Research Support Instruments, Inc., Boulder, CO b. Goddard Space Flight Center, Greenbelt, MD c

Transcription

1 SR-935 Characterization of a hardened ultrastable UV linear variable filter and recent results on the radiometric stability of narrow band interference filters subjected to temperature/humidity, thermal/vacuum and ionizing radiation environments Donald F. Heath, a Ernest Hilsenrath, b and Scott Janz c a Research Support Instruments, Inc., Boulder, CO b Goddard Space Flight Center, Greenbelt, MD c Joint Center for Earth System Technology, UMBC, Baltimore, MD Remote Sensing of the Atmosphere and Clouds, SPIE 3501, (1998) September 1998, Beijing, China Copyright 1998 Society of Photo-Optical Instrumentation Engineers This paper will be published in Optical Remote Sensing of the Atmosphere and Clouds and is made avaialable as an electronic preprint with permission of SPIE. Single print or electronic copies for personal use only are allowed. Systematic or multiple reproduction, or distribution to multiple locations through an electronic listserver or other electronic means, or duplication of any material in this paper for a fee for commercial purposes is prohibited. By choosing to view or print this document, you agree to all the provisions of the copyright law protecting it.

2 Characterization of a hardened ultrastable UV linear variable filter and recent results on the radiometric stability of narrow band interference filters subjected to temperature/humidity, thermal/vacuum and ionizing radiation environments Donald F. Heath, a Ernest Hilsenrath, b and Scott Janz c a Research Support Instruments, Inc., Boulder, CO b Goddard Space Flight Center, Greenbelt, MD c Joint Center for Earth System Technology, UMBC, Baltimore, MD ABSTRACT A series of in-band and out-of-band transmittance measurements of filters covering the wavelength range from 317 to 1019 nm and a linear variable filter for the 300 to 400 nm range have been made. The bandpass filters and the linear variable were fabricated using the ion-assisted-deposition or similar processes. The radiometric stability of the central wavelength, bandpass (FWHM), and peak transmittances were measured before and after exposures to combined high temperature and humidity, a thermal vacuum cycle, an ionizing particle radiation environment, flight on the Space Shuttle and at two temperature ranges. Representative radiative signal-to-noise ratios are given for solar irradiance observations with a silicon photodiode detector. Keywords: Environmental effects on interference filters, linear variable filters, ultrastable interference filters 1. INTRODUCTION Prior to the 1990s, it was extremely difficult to utilize the many advantages of non-dispersive narrow-band interference filters for precision spectroradiometric measurements from space because of their inherent unstable characteristics. Some of these unstable characteristics are associated with changes in central wavelength, spectral transmittance and bandpass associated with the absorption or desiccation of water vapor, ageing, temperature changes, ultraviolet and ionizing particle radiation and thermal-vacuum effects. Their utilization in space- and surface-based spectroradiometric monitoring of direct and scattered solar radiation was also adversely affected by low peak transmittance, especially in the ultraviolet, low spectral resolution, and moderately high levels of out-of-band transmittances in addition to their unstable characteristics previously mentioned. A good description of problems associated with the use of narrow-band interference filters in spectroradiometric determination of total column ozone is given by Basher 1 and Basher and Matthews. 2 In recent years, a new narrow-band interference filter fabrication process has evolved which is often referred to as the ion-assisted deposition (IAD) process. This process or variations of it such as the microplasma process of Optical Corporation of America (OCA) combined with the use of refractory metal oxides such as TiO 2, Ta 2 O 5, HfO 2, etc. combined with SiO 2 provide coatings of high chemical and thermal stability. The normal vacuum deposition process produces multilayers of thin films which are characterized by a porous structure which is severely affected by humidity variations and water vapor absorption. Further author information D.F.H.: dheath@csn.net; (303) E.H.: hilsen@ssbuv.gsfc.nasa.gov; (301) S.J.: janz@ssbuv.gsfc.nasa.gov; (301)

