to be measured before the cham-ber is closed. In addition, the beam-pipe includes a deployable45 ◦ mirror, which can be inserted in order to obtain a visible-light view into the instrument (while blocking the VUV path,however). The vibration test campaigns on the SPICE SOU Flight Spare(FS) and Flight Model (FM) were performed and successfullycompleted in March 2016 and February 2017, respectively. Dur-ing the FS vibration test an additional optical alignment mea-surement was carried out on the non-adhesive mirror mount inorder to verify the mechanical stability for each load case. Dur-ing both test campaigns the instrument was double bagged andpurged to comply with cleanliness and contamination control re-quirements. Responses from the applied accelerometers (19 forFS, 13 for FM) were acquired, stored and analysed for each loadcase to verify the applicable requirements in terms of structuralintegrity and minimum resonance frequency as well as to im-prove correlation with the numerical models (both NASTRAN
Article number, page 19 of 26 & A proofs: manuscript no. SO_Book_SPICE_paper and ANSYS finite elements models). The tests were carried outin the order foreseen by the test procedure, namely: – Initial sine survey – -12 dB bedding-in random vibration – Bedding-in sine survey – Random vibration – -12 dB, -9 dB, -6 dB, -3 dB, sine survey, 0 dB – Intermediate sine survey – High-level sine – Final sine survey.During the FM test campaign, for the Z axis testing only, an addi-tion -6 dB random vibration run was conducted in order to verifythe e ff ectiveness of the notching applied to protect the SCM unit. To implement the needed thermal cases, the instrument was po-sitioned on a mounting plate, and within a thermal enclosure,both of which are heated / cooled by a fluid loop heat exchanger.The cooling for the detector cold-element is by a separate fluidloop.The thermal test campaign consisted of a thermal vacuumtest and a thermal balance test. The objectives of these testswere to demonstrate the operation of the SOU across the pre-dicted flight temperature range, to provide environmental stressscreening and to provide well-characterised data for validationand correlation of the thermal model.All the objectives of the test were ultimately satisfied. In ad-dition, the test provided useful information regarding many op-erational aspects of the instrument, such as the PID control ofthe heaters used within the detector assembly. The data acquiredduring the thermal balance test enabled the thermal model to becorrelated to a high level of accuracy, thus giving confidence tothe flight thermal predictions that were subsequently produced. The functional elements of SPICE (motors, detectors, high volt-age supplies, etc.) are only fully operational in a vacuum envi-ronment. In order to demonstrate this functionality before, dur-ing and after environmental testing (vibration and thermal vac-uum tests), full functional tests were carried-out to exercise allcapabilities of the instrument under vacuum. Reduced tests werecarried-out in the hot and cold phases of the thermal vacuumtest, to demonstrate compliance with requirements on the max-imum and minimum operating temperatures of the instrument.The tests were also designed to provide engineering calibrationdata, such as the sensor reading versus angle relationship for thescan mirror.All tests were successful, demonstrating that SPICE func-tionality was una ff ected by exposure to the instrument level vi-bration and in-flight thermal environments. The tests also al-lowed the team to develop and debug the operating sequencesrequired for in-flight commissioning, such as opening the doorsand switching on the detectors for the first time. The main requirements for these tests were to show su ffi cientlylow field e ff ects (on the other Solar Orbiter instruments) for: (a)Magnetic dipole (i.e. direct-current e ff ect), due in particular tothe motors in the SPICE mechanisms, which have a permanent dipole, even when not operating; and (b) AC magnetic fields, dueto motor operations and devices in the electronics box.For the magnetic-dipole, the main measurements were atsubsystem level on FM or FS parts, and made at the NationalPhysical Laboratory (NPL, London, UK). For the AC-magnetics,the measurements were at instrument-level on the optics unit(EM model) and electronics box (FS model), made at the sametest house as used for EMC testing. These measurements werewith AC magnetometers placed at di ff erent locations and dis-tances around the SPICE units.The dipole results were within specifications for all subsys-tems apart from the Slit Change Mechanism (SCM). This resultmatches the EM model, where it was identified that magneticshielding was required for the SCM. The shielding did not per-form as well as anticipated, and the overall dipole results for theinstrument were subject to a waiver. However, since the dipoleis constant, this can be accounted for in the operations of themagnetometer instrument.The AC results showed that the SCM and focus motors gen-erate emissions at 25 Hz, which was expected. Future spacecrafttesting will determine the level of interference that this causes,but it is not expected to be a significant issue. The timing ofany motor operation is recorded by the SPICE flight software,allowing the interested parties to identify the source of the dis-turbance, and to account for it in their own measurements. Thiswill minimise the impact on science operations. The main EMC tests were performed on the EM SEB and op-tics unit. A full suite of susceptibility and emissions tests wascarried-out at a test house (Element, Dorset, UK), which pro-vided the anechoic chamber and equipment needed to meet therequired test standards. Further tests were done on the FM andFS hardware, in agreement with ESA. A subset of conductedemissions and other electrical tests was performed on the FMoptics unit at RAL, since the test house could not meet the re-quired cleanliness standards.The susceptibility testing was very successful, and provideda high level of confidence that SPICE will not be disturbed byemissions from other Solar Orbiter instruments or the spacecraftitself. Emissions testing on the EM identified several areas ofconcern, and updates to the FM design were implemented basedon the lessons learned. The FM results confirmed that motor op-erations will be considered as ‘EMC noisy’, but normal sciencestudies (with mirror scanning) are not expected to have any sig-nificant impact on the rest of the spacecraft. As with the mag-netic testing, future tests at spacecraft level and in-flight willcharacterise this further.
