Laboratory Experiment of a High-contrast Imaging Coronagraph with New Step-transmission Filters Jiangpei Dou *a,b,c, Deqing Ren a,b,d, Yongtian Zhu a,b & Xi Zhang a,b,c a. National Astronomical Observatories/Nanjing Institute of Astronomical Optics & Technology, Chinese Academy of Sciences, Nanjing 210042,China b. Key Laboratory of Astronomical Optics & Technology, Nanjing Institute of Astronomical Optics & Technology, Chinese Academy of Sciences, Nanjing 210042,China c. Graduate University of Chinese Academy of Sciences, Beijing 100049, China d. Physics & Astronomy Department, California State University Northridge, 18111 Nordhoff Street, Northridge, California 91330-8268 ABSTRACT We present the latest results of our laboratory experiment of the coronagraph with step-transmission filters. The primary goal of this work is to test the stability of the coronagraph and identify the main factors that limit its performance. At present, a series of step-transmission filters has been designed. These filters were manufactured with Cr film on a glass substrate with a high surface quality. During the process of the experiment of each filter, we have identified several contrast limiting factors, which includes the non-symmetry of the coating film, transmission error, scattered light and the optical aberration caused by the thickness difference of coating film. To eliminate these factors, we developed a procedure for the correct test of the coronagraph and finally it delivered a contrast in the order of 10-6 ~10-7 at an angular distance of 4λ/D, which is well consistent with theoretical design. As a follow-up effort, a deformable mirror has been manufactured to correct the wave-front error of the optical system, which should deliver better performance with an extra contrast improvement in the order of 10-2 ~10-3. It is shown that the step-transmission filter based coronagraph is promising for the high-contrast imaging of earth-like planets. Keywords: extra-solar planets, high-contrast imaging, coronagraph, step-transmission filters, deformable mirror 1. INTRODUCTION With over 300 extra-solar planets found by using indirect techniques, the search of earth-like planets around these bright stars using direct imaging technique is increasingly important, since the search for life requires the ability to detect photons directly from earth-like planets and the use of spectroscopy to analyze physical and atmospheric conditions. The direct detection of earth-like low-mass planets is extremely challenging since the brightness ratio between an earth-like planet and its parent star is in the order of 10-9 in the visible, and the diffraction of starlight is much stronger than the * jpdou@niaot.ac.cn; phone 86-025-85482285 Techniques and Instrumentation for Detection of Exoplanets IV, edited by Stuart B. Shaklan, Proc. of SPIE Vol. 7440, 744019 2009 SPIE CCC code: 0277-786X/09/$18 doi: 10.1117/12.825253 Proc. of SPIE Vol. 7440 744019-1
nearby planet image. NASA has planed to build and launch a space-based telescope called Terrestrial Planet Finder (TPF). For the TPF Coronagraph, a contrast of 10-10 at an angular distance better than 4λ/D is required in the visible wavelength [1]. Recently, many high-contrast imaging coronagraphs have been proposed which can theoretically achieve 10-10 contrast at a few λ/d from the central star [2]. Ren and Zhu (2007) [3] proposed an apodized pupil that is based on step-transmission filters in which the intensity is apodized with a finite number of steps of identical transmission in each step. In our recent paper (2008) [4], we presented the results of the first laboratory experiment based on two 13-step-transmission filters, with visible light. The target contrast to be demonstrated in that experiment was set to be 10-4 ~10-5, and they indeed delivered a 10-5 contrast for the first test. Considering the satisfactory results, we still adopted a coronagraph employing step-transmission filters in this work since the manufacture is relatively simple and a precise transmission can be made based on current technique. A further improvement has been made to the newest filters from the design as well as the manufacture and the corresponding test procedures. In this paper, we present the latest laboratory results for a coronagraph that employs 13-step-transmission filters and it is optimized in the visible wavelength. The target contrast to be demonstrated in this experiment is 10-6 ~10-7. Finally it delivered a contrast in the order of 2.0x10-7 at an angular distance of 4λ/D, which is well consistent with our theoretical design. The manufacture of a deformable mirror to correct the wave-front error of the optical system is also discussed, which is expected to deliver better performance with an extra contrast improvement of 10-2 ~10-3. The coronagraph with a contrast of 10-10 is our long term goal, which will allow for the direct imaging of earth-like planets. 