1. Time Series Observations with MIRI
Sarah Kendrew (ESA, Baltimore)
Transiting Exoplanets with JWST workshop, July 2017

2. What is MIRI?
The only JWST instrument to cover λ > 5 µm; only actively cooled part of JWST (T < 7K)
Built by European/US consortium, PI: Gillian Wright/George Rieke
Observing modes:
Imaging 5-28 µm
Low-resolution spectroscopy (slit or slitless) 5-12 µm
Coronagraphy
Medium-resolution integral field spectroscopy 5-28 µm
Wright et al, 2015

3. MIRI Mechanical Layout
Spectrometer pre-optics
Spectrometer main optics
Spectrometer focal plane module
Carbon fibre hexapod
Imager module
Pick-off mirror
Wright et al, 2015
Details not too important here but good tp remember that MIRI has two main parts to it – the spectrometer is its own module, imager, coronagraphy and LRS are in the other module.

4. MIRI capabilities & coverage vs. NIR instruments

5. Available Time Series Modes for MIRI (I)
Low-resolution slitless spectroscopy (single-object)
R~100 from 5-12 µm, dispersion via double prism in imager filter wheel
Read out in a dedicated subarray for increased dynamic range
LRS slitless is always assumed to be TSO
TSO fully supported:
no dithering allowed
TA mandatory
> 10 ksec exposures allowed (through high gain antenna moves)
Processing via TSO-optimised pipeline
Imager focal plane
LRS slitless TA region

6. Available Time Series Modes for MIRI (II)
Time Series Imaging: limited availability
no-dithering option for SUB64 subarray
> 10 ksec exposures allowed for SUB64
TA not supported (blind pointing accuracy ~0.65” rms radial)
Data not processed in TSO-optimised pipeline (… but user can re-process)
all 9 standard imaging filters available
Integral Field Spectroscopy: not currently supported

7. MIRI Detectors (Ressler+ 2015, Rieke+ 2015)
3 Si:As focal plane arrays, 1024 x 1024 pixels (Raytheon)
1 for imager, coronagraphy, low-res spectoscopy
2 for MRS (short- & long-wavelength channels)
25 µm pixel pitch
Most closely related to the 256×256 pixel Si:As arrays in Spitzer, with 10 extra years of development
Redundancy in electronics - side A or B
Uniform response: pixel-to-pixel variations no more than 3% rms
Small proportion of inoperative pixels: either hot or dead <200 (0.1%)

8. MIRI Readout Data Number (DN) Frames (groups) Time Exposure
Integration
Time
Exposure
FAST vs. SLOW mode
FAST: 1 sample per pix (2.8 s for FULL frame)
SLOW: 8 samples per px (23.9 s for FULL frame)
SLITLESSPRISM subarray read time: 0.16 s (factor ~17 faster than full array)
Frame = single clocking scan through array Integration = a number of non-destructive readout frames where photons are allowed to integrate Exposure = is a number of sequential integrations MIRI does not explicitly have “groups” By definition, MIRI has one frame per group 8

9. MIRI Detectors non-ideal behaviour
Reset Switch Charge Decay (RSCD)
Seen as a non-linear hook in the first frames of the ramp
very predictable
Last frame effect
The array is reset sequentially by row pairs. The last frame of an integration ramp on a given pixel is influenced by signal coupled through the reset of the adjacent row pair
Column and row pull down/up
Does not carry a noise penalty – modelled and removed in pipeline
Frames (groups)
Integration
Last frame
RSCD

10. Non-ideal behaviours (continued)
Persistence - memory of previous illumination pattern <<0.5% of previous signal after a few minutes Seen as an weak positive image that decays in strength exponentially with time Time to recover is a function or previous illumination and integration time Imprints – memory effects that do not fade in time Seen in similar detectors (IRAC) and removed by annealing Drifts – signal drifts with amplitudes roughly proportional to signal level Persistence decay is likely one component of the cause of signal drifts Memory effects are common to these type of detectors
Image
5 min exposures
Dark after 90 seconds
Dark 2
Dark 3

11. Characterising the MIRI detectors
Has been a major part of MIRI team’s efforts in recent years
Several test campaigns on flight-similar detectors at JPL, led by Mike Ressler
Testing mid-IR detectors to the level of precision required for TSO observations is v challenging on the ground particularly in the high-flux regime
MIRI team looking to gain much experience during commissioning & cycle 1 calibration, e.g.:
stability of short ramps
mitigating strategies for persistence

12. Detectors summary: MIRI Exposure Rules of Thumb
Long ramps good, short ramps bad
Min. recommended number of groups per integration: 5. NGROUPS = 2-4 is allowed but performance not v well tested and quality of calibration uncertain
If background subtraction is a potential issue then consider taking dedicated off-source background
Current saturation limits are conservative (assume 70% full well).
Settling time seems shorter and behaviour reasonably well understood

13. Time Series Observations with slitless LRS
Spectral resolving power R~100, variable with wavelength: 5 µm, µm
Dispersion relation shows turnover below 4.5 µm
wavelength calibration challenging below 5 µm
Sensitivity ~ 10x worse than for LRS slit
Bright limit ~ 17x better than for LRS slit (smaller subarray) µm px

14. LRS Continuum Sensitivity (Glasse et al, 2015)
SLIT
Reminder:
Maximise ramp length!
SLITLESS sensitivity: Factor ~10 worse

15. LRS Bright limits (SLIT)
SLITLESS MODE: Factor 17 brighter
Bright limit:
~ µm
(conservative!)
ETC has best estimate (to 70% full well)
Glasse et al, PASP 2015
Assumes 2-frame reads, < 60% full well
Full array FAST mode (frame time ~2.8 s)
NB: does not yet include revised detector quantum yield

16. LRS slitless target acquisition
Uses the same subarray as science observations
TA filters for all modes: F560W, F1000W, F1550W, FND (neutral density, ~2e-3 suppression over 8-18 µm)
ETC has TA calculating capability
Verification image will be taken per default to verify TA
For most v bright cases F1550W filter is recommended as should avoid saturation & persistence on filter wheel move after TA.
NOTE: min. NGROUPS for TA = 3

17. Time series observations with MIRI Imager
Limited time series capabilities with SUB64 subarray:
no-dither option allowed (though not yet implemented)
target acquisition not available
Hope to offer TSO with all subarrays in future cycles
SUB64 Frame time = s, or gain in bright source limit compared with FULL frame

18. Imager performance (SUB64)
Assumes NGROUPS=2, 60% full well

19. Conclusions
MIRI is building on the heritage of Spitzer, with 10 extra years of technology development
Slitless low-resolution spectroscopy (R~100, 5-12 µm) is the workhorse TSO observing mode. Hope to add full support for TSO imaging for cycle 2.
Understanding the noise and developing removal strategies to reach the photon noise-limited spectro-photometry is complex, under active investigation for ground testing and high priority for commissioning & calibration
Findings will feed into the TSO pipeline – also work in progress.
Community feedback is very valuable!