Instrument Calibration of the Interface Region Imaging Spectrograph (IRIS) Mission

Abstract

The Interface Region Imaging Spectrograph (IRIS) is a NASA small explorer mission that provides high-resolution spectra and images of the Sun in the 133 - 141 nm and 278 - 283 nm wavelength bands. The IRIS data are archived in calibrated form and made available to the public within seven days of observing. The calibrations applied to the data include dark correction, scattered light and background correction, flat fielding, geometric distortion correction, and wavelength calibration. In addition, the IRIS team has calibrated the IRIS absolute throughput as a function of wavelength and has been tracking throughput changes over the course of the mission. As a resource for the IRIS data user, this article describes the details of these calibrations as they have evolved over the first few years of the mission. References to online documentation provide access to additional information and future updates.

Keywords: Instrumentation: calibration.

Figure 1

Average data number (DN or ADU) of the FUV and NUV long-term trends – the residuals between mean 0 s $1\times 1$ summed dark levels in each read port and the dark model (as defined at launch) – are plotted versus time as symbols. The current long-term trend models are overplotted as lines; they fit the data well and reduce the errors.

Figure 2

Plots showing the trend in the number of FUV hot pixels over time at the $>90$% threshold (NUV/SJI trends are similar, with 100 – 200 fewer $n_{\mathrm{hot}}$). The vertical dashed lines indicate the times of CCD bakeouts. The ports are shown separately, but follow a similar trend.

Figure 3

Typical long exposure from the FUV-S CCD on the left and the FUV-L CCD on the right showing the scattered light background. The dark area on the short-wavelength side of the FUV-L detector is not illuminated by the spectrograph optics.

Figure 4

Color-coded pointing positions on a coarsely interpolated map of the background intensity.

Figure 5

Background intensity as a function of solar radius with the same color-coding as used in Figure 4.

Figure 6

Median level of the FUV background as a function of radius from position $X_0$, $Y_0$ for each filter wheel position used for science observations. The fit in red is a broad-band limb-darkening function smoothed with a Gaussian.

Figure 7

Example of a Reuleaux dither pattern, as used for the SJI NUV flat-field observations, with 30 pointings and a pattern size of 70 arcsec.

Figure 8

Example of the resulting flat field and object frame from Chae’s method applied to the 2832 Å channel of the SJI. The two bottom frames show an original image from the observed sequence and its correction by the flat field.

Figure 9

Example of the resulting flat field and object frame from Chae’s method applied to the 1330 Å channel of the SJI. The two bottom frames show an original image from the observed sequence and its correction by the flat field.

Figure 10

Level 2 data from the 2796 Å channel of the SJI. The slit prism is shifted in the science observation with respect to the flat field so that dust specks on the surface of the slit prism show bipolar signatures that are indicated by the arrows. The removal of the slit in the vicinity of a dark dust speck on the detector produces a streak across the slit (circled).

Figure 11

Depth of the 2D Gaussian fitted in the flat field of each month. This is the depth relative to the “continuum” fit for the Gaussian, which translates into a measure of the steepening of the profile. The horizontal bars at the bottom of the plot indicate the eclipse seasons. The vertical dashed lines indicate detector bakeouts.

Figure 12

Lower half of the 1330 Å flat from January 2016. This is a raw flat, i.e. the slit was left untouched. In some places, the slit has become brighter than the surrounding areas. Dark dust specks on the slit remain dark. Dosage burn-in is visible as a large-scale pattern.

Figure 13

NUV spectrograph flat fields. See text for details. The two prominent horizontal lines in the top panels are caused by the fiducial marks in the slit. The spatial flat (bottom left panel) has the fiducial marks removed; the remaining dark horizontal line near the bottom is caused by a small dust particle on the slit.

Figure 14

Blue LED images taken shortly after launch (top) and in October 2014 (bottom). The second image shows the burn-in pattern from the two C ii lines on the left. The LEDs have not been designed to fully cover the CCDs, hence the large non-illuminated area in the middle.

Figure 15

Burn-in profile from the C ii lines in selected blue LED images. The amplitude of the burn-in at the actual FUV wavelengths is about six times larger than it is at the blue wavelengths shown here.

