NIRSpec Operations Concept D R A F T

1. NIRSpec Operations ConceptD R A F T Michael Regan, Jeff Valenti (STScI) Wolfram Freduling, Harald Kuntschner, Robert Fosbury (ST-ECF)

2. Operations Concept: purpose • End-to-end view of instrument operation • Used to describe: • Instrument operational modes • Target acquisition strategy and spacecraft pointing accuracy and stability needs • Influence of instrument stability and mechanism repeatability • Calibration implementation • Data-flow model and ground/flight s/w tasks • Mechanism and lamp usage estimates NIRSpec Operations Concept

3. Instrument schematic • Optical and mechanical subsystems NIRSpec Operations Concept

4. Observing strategies • NIRSpec lexicon • An Activity is a set of basic instrument operations that results in the completion of a clearly identifiable task such as: a target acquisition, a wavelength calibration, science observation etc. • A spacecraft Visit is the contiguous period during which the pointing is controlled by a single guide star (set) • A visit will generally contain a number of different activities • NIRSpec field: all instantaneous FOV that can be accessed with a single guide star (set) • Target set: a group of sources that can be observed simultaneously • There can be multiple target sets within a NIRSpec field • Target sets for acquisition and science can be different • Association: a set of data that can be processed as a unit by the pipeline • Each move of a NIRSpec spectral element defines a new association • Each scientific target set implies a new association NIRSpec Operations Concept

5. Obs. strat. • MSA configuration: a pattern of open/closed MSA facets designed to match a given target set or calibration requirement • An associationmay include multiple MSA configurations • An Aperture pattern is a group of open facets comprising a single spectroscopic slit/aperture • Sub-aperture dithering may be used within an association to move targets within an aperture for each MSA configuration • Large-angle dithering with MSA reconfiguration may be used to span detector gaps and ameliorate the effects of bad pixels • Between two consecutive detector resets, non-destructive reads yield an Image sequence • At each dither location, multiple image sequences may be recorded • A single image sequence may not span two sub-aperture dither locations • A NIRSpec observing program will contain hierarchical arrangements of these elements NIRSpec Operations Concept

6. Obs. strat. • Types of science observations • Slitless spectroscopy • no TA required • Spectroscopy with fixed-slits • Multi-object spectroscopy with the MSA • Imaging spectroscopy with the integral-field unit (IFU) • NIRSpec imaging (mirror) will be used primarily for target acquisition and calibration • An important requirement is the retention and archiving of images taken through an MSA aperture pattern — used to analyse the positions of objects within the apertures NIRSpec Operations Concept

7. Obs. strat. • MSA apertures • The MSA will be n x 512 facets in the spectral direction by n x 256 facets in the spatial direction. Currently, n is expected to be either 2 or 4 • Aperture width in the spectral direction is set by the spectral resolution requirement and will typically be 1.5/2 times as large in Band III (long wavelength) as in Band I • Aperture length in the spatial direction is set by the requirements for sky subtraction (small sources) or by the angular size of larger sources. It is restricted by the requirement for a high multiplex factor to avoid overlapping spectra NIRSpec Operations Concept

8. Obs. strat. • MSA configuration • can be specified as an n2 x 512 x 256 bit array that can be compressed using either a generic or a tailored compression/encoding scheme • A standard (limited) set of aperture patterns for spectroscopy will have primary calibration data. Other aperture patterns will have secondary calibration data provided on a best effort basis NIRSpec Operations Concept

9. Obs. strat. • Dithering • Dithers are needed to obtain maximum spectrophotmetric accuracy, maximum sensitivity, and to deal with focal plane defects • Why? • Bad pixels • Uncertainty in the dark current • Fill in detector gaps • Slit losses • Flat field uncertainties • Spectral sampling NIRSpec Operations Concept

10. Obs. strat. • Sub-aperture dithering • The (point-)source throughput for a given aperture is determined by wavelength — relative size of PSF — and position within the aperture • Due to quantization of aperture locations, sources will not generally be optimally placed wrt either the aperture boundary or the inter-facet bars • Sub-aperture dithering is a method of improving flux calibration by providing empirical constraints on source extent and position. Images of the field with NIRSpec or NIRCam can further improve accuracy • Dither patterns are chosen to optimise S/N subject to a constraint on the required flux calibration accuracy NIRSpec Operations Concept