2. EXPERIMENT 2.1 A coronagraph based on 13-step-transmission filters Ren and Zhu [3] proposed a coronagraph using step-transmission filters. Although it can deliver a high contrast better than 10-10 in theory, its actual performance is limited by the transmission error in each step [4]. With such a high contrast design, the transmission of the last step will decreased to be 0.0229%. In practical, such a low transmission is extremely difficult to precisely test so a precise transmission for this step can not be assured. To overcome this problem we re-designed the filter with a target contrast to be set in a lower order of 10-6 ~10-7, in which case the corresponding transmission of the last step changed to be 1.22%, which is much easier to be made. Another advantage for such a design is its high system throughput of 41.4% when using one filter only for the coronagraph. The inner working angle (IWA) has been changed to 3~4λ/D and the outer working angle (OWA) stays the same since the total step number is also 13 for this work. Table 1 summarizes the specifications of the newly designed apodized pupil filters. Although using the two-filter configuration we can gain a 10-8 ~10-9 high contrast theoretically, however, we found it is difficult to reach such a performance because of the multi-reflection between two filters and possible wave-front error. The multi-reflection will introduce scatted lights. So in the following sections, we only discuss the test results of the coronagraph that uses one filter only. Figure 1 shows the transmission amplitude pattern of the newly designed 13-step filter and the theoretical coronagraph point spread function (PSF). The PSF contrast plot along the diagonal direction is shown in Figure 2, a contrast in the order of 10-6 ~10-7 can be achieved at an angular distance larger than 3λ/D. Proc. of SPIE Vol. 7440 744019-2
Table 1. New design of the apodized pupil filter. Apodized-pupil Filter IWA(λ/D) 3~4 OWA(λ/D) 13 Contrast 10-6 ~10-7 Throughput (%) 40 Fig.1. Left: New transmission amplitude pattern of the apodized pupil. Right: the PSF intensity distributions on the focal plane. The intensity is in linear scale. Fig.2. The theoretical contrast for one filter apodized coronagraph (solid line) and that without any apodization (doted line). A contrast in the order of 10-6 ~10-7 should be achieved at an angular distance of 3λ/D in theory. Proc. of SPIE Vol. 7440 744019-3
Based on current design we manufactured three 13-step-transmission filters, with an improvement by eliminating problems occurred in the previous one. The filters were manufactured with metallic Cr coated on a glass substrate, with a 100% normalized transmission in the central step. The transmission in each step is controlled by the thickness of the Cr film. Each step of the film is 4mm wide and allows an overlap of 0.02mm with its neighboring steps. Transmission accuracy is controlled within +/-3% for the first two steps and +/-5% for the remaining ones. The measurement data of the transmission for each filter is provided by the manufacturer. At this time, the glass substrate possesses a much higher surface quality with a P-V value better than λ/10. Therefore, no evident spherical aberration occurred in the tested PSF images shown in the following sections. Figure 3 shows the photograph of the actual filter and one photo of its surface under microscope. Each step of one such filter was tested under a microscope after receiving it. Unfortunately, pinholes were common for all three filters. Although an overlap of 0.02mm is allowed, a gap between the last two neighboring steps was also found due to bad manufacture of coating masks, which will introduce unwanted diffraction lights. Fig.3. Left: The photograph of the tested filter. Cr film was coated on the upper surface of the glass substrate and an anti-reflection coating was coated on the opposite surface. The clear aperture of the substrate is 62mm x 57mm with a thickness of 10mm. Right: Pinholes and the gap between the last two neighboring steps. Red circles indicate the pinhole on Cr film. Rectangular bar shows the gap between last two neighboring steps. 2.2 Optical Setup As shown in previous laboratory experiment that a great amount of scattered lights were introduced by multi-reflection of the CCD window, we re-adjusted the optical configuration this time. Figure 4 shows the optical setup for this work. We enlarged the PSF image to a proper size by relay optics. The long optical path provides a means to remove scattered light. Correspondently, the tested contrast has been improved from 10-4.5 to 10-5 at 4λ/D for the first filter (Filter 1), which is shown in Figure 5. Comparing the two observed PSF images, it is obvious that scattered lights were greatly eliminated around the central parts of PSF image (see Figure 5). Proc. of SPIE Vol. 7440 744019-4
Fig.4. Optical Configuration for the experiment. Before collimating lights from pinhole, we use an extra light baffle to block unwanted lights to the system and a biconvex lens with focal length of 250mm was introduced to enlarge the PSF image. Filter 1 Filter 1 Fig.5. Up: None enlarged PSF image (left) and enlarged PSF image (right) using Filter 1. Scattered lights occurred around the central PSF introduced by CCD window which can be further removed by using relay optics. Bottom: Corresponding contrast plot along the diagonal direction. The contrast has been improved from 10-4.5 to 10-5 at angular distance of 4λ/D, which shows that the relay optics can effectively reduce the scattered lights around the central region of PSF image. Proc. of SPIE Vol. 7440 744019-5
