Project Characterization
Instrument characterization is a very important part of the project. Since the goal is to create an instrument grade device, the instrument must be well characterized. In other words, we need to know how the instrument performs under standard conditions. In this way, we can know what features in the data are from the system, and what features are from the object we are trying to measure.
The first part of the characterization is understanding how the detectors respond to different polarization states. The optical front end is selecting different polarization states of light, so characterizing the detector's response to this is important. Different methods were used to isolate and reduce environmental and instrument error. These methods are described below.
Initial testing with OceanOptics light source
ASD and OceanOptics test with Integrating Sphere
OWL (Outrageously Wide Lambda) with Integrating Sphere
OWL test with dual Integrating Spheres
Solar Characterization
Calibration
Initial Testing using OceanOptics light source
The first test was done with an OceanOptics light source, a mechanical stage, a linear polarization stage, a diffusion plate, and both ASD and OceanOptics spectrographs. The rotation stage was manually turned to every 45° to determine spectrograph response. The data showed roughly 5% polarization sensitivity. A factor contributing to this sensitivity may have been the non-uniformity of the light source, or a response due to the diffusion plate. It was decided to move to an integrating sphere to improve the uniformity of the light source for the next test.
ASD and OceanOptics test with Integrating Sphere
The integrating sphere showed a substantial improvement over the OceanOptics light source. The ASD spectrometer showed a polarization sensitivity of 0.3%, and the OceanOptics spectrometer showed a polarization sensitivity of 1.5%. At this point it was decided to go the the next phase of testing, which required a polarization plate with a larger spectral range.
OWL (Outrageously Wide Lambda) with Source Integrating Sphere
The Outrageously Wide Lambda Polarizer is a wire-grid technology polarizer with a specialized coating to improve contrast in the visible range. This allows for a very large spectral range, from 350 to 2500nm. This test was to see how the system would behave at the near-IR, since the previous polarization plate was not responding in this range.
This test showed an extremely large polarization sensitivity, up to 30%. With this response characterized, the design problem became much harder.
OWL test with Receiver Integrating Spheres
With the polarization sensitivity of the detector confirmed, the decision was made to attempt to feed the detector randomly polarized light. In order to accomplish this, an integrating sphere needed to be placed between the polarization stage and the entrance to the spectrograph's fiber. Testing commenced immediately do to a generous donation of an integrating sphere by ILX Lightwave. In using an integrating sphere the amount of light going into the fiber was dramatically reduced. This caused huge noise levels after 2000nm, since there was insufficient light to have a stable reference. However, the amount of polarization response seen in the detector was reduced by a factor of 3. This proved that the integrating sphere was working as planned, however more light was needed for this method to work.
Solar Characterization
After many iterations, the source integrating sphere was determined to not have enough power for the characterization if the front end had an integrating sphere before the detector. The advantage of using a source sphere was that it was well characterized radiometricly from 350 to 1100nm. However, the measurement required did not need to be radiometricly accurate. The relative intensity was the only information being used, so disregarding radiometric calibration was not a huge loss. Since a direct path to the sun is inherently free of polarization effects, looking directly at the sun with the optical front end also gave a fairly good characterization. The only parts in the data that would be effected are the data at the atmospheric absorption lines. Since the data we were trying to characterize were not in these regions, the sun was a valid characterization target. Below are plots from looking directly at the sun with the system, both with a sphere and without a sphere.
The characterization with the integrating sphere shows a deviation of no more than 1% for areas of interest from 400 to 1750nm. The characterization with the fiber shows systematic deviation of up to 13%, but this should be able to be calibrated out if it proves repeatable.
Pictured below is the setup for the measurement taken above using the integrating sphere.
Project Calibration
The ASD spectrometer has been calibrated with standards traceable to NIST (National Institute of Standards and Technology) at an off site calibration facility. The optical front end will be characterized with an integrating sphere followed by a known polarization plate and a calibrated power meter. The optical front end will then be put in between the integrating sphere and the power meter, and the response of the optical front end will then be characterized by a transfer of accuracy of the polarization plate and power meter.The pointing accuracy of the telescope mount will be based upon manufacturer specifications, and will not be characterized.