Design and test of a high-contrast imaging coronagraph based on two. 50-step transmission filters

Transcription

1 Design and test of a high-contrast imaging coronagraph based on two 50-step transmission filters Jiangpei Dou *a,b, Deqing Ren a,b,c, Yongtian Zhu a,b, Xi Zhang a,b,d, Xue Wang a,b,d a. National Astronomical Observatories/Nanjing Institute of Astronomical Optics & Technology, Chinese Academy of Sciences, Nanjing ,China b. Key Laboratory of Astronomical Optics & Technology, Nanjing Institute of Astronomical Optics & Technology, Chinese Academy of Sciences, Nanjing ,China c. Physics & Astronomy Department, California State University Northridge, Nordhoff Street, Northridge, California d. Graduate University of Chinese Academy of Sciences, Beijing , China ABSTRACT We propose a high-contrast coronagraph based on the step transmission filters for the direct imaging of an Earth-like exoplanets. To demonstrate the performance of the coronagraph, two 50-step transmission filters were manufactured and several experiments have been performed. At present, the coronagraph can reach a high contrast around 10 7 at an inner angular distance of ~2λ/D in the visible wavelength. Such a coronagraph should be installed on an off-axis space telescope which will be promising for the direct imaging of an Earth-like exoplanet in the future. Keywords: Earth-like exoplanets, high-contrast imaging, coronagraphy, step-transmission filter, space telescopes 1. INTRODUCTION The detection of an Earth-like exoplanet is one of the most exciting topics in modern astronomy and public domain. Over 680 exoplanets have been discovered mostly by the indirect detection methods, such as the radial velocity methods or transiting approach; however, none Earth-like exoplanet has ever been confirmed up to date. Furthermore, little is known about the discovered exoplanet s physical characteristics, such as the atmosphere conditions and chemical compositions. The characterization of an exoplanet s atmosphere requires the direct detection of photons from the exoplanet. In recent years, direct imaging of exoplanets is receiving increasing attention, which will allow the astronomical community eventually to achieve the most critical scientific goals of astrophysics. The direct imaging and the follow-up spectroscopic characterization of Earth-like planets would allow us to fully characterize a planet's habitability and may detect signs of life. Such detection will finally help to determine another Earth-twin, which is one of the most fundamental scientific questions. In a broader scientific context, the direct imaging technique will eventually allow us to greatly expand our understanding of all types of planetary properties. In addition to Earth-like planets, it would be used to study the orbital and physical properties of Jupiter-like exoplanets and debris disks. Such data will be crucial to the refinement and validation of planetary system models. * jpdou@niaot.ac.cn; phone ; fax ; niaot.ac.cn 2011 International Conference on Optical Instruments and Technology: Optical Systems and Modern Optoelectronic Instruments, edited by Y. Wang, Y. Sheng, H.-P. Shieh, K. Tatsuno, Proc. of SPIE Vol. 8197, 81970F 2011 SPIE CCC code: X/11/$18 doi: / Proc. of SPIE Vol F-1