9. Characterisation and calibration of performance
Because the detector assembly is a new development, and its per-formances are central to the final performance of the instrument,extensive tests were made at detector level. These were aimed atverifying these aspects: – Spatial resolution: by recording images of test-targets andvarious size pinholes on each detector, including through-focus.
Article number, page 20 of 26PICE Consortium: The Solar Orbiter SPICE instrument
Table 4.
Detector quantum e ffi ciency (QE) measured on the micro-channel plate wafer prior to detector assembly. Results are given forshort-wavelengths (SW) and long-wavelengths (LW) channel. λ QE (%)(nm) SW LW49 . . . . . . . . . . . . . . . . . . . . . . . . . .
22 0 . – Radiometric calibration: by including in the set-up a cali-brated VUV photo-diode, to measure the beam power, andcombining with detector quantum e ffi ciency measurementsmade on the micro-channel plate wafer prior to detector as-sembly (Table 4). – Detector throughput: performing radiometric calibrationmeasurements at a variety of MCP voltages allows the op-timal voltage to be determined. – Linearity: controlling the intensity of light incident on the de-tector, combined with varying the exposure time, determinesthe response curve. – Flat-field: the detector is mounted on an x - y stage so thatimages can be taken versus varying lateral position, in orderto allow the e ff ect of beam-shape to be removed from theimage data to produce the true flat-field response. – Noise: the statistical properties of the detector are exploredby taking multiple images under the same observing condi-tions. – Dark signal: Images are taken at varying exposure times withno light, to assess the thermally generated signal in the APSas a function of temperature.The tests were made on di ff erent models of the detector; first onearly prototypes, then on EM and finally on FM.In the detector test set-up, developed and made by GSFC,the Detector Assembly is placed in the RAL test chamber, andilluminated with a test beam from a commercial krypton lamp(123 nm). The lamp output is collimated to a beam diameter ofapproximately 1 cm, and this illuminates the detector either di-rectly (to fill the array for flat-field tests), or via a focusing mirror(for imaging tests). The collimator has a series of exchangeablefield-stop targets (pinhole apertures) to image on to the detector,and also a series of pupil apertures to vary the intensity (imagingf-number). These detector-level tests cover the properties dis-cussed above.For the FM, the final flat-field patterns are as shown inFig. 20. Each detector shows a hexagonal fine scale structure dueto the micro-channel plate intensifiers and the fibre optic blocks,with a 1 σ variance of 1.6% for the SW channel, and 2.3% forthe LW channel. In addition, there are large scale di ff erences inthe response which range from about 0.8 to 1.5 for the SW chan-nel, and 0.9 to 1.3 for the LW channel, relative to the median,with the highest values near the detector edges. Both the largeand small scale variations in the flat field are removed duringstandard data processing. Fig. 20.
Flat-field (relative responsivity across the arrays, shownin grey-scale), for SW and LW detectors, in the APS image(1024 × Fig. 21.
Telescope through-focus wavefront error (focus term), showingthat optimum focus is at ~100 µ m from mid-range (range of mechanismis ± µ m). The single mirror telescope consists of the entrance-aperture, themirror and the SPICE slit. The telescope optical performanceand alignment to instrument cube was tested in visible light dur-ing the instrument build. For the image-quality test, a wavefronterror (WFE) measuring interferometer was used, probing thetelescope in double-pass, by using a reflective-sphere located atthe slit position, for the on-axis FOV position.In final alignment, the telescope achieved a WFE of ~0.2waves peak-to-valley at 633 nm (centre of FOV, best-focus),which is adequate for required spatial image quality. Thethrough-focus WFE was also measured, to verify that the tele-scope is set at best-focus (at mid-range of focus-mechanism),and the result is shown in Fig. 21.