3. RESULTS AND DISCUSSION Although the contrast can be improved by using relay optics as discussed in Section 2, however, it does not reach the target contrast at least of 10-6. We found it is mainly caused by the dissymmetry of the coating film since this is the first time for the manufacturer to make this filter, which is in our estimation. Another two filters were again manufactured and the corresponding test results will be mainly discussed in this section. Figure 6 shows the observed PSF images and corresponding contrast using Filter 2 and Filter 3, respectively. Filter 2 Filter 3 Fig.6. Up: Observed PSF images using Filter 2 (left) and Filter 3 (right). Bottom: A contrast of 10-5.7 and 2x10-7 are achieved at an angular distance of 4λ/D for Filter 2 (left) and Filter 3 (right) respectively. For Filter 2, the maximum intensity of the central spot light (in red rectangle) is even higher than that of sub-peak (in blue circle), which is inconsistent with the theoretical design (see Figure 6). And the corresponding contrast will be limited by the diffraction from such a bright spot light along the vertical directions. That is why only a contrast in the order of 10-5.7 was achieved at an angular distance of 4λ/D. Such a result was caused by the large relative transmission error which reaches 10% in some steps according to our transmission test. Such a transmission error is needed to be Proc. of SPIE Vol. 7440 744019-6
further controlled for our coronagraph that uses these filters. Figure 7 shows the simulation results using the tested transmission data of the filter, and the simulated PSF image is totally consistent with the observed PSF image shown in Figure 6, which confirmed the contrast limitation is from the transmission error. Therefore we manufactured another filter (Filter 3) with a better transmission precision for each step. Using Filter 3, the bright spot in red rectangle reduced effectively and the corresponding contrast was improved to 2x10-7 at an angular distance of 4λ/D (see Figure 6). Fig.7. Photograph of simulated PSF image using the tested transmission data. The intensity distribution is totally consistent with the observed PSF in the discovery region. Three bright spots occurred in the horizontal arm that is consistent with the observed PSF (see Figure 6 up-left). Fig.8. The observed PSF image of using Filter 3 with an over-exposure time. Several speckle noise occurred in the detect region, which was mainly caused by wave-front distortion of the optical system. Figure 8 shows the observed PSF of using Filter 3 with an over-exposure time of 250s. The tested contrast along diagonal direction is well consistent with the simulation, however, we found the PSF image is not symmetrical and some speckle noise randomly occurs around the central parts of PSF image. These speckles were mainly caused by the wave-front error of the optical system [5]. Our later simulation shows that the wave-front distortion originates from the optical aberration due to the thickness difference of coating film among different steps, and maybe a limited factor for Proc. of SPIE Vol. 7440 744019-7
contrast better than 10-7. To overcome such a problem, we will introduce a deformable mirror (DM) to correct the wave-front distortion of the optical system. At present, a DM with a 12x12 actuators has been manufactured. We expect that the DM approach [6] will also provided an extra contrast gain of 10-2 ~10-3. 4. CONCLUSION In this work, we report the latest test results from our laboratory experiment using a coronagraph that is based on 13-step-transmission filters. In current test a contrast better than 10-6 has been achieved at an angular distance of 4λ/D, which confirms the stability of the coronagraph using step-transmission filters. As well we have identified several factors that limit the performance of the coronagraph. As a follow-up effort, we will introduce the DM manufacturing to correct the wave-front error of the optical system, which should deliver a better performance. Meanwhile, we will consider changing the coating materials from Cr film to Al. Since the Al film processes a similar refraction to air in visible wavelength the optical aberration due to thickness difference of coating film should be effectively reduced. We will report further results in future publications. ACKNOWLEDGEMENTS This work was supported by the National Natural Science Foundation of China under grants 10873024. We would like to thank the Mirror Laboratory of our institute for assistance with the optical aberration test and the substrate manufacture. We will also thank Changchun Institute of Optics, Fine Mechanics and Physics for their manufacture of the filter. This work is partly supported by NSF under grants AST-0501743 and ATM-0841440, as well as CSUN Probationary Faculty Support Program (SPRING 2009). REFERENCES [1] Brown R. A. and Burrows C. J., On the feasibility of detecting extrasolar planets by reflected starlight using the hubble space telescope, ICARUS 87(2), 484-497(1990). [2] Guyon O., Pluzhnik E. A., Kuchner M. J., Collins B. and Ridgway S. T., Theoretical Limits on Extrasolar Terrestrial Planet Detection with Coronagraphs, ApJ 167(1), 81-99(2006). [3] Deqing Ren and Yongtian Zhu, A Coronagraph Based on Stepped-Transmission Filters, PASP 119(859), 1063-1068(2007). [4] Jiangpei Dou, Yongtian Zhu, Deqing Ren and Xi Zhang, Laboratory experiment of a coronagraph based on step-transmission filters, Proc. SPIE 7010, 70104J-1-10 (2008). [5] Deing Ren and Haimin Wang, Spectral subtraction: a new approach to remove Low and high-order speckle noise, ApJ 640(1), 530-537(2006). [6] Jiangpei Dou, Deqing Ren, Yongtian Zhu and Xi Zhang, Focal plane wave-front sensing algorithm for high-contrast imaging, Science in China Series G 52(8), 1284-1288 (2009). Proc. of SPIE Vol. 7440 744019-8