HASP SPARTAN V Optical Subsystem

Transcription

1 HASP SPARTAN V Optical Subsystem Anthony Cangelosi Chris Nie Colorado Space Grant Consortium Academic Advisor: Kendra Kilbride Anthony.Cangelosi@Colorado.edu March 29 th, 2010 Abstract With the high cost of launching telescopes into orbit and the similarly high cost of building large ground based telescopes with resolution limited by atmospheric conditions, the search for small transiting exoplanets has been a costly endeavor. With the limited time available on these large telescopes and the numerous observations required to observe a transiting planet, other methods must be sought to continue the search. HASP Spartan V will be flying a telescope/ccd payload on a weather balloon to the edge of space to attempt to perform defocused photometry on a magnitude 4 to 0 star. This payload is built around a small diameter, 75mm custom built folded refractor telescope coupled with a QSI 504ME CCD. The telescope will be calibrated to defocus the star s image over 10x10 pixels Defocusing the star over a larger area will allow more photons per image to be acquired thereby attaining an overall shorter total integration time required to achieve a signal to noise ratio of 10 5 :1. Methods to calibrate the optical system, choose correct exposure times and properly analyze the data need to be developed on the ground for an automated system with no direct human feedback during its flight. This paper will address the issues, challenges and solutions of the scientific objective of the HASP Spartan V payload. 1. Introduction HASP Spartan V s scientific mission of observing and imaging a star for a sufficient amount of time as to attain a sufficient signal to noise ratio (SNR) to definitively detect an exo-planet transit. As a planet transits its parent star along our line of sight, a small dip in the star s flux can be observed. Due to the ratio of the cross section of a star vs. an Earth sized planet a SNR of roughly 10 5 :1 is needed to confirm that a transit has occurred and not a statistically allowable variation in flux. Spartan V has several challenges facing the Optical Subsystem in order to meet this goal. The first of which is acquiring a telescope that can fit within the payload dimension requirements while providing a sufficient focal length and aperture. The second calibrating and focusing the telescope for operation at 120,000 ft. Other challenges involve choosing correct exposure times as to not saturate or under expose the pixels, selecting and modifying a CCD, controlling the thermal environment, finding a target star during flight and analyzing the data post flight. 2.1 Choosing a Telescope HASP Spartan V s dimension limitations require a relatively short telescope, roughly 10 to 11 inches in length including the CCD. It is also important to mention that an on board auto-focusing mechanism could not be fabricated due to the complexity and cost involved. Originally the Maksutov-Cassegrain telescope design was the primary candidate to meet these requirements. The design combined refractor and reflector telescope design to provide a high focal length in a relatively short length, mm long with a 1200 mm focal length and 90mm aperture. While this design was looking promising it came to our attention that the thermal 2010 COSGC Space Research Symposium Page 1

2 expansion and contraction that would be experienced during flight, only a few millimeters, but this would be amplified by about 25 times due to the magnification effect of the secondary mirror. While this change in the focal point could be corrected for, any discrepancy in the operating flight temperature from what the telescope was focused for would lead to a difference in the focal point significant enough to possibly render any images too far out of focus and unusable. Once the conclusion was reached that the Maksutov- Cassegrain design would not work we began looking at refracting telescopes due to their linear behavior with thermal expansion but unfortunately even significantly reducing the aperture size still left the overall length beyond the dimensions allowed. The solution to this was brought to our attention by Russ Mellon, President of Equinox Interscience in Golden, Colorado. Russ recommended a folded refractor design that used flat mirrors to bend the light beam from the primary lens in a zigzag pattern to reduce the overall length to about 1/3 of the original. It did however add height to the design, giving it the shape of a shoe box, but this fit into the dimension requirements. Below, in figures 1, 2, and 3 are the external telescope design, internal optical bench layout of the mirrors and lens, and the Spartan V current design as of March 29 th, Figure 2. Internal Telescope Design/Optical Bench, Credit: Chris Nie, HASP Spartan V Mission Specialist Team Figure 3. HASP Spartan V payload. Credit: Spartan V Structures Team (Jeff Byrne and Sreyas Krishnan) Figure 1. External Telescope Design. Credit: Chris Nie, HASP Spartan V Mission Specialist Team This custom built telescope will have an aperture of 75mm with a focal length of 800mm contained within an overall length of 266mm or 10.5 inches. Figure 2, contains the optical bench layout of the components being used to accomplish this. The primary lens is an achromatic 75mm diameter lens with a 400mm focal length. The mirrors are 75x75 and 40x40 mm TECHSPEC ¼ Wave First Surface Mirrors. In order to achieve a focal length of 800mm the Orion 2x Ultrascopic 3-element Barlow Lens is used. The CCD is the QSI 504ME, its details will be discussed in a later section. It should be noted that there is no filter being used in this system in order to maximize usage of the available light. The fabrication of parts and assembly will be performed with the assistance of Russ Mellon at his prototyping shop near Boulder, Colorado COSGC Space Research Symposium Page 2

