Thermal Stability and SOFIE Measurements

Changes in instrument temperature can affect the measurements in a variety of ways. Some important considerations are outlined below.

1. Detector response

Changes in detector temperature can affect their performance.

2. Spectral filter response

Temperature changes can cause changes in filter position, opacity, and width.

3. Beam Steering

Temperature changes in the optical components can alter their characteristics and cause the beam to move slightly out of alignment. Known as "beam steering", this effect can induce drift across a non-uniform detector, or intensity variation due to non-uniform optics.

4. Chopper Stability

Thermal changes can affect the behavior of the chopper.

The above affects generally manifest themselves as signal drift that is correlated to instrument temperature. Signal drift can be understood through analysis of flight data and removed during data processing.

Analysis of instrument temperature effects using HALOE flight data

HALOE performs exo-atmospheric solar scans as part of each measurement event. Because the distance and angle between satellite and sun vary over time, the instrument experiences varying solar intensity which results in a range of instrument temperature. Because the target solar intensity is fairly constant over short periods of time, changes in the measured peak solar intensity are assumed to result from changes in instrument temperature. Thus, the exo-atmospheric solar scan data provide a means for understanding the effects of instrument temperature on instrument response.

Figure 1 below (Figure numbers on top) shows the relationship between filter temperature and peak signal in the CH4 V channel, for all HALOE events during 1992. These data indicate a linear relationship between filter temperature and detector voltage. The induced transmission due to filter temperature change was determined using the HALOE CH4 V measurements (Figure 2). For this purpose, the induced transmission is determined as the ratio of voltage at some temperature to the voltage at the initial temperature. These results indicate that a thermal change of about 2e-4 degrees C will induce a transmission equal to the expected SOFIE measurement noise (1e-6). Figures 3 and 4 shows similar results as in Figures 1 and 2, except for the HALOE HCl V channel. The HCl channel response to thermal drift is very similar to that of the CH4 channel.

Thermal stability for the SOFIE difference signal measurements will depend on how well the thermal dependence of filter pairs are matched. Figure 5 shows a theoretical calculation of the thermal drift that will induce a difference signal of 1e-6, as a function of the thermal mismatch between filter pairs. For this purpose filter response mismatch is defined as the difference in slope of the signal vs. temperature curves (e.g. Figure 1). An important point here is that the thermal stability of a single filter does not matter as long as band pair filters have similar thermal response characteristics.

The HALOE HCl and CH4 channels were used as proxies to estimate expected thermal drift induced difference signals for SOFIE band pairs. The HCl and CH4 channels are close in wavelength, with band centers at 3.40 and 3.46 microns, respectively. This separation is generally smaller than the separation between most SOFIE band pairs. HALOE data for 1992 (Figures 1 and 3) were used to derive the CH4 - HCl difference signal induced by thermal drift from an initial temperature of 20 deg. C (Figure 6). The induced difference signals are less than 1e-6 for temperature changes of less than about 0.02 degrees C. The thermal response mismatch between the HALOE CH4 and HCL channels is shown for 10 years of the HALOE mission in Figure 7. These results indicate that the HALOE filter characteristics have remained very stable, with thermal mismatch generally less than 1% between these two channels.

The HALOE HF and CO2 channels are farther apart, at 2.45 and 2.80 microns, respectively. These channels were also used to characterize pseudo band pair difference signals induced by thermal drift, and should represent a worst case scenario. For these channels a difference signal of 1e-6 is induced by roughly 1e-4 degrees C of thermal drift.

Signal and filter temperature drift rates were examined for the exoatmospheric portion of HALOE occultations, for high and low beta angles (the angle between the orbit plane and earth-sun vector). SOFIE will experience beta angles less than +/- 6 degrees. Heating rates are highest for high beta angles, and thus represent a worst case scenario that SOFIE will not experience. Exoatmospheric HCL channel signal and filter temperature profiles are shown in Figures 8 and 9 for one HALOE event at high beta angle (-60 degrees). These data are shown versus elapsed scan time, which is equivalent to altitude. The exoatmospheric drift rates in V and filter temperature were -3.3e-5 V/s and 1.2e-3 K/s, respectively. Figure 10 shows that the dependence of V on T (about 0.027 V/K) for this scan was similar to the 1992 average shown in Figure 3 above. Exoatmospheric HALOE data for one event at low beta angle (-3 degrees) are shown in Figures 11-13. These measurements show negligible drift in V and filter temperature, and no discernible relationship between V and T (Figure 13). These results should be more representative of SOFIE occultations.

HALOE measurement characteristics were considered in simulations of the signal drift induced by thermal change during an occultation. In this exercise, the important instrument characteristics are 1) the thermal response of a filter (V vs. T, e.g., Figure 1), thermal response mismatch for a filter pair (e.g., Figure 7), thermal drift rate for a filter, and thermal drift rate difference for a filter pair. HALOE scan data show that the exoatmospheric drift rate for the HCL filter temperature is on the order of 2e-4 K/s, and that the CH4 filter temperature drift rate is roughly within -30% to 30% of the HCL filter drift rate. As shown above, the HCL and CH4 channel thermal response is on the order of 0.02 V/K, and that these channels exhibit thermal response mismatch of less than 1%. These characteristics were used to construct a theoretical scan, with the goal of determining the signal and difference signal drift from their initially balanced conditions. Results considering thermal drift of 0.002 K/s and 0.0025 K/s for filters 1 and 2 (Figures 14 and 15) show that after about 1 minute of exoatmospheric scan time the single detector signals drift by about 1e-3 of the balanced signal, and that the difference signal drifts by about 1e-4 of the balanced signal. Using filter temperature drift rates of 0.0002 and 0.00025 for filters 1 and 2 (closer to HALOE measurements at low beta), reduces the signal drifts by roughly an order of magnitude (Figures 16 and 17).

The difference signal drift after 60 seconds exoatmospheric scan time versus the thermal drift rate of filter 1 is shown in Figure 18. For these calculations, filter 2 had a thermal drift rate equal to 25% of that for filter 1. The thermal response of filter 1 was set to 0.02 V/K, and the filter pair thermal response mismatch was 1%. The difference signal drift is reduced when the thermal drift rate is low, and is near the desired level for SOFIE (1e-4 of balance) for a drift rate of 0.001 K/s (0.06 K in one minute).

If the SOFIE filter characteristics are similar to HALOE, then the results presented here provide some guidance in predicting the effects of thermal instability on SOFIE measurements. Based on the analysis below, the temperature change inducing a difference signal of 1e-6 is probably less than 0.02 deg C. A single detector signal of 1e-6 is induced by a thermal change of roughly 2e-4 deg. C. In any case, signal drift induced by thermal instability can be characterized in flight data by analyzing the exo-atmospheric portion of a scan to determine a drift rate which can be removed from signals before the retrieval of science data.