The tests for VUV performance were performed in the final testsetup in a thermal-vacuum chamber, as described in Sect. 8.3.2.For this the test-beam was a collimated VUV calibration sourceoriginally developed for use on the SOHO / CDS instrument(Lang et al. 2000). This source has some geometric limitationsfor SPICE testing, in that it has (a) a small aperture of ~5 mm di-ameter, compared to the SPICE aperture of ~43 mm, and (b) anangular range of ~1.7 (cid:48) diameter, which is larger than the SPICE
Article number, page 21 of 26 & A proofs: manuscript no. SO_Book_SPICE_paper
Fig. 22.
Composite SPICE spectral image (argon spectrum), of the threebeam positions used. On the left, the slit image is displayed on the samescale. The identification of the hollow-cathode-source argon lines wave-lengths is derived from the literature. It should be noted that the slitimages appear at di ff erent row positions on the SW and LW detectors. Fig. 23.
Measured spectral-line profiles of all 4 slits (2 (cid:48)(cid:48) , 4 (cid:48)(cid:48) , 6 (cid:48)(cid:48) , and30 (cid:48)(cid:48) ), at 74.03 nm wavelength (detector column ~450). angular resolution of ~4 (cid:48)(cid:48) . The source was operated to obtainthe VUV emission line spectrum of argon, known lines of whichoccur in the SPICE spectral range. The instrument performancewas characterised at three beam positions in the along-slit FOV,as shown in Fig. 22.To verify that SPICE meets its spectral resolution require-ment, the 2 (cid:48)(cid:48) line images were analysed, for their LSF width(FWHM) versus spectral position (di ff erent argon lines), FOVposition, and instrument temperature. It is found that the in-strument meets its requirement of 4-pixel FWHM over most ofthese ranges, and the worst-case value at the edge of the FOV is~4.5 pixels. The LSF was also checked versus slit number, andthe plot in Fig. 23 shows the spectral line profile of all 4 slitsoverlaid.Other tests were imaging tests versus temperature (instru-ment hot and cold operational cases), and a test of radiometricsensitivity, made by scanning the wide slit across the test beam,to measure total beam signal. This was combined with the knownsource beam power (in the argon lines 71.8 to 74.5 nm, for whichPTB have a power measurement). Table 5.
Measured SPICE wavelength ranges and related dispersion.
De-tector NominalRange(nm) Measured / Useable Range(nm) DerivedDispersion(nm / pixel)SW 69.8593 –79.5931 69.7008 –78.9280 0.009562LW 96.8704 –105.361 96.8783 –104.919 0.008307 The wavelength calibration of SPICE has been derived usingdata described in Sect. 9.3.The measurements have been taken at room temperature(with a grating temperature of 18 . (cid:48)(cid:48) slit respectively. Spectralimages were formed around the middle of the slits and spectralintensity curves were made by averaging over 29 pixels (SW)and 39 pixels (LW) along the slit length. The spectral line pro-files were fitted using Gaussian shapes to derive the centroidsand the FWHM.The pixel-to-wavelength calibration was made for the 2 (cid:48)(cid:48) slit by comparing the list of measured lines and their centroidsagainst a list of standard lines (Hollandt et al. 1996), usinga linear fit. The dispersion derived by the fit coe ffi cients was0.009562 nm / pixel for the SW and 0.008307 nm / pixel for theLW channels. These values were compared to the design valuesof 0.009515 nm / pixel and 0.0083 nm / pixel, showing satisfactoryagreement.The wavelength ranges established by this wavelength cal-ibration are shown in Table 5 and are well within the require-ments. Only the e ff ective useable pixels are included to deter-mine the wavelength ranges in Table 5. These are 10 to 975 forSW and 35 to 1003 for the LW detectors, as derived from flatfield measurements.This calibration is also valid for the 4 (cid:48)(cid:48) and 6 (cid:48)(cid:48) slits. The 30 (cid:48)(cid:48) slit is much wider and contains also spatial information, but theresults are still consistent with the other slits.The wavelength calibration varies with north-south locationon the detector due to relative roll of the slit and the detector.The roll angles are 0.017 radian for the SW, giving a maximumshift of 5 pixels in the spectral direction at each end of the slit,and 0.007 radian for the LW giving a slightly smaller shift of2 pixels.VUV measurements were also performed at di ff erent gratingtemperatures to analyse the thermal cycle variations and spectralshift, which will a ff ect the instrument along the orbit. Two maincases are reported: a cold case where the temperature of the grat-ing was − (cid:48)(cid:48) slit was used for this investigation.We have found a shift of around 10 pixels towards longerwavelengths for the cold case and a shift of around 15 pixels to-ward shorter wavelengths for the warm case, which should sim-ulate the Solar Orbiter perihelion approach. This analysis will bedone again in-flight, as the conditions might change and a curvefor this shift will be provided along the orbit. The instrument responsivity is quantified in terms of the total de-tected signal, per incident radiance L (in W / m steradian) within Article number, page 22 of 26PICE Consortium: The Solar Orbiter SPICE instrument
Fig. 24.