3 2.2 Telescope Calibration and Exposure Time In order to calibrate and focus the telescope for a 10x10 defocus at 120,000 ft. without an active focusing system on board, extensive testing and characterization must be done on the ground prior to flight. The first calibration to be accomplished after assembly is to properly defocus the telescope to 10x10 pixels at room temperature. This will be accomplished by creating a dark room around the telescope which will consist of box with a pinhole LED at one end with the telescope aimed at it. The LED will then be imaged and using shims to move the CCD through or beyond the focal point until a 10x10 defocus is achieved. Using the room temperature defocus and the known wattage of the LED, exposures will be taken of the LED to extrapolate saturation times for various magnitudes of stars based on the known flux associated with magnitude and the flux from the LED. This data will also determine the amount of photons that can be actually acquired per image based on the light pattern which will then determine a precise number of exposures required to attain a signal to noise ratio of Preliminary calculations have been made for the exposure times and number of frames required based on an idealized situation of light being evenly spread across the entire 10 pixel diameter image where in reality, the image will be shaped more like a doughnut. These calculations can be found in appendix A. During this time dark exposures will be taken in groups of five with exposure times and temperatures equal to the time and CCD temperatures for the LED exposures. The median combined dark frame (which will include the bias) will be subtracted from the raw LED exposures. The final telescope-cdd calibration test will consist of focusing the telescope for operation at 120,000 ft and the associated thermal environment. This will be accomplished by using the room temperature defocus and varying the ambient temperature using dry ice to provide a wide variety of temperatures from below the ambient temperature at altitude to above it. Throughout the temperatures exposures will be taken in groups of 3 to 5 and mean combined to provide a plot of temperature vs. defocus amount. Because the defocusing is linear with distance from the focal point and the expansion of the aluminum is linear with temperature a proper location for the CCD using shims can be determined and implemented for the desired temperature. Once this has been accomplished, the temperatures will be again varied around the flight temperature during our observations to quantify the focus sensitivity near the flight temperature. 2.3 Choosing and Modifying a CCD The CCD camera is the most expensive and critical part of the optical subsystem, its specifications defines the telescope s requirements and is the sole instrument for taking images. The Quantum Scientific Instruments QSI-504ME CCD best fit the project s budget, size and scientific needs. Due to Spartan V s mission s relevance to future projects, Dr. Eliot Young PhD, is supplying the funds to cover the cost of the CCD. The one drawback however is the sealed argon gas chamber surrounding the CCD chip. HASP requires any pressure vessel to go through a comprehensive testing and certification process in order to be flown on the platform. Unfortunately no preexisting data exists for vacuum testing on the QSI 504ME, and after speaking with representatives from QSI, the chamber is not even designed for operation in a vacuum. When exposed to near vacuum conditions the seal on the gas refill port will vent the argon gas, but there has been no official testing of this and merely venting the gas is not sufficient to prove structural stability. The argon gas is present in order to prevent the accumulation of condensation on the window to the CCD. However, the conditions at altitude being sufficiently dry and cold can allow for the window to be removed completely, thereby eliminating the presence of a pressure vessel while not jeopardizing the CCD s capabilities. Therefore during the telescope fabrication process the CCD window on the camera will be removed, in doing so however, precautions must be made to avoid dust and moisture accumulation. This will be accomplished by covering the nose and using silica packs or a moisture prevention bag to store the camera. Prior to flight using compressed air, any accumulated dust should be removed. 2.4 Thermal Analysis of Charged-Coupled Device: The chosen CCD for this mission is the QSI 504 Scientific Cooled CCD Camera. It has an optimum operating temperature of -20 C and has a fan cooling system included. The problem with the cooling system is that it is based off of convection heating and cooling and at the altitude of our mission convection is nearly nonexistent because the platform will be above 99.5% of the atmosphere. Due to this fact we must rely on conduction as the main source of thermal control in our payload. This will be accomplished through a technique called 2010 COSGC Space Research Symposium Page 3

