Abstract
Interference filters assembled in ‘butcher-block’ arrays are used close to focal plane arrays for spectral definition in sensors such as VIIRS. Ideally only light incident on a particular filter above a detector element which is within the specified band will reach the detector. Optical cross-talk is in-band or out-of-band light incident on an adjacent filter or adjacent region of the same filter reaching the detector. Out-of-band optical cross-talk results in spectral and spatial ‘impurities’ in the signal and consequent errors in the calculated environmental parameters such as ocean color that rely on calculations based on combinations of signals from more than one band. Out-of-band signal and out-of-band optical cross-talk was observed in testing of VIIRS Flight 1 despite the apparent lack of out-of-band signal in the filters as previously specified and measured during the procurement and build process. To correct this for VIIRS Flight 2 and subsequent, new laboratory measurements, Total Transmission and Slit Modified Transmission, using standard laboratory spectrophotometers with commercial attachments were developed at Raytheon and are discussed. Measurements with these techniques demonstrated that the out-of-band features observed at the sensor level are also seen at the individual filter level. These features are due to the Stierwalt effect where out-of-band features are observed when a detector is in proximity to a filter (near field), as the out-of-band light exits the filter assembly at discrete large angles from normal, to which the term “Angle Resolved Scatter” (ARS) has also been applied. Thus filters can be characterized using commercially available spectrophotometer techniques at the filter level prior to assembling into butcher-blocks or testing at a higher integration level for optical cross-talk. The simplicity of these techniques allows the filter manufacturer to incorporate them in their testing. This capability was incorporated and implemented into Raytheon specifications for filters and filter assemblies for the VIIRS VIS/NIR filters for VIIRS F2 and subsequent flights. Testing of these new filters at the component level using the Total Transmission and Slit Modified Transmission techniques showed the absence of these ARS optical cross-talk features. In-band scatter measurements showed significantly reduced scatter compared with Flight 1 as expected as the out-of-band light observed for the Stierwalt effect is a combination of scattering and optical design. Further measurements at the butcher-block level were made demonstrating that the filter-level measurements were a good indicator of optical cross-talk performance at the next higher level. This presentation will present the techniques developed and the correlation between filter-level and butcher-block-level performance for out-of-band features.
Techniques for Characterizing Filters for Sensor Optical Cross-Talk
Interference filters assembled in ‘butcher-block’ arrays are used close to focal plane arrays for spectral definition in sensors such as VIIRS. Ideally only light incident on a particular filter above a detector element which is within the specified band will reach the detector. Optical cross-talk is in-band or out-of-band light incident on an adjacent filter or adjacent region of the same filter reaching the detector. Out-of-band optical cross-talk results in spectral and spatial ‘impurities’ in the signal and consequent errors in the calculated environmental parameters such as ocean color that rely on calculations based on combinations of signals from more than one band. Out-of-band signal and out-of-band optical cross-talk was observed in testing of VIIRS Flight 1 despite the apparent lack of out-of-band signal in the filters as previously specified and measured during the procurement and build process. To correct this for VIIRS Flight 2 and subsequent, new laboratory measurements, Total Transmission and Slit Modified Transmission, using standard laboratory spectrophotometers with commercial attachments were developed at Raytheon and are discussed. Measurements with these techniques demonstrated that the out-of-band features observed at the sensor level are also seen at the individual filter level. These features are due to the Stierwalt effect where out-of-band features are observed when a detector is in proximity to a filter (near field), as the out-of-band light exits the filter assembly at discrete large angles from normal, to which the term “Angle Resolved Scatter” (ARS) has also been applied. Thus filters can be characterized using commercially available spectrophotometer techniques at the filter level prior to assembling into butcher-blocks or testing at a higher integration level for optical cross-talk. The simplicity of these techniques allows the filter manufacturer to incorporate them in their testing. This capability was incorporated and implemented into Raytheon specifications for filters and filter assemblies for the VIIRS VIS/NIR filters for VIIRS F2 and subsequent flights. Testing of these new filters at the component level using the Total Transmission and Slit Modified Transmission techniques showed the absence of these ARS optical cross-talk features. In-band scatter measurements showed significantly reduced scatter compared with Flight 1 as expected as the out-of-band light observed for the Stierwalt effect is a combination of scattering and optical design. Further measurements at the butcher-block level were made demonstrating that the filter-level measurements were a good indicator of optical cross-talk performance at the next higher level. This presentation will present the techniques developed and the correlation between filter-level and butcher-block-level performance for out-of-band features.