SPICE e ff ective area, derived from the measured performanceof each subsystem. The horizontal bars show the wavelength coverageof the SW and LW channels for second order (lower left) and first order. a chosen instrument spatial-FOV of solid-angle Ω S , where thisis given by the product of the angular sizes of the chosen slitwidth and along-slit binning. This instrument signal in total DN,summed over all pixels, for exposure time t exp , is given by: N = Lhv A ape Ω S R ( λ ) t exp . Here A ape is entrance aperture area. R ( λ ) is the instrument re-sponsivity, in detected signal DN per photon entering the instru-ment etendue A Ω .Theoretically R ( λ ) is given by: R ( λ ) = R mir ( λ ) η gra ( λ ) QDE det ( λ ) g det , where R mir is the mirror reflectivity, η gra is the grating absolutee ffi ciency, QDE det is the detector photocathode quantum detec-tion e ffi ciency (electrons created per incident photon) and g det is the detection system gain (DNs per photo-cathode electron).These e ffi ciency parameters of each component or subsystemare separately measured during its development, for performanceverification.The on-ground radiometric calibration at instrument-levelthen has the aim of measuring the overall responsivity R ( λ ). Inground calibration of the FM, this is made using the hollow-cathode source that generates several spectral lines of argon, forwhich the total power P in the test beam is approximately known(from characterisation of the source on previous projects). In theabove equation, for this test the known power P replaces theproduct LA ape Ω . In obtaining the total signal N in this test, itis necessary to integrate not only over the whole SPICE imageplane, but also over multiple images, while scanning the SPICEpointing across the test beam. This is because the angular size ofthe test beam is approximately 2 (cid:48) , which is larger than what canbe captured with the widest SPICE slit (0.5 (cid:48) ).This knowledge of source power is so far only for the argonlines in the range λ = ff ective area (as de-rived from subsystem level measurements) is shown in Fig. 24.The first order responsivity is mainly a function of QDE det , whilethe second order responsivity is strongly a ff ected by the gratinge ffi ciency. There is no responsivity in the SW channel for secondorder. The Spectral Hybrid Compression (SHC) scheme adopted bySPICE is a lossy scheme that takes full advantage of the sim-ilarities of spectral image features in all three independent di-mensions of the data set (DeForest 2015). SHC works by Fouriertransformation of the spectral axis in a narrow spectral windowaround the line of interest, which allows the line profile to becoded as a set of Fourier coe ffi cients. This coding scheme con-centrates the information content of the instrument’s spectral im-ages, into the first few Fourier coe ffi cients of the spectral pro-file. In particular, the key features of the line for plasma diag-nosis – intensity, width, and central wavelength – are capturedin the zero-frequency amplitude, the amplitude ratios of the firstfew coe ffi cients, and the phase relationship of the first few co-e ffi cients, respectively. Spatial redundancy is exploited also, asthe Fourier coe ffi cients themselves are treated as individual im-age planes that are subjected to lossy compression via a wavelettransform (similar to JPEG2000). The data volume allocated tohigher spectral frequency Fourier planes is less than that allo-cated to the low spectral frequency planes, accounting for thelower information content of the higher spectral frequencies. Ametadata tag indicates compression quality of each line profile;this tag is itself losslessly encoded to minimise impact on thedata stream while still providing a reliable quality indicator tohighlight any regions where compression artifacts may be sig-nificant.Because it is a lossy scheme, SHC was extensively validatedin the early design phases of SPICE development. An importantcomponent of that process was testing via compression of severalhigh quality (long exposure) spectral scans from the Hinode / EISspectrograph, degraded by the addition of Poisson noise to cor-respond to the noise characteristics of typical SPICE observa-tions. The spectral data sets included both simple line structureand multi-component line structure elements. In each case, wecompared post-facto line fits of the “original” degraded EIS datato similar fits taken from the compressed-and-expanded data.A benchmark of compression to 16 bits per profile (28 × ) wasused for validation, compared to the less aggressive 22-28 bitsper profile anticipated for flight campaigns. Typical performancewas: (1) line intensity was reproduced to within a small fraction(typically under 25%) of the photon counting noise in typicalobservations; (2) line centre wavelength (Doppler shift) was re-produced to within 0.1 pixel RMS across all cases where theline fit converged; (3) line full-width was reproduced to within0.2 pixel RMS across all cases where the line fit converged; (4)sidelobes with intensity ratios as little as 5% compared to theline core were reproduced. Throughout testing, errors inducedby lossy compression proved to be (a) uncorrelated to the signaland (b) well within the error budget driven by the SPICE sciencerequirements.