11. Obs. strat. • Large-angle dithering • Dithering with MSA reconfiguration for a given target set will be carried out to ameliorate the effects of bad detector pixels and gaps between detector sub-arrays • The required telescope slews will require neither a new guide star acquisition nor a new target acquisition NIRSpec Operations Concept

12. Obs. strat. • Overhead cost of dithering • For long observations there will effectively no cost associated with dithering • Individual exposure times are limited by cosmic rays and dark current • For short observations there will be a tradeoff • MSA reconfigurations reduce systematic errors but result in lower signal-to-noise • Sub-aperture dithers have a lower cost NIRSpec Operations Concept

13. Obs. strat. • NIRSpec Activities • An activity encodes a combination of different low-level instrument operations in order to enable identifiable tasks such as: target acquisition, contemporaneous calibrations, science observations… • Envisaged activities include: • Measure mirror position • Purpose: map positions on FPA to positions in AFP • Method: use calibration lamp to image ‘L-shaped’ MSA apertures for x and y centroid determination • Goal: to avoid having to do this by maintaining stable mirror orientation to avoid global shifts of more than 0.2 detector pixels (2s) along the dispersion direction and more than 0.2 detector pixels (2s) along the spatial direction after a grating wheel move NIRSpec Operations Concept

14. Obs. strat. • Select imaging mode • Purpose: To enable observer choice between unprotected and protected imaging modes depending on the absence or presence of bright objects (capable of increasing detector dark current) in the NIRSpec FOV • Method: Requirements derived from observer’s use of Observation Design Tool (ODT — this is not an instrument safety issue). MSA shutters closed in zones around bright sources while light path blocked by filter wheel • Target acquisition • Purpose: To position the reference target set accurately in the AFP • Method: Requires multiple reference sources to achieve desired accuracy. Alternatively, dithered observations of fewer sources can be used. Requires 3 reads of either a full-frame image or a set of up to 16 sub-images centred on nominated acquisition targets. Needs CR removal, flat fielding, multiple image centroiding, application of a failure criterion (and retention of diagnostic information), application of mirror orientation correction, application of telescope offset • Direct image • Purpose: To confirm the presence of science targets in their correct positions in the AFP after a target acquisition sequence. The accuracy of flux and wavelength calibration depends on how well the target positions are known. • Method: Use either a protected (by the filter wheel) on an unprotected (no filter wheel change) MSA reconfiguration to prepare for a direct image through the science apertures. At least 3 detector reads are needed to allow CR rejection NIRSpec Operations Concept

15. Obs. strat. • Disperser selection • Purpose: To configure NIRSpec for spectroscopic observations of science targets • Method: Command the appropriate dispersing element into the optical path. The correct filter will already be in place after the confirmatory direct image is obtained • Wavelength calibration • Purpose: To determine the zero point of the wavelength scale for each disperser element to an accuracy of better than 0.2 detector pixels (2s). (The wavelength dispersion relation, i.e., the higher terms, is obtained as part of the annual calibration program) • Method: Image a calibration line source through a set of science apertures (and fixed slits) and a selected disperser element • Goal: To avoid having to do this if the dispersing element mechanism repositions to sufficient accuracy • Science observations • Purpose: To obtain a science observation of a target set with a given MSA configuration and/or fixed slit and a given dispersing element. Note that a given target set can (and generally will) be observed with different dispersing elements and with MSA reconfigurations (large angle dithers) • Method: Carried out after all preparatory sequences. Can include sub-aperture dithers (<0.5 arcsec). MSA reconfigurations for large angle dithers will generally be carried out in an unprotected mode since a dispersing element is in place. However, a protected MSA reconfiguration, involving a filter wheel move, will also be available NIRSpec Operations Concept