2 The direct imaging of an Earth-like exoplanet remains extremely challenging due to the large flux ratio contrast that is 10-9 in the visible wavelength and its location very close to the primary star [1], which should be done in space. Furthermore, the diffraction pattern from the starlight will contaminate the exoplanet image. A high-contrast imaging provided by a coronagraph that is designed to remove or suppress the diffraction of the starlight is critical for such a direct detection. For NASA s Terrestrial Planet Finder Coronagraph (TPF-C), a contrast of at an angular distance better than 0.1 arcsec is required in the visible wavelength. Recently, a series of high-contrast coronagraphs have been proposed for the direct imaging of an Earth-like exoplanet, which can theoretically provide a contrast of at an inner working angular (IWA) distance of few λ/d (where λ/d is the diffraction beam width, λ and D are the wavelength and the aperture diameter, respectively) [2]. Because of the manufacture error of the coronagraph critical components and the wave-front error induced speckle noise that originates from imperfections of the optics, few coronagraphs can reach such a contrast in 10-10, even in the lab. Pupil apodizing that modifies the light transmission on the pupil is a promising technique for high-contrast imaging. Traditionally, pupil apodizing is optimized by the using of a continuous transmission pupil [3]. The achievable transmission precision for such continuous transmission pupils are on the order of 5% for commercial grade techniques. High precision transmission apodizing pupil can be achieved by the using of step-transmission filters, in which only a discrete number of transmission steps are needed and the transmission in each step is identical. This greatly improves the precision of the transmission of an apodizing pupil since the coating can be done on one step each time and the transmission in each step can be measured in the manufacture process until it reaches the required precision [4]. In recent years, we have developed several coronagraphs that deploy the step-transmission filters [5]. The contrast of the previous systems could reach ~10-6 at IWA of 4λ/D. And the out working angle (OWA) was only of 13~15λ/D, which is proportional to the step number along radius. The total clear aperture of these filters were relatively large, and the associated focal ratio of the system was relatively small, which will increase the optical phase error along the off axis optical path. To increase the OWA of the system, and reduce phase error along the off axis optical path, here we propose a coronagraph that is based on two 50-step transmission filters, with a relatively small aperture size. The transmission variation is realized by coating with a metallic film of variable thickness along radial directions. In the design of the filters, the phase error that caused by the film thickness difference is included, which has been discussed in our previous paper [6]. The target contrast of the coronagraph is set to be ~10-7 at an IWA of 2λ/D for each filter. The theoretical contrast of the coronagraph that deploys the two filters perpendicular to each other should reach 10-9 ~10-10 at 2λ/D. To demonstrate its feasibility and evaluate the performance of the coronagraph, we manufactured two 50-step transmission filters and the transmission error for each step on the filter should be controlled within ±3% of each step s nominal transmission. Based on these two filters, several tests have been repeatedly performed on the high-contrast imaging test-bed in the laboratory. The point spread function (PSF) images of the coronagraph that deploy two filters were taken and associated contrast was calculated which could reach 10-7 at 2λ/D. It is the first time for our step-transmission filter based coronagraph to achieve a high contrast at an IWA smaller than 4λ/D. One of the probable reasons for this is that we have included the thickness difference induced phase error in the design procedure to eliminate its influence on the small IWA regions. Another reason is that we have taken 1000 PSF images under each exposure time and each PSF image has been calibrated to eliminate the probable drifting. Finally, it is found that the main limitation for such a coronagraph is from the transmission error in the steps with relatively low transmissions. Higher transmission precisions in the low transmission steps may guarantee further contrast improvement. We will work closely with coating companies to future improve the transmission precisions in associated steps of the filters, which should deliver better performances. Proc. of SPIE Vol F-2

3 2. CORONAGRAPHIC SYSTEN DEVELOPMENT IN THE LABORATORY 2.1 Step-transmission filter design for an off-axis telescope The design of the step-transmission filters that is composed of a finite number of transmission steps with the same transmission in each step was firstly proposed by Ren & Zhu in 2007 [4]. Here we increase the step number along radius to be 50 to increase the OWA of the coronagraph system. Comparing with previous coronagraphs, it could reach a high contrast at a large detection region from 2λ/D to 50λ/D. Furthermore, we have employed a discrete optimization algorithm that has included the thickness difference induced phase error in each step in the design, which could further eliminate the influence of the optical aberrations from the filters themselves [6]. A numerical simulation is performed to demonstrate the theoretical performance of the coronagraph using one and two filters. Here we briefly present the theoretical design of the coronagraph and its associated result. For the coronagraph using one filter, the theoretical contrast should reach 10-7 at 2λ/D. Well, contrast could reach if the coronagraph is apodized along two directions with two filters perpendicular to each other. However, in practice, due to the thickness difference along radius that caused by the metallic coating film, which will induce phase error to the theoretical system, there will be a serious contrast degradation in the small IWA regions. Figure 1 shows the transmission amplitude pattern of the two 50-step filters and its associated PSF images. A contrast comparison is shown in Figure 2, where the real line represents the contrast when considering the thickness difference induced phase error in the design, the doted line represents the contrast with induced phase error. It is clearly shown that the thickness difference induced phase error will degrade the contrast performance in the IWA of 2~3λ/D regions. In practice, the optimization algorithm that has considered such a phase error should effectively eliminate the influence on the contrast. Table 1 summarizes the theoretical design of the two 50-step transmission filters. Figure 1. Top: Transmission amplitude pattern of the transmission filters; Bottom: Associated coronagraph PSFs. Proc. of SPIE Vol F-3

4 Figure 2. A theoretical contrast comparison with and without considering the thickness difference induced phase errors. Table 1. Theoretical design of the two 50-step transmission filters. Filter 1 Filter 2 Filter1+Filter2 IWA(λ/D) 2 OWA(λ/D) 50 Contrast Throughput (%) Manufacture of the two 50-step transmission filters To demonstrate the feasibility of the system and evaluate the performance of the coronagraph, we manufactured two 50-step transmission filters which were optimized at the wavelength of 632.8nm by the Reynard Corporation. These two step-transmission filters were based on the metallic coating technique and should deliver a good performance by careful selection of the coating material. For these filters, a metallic film of Inconel was deposited on one surface of a BK7 substrate with a diameter of 40mm and thickness of 5mm. The surface quality of the substrate was controlled within a P-V value of λ/4 (where λ=632.8nm). The filters have a clear aperture size of 25mm with totally 50-step Inconel film coated in such a clear aperture. The region out of the clear aperture (25~40mm) on the substrate is coated with a none-transmission metallic film to block unnecessary scattered lights induced to the system. The transmission of the so called none-transmission area is lower than 0.03% according to the current technique. On the opposite surface of the substrate, a single-layer of anti-reflection coating is applied to reduce possible multiple reflections. Figure 3 shows the two manufactured filters. The transmission requirement for these two filters is presented in Figure 4. Proc. of SPIE Vol F-4