10. Operations concept
Like for the other instruments of Solar Orbiter, the SPICE sci-entific observations are planned as elements of a succession ofSOOPs (Solar Orbiter Observing Programs) forming the longterm plan agreed upon on a per-orbit basis by the Science Work-ing Team (SWT). Each SOOP is designed to address one or moreof the scientific objectives described in the Solar Orbiter ScienceActivity Plan (Zouganelis et al. 2019). The actions to be per-formed by the instrument are planned as a timeline, meaning asequence of telecommands, encapsulated in several BOPs (BasicObserving Programs). A BOP can conveniently group together
Article number, page 23 of 26 & A proofs: manuscript no. SO_Book_SPICE_paper commonly used sequences of telecommands, like the selectionof a slit followed by the execution of a study, or the preparationof the instrument from standby to operating mode. Most of thescientific observations will consist of ‘studies’, as introduced inSect. 7.7. The commanding software (or planning tools) imple-ments these three concepts (timeline, BOPs and studies) in threemodules, The study generator, the BOP generator, and the time-line tool.The planning tools are written using the Django Python webframework for the back-end, while the front-end was built withBootstrap and JQuery. We use Selenium (functional testing) toensure everything works as expected and pytest (unit testing) tomake sure the code doesn’t break. We also have a Jenkins in-stance that fires up at every git commit to launch tests and gen-erates a SonarQube report that gives us insight on what could beimproved or refactored. The whole project can be deployed froma Debian Docker image. The planning tools are accessed via useraccounts with hierarchical privileges (observer, planner, admin-istrator) and allows several users to work simultaneously whileavoiding conflicts. Observers can design studies, BOPs and time-lines. Planners have the additional rights to approve them fortransfer to Solar Orbiter’s Mission Operations Centre (MOC)or Science Operations Centre (SOC). Administrators have fur-ther rights to manage the database. In the following, we give anoverview of the main functionalities of the planning tools.
The study generator allows a user to design studies compliantwith all the instrumental constraints. The top-level element ofits graphical user interface (GUI) is the study chooser (Fig. 25).The table lists all the studies present in the database (not onlythose on board) and their main parameters and allows to searchamong existing studies. The two images at the bottom are typicalspectra corresponding to the short and long wavelength detectorsrespectively. From there the planner can either edit an existingstudy or create a new one in the study editor. Studies can beof four types: full spectrum, spatial scan (raster), time series (sitand stare), or scanned time series. The planner can then set all thestudy parameters, like exposure times, start, stop and size of thespatial steps, or the spectral windows necessary for the wantedobservations. If the desired spectral window does not exist, theplanner can create a new one using the window editor.With the study editor and window editor, the planner is ableto set all the parameters of the on-board LUTs (see Sect. 7.7)necessary to configure SPICE for the desired observations. Thestudy editor and the window editor enforce all known instrumen-tal constraints. For example, exposure times can only take thevalues predefined in the on-board LUTs; spectral windows cannot overlap; for scanned studies, scans must contain a multiple of32 positions if there are any narrow windows ( <
32 pixels width)or if SHC compression is used.The on-board LUTs allow the storing of 64 studies of which16 are reserved for engineering purposes. These 64 studies alluse the same LUTs to define, for example, their spectral win-dows. Since the LUT that defines the spectral windows has 256entries, the combined 64 studies must not use more than 256windows. It is worth noting that several study parameters aredefined as attributes of the spectral windows: type (intensity orprofile), width, compression, binning. Therefore, observing thesame spectral interval with two di ff erent binnings requires twodi ff erent windows and uses two entries of the correspondingLUT. While each study is defined independently, the planningtools must therefore ensure that the 64 studies selected for up- load to the instrument fit in the existing number of LUT entries.Conflicts (e.g. too many spectral windows) are resolved either apriori by preventing the planner from making modifications or a posteriori by asking the planner to make choices. In addition,each study is tested on the software instrument simulator beforeit can be used in a set of 64 selected for upload. As previously described, BOPs are used as a convenient encap-sulation for groups of sequences of telecommands that are oc-curring frequently, and the BOP generator provides a simple in-terface for assembling and manipulating sequences of telecom-mands with their formal parameters and relative time delays.The timeline tool provides an interface for assembling BOPsinto timelines. The tool enforces the mission constraints (rolls,EMC quiet periods, etc.) as described in the E-FECS (EnhancedFlight Events and Communications Skeleton) files provided bythe SOC. After validation, the tool automatically converts thetimeline into XML files describing the corresponding Instru-ment Operations Request (IOR) to be sent to the SOC. In Ta-ble 6, we provide two examples of observing sequences designedfor the study of Quiet Sun dynamics and elemental composi-tion. The Quiet Sun dynamics sequence fits within the maxi-mum data rate allocated to SPICE and could thus be run contin-uously. This demonstrates the ability of SPICE to provide high-cadence, high-resolution data despite the orbit-related telemetryconstraints of Solar Orbiter.
