----------------------------------- ----------------------------------- Calibration reports and long term planning A NUMBER OF WARNINGS AND RACCOMANDATIONS CONCERNING THE USE OF CALIBRATED PFS DATA -------------------- -------------------- Index -------------------- -------------------- 1- THE EXPERIMENT 2- INTERFEROGRAMS 3- THE LWC SPECTRA 4- THE SWC SPECTRA -------------------- -------------------- 1 - THE EXPERIMENT -------------------- -------------------- 1.1 – PFS is a double pendulum interferometer. Data have been acquired moving the double pendulum in both directions : forward and reverse. For a number of technical details, not to be discussed here, the interferometer behaves as a different instrument in the forward or in the reverse motion, therefore the two sets of measurements need to be handled and calibrated separately. 1.2 – PFS has the possibility to change the speed of the double pendulum, which implies changing the operating frequencies. In the Near Earth Verification the study of the frequency response to the spacecraft microvibrations was performed. The decision was taken to adopt the speed 2500 Hz, which allowed to move strong disturbances from the spacecraft present in the range 2000 – 2500 cm-1 in the range 1600-2000 cm-1. The measurements at Mars have always been taken with this new speed, which has the consequences described below. 1.3 – PFS works as a double interferometer with two spectral channels the LWC and the SWC. The official wavenumber range of the two channels is: LWC = 275 – 1600 cm-1 and SWC 1800- 8400 cm-1. This is valid for single measurements, while for large averages of spectra it seems that the range can be extended, as the effects of the spacecraft microvibrations are with vanishing average. -------------------- -------------------- 2 - INTERFEROGRAMS -------------------- -------------------- 2.1 – The first action on the interferogram is to subtract to its measurements its average value, so that it becomes around zero ( positive and negative). 2.2 – Spikes or other disturbances in the interferograms. Interferograms may present occasionally some bad measurements, either because of motion problems or because of laser diodes malfunctioning. The Fast Fourier Transform of these interferograms would produce, if not cured, wrong spectra. We therefore recommend to check the interferograms against spikes : a) a single spike case. In this case the single measurement can be replaced by the linear interpolation of the two adjacent ones. b) Sometimes the interferogram has the last ( or the first) set of measurements saturated or out of range , the possible cause being low quality motion or impossibility to move. In these cases the interferogram can be recovered by replacing those points with zeroes. 2.3 – Interferogram symmetrization. a. In the 3 calibration samples provided in the Archive, the computation of the spectrum ( FFT transform) has been performed without previous interferogram symmetrisation. The handling of the interferogram to compute the spectrum (before calibration) is well described in the book “ Fourier Transform Spectroscopy” by S.P. Davis, M.C. Abrams and J.W. Brault, Academic Press, London 2001. The process is not unique and if done conveniently can solve a number of problems. In particular the symmetrization of the interferograms is needed if we want to allow the radiance to go negative for a correct calibration. Symmetrization of the interferograms is done by means of the phase correction transformation. b. As the calibrated data samples are computed without interferogram symmetrization, the radiance inside the CO2 absorption bands could not be correct as it requires a more accurate study. -------------------- -------------------- 3 - LWC SPECTRA -------------------- -------------------- 3.1- Single spectra can be considered valid only between 275 cm-1 and 1600 cm-1. Average of 10 spectra improves SNR but does not change this limit. Average of several thousands of spectra can also change this limit and allow the spectrum between 1600 and 2000 cm-1 to become valid. 3.2- Low level radiance spectra show occasionally an instrumental effect caused by the spacecraft microvibrations: tilting of the spectrum toward high or low radiance levels , with the CO2 band (600-670 cm-1) being the central point for the tilting. This means that the CO2 band is not affected, while the other portions of the spectrum is affected. The average of 10 spectra is usually sufficient to eliminate this effect. 3.3- PFS is a differential instrument, namely the temperatures are measured with respect the temperatures of the PFS detectors. The LWC detector is kept at 284 K. If Mars temperature goes above the 284 K, we have a thermal inversion. The interferogram phase function is very sensitive to this thermal inversion, and much care must be put in considering the spectrum, as the calibration procedure must change, or wrong anomalous structures will appear in the spectrum. 3.4 – A small spurious feature is occasionally present in the spectrum at 345 cm-1, which is caused by the responsivity having a sharp thin feature there. -------------------- -------------------- 4 – SWC SPECTRA -------------------- -------------------- 4.1 – In all the Nadir measurements the portion of the spectrum between 4500 and 6000 cm-1 must be considered as non valid. This portion of the spectrum, often called “the 2 microns modulation” also when we average thousands spectra will not be useful, as the radiance is destroyed by some unknown phenomena, presently not understood. The 2 microns modulation disappears most of the time when we take measurements at limb. 4.2 – Also the SWC is a differential instrument, with the detector at 210 K. With respect to the LWC the Martian temperatures are often above and below this temperature. This creates often a problem of temperature inversion. In particular inside the 4.3 microns band the Martian temperature of the top layers of the atmosphere can easily be at temperature lower than 210 K, with consequent problems in getting a reliable radiance value in the bottom of the band. Ad hoc studies need to be performed to obtain the correct radiance in the bottom of band present in the thermal part of the SWC. 4.3 – Sometimes the spacecraft microvibrations generate a low level ghost of the 4.3 microns band of the CO2. As the ghost phase can be positive or negative, it can be easily identified in looking at the standard deviation of an average spectrum. In any case is sufficient to average 10 spectra to eliminate these ghosts. 4.4 - A pointing mirror contamination is present between 2850 and 2980 cm-1 . This contamination was absent on ground, and present immediately when in space. With ad hoc studies this contamination can be eliminated and included in the responsivity. 4.5 – A basic difference in the SWC spectra is present between before orbit 634 and after it. The difference has been caused by the switching off of the SWC laser diode, and the use of the LWC laser diode to control the motion of both motors for both channels. Before orbit 635 both laser diodes were on, after it only the LWC laser diode was on to drive double pendulum motion and interferogram sampling of both channels. In this way the quality of the SWC spectrum improved a lot as satellites of the SWC laser diode disappeared from the spectrum. However the sampling path changed, and the wavenumber range covered changed. Consequently new calibration files must be used. 4.6 – In the first 634 orbits , when the SWC laser diode was on, all the spectra from the 4300 cm-1, to 8200 showed many strong spikes due to vibration satellites of the strong laser diode intensity observed at 8380 cm-1. These satellites were caused by the 4 reaction wheels and by the 3 (+3) laser gyroscope. Occasionally up to 12 sets of spikes were observed. They also have vanishing average. 4.7 – After orbit 635 : SWC laser off: everything is much better and , however , we must keep in mind that the sampling grid is slightly different and must be reconsidered.