Figure 16

Portion of the deuterium lamp spectrum observed with the FUV-S spectrograph channel during ground testing. Fitted line positions and fiducial positions are shown in red, while the desired rectilinear coordinates are shown in blue. The fiducial/spectral line crossings are indicated by the diamonds, while the line centroids are plotted as small points. The 2D fitting of the red and blue points yields the geometric transformation. The image is upside down with respect to typical IRIS Level 1 data and is displayed in inverted grayscale to show emission lines as dark features.

Figure 17

Example of wavelength measurements and best-fit calibration function.

Figure 18

Offset measurements between the FUV-S and FUV-L detectors. The solid line is a 90-day running average. There is no systematic drift over the course of the mission.

Figure 19

Difference between the primary and alternate wavelength calibration methods for the FUV SG over the first several years of the mission.

Figure 20

Estimates of the pre-launch effective area. The spectral windows of the spectrograph detector are indicated by vertical dotted lines. The two curves in each of the bottom panels indicate the response of each of the two FUV and NUV slit-jaw imager channels.

Figure 21

Full-Sun FUV spectra for 20 October 2014. Top panel: IRIS, middle panel: SOLSTICE, and bottom panel: derived IRIS effective area at four wavelengths (star symbols). Vertical dotted lines in the top and middle panels indicate the spectral regions used for the cross-calibration.

Figure 22

Model used for the calibration of the FUV SG (shown here with values for 1 March 2015). See text for details.

Figure 23

Full-Sun NUV spectra for 20 October 2014. Top panel: IRIS, middle panel: SOLSTICE, and bottom panel: derived IRIS effective area. Vertical dotted lines in the top and middle panels indicate the spectral regions used for the cross-calibration.

Figure 24

Trending of the IRIS SG response over the course of the mission. Top panel: FUV near 1335 Å, middle panel: FUV near 1400 Å, and bottom panel: NUV. Colored diamonds and crosses indicate absolute cross-calibrations with SOLSTICE, black dots are scaled running averages of IRIS quiet-Sun observations, and colored solid lines are the final parametric fits. See text for details.

Figure 25

Trending of the response of the four IRIS SJI channels over the course of the mission. Diamonds indicate cross-calibrations with SOLSTICE, black dots are scaled running averages of IRIS quiet-Sun observations, and colored solid lines are the final parametric fits. See text for details.

Figure 26

Photon transfer curve for the SJI CCD measure with a blue LED.

Figure 27

Photon transfer curve for the SJI CCD measured with a deuterium lamp at 1350 Å.

Figure 28

Orbital wobble correction. Top panel: IRIS orbital wobble in units of IRIS 0.16 arcsec pixels, for the $x$- (solid lines) and $y$- (dashed lines) directions, and three roll angle values: 0 (black), $+90$ (red), and −90 (dark blue). Bottom panel: Same curves as in the top panel, but with a phase shift of the wobble curves for $+90$ and −90 of $+0.25$ and −0.25, respectively. The shifted curves are very similar to the wobble curves for $0^{\circ }$ roll angle. The orbital phase is zero at the time when IRIS passes through its ascending node.

Figure 29

Variation of the wobble over the mission at a roll angle of $0^{\circ }$. Measurements for each year are shown in different colors. The wobble magnitude is defined as the absolute value (in Euclidean space) of the drift within an orbit. The asterisks indicate where orbital calibration data were taken with dashed lines to guide the eye on the annual variation. There were only three calibrations in 2014, so we have not connected these data points with lines.

Figure 30

Intensity of the laser spectrum as a function of the half-waveplate orientation for the four measurement series (colored points). The fit to the data is shown by the solid black line.

Figure 31

Evolution of the setting for best focus (in focus motor steps) over the mission. The changes are primarily due to variations of the thermal environment. The annually recurring jumps are caused by the eclipse seasons.

Figure 32

Flow chart of various IRIS data levels and associated pipeline processing.

Figure 33

IRIS spectral data layout for various data levels. Left: Levels 0 and 1 spectral data have up to eight windows appropriately placed within a pixel array matching the CCD detector. Middle: Level 2 data have extracts of the eight windows, assembled into rasters based on slit position $x$. Right: Level 3 data assemble the Level 2 rasters into time-series datacubes.