11. Data processing and tools
A SPICE FITS file will contain a primary Header / Data Unit(HDU), and it may contain one or more additional HDUs (im-age extensions). All primary and image extension HDUs (ob-servational HDUs) will contain observational data with com-plete, self-contained headers, meaning there is no distinctionbetween primary HDUs and image extensions other than thoserequired by the FITS standard. In addition, the file may con-tain a binary table extension that will hold the values of FITSkeywords that vary with exposure number. Every observationalHDU in LL01, L1 and L2 files (see sections below for details)will store data from a single readout window, dumb-bell regionor intensity window, or from a single detector in the case of a fullframe readout. The data cubes of all SPICE FITS files will be 4-dimensional, with dimensions (X,Y,dispersion,time) . Oneor more of the dimensions may be singular, for example rasterscans have a singular time dimension and sit-and-stare observa-tions have a singular X dimension. Each observational HDU ofan L3 FITS file will contain a single derived data product (e.g.intensity, velocity). The SPICE Low Latency Pipeline will run in a virtual machine atSOC, with low latency telemetry data as input, and FITS files oflevel LL01 as output. LL01 files will contain uncalibrated dataexpressed in engineering units. The time will be given in on-board time and the pointing will be relative to the Solar Orbiterboresight. The SOC will provide a simple web-based visualisa-tion tool for the low latency data for observation planning pur-poses. The LL01 files will be stored in the Solar Orbiter Archive.
Article number, page 24 of 26PICE Consortium: The Solar Orbiter SPICE instrument
Fig. 25.
The graphical user interface of the SPICE study generator.
Table 6.
Examples of SPICE observations. The assumed compression rates are 20:1 for spectral windows (SHC) and 10:1 for intensity windows(marked with a ∗ ). In these observations, the raster step in the x direction is equal to the nominal slit width. Objective Spectral lines Slit FOV Exp. time Duration Data rate Data volumeQS dynamics C iii viii i ii ∗ , O iv ∗ ,Ne vi ∗ , Mg ix ∗ , Mg viii ∗ , Mg viii ∗ , O vi × −
12 MbComposition Ar viii v ix vi viii viii vi ∗
6" 954” × − The SPICE Science Data Pipeline will run in Oslo, with teleme-try data as input, and files of multiple levels and file formats asoutput. All output files from the Science Data Pipeline will bestored in the Solar Orbiter Archive.
Level 1 FITS files (L1) will be uncalibrated data expressed inengineering units. Production of L1 files will include time con-version from on-board time to UTC and transformation of co-ordinates from being given relative to the spacecraft boresightto being relative to Sun centre. L1 files will also include addi-tional metadata taken from the study definition database and thetimeline. This information will typically be strings that either de-scribe all files observed with a particular study, or that describea particular instance of a study.
Level 2 FITS files (L2) will be calibrated files that are ready forscientific analysis: the data will be calibrated to physical units, and will be corrected for flat-field, dark current etc. Geometriccorrections that account for slit tilts, spectral slant, detector mis-alignments, non-uniform dispersion, and other geometric distor-tions will be applied to the data cubes by interpolation onto aregular grid. The rotation of the field of view due to spacecraftroll will not corrected for, instead it will be described by the
PCi_j transformation matrix of the World Coordinate System(WCS) coordinates.
