- The NIRSpec Exoplanet Exposure Time Calculator

The NIRSpec Exoplanet Exposure Time Calculator, NEETC, has been developed by the ESA instrument team to enable close investigation of the instruments capabilities in terms of exoplanet observations. It is optimised to deal with observations of transiting exoplanets, i.e.. photon-noise limited observation of bright point sources and is based on radiometric models of NIRSpec. A full description of the Bright Object Time Series mode is given under BOTS. Most results and simulations presented in these webpages regarding exoplanets have been produced with NEETC.

Number of groups per integration & Saturation

The number of groups per integration, which will give the best duty cycle without saturating the detector, if possible, is computed automatically. We call this the optimal number of groups but what is actually optimal will depend on the specifics of the observations and science case at hand. Although at least two groups per integration (reset- read-read-...) is usually preferred, for very bright sources it is possible to obtain integrations with a single group readout (reset-read). The NEETC supports read-reset SNR calculations.

If the number of groups per integration is higher than the optimal number of groups, some pixels will saturate as some point during the integration. Due to the non-destructive readout of NIRSpec's IR-detectors, the groups before saturation are used to calculate the signal (and SNR in the NEETC). For example, for an exposure consisting of four groups per integration, if part of the spectrum is saturated after three groups, the signal and SNR in these pixels will be calculated based on three groups only. This will lead to a lower efficiency (duty cycle) in this part of the spectrum, as the forth group will not be used due to saturation. For the rest of the spectrum all four groups are used.

A worst-case scenario is used when checking for saturation, assuming that the peak of the PSF of the target always falls at the center of a pixel, as opposed to in the corner where 4 adjacent pixels meet (yielding a maximum number of accumulated electrons typically 20-30% lower). The PSF-core size varies throughout the wavelength range from 0.5 to 1.5 physical detector pixels, which is not taken into account in the NEETC.

Noise model


Noise parameters

Detector parameters for BOTS and the NEETC

Read out noise 16.8 e-/pixel Saturation limit 65 000 e-
kTC noise 39 e-/pixel Number of resets between integrations 1
Dark current 22.86 e-/hr/pixel Gain 2 e-/ADU

The NEETC computes the detected signal using photon-conversion-efficiency and dispersion curves for the 9 instrument configurations supported in BOTS mode. The NEETC is based on radiometric models of NIRSpec and takes only the noise floor consisting of shot noise (from the combined stellar signal and dark current), read out noise (RON) and kTC noise (only relevant when when using the reset level to sample up the ramp) into account. A detailed description of the NEETC and the noise model is given in Nielsen et al. 2016.

The values for the noise properties used in the NEETC are listed in the table to the right and are based on several test campaings. The saturation limit has been set at 80% of the range for which linearity corrections hold up, when using the gain=2 e-/ADU mode. An analysis of the instrumental effects and calibration uncertainties have been carried out during the ISIM-CV3 test campaign. The results are presented here.

Last - first

Up the ramp fitting (Rauscher et al. 2007) is omitted in favour of using only first and last group of an integration which is suitable for photon-limited observations. This yields:

noise floorlast-first = SQRT(2 x RON2 + photon noise2) ,

where the photon noise includes both dark current and the stellar signal. The factor two in front of the RON comes from the fact that two measurements are used.

Last - zero

The bias/zero level of the detector can be used to compute the signal as well. For read-reset there is no other option, but as the duty cycle goes up when the zero level is used, this scheme can also be useful for integrations with few groups. The kTC noise which is affiliated with the zero-level must be taken into account:

noise floorlast-zero = SQRT(RON2 +kTC2 + photon noise2) .