16. Imaging and spectroscopic modes NIRSpec Operations Concept

17. Detector Operations • NIRSpec will be highly detector noise limited in R > 1000 modes • Up-the-rump/MULTIACCUM sampling has been shown to be better than Fowler for detector noise limited observations • In addition, up-the-ramp sampling is more robust against cosmic rays NIRSpec Operations Concept

18. T2 T2 T2 T2 T2 Baseline Readout Mode Signal Level Reset Samples Groups TIME NIRSpec Operations Concept

19. T2 T2 T2 T2 T2 Alternative Readout Mode(depends on noise characteristics of flight electronics & detector) Signal Level Reset Samples Groups TIME NIRSpec Operations Concept

20. Readout summary • Only one detector mode, MULTIACCUM, is needed for the full operation of NIRSpec NIRSpec Operations Concept

21. Other parameters • Sub-array readout • Minimum 12 second exposure time is too long for many sources • Sub-array readout will be needed • Sub-array must not be limited to a square (e.g., rectangle needed for bright, fixed-slit targets) • Only one sub-array at a time • Readout time = 12 x (number of pixels in subarray/8 million) seconds NIRSpec Operations Concept

22. Electronic Gain • Goal is to have only one gain setting for NIRSpec • Maximum gain is set by Nyquist sampling single sample read noise (~9e-) or ~4e-/ADU • Would like to be able to use entire full well ~ 90K – 200K e- • 16 bit A/D values lead to 64K dynamic range • Saturated values can be reconstructed from early reads in up-the-ramp • A single gain of 1.5e- to 2.5e- will work NIRSpec Operations Concept

23. Integration Times NIRSpec Operations Concept

24. Target acquisition — outline • Need for robust and completely autonomous TA • Assumptions: • independent of NIRCam • uses pre-defined slit mask (i.e. not computed on-board!) • absolute pointing good enough to place objects initially close to slits • TA determines x, y and roll angle offsets • Requirements for input from observers: • Spacecraft roll requirement • Target set descriptions • Analysis of TA image: • is a challenge because of big pixels • impossible to achieve required accuracy using a single reference star • on-board analysis of images must be able to: • identify reference stars • compute offsets • flat fielding will be necessary • Error budget: • Inherent in target set description • Introduced by spacecraft/instrument NIRSpec Operations Concept

25. TA input from observer • Desired spacecraft roll angle (f) • Astrometry and photometry for a reference target set • The reference target set may contain science targets • A MSA configuration designed for this target set • Optional MSA configuration to protect detector from bright sources during TA • Indication of whether the reference targets should be used to determine (d x, d y, d f) or only (d x, d y) • Exposure time for TA images NIRSpec Operations Concept

26. TA goals • The MSA configuration requires targets within a zone of half a slit-width (spectral direction) centred on the aperture pattern • This is called the Nearest Facet Trigger Zone (NFTZ) • Errors in the TA or the target set astrometry can place targets outside their intended NFTZ • A successful TA places more than 75% of targets within the NFTZ in more than 95% of cases • With a typical Band I slit-width of 200mas, this leads to a requirement of a TA error of 25 mas (2s) or 12 mas (1s) NIRSpec Operations Concept

27. Throughput impact of TA errors • Typical TP variation is 10% for 12mas offset • Average TP loss for a target set for a 12mas offset is ~ 5% at 2µm NIRSpec Operations Concept

28. Telescope is commanded to go to NIRSpec field Telescope slews During the slew, NIRSPEC is configured for target acquisition and the position of the MSA relative to the detector is calibrated. This consists of the following steps: Rotate the grating wheel into mirror position. NIRSPEC is now in imaging mode Rotate filter wheel to diffuser/dark position Turn on appropriate continuum lamp Take short exposure Read a window on the detector which is centered on target acquisition aperture Send image to computer Turn off lamp Computer analyses the images and computes position of MSA relative to detector Configure MSA for imaging of reference target set: all MSA shutters are opened (optionally shutters around bright objects are closed) After the telescope arrives at the target field, a MULTIACCUM 3 exposure is taken. This is nominally a reset followed by three reads of the detectors Images are sent to computer For each image, computer Divides by flat (TBC) Takes the minimum of the read 1 – read 0, and read 2 – read 1 differences (CR reject) Determines CoG of targets Computes position of objects from CoGs Computes dx pixel, dy pixel, df angle (TBC) If specified by observer, df = 0 Using the previously computed position of the MSA relative to the detector, the xy offset is computed Commands offset dx pixel, dy pixel, df (TBC) TA procedure (MSA) NIRSpec Operations Concept