3 IAD and similar processes have been investigated by a number of laboratories/companies as a means for overcoming this problem of achieving a unity fill factor by filling in the pores. Filters fabricated with refractory metal oxides using the IAD or a similar process show vastly superior stability under all conditions of ordinary filter degradation. This work summarizes and describes the most recent results on the effects of combined high temperature and humidity on filter central wavelength, spectral bandpass (FWHM), and peak transmittance on a series of filters fabricated using the IAD or similar process on filters from 317 to 1019 nm with FWHM which range from 0.6 to 17 nm. Results are given for pre-flight interference filter characterization of the Limb Ozone Retrieval Experiment (LORE) which was flown on the Space Shuttle STS-87 flight in November The effects of a temperature change of 26C, a 7 day thermal vacuum cycle over the temperature range from -135 to +100C, and an ionizing particle radiation environment of 30 krad exposure on the central wavelength, spectral bandpass and peak transmittance are described. Results on preliminary characterization measurements on a linear variable filter (LVF) for the wavelength range of 300 to 400 nm over 25 mm length are given for in-band and out-of-band transmittance. This LVF was fabricated by Barr Associates using the IAD process. 2. STABILITY OF IAD BANDPASS INTERFERENCE FILTERS SUBJECTED TO ENVIRONMENTAL STRESS Previously published investigations 3,4,5 of the stability of narrow-band interference filters produced with an IAD-type process have indicated that the central wavelength, FWHM and transmittance curves exhibit a higher degree of insensitivity to large changes in temperature and relative humidity. In this work we report the results on the stability of central wavelength, FWHM and peak transmittance of a series of eleven IAD-type interference filters in the wavelength region of 317 to 1019 nm after exposure to a relative humidity of 95% at a temperature of 85C for 48 hours and 15% relative humidity at 7C for 24 hours. Filter characterization measurements were made at 24 and 50C for a group of five filters (345 to 1000 nm) which are of the type used in the Limb Ozone Retrieved Experiment (LORE) flown on the Space Shuttle flight STS-87 in November Sets of these filters were also measured before and after a 7-day thermal vacuum cycle of -135 to +100C and before and after exposure to 30 krad of ionizing radiation from a Co 60 source. It is important to note that all filter transmittance measurements reported in this work were made with the filter located between the exit slit of the monochromator and a silicon photodiode which was sensitive to radiation longer than 1100 nm. The light passing through the filter was in a f/7 diverging beam. In-band filter transmittances were measured with light passing through each of the two sides of the filter. The measured in-band transmittances were invariant at the 1% level or less. The out-of-band transmittances of the visible and near-ir filters were not invariant with the direction of light through the filter due to fluorescence into the passband of the filter. This is the source of high out-of-band transmittances for the 400 nm to the 650 nm region in Figure 8. Inverting the direction of light passing through this filter will change the apparent outof-band transmittances by a factor of Fluorescence into the passband of the filter is also the source of the differences observed in the measured transmittances on the short wavelength side of the LORE filter shown in Figure Combined effects of high temperature and high humidity on interference filter characteristics A series of eleven IAD-type filters were subjected a relative humidity of 95% at a temperature 85C for 48 hours. This was followed by a 24 hour exposure at a relative humidity of 14% at a temperature of 7C for 24 hours in an environmental chamber at Orbital Sciences Pomona, CA facility. Comparisons of measurements of central wavelength, FWHM, and peak transmittance for eleven filters are given in Table 1. The 317 and 329 nm filters were fabricated by Barr Associates and the other 9 filters were produced by OCA. The elements of the Barr filters were cemented together with epoxy, whereas the elements of the OCA filters were not sealed together. The high in-band transmittances and extremely low out-of-band transmittances are typical of IAD filters produced using multilayers of refractory oxides and SiO 2, shown in Figures 1 through 10. The in-band transmittances are shown for the preand post-exposure to combined high temperature and humidity for several representative filters of different bandwidths and central wavelengths from the UV to near IR. The small differences in the pre and post in-band transmittance measurements shown in Figures 1, 3, 5, 7, and 9 are comparable to the measurement uncertainties. 2

4 Table 1. Effects of High Temperature and High Humidity on Filter Transmittances Filter Central Wavelength (nm) FWHM (nm) Peak Transmittance (T P ) Pre Post Pre Post Pre Post Ratio T P Post/Pre Figure 1. In-band transmittance before and after an extended period of exposure to high temperature and high humidity for a narrow-bandwidth filter (FWHM = 0.6 nm). 3