5 Figure 3. The photograph of the two manufactured 50-step transmission filters. Figure 4. The transmission requirement for the two 50-step filters. Figure 5. The tested transmission profile provided by the manufacturer. Proc. of SPIE Vol F-5

6 For the filters, the transmission error in each step is required to be controlled within ±3% of the nominal transmission in corresponding steps. And the measurement data of the transmission was provided by the manufacturer when we received the two filters. However, it is found that the transmission error was actually controlled in an absolute value of ±3% rather than a relative of its nominal value in each step. Figure 5 shows the tested transmission profile that was provided by Reynard Corp, which did not meet the requirements. Later test shows that such a transmission error will have a serious influence on these low transmission steps, which will degrade the performance of the system, since the final performance of the coronagraph is mainly depended on the transmission precision on the low-transmission steps. We will work closely with Reynard Company to figure out this problem for future filters. 2.3 Configuration of the experimental optics Figure 6 shows the optical configuration of the test coronagraphic system in the laboratory. The step-transmission filters that serve as the pupil amplitude apodization plate are located just in the collimated beam of the optics system. One doublets that is designed and optimized for the coronagraph is used as the collimator. The starlight that has been modulated by the transmission filters will be focused by another doublet, and the associated PSF image will be formed on the coronagraph focal plane. To increase the focal ratio of the system to reduce the optical aberrations along the off axis optical path, here we use the replay optics including one lens and a flat mirror to slow the optical light. Point source Collimator Imager Focus Relay optics Filters CCD Figure 6. The optical configuration of the experimental coronagraph optics. All the coronagraph components (shown in Figure 6) were installed on a vibration isolated test-bed in a totally dark room in the lab. The point light source that simulated the star image is created by a μm He-Ne laser as the light source combined with a 5um pinhole and a 40x microscope objective lens. The laser light will be focused on the pinhole through the microscope objective lens. The input focal ratio of the system is set to be f/33.45, the clear aperture size of the entrance light is 25 mm. And the output focal ratio is set to be f/74. In that case, the optical light will be very slow and optical aberration along the off axis could be reduced to an expected extent. 3. LABORATORY TEST RESULTS AND DISCUSSIONS To evaluate the actual performance of the coronagraph that uses the two 50-step transmission filters, here we used a commercial 16-bit CCD camera (SBIG Company, USA) to take the PSF image on the focal plane of the whole system. The camera has 1530 by 1020 pixels at a 9 by 9 μm pixel size. Since one 16-bit CCD camera is not sufficient enough to achieve a high dynamic range measurement such as a contrast in the order of 10-7 in one single exposure, we employed a Proc. of SPIE Vol F-6

7 combination of PSFs taken under three exposure times of 0.12s, 12s and 1200 s, respectively. In each exposure, a partial of the contrast profile will be fitted to finally generate the contrast profile in the whole dynamic range. A set of neutral density (ND) filters was inserted between the laser light source and the microscope objective lens. Firstly, these ND filters can adjust the peak intensity in the first PSF image that is taken under 0.12s to be ~45,000, which is suitable for the fitting of the whole contrast profile. Secondly, for extreme long time exposures such as 1200s, it could be realized by reducing pieces of the inserted ND filters rather than increasing the exposure time, in which case the readout noise from the CCD could be reduced to a great extent. Here in this test we took 1000 PSF images under each exposure time. After taking each PSF image, an opacity mask was inserted to block the laser light to enter the system to get a reference background image. The reference image has the same exposure time with each PSF image and will be subtracted from the PSF image to eliminate the influence of both effect of the dark current of the CCD and the background light. Although we took this experiment at an environment temperature lower than 10 C, and all the tests were taken through a remote-control, there was still some turbulence which would cause the drift of the PSF images. To eliminate such an influence, we need to calibrate each PSF image carefully. For the PSF images under exposure time of 0.12s, there is only one peak intensity around 45000, which is not saturated and can be as the central position of each image for calibration. Then the 1000 images could be combined by co-centering each image to this peak intensity position and the associated averaged PSF was achieved. Well for the PSF images taken under exposure times 12s and 1200s, it became a little bit complicated since it was saturated in the central part of each PSF. To calibrate each image under long exposure times, we used the sub-peak intensity pattern on the horizontal beam of each PSF image and to align each PSF image according to these patterns. Finally the averaged PSF images could be achieved through the above approaches. Figure 7 shows the calibrated PSF images under three different exposure times. The strong bright vertical pattern in the two long exposured PSF images was caused by the CCD image bloom because the CCD does not have an anti-bloom function. Figure 7. The tested PSF images under different exposure times after calibration to eliminate the environment turbulence induced PSF drifting (in the red circle, the sub-peak intensity pattern on the horizontal beam is used for image calibration). Figure 8 shows the fitted contrast profile based on the three calibrated coronagraphic images under different exposure times. The coronagraph has delivered a contrast of 10-7 at an IWA of 2λ/D along the diagonal directions. In Figure 8, the doted line represents the contrast profile that is generated from the PSF images taken with no transmission filters Proc. of SPIE Vol F-7