29. Telescope slews to final position MSA is configured to slit mask for science target set Science observations start The first image in the science sequence will in most cases be an undispersed image to be used to aid the extraction of spectra The TA procedure for the fixed slits will be identical — with the open MSA being used to determine offsets from a reference target set TA procedure - continued NIRSpec Operations Concept

30. TA error budget • The final positioning of a target set within the set of aperture patterns (slit mask) will depend on the accuracy of the supplied target coordinates relative to the reference target set (random?) and the accuracy of the TA procedure (bulk offset) • The TA accuracy depends critically on the number of reference targets and the requirements cannot be met with a single target NIRSpec Operations Concept

31. Microshutter grid/detetector pixellation will lead to biases in the centroid of an individual point source ~14mas More sophisticated algorithms can reduce this Only by dithering one source or using multiple reference objects can this be averaged out With 9 reference targets, a final error of 5 mas can be achieved Centering through the MSA NIRSpec Operations Concept

32. TA error budget • This represents a bulk shift/rotation between the MSA aperture patterns (slits) relative to the mean target set position NIRSpec Operations Concept

33. Errors in placement of individual science targets on associated slit Arise from: Errors in knowledge of target positions Difference between actual roll angle and that assumed for the mask design Total roll angle error budget assumed to be 10 arcsec NIRSpec Operations Concept

34. Source of target coordinates • Generally from NIRCam images • Roll angle is most critical factor • This means that a JWST roll angle error enters twice into a NIRSpec observation — once from Cam and once in Spec => absolute roll requirement for JWST is 10/√2 = 7 arcsec • This absolute roll requirement is alleviated if the TA procedure can compute and apply a df • Non-JWST target set sources are possible but difficult NIRSpec Operations Concept

35. Image Stability • Around 1/3 of the science will be one day per grating selection • Need to be stable on this timescale • Otherwise, will have to reacquire and recalibrate • Spacecraft roll about FGS star will need to be stable to within ~3 arcsec per day • Smaller due to larger radius to FGS star • It is vital that a series of small offsets (for sub-aperture and large-angle dithering) do not produce a cumulative error (JWST Lev 2) NIRSpec Operations Concept

36. Calibration goals • To allow the determination, for each observed target, the intensity of radiation as a function of wavelength and position along the spatial direction of the slit • Provide reference files for all standard approved science modes • Provide reference files to enable all science operations such as TA • Monitor instrument status and performance • Minimise on-orbit calibration time investment • Maximise utility of general science calibrations • Minimise science programme-specific calibrations NIRSpec Operations Concept

37. Calibration — science requirements • Derived from reference science programmes (Kuntschner et al. 2003) • Wavelength: The combination of systematic and relative errors in the wavelength calibration will be smaller than 1/10 (rms) of the characteristic resolution element (FWHM) for a given grating/prism, over the full wavelength range and FOV • Spectrophotometric: Assuming no Poisson noise in the signal, multiple observations of the same target with different MSA, fixed slit or IFU configurations will provide a repeatability for the overall throughput (with respect to a given standard source) of better than 5% (rms) for the full FOV. The throughput uncertainty as a function of wavelength will be below 5% (rms) • Spatial: The spatial coordinate along the slit will be known to better than 20mas (rms) with respect to the coordinate frame defined by the target acquisition reference objects. This accuracy will be met at all wavelengths, over the full FOV. • Spatial PSF: The spatial PSF shape (relative intensity of a point source along the spatial direction of the slit) will be known to better than 3% (rms) at all wavelength and the full FOV • Spectral PSF: The line spread function (relative intensity distribution of a delta function along the dispersion direction) will be known to better than 3% (rms) along the dispersion direction at all wavelengths and the full FOV NIRSpec Operations Concept

