Principal Components Analysis of a JWST NIRSpec Detector Subsystem

We present principal component analysis (PCA) of a flight-representative James Webb Space Telescope Near Infrared Spectrograph (NIRSpec) Detector Subsystem. Although our results are specific to NIRSpec and its T ~ 40 K SIDECAR ASICs and 5 μm cutoff H2RG detector arrays, the underlying technical approach is more general. We describe how we measured the system’s response to small environmental perturbations by modulating a set of bias voltages and temperature. We used this information to compute the system’s principal noise components. Together with information from the astronomical scene, we show how the zeroth principal component can be used to calibrate out the effects of small thermal and electrical instabilities to produce cosmetically cleaner images with significantly less correlated noise. Alternatively, if one were designing a new instrument, one could use PCA to determine a set of environmental requirements (temperature stability, electrical stability, etc.) that enabled the planned instrument to meet performance requirements.