8 apodiztion, which is used for the scaling purpose. Compared with the theoretical profile (shown in Figure 2), the tested PSF has a relatively large deviation. Such a deviation may be mainly caused by the filter transmission error especially in these low-transmission steps as well as possible residual wave-front error from the coronagraph optics, which can be improved by using higher-quality filter and inducing wave-front correction techniques. Figure 8. The test contrast profile of the coronagraphic image along the diagonal directions. 4. CONCLUSIONS In this paper, we propose a high-contrast imaging coronagraph with two 50-step transmission filters. The coronagraph is designed for an off axis telescope and optimized in the visible wavelengths. To demonstrate the feasibility and evaluate the performance of the coronagraph, we have manufactured two 50-step transmission filters. Based on the high-contrast imaging test-bed in the laboratory, we have performed series of test for the coronagraph that uses the two 50-step transmission filters. Finally, it has delivered a contrast of 10-7 at an IWA of 2λ/D, without the wave-front correction by using a deformable mirror (DM). It is the first time for our step-transmission filter based coronagraph to achieve a high contrast at an IWA smaller than 4λ/D. One of the probable reasons for this is that we have included the thickness difference induced phase error in the design procedure to eliminate its influence on the small IWA regions. Another reason is that we have taken 1000 PSF images under each exposure time and each PSF image has been calibrated to eliminate the probable drifting. As a follow-up effort, the transmission error in each step of the filter should be critically controlled under ±3% of its nominal transmission in corresponding steps, especially in the low-transmission steps, which should guarantee a better performance. We will work closely with the coating companies for future filters manufacture. A 12x12 actuator DM from Boston Micromachines will also be added on the pupil plane of the coronagraph to further correct the quasi-static phase errors of the coronagraph optics. The measurement and correction of the static phase errors of an optics system could be found in our recent papers [7-9]. Later results will be discussed in our future publications. Proc. of SPIE Vol F-8

9 ACKNOWLEDGEMENTS This work was funded by the Advanced Research of Space Science Missions and Payloads of the Space Science Strategic Pioneer Program-CAS and the National Natural Science Foundation of China (Grant Nos and ), as well as the National Astronomical Observatories Special Fund for Astronomy Part of the work described in this paper was carried out at California State University Northridge, with the support from the National Science Foundation under grant ATM 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, (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, ApJS 167, (2006). [3] Kasdin, N. J., Vanderbei, R. J., Spergel, D. N. and Littman, M. G., Extrasolar planet finding via optimal apodized-pupil and shaped-pupil coronagraphs, ApJ 582, (2003). [4] Ren, D. Q. and Zhu, Y. T., A coronagraph based on stepped-transmission filters, PASP 119, (2007). [5] Dou, J. P., Ren D. Q., and Zhu Y. T., High-contrast coronagraph for ground-based imaging of Jupiter-like planets, RAA 10 (2), (2010). [6] Ren, D. Q., Dou, J. P., and Zhu, Y. T., A transmission-filter coronagraph: design and test, PASP 122, (2010). [7] Dou, J. P., Ren D. Q., & Zhu Y. T., An iterative wavefront sensing algorithm for high-contrast imaging systems, RAA 11 (2), (2011). [8] Dong, B., Ren, D. Q. and Zhang, X., Stochastic parallel gradient descent based adaptive optics used for a high contrast imaging coronagraph, RAA 11 (8), (2011). [9] Dou, J. P., Ren D. Q., Zhu Y. T., and Zhang, X., Focal plane wave-front sensing algorithm for high-contrast imaging, Science in China Series G 52 (8), (2009). Proc. of SPIE Vol F-9