38. Calibration types • NIRSpec will be able to produce all required performance and calibration data using a combination of the following on-board calibration types: • Pointed calibrations: Dedicated observations of astronomical objects. For some observations an accurate TA may be needed • Sky calibrations: Observations of typical sky (i.e. background) regions. Normal telescope pointing will be sufficient • Lamp calibrations: Internal continuum and line lamp observations • Dark calibrations: Observations with the light path blocked. Auto and Opportunistic calibrations: The calibrations make use of the science data itself (auto) or are extracted from another set of suitable science observations (opportunistic). These calibrations do not require specific observations • Pointed and sky calibrations require NIRSpec to be the primary instrument, while dark and lamp calibrations are suitable to be carried out in parallel mode NIRSpec Operations Concept

39. On-board lamps • NIRSpec will be equipped with internal line and continuum lamps • The brightness of the signal produced by the calibration system must be stable (limits TBC) and predictable • Pre-planned power settings and exposure times must result in data quality that satisfies the needs of the exposure • It will be possible to use the calibration system with any allowed spectroscopic mode and produce a signal which satisfies the calibration requirements • Detector over-illumination (producing persistence) will be avoided • The calibration subsystem should be capable of use in parallel with other JWST science instrument operations NIRSpec Operations Concept

40. The exposure time required to produce the signal needed for calibration purposes will ≤ 60s for the wavelength calibration system and ≤ 10 x 60s for the continuum calibration • The line lamps will produce > 10 lines uniformly covering the full wavelength range of each dispersing element with a S/N > 30 in each of the lines • The continuum lamps will provide a S/N > 500 over the full wavelength range and FOV of each dispersing element • The stability of the lamps will be such that after the nominal lifetime of JWST the required signal is still achieved within the nominal exposure times listed above NIRSpec Operations Concept

41. Science program specific calibrations • Will consist of the following types: • “Through slit” images of the target(s) after a successful target acquisition • Wavelength zero-point calibration • It is assumed that the spacecraft and instrument stability are such that these do not need to be repeated during observations of a particular target set with a given disperser within a particular visit NIRSpec Operations Concept

42. Monitoring calibrations • Monitoring calibrations scheduled at regular time intervals for use by all regular science programs • NIRSpec offers a limited number of observing modes • But monitoring calibrations needed for large set of MSA • Implies need for highly stable and well behaved spectrograph: • geometric distortions, sensitivity and wavelength solution should change only smoothly across the FOV. • Small scale flat-field (FF) calibration strategy relies on an extensive pre-flight calibration of the detectors. A determination of the full pixel-to-pixel FF as a function of wavelength is impossible in orbit (e.g., lack of narrow band filters, Kuntschner et al. 2003). • For efficiency, must separate monitoring calibrations into those that can be done in parallel and those where NIRSpec is required to be the primary instrument • The loss in science time without parallel capabilities is estimated to be at least 5% of total NIRSpec time NIRSpec Operations Concept

43. Parallel-capable calibrations NIRSpec Operations Concept

44. NIRSpec primary calibrations NIRSpec Operations Concept

45. Data-flow model • Need for a coherent environment to handle complex observation design process and the resulting properly-described data structures • Outlines needs for ground and flight s/w NIRSpec Operations Concept

46. Usage estimates • Thermal worst-case day • Mission lifetime NIRSpec Operations Concept

47. Thermal worst-case day • Assumptions • Spectroscopic survey of bright sources • Lowest ratio of integration time to mechanism moves • Assume that only the load on 24 hr period is important • 30 minutes of integration time per source • Subsequent fields are nearby so slew time is small • Three MSA configurations per visit NIRSpec Operations Concept

48. Visit setup after first visit NIRSpec Operations Concept

49. Science tasks NIRSpec Operations Concept

50. Worst-case daily load • How long for one visit? • 45.6 min or 0.032 day • 33 visits per day • What does this mean? • Filter wheel moves – 63 times/day (every 1400 seconds) • Grating wheel moves – 63 times/day (every 1400 seconds) • MSA configures – 126 times/day (every 720 seconds) • Line Lamp – 63 times/day (every 1400 seconds) NIRSpec Operations Concept