Level 3 files (L3) will come in di ff erent types and file formats:regular L3 FITS files, concatenated L3 FITS files, and quick-look L3 files. Regular L3 FITS files will contain data productsthat are derived from a single L2 file: line intensity, velocity andline width resulting from line fitting, and other data products de-rived from the fit parameters, like densities, temperatures, andFIP bias. Each HDU of an L3 file will contain a single deriveddata product for a single automatically detected and identifiedemission line. It should be noted that the pipeline may detectmore than one emission line in each readout window. Concate-nated L3 FITS files will contain time series of derived data prod-ucts, meaning a single concatenated L3 FITS file contains the Article number, page 25 of 26 & A proofs: manuscript no. SO_Book_SPICE_paper derived data products stemming from multiple regular L3 FITSfiles (and hence stemming from multiple L2 files).L3 files can also be JPEG images or MPEG movies meantfor quicklook purposes. Quicklook L3 files will not contain themetadata that are present in L1, L2, and L3 FITS files. A quick-look L3 JPEG file will be either an image of a derived data prod-uct, a wide slit image, or a dumb-bell image. A quicklook L3MPEG file will be either a movie of a derived data product, or amovie of wide slit / dumb-bell multi-exposure observations. The processing software, written in IDL, will be provided viaSolarSoft. The software will enable users to manually processthe L1 data to L2 data by applying calibrations, and to convertL2 to L3 by performing line fitting and creating data productsderived from the fit parameters. During manual processing fromL1 to L3, the user will be able to choose which of the steps de-scribed in Sects. 11.3.2 and 11.3.3 should be applied or not, andto tweak the parameters involved in each step. If the user choosesto omit the step performing the geometric correction on the datacube, the geometric distortions will instead be fully described byFITS WCS keywords.
Visualisation and analysis tools written in IDL will be provided.The visualisation tools (GUI) will allow for inspection of FITSfile data products at all levels. The analysis tools will also pro-vide an interface to the manual processing from L1 to L2 andL3 data, as described in Sect. 11.4. It is also foreseen to gen-erate browse data for multi-instrument visualisation with JHe-lioviewer (Müller et al. 2017).
12. Summary
The SPICE instrument is a high-resolution imaging spectrome-ter onboard the Solar Orbiter mission, operating at EUV wave-lengths from 70.4 − − Acknowledgements.
The development of the SPICE instrument has been fundedby ESA member states and ESA (contract no. SOL.S.ASTR.CON.00070).The German contribution to SPICE is funded by the Bundesministerium fürWirtschaft und Technologie through the Deutsches Zentrum für Luft- und Raum-fahrt e.V. (DLR), grants no. 50 OT 1001 / / ffi ce (SSO). TheUK hardware contribution was funded by the UK Space Agency. The work of S.K. Solanki has been partially supported by the BK21 plus program throughthe National Research Foundation (NRF) funded by the Ministry of Educationof Korea. Finally, the authors would like to thank the referee, Dr Shimizu, whosecomments help to improve the quality of this paper. References
Auchere, F., Andretta, V., Antonucci, E., & et al. 2019, A&A, this volumeBanerjee, D., Gupta, G. R., & Teriaca, L. 2011, SSRv, 158, 267Culhane, J. L., Harra, L. K., James, A. M., et al. 2007, Sol. Phys., 243, 19Curdt, W., Brekke, P., Feldman, U., et al. 2001, A&A, 375, 591Curdt, W., Landi, E., & Feldman, U. 2004, A&A, 427, 1045De Pontieu, B., Title, A. M., Lemen, J. R., et al. 2014, Sol. Phys., 289, 2733DeForest, C. E. 2015, Systems and Methods for hybrid compression of spectralimage data, U.S. Patent No. 9,031,336Del Zanna, G. & Mason, H. E. 2018, Living Reviews in Solar Physics, 15, 5Fludra, A. 2001, A & A, 368, 639Fludra, A., Gri ffi n, D., Caldwell, M., et al. 