5 Figure 2. Out-of-band transmittance for filter in Figure Effect of a Space Environment A filter from the same lot as the one shown in Figure 3 was flown on the Space Shuttle STS-72 as part of the Rayleigh Scattering Attitude Sensor (RSAS) experiment. The pre- and post-flight transmittances are shown in Figure 11. There appears to have been a shift of the central wavelength to shorter wavelengths of about 0.13 nm. Five LORE IADtype filters were flown on the Space Shuttle STS-87 in November 1997; however, the LORE instrument was flown in a pressurized container. These filters are expected to be available for post-flight measurements in July. 2.3 Temperature dependent characteristics In-band transmittances were made with the filters at 24 and 50C. The results of these measurements for the five LORE filters are given in Table 2. Over this temperature range there appears to be a small shift to the red of about nm/c, which is probably not significant. Quite possibly this temperature increase caused about a 1.0% decrease in peak transmittance and no significant change in FWHM. 2.4 Thermal vacuum effects Figure 3. Transmittance before and after high temperature and humidity exposure. The observed effects of a 7-day thermal cycling from -135 to +100C on five LORE filters are summarized in Table 3. During this cycle the laminated Barr 675 nm filter shattered, which most likely was due to differences in thermal expansion coefficients. The 7-day thermal vacuum cycling appears to have induced an average shift in the central wavelengths to shorter wavelengths of about 0.10 nm, which probably is not statistically significant. The FWHM appears to have decreased by about 0.07 nm, which is also most likely statistically insignificant. Surprisingly, the peak transmittance of the 345 nm filter increased by 13% and decreased by about 11% for the 600 nm filter. 4

6 Figure 4. Corresponding out-of-band transmittance for filter shown in Figure 3. Figure 5. In-band transmittance before and after an extended exposure to high temperature and humidity for a narrowbandwidth filter (FWHM = 2.2 nm). The extremely large temperature range of 235C was not intentional. The flight filters in the LORE instrument only experienced a change in temperature of 40C. The reason the test filters experienced such a wide dynamic range of temperature is because they were mounted on a tray in the thermal vacuum chamber, and were not mounted in the instrument. 2.5 Ionizing radiation effects The LORE filters were irradiated with 30 krad of ionizing radiation from a Co 60 radioactive source at the Goddard Space Flight Center. The effects on the filter characteristics are given in Table 4. These measurements indicate a statistically insignificant shift of the central wavelength to shorter wavelengths of 0.06 nm, and a negligible change in FWHM. There was a significant decrease in filter peak transmittance of 45% for the 345 nm filter and an average of 10% decrease for the 525, 600, and 675 nm filters. The most probable explanation for the decrease in the peak transmittance is the darkening of the color glass blocking elements by the ionizing radiation. 5

7 Figure 6. Corresponding out-of-band transmittance for filter shown in Figure 5. Figure 7. In-band transmittance before and after an extended exposure to high temperature and humidity for a narrowbandwidth filter (FWHM = 2.2 nm). 3. CHARACTERIZATION OF AN IAD LINEAR VARIABLE FILTER Preliminary characterization measurements have been made of a LVF fabricated with the IAD process by Barr Associates. The LVF covers a nominal wavelength range of 300 to 400 nm over a distance of 25 mm. The LVF was evaluated by imaging a 10 micron wide slit illuminated by a low-pressure Hg lamp onto the fused silica substrate with the deposited linear variable interference coating. A silicon photodiode is mounted behind the LVF. The spectral scan which is shown in Figure 12 was obtained by translating the filter perpendicular to the optical axis in 0.1 mm steps. The spectral scan shows three groups of Hg lines at 365, 334.1, and nm. The latter two lines, separated by 0.6 nm, while not resolved, begin to indicate the presence of two adjacent spectral lines. The FWHM of the group of lines is 2.6 nm. The peak transmittance of this group of lines is 15.8%. The LVF uses a combination of coatings on fused silica and uncoated UG11 color glass for the wavelength region < 270, > 400 nm. A purple filter which blocks < 300 nm and > 480 nm consists of two coated fused silica substrates laminated to uncoated BGG28 and BG3 color glasses. A measurement of the out-of-band transmittance was obtained by isolating the Hg line with filters and illuminating a 1.0 mm wide slit which was imaged onto the LVF. This scan of the out-of-band 6

8 Figure 8. Corresponding out-of-band transmittance for filter shown in Figure 7. Figure 9. Superimposed LORE filter transmittances measured at 24C and 50C. transmittance at nm is shown in Figure 13. Typically, the out-of-band transmittance is about , except at the intersection of the two types of blocking elements, which is about a factor of 10 higher. 4. RADIATIVE SIGNAL-TO-NOISE RATIOS The radiative signal-to-noise ratio is defined as the ratio of the in-band signal to the out-of-band signal for a particular filter, source and detector. The in-band signal is defined as the T()E()R()d and the out-of-band signal is defined as 1 2 7

9 Figure 10. Out-of-band transmittance for LORE 525 nm filter for two filter orientations at the exit slit of the monochromator. The dots are for measurements with light entering the filter face as recommended by the manufacturer. The crosses are for the orientation reversed. Figure 11. Superimposed RSAS filter transmittances measured before and after flight on the Space Shuttle. < 1 0 T()E()R()d > 2 T()E()R()d (1) where T() is the measured filter transmittance, E() is the solar irradiance, and R() is the response function in amps/watt for the silicon photodiode. In practice the above integrals are replaced by summation of the products over a finite number of wavelength intervals. A listing of some typical radiometric signal-to-noise ratios (SNR) for the IAD filters evaluated is given in Table 5. In deriving these radiometric SNRs, the apparent filter transmittance is used, which includes the filter transmittance and the fluorescence into the pass band of the interference filter and the detector radiant sensitivity at the central wavelength of the interference filter. A previous study has shown that the low SNR for the 710 nm filter is due to fluorescence effects. 5 This effect can result in orders of magnitude differences in out-of-band filter transmittances depending 8

10 upon whether the interference filter is placed between the light source and the entrance slit of the monochromator or between the exit slit and the detector. Unfortunately, in many space applications the detector views the non-dipsersive element without any means of spectral separation between them. Table 2. Temperature Effects on Filter Characteristics Filter Central Wavelength (nm) Change in Central Wavelength (nm) 24 to 50C Transmittance (Peak, 24C) Ratio T P (50C)/T P (24C) FWHM (nm) 24C Ratio FWHM 50C/24C Table 3. Thermal Vacuum Effects (-125 to +100C) Central Wavelength (nm) FWHM (nm) Transmittance (T P ) Filter Pre Post Pre Post Pre Pos Ratio 345# # # # Broken 4.53 Broken Broken 1000# Table 4. Ionization Radiation Effects (30 krad Exposure) Central Wavelength (nm) FWHM (nm) Transmittance(peak) Filter Pre Post Pre Post Pre Pos Ratio 345# # # # #

11 Figure 12. Translational scan of linear variable filter of Hg lamp. Figure 13. Out-of-band transmittance of LVF at SUMMARY AND CONCLUSIONS The in-band and out-of-band transmittances have been measured for 16 IAD-type narrow-band interference filters which cover a wavelength range of 317 to 1019 nm, with FWHM which range from 0.6 to 17.6 nm. Transmittance measurements were made before and after a long-term exposure to combined high temperature and humidity (95% relative humidity at 85C for 48 hours) a range in thermal vacuum cycling from -135 to +100C, an exposure to 30 krad of ionizing radiation from Co 60 gamma rays, and during temperature changes from 24 to 50C. The most sensitive filter parameter affected by 10

12 environmental tests was peak transmittance, which exhibited changes by as much as ±10% from the 7-day thermal vacuum cycle. Table 5. Typical Radiometric Signal-to-Noise Ratios (Ratio In-Band / Out-of-Band Signals) Wavelength (nm) Peak Transmittance FWHM (nm) Weighted Average Out-of-Band Transmittance Radiometric SNR (LVF) The change in peak transmittance observed after exposure to ionizing radiation is most probably due to darkening of the color glass filters. The uncertainty of the central wavelength of the filters is about 0.1 nm, of FWHM is about 0.05 nm and of peak transmittance is about 1.0%. One of the IAD filters was flown on the Space Shuttle STS-72 and after a nominal one to two weeks in space its central wavelength exhibited a 0.13 nm shift to shorter wavelengths. No detectable changes in peak transmittance or FWHM were observed. Preliminary characterization of in-band and out-of-band transmittances of a 300 to 400 nm LVF indicated a FWHM of 2.6 nm at 365 nm and a peak transmittance of 16%, and a very small indication of beginning to resolve two spectral lines separated by 0.6 nm. The average out-of-band transmittance in regions not blocked by color glasses is , which yielded an approximate radiative SNR of 30 for solar observations with a silicon photodiode detector. 6. REFERENCES 1. R. E. Basher, The Effect of bandwidth on filter instrument total ozone accuracy, J. Appl. Meteor. 16, , R. E. Basher and W. A. Matthews, Problems in the use of interference filters for spectrophotometric determination of total ozone, J. Appl. Meteor. 16, , J. Potter and J. Simons, Stability of IAD refractory oxide narrowband interference filters, Proc. SPIE 1952, , M. Scobey and P. Stupik, Stable ultra-narrow bandpass filters, SPIE International Symposium 2262, 37-46, D. F. Heath, Z. Wei, E. Hilsenrath, and S. Janz, Calibration and characterization of remote sensing using ultrastable interference filters, EUROPTO Series, H. Fujisada, ed., London, Proc. SPIE 3221, , T. A. Mooney, Focal plane filter microassemblies: A status report, EUROPTO Series, H. Fujisada, ed., London, Proc. SPIE 3221, ,