2013, in Proc. SPIE, Solar Physicsand Space Weather Instrumentation V, Vol. 8862, 88620F–1 – 88620F–15Fludra, A. & Landi, E. 2018, in Proceedings of the International AstronomicalUnion, Vol. 13, Symposium S335 (Space Weather of the Heliosphere: Pro-cesses and Forecasts), 87FFludra, A. & Schmelz, J. T. 1999, A & A, 348, 286Fludra, A. & Warren, H. 2010, ApJ, 523, 47Garcia-Marirrodriga, C. & et al. 2019, A&A, this volumeHollandt, J., Huber, M., & Kuhne, M. 1993, Metrologia, 1993, 381Hollandt et al. 1996, NIST Atomic Spectra Database, 35, 5125Horbury, T. & et al. 2019, A&A, this volumeKrucker, S. & et al. 2019, A&A, this volumeLang, J., Kent, B. J., Breeveld, A. A., et al. 2000, J. Opt, 88Lang, J., Kent, B. J., Paustian, W., et al. 2006, Applied Optics, 45, 8689Maksimovic, M., Bale, S. D., Chust, T., & et al. 2019, A&A, this voumeMüller, D., Marsden, R. G., St. Cyr, O. C., & Gilbert, H. R. 2013, Sol. Phys.,285, 25Müller, D., Nicula, B., Felix, S., et al. 2017, A&A, 606, A10Müller, D., Zouganelis, I., St. Cyr, O. C., & Gilbert, H. R. 2019, A&A, thisvolumeOwen, C. & et al. 2019, A&A, this volumeRochus, P. & et al. 2019, A&A, This VolumeRodríguez-Pacheco, J., Wimmer-Schweingruber, R. F., Mason, G. M., & et al.2019, A&A, this volumeSchühle, U., Uhlig, H., Curdt, W., et al. 2007, in The Second Solar Orbiter Work-shop, Vol. 641Schwenn, R. & Marsch, E. 1990, Physics and Chemistry in Space, Vol. 20,Physics of the Inner Heliosphere I. Large-Scale Phenomena, ed. R. Schwenn& E. Marsch (Berlin: Springer-Verlag)Solanki, S. K., del Toro Iniesta, J. C., Woch, J., & et al. 2019, A&A, this volumeThomas, R. J. 2003, in Proc. SPIE, Vol. 4853, Innovative Telescopes and Instru-mentation for Solar Astrophysics, ed. S. L. Keil & S. V. Avakyan, 411–418von Steiger, R., Schwadron, N. A., Fisk, L. A., et al. 2000, J. Geophys. Res., 105,27217Wilhelm, K., Curdt, W., Marsch, E., et al. 1995, Sol. Phys., 162, 189Zouganelis, I., De Groof, A., Walsh, A., & et al. 2019, A&A, this volumen, D., Caldwell, M., et al. 2013, in Proc. SPIE, Solar Physicsand Space Weather Instrumentation V, Vol. 8862, 88620F–1 – 88620F–15Fludra, A. & Landi, E. 2018, in Proceedings of the International AstronomicalUnion, Vol. 13, Symposium S335 (Space Weather of the Heliosphere: Pro-cesses and Forecasts), 87FFludra, A. & Schmelz, J. T. 1999, A & A, 348, 286Fludra, A. & Warren, H. 2010, ApJ, 523, 47Garcia-Marirrodriga, C. & et al. 2019, A&A, this volumeHollandt, J., Huber, M., & Kuhne, M. 1993, Metrologia, 1993, 381Hollandt et al. 1996, NIST Atomic Spectra Database, 35, 5125Horbury, T. & et al. 2019, A&A, this volumeKrucker, S. & et al. 2019, A&A, this volumeLang, J., Kent, B. J., Breeveld, A. A., et al. 2000, J. Opt, 88Lang, J., Kent, B. J., Paustian, W., et al. 2006, Applied Optics, 45, 8689Maksimovic, M., Bale, S. D., Chust, T., & et al. 2019, A&A, this voumeMüller, D., Marsden, R. G., St. Cyr, O. C., & Gilbert, H. R. 2013, Sol. Phys.,285, 25Müller, D., Nicula, B., Felix, S., et al. 2017, A&A, 606, A10Müller, D., Zouganelis, I., St. Cyr, O. C., & Gilbert, H. R. 2019, A&A, thisvolumeOwen, C. & et al. 2019, A&A, this volumeRochus, P. & et al. 2019, A&A, This VolumeRodríguez-Pacheco, J., Wimmer-Schweingruber, R. F., Mason, G. M., & et al.2019, A&A, this volumeSchühle, U., Uhlig, H., Curdt, W., et al. 2007, in The Second Solar Orbiter Work-shop, Vol. 641Schwenn, R. & Marsch, E. 1990, Physics and Chemistry in Space, Vol. 20,Physics of the Inner Heliosphere I. Large-Scale Phenomena, ed. R. Schwenn& E. Marsch (Berlin: Springer-Verlag)Solanki, S. K., del Toro Iniesta, J. C., Woch, J., & et al. 2019, A&A, this volumeThomas, R. J. 2003, in Proc. SPIE, Vol. 4853, Innovative Telescopes and Instru-mentation for Solar Astrophysics, ed. S. L. Keil & S. V. Avakyan, 411–418von Steiger, R., Schwadron, N. A., Fisk, L. A., et al. 2000, J. Geophys. Res., 105,27217Wilhelm, K., Curdt, W., Marsch, E., et al. 1995, Sol. Phys., 162, 189Zouganelis, I., De Groof, A., Walsh, A., & et al. 2019, A&A, this volume