4 heat sinking. The heat from the CCD will be conducted into the contacted surfaces to a cooler material (the outer shell of the telescope). The equation for thermal conductance between a substance and a conductive wall is: Where is the heat flow in watts, k is the thermal conductivity in watts per meter degree Kelvin, A is the surface area being contacted in meters squared, L is the thickness of the contacted surface in meters, and is the temperature difference of the heat sinking surface and the source of the heat in degrees Celsius. With this equation we can assume that the heat flow is the amount of heat given off by the CCD during use (9 watts) and k and L are known as set constants for aluminum. We find the area of contact between the CCD and the telescope to be meters squared. If we assume that the source temperature is the temperature at which we want the CCD to be operating at (-20 C) and the surrounding temperature to be the ambient of the atmosphere (-40 C) then we can solve for the amount of heat conduction in watts. If this value is greater than nine, the amount of heat being produced by the CCD, then the CCD shall be properly cooled by the heat sink created in the telescope design. This value is found to be well above 9 W. This means that the limiting factor for cooling the CCD will be the critical heat capacity of the telescope housing that is conducting heat away from the CCD. Part of this heat will be radiated into the surrounding space, the specific amount can be found by using the equation for radiation heat transfer: Where q is the amount of heat in watts being dissipated away, sigma is the Stefan-Boltzmann constant ( ) in watts per meter squared per degree Kelvin to the fourth, epsilon is the emissivity value of the substance in question, A is the area that is exposed in meters squared, and T is the temperature of the surroundings in Kelvin. It was found that the telescope housing will radiate W while the CCD is in use. Also, the CCD will radiate a small amount (0.106 W) as well. This means that during the mission the heat sink may reach heat capacity. It has been determined, however, that the mission can be completed before this equilibrium of heat transfer is reached. 2.5 Acquiring a Target Star The Spartan V payload s ability to find a target star is limited by its relative pointing accuracy. Due to electromagnetic interference present in previous HASP flights, compasses have not been able to provide highly accurate pointing precession. Instead the accuracy is more to the order of North, Northeast, East, Southeast, etc. The elevation of the telescope and the telescopes rotation relative to the platform can and will be known to less than a degree. This combination of abilities will require the telescope to perform a Raster Scan of a region of sky, specifically the region of the galactic center where the star density will be the highest. During the scan the images will be binned to provide faster downloading and processing. When an image is found to contain a sufficient amount of counts the system will begin tracking the star and start imaging with an exposure time calculated based on the binned counts. 2.6 Post Flight Data Analysis The data from the flight will consist of at least 2,000 images of what is planned to be the same star. Due to our tracking method, it is unlikely that the star will be located in the exact same position in every image. It is also unlikely that the temperature conditions will be the same in each image. Therefore the best method to analyze the images will be on an individual basis and combine the final results. The most useful tools for accomplishing this will either be IRAF or IDL performing matrix operations on the arrays of pixel values (counts or ADU). 1. Determine the CCD temperature associated with each image 2. Take the temperature and convert it to a dark current value 3. Subtract the dark current from each pixel in the image 4. Determine the average local background per pixel around the target star 5. Subtract this value per pixel from every pixel in the image. 6. Determine the remaining number of counts from the target star (not per pixel) and take CAREFUL note of the actual area of the star as it will be needed for SNR calculations. 7. Use the area used to calculate the target star s counts and attain the total amount of dark current and background within that area COSGC Space Research Symposium Page 4

5 8. Repeat for each image 9. Sum results and calculate SNR and total star counts The first, second, and third steps will be to remove the dark current/noise. First you will need to associate a CCD temperature with each image; this should be included in an information stamp embedded within each image. The temperature of the CCD will have an associated dark current based on previous calibration and testing results, so convert this temperature to a dark current per pixel value. If an exact match for this value doesn t exist then round up to the nearest higher temperature and subtract the read noise. Make sure to record this value per pixel for SNR calculations and then subtract it from each pixel in the image. It is unlikely that a flat field will be able to be taken during flight and after landing any flat fields will prove to be useless due to contaminate introduced post imaging therefore the normal flat field normalization will be skipped. The fourth and fifth steps will be removing the background. First determine the average background counts per pixel around the star and then subtract this value per pixel from every pixel in the image. We can subtract the localized background from the entire image because we are only interested in the target star. Record this average background value per pixel for later SNR calculations. The sixth step will be to measure the amount of counts from the target star. In IRAF a method using a circular aperture with specified radius can be used to sum its contents. Record the area used to sum the star counts for SNR calculations. The seventh step will be to calculate the total background and dark current present in the area used to sum the star counts. Although they have been subtracted these values play a role in the SNR calculations. Record the total background and dark noise counts for the star area separately. Repeat this process for every image and when complete there should be a table resembling figure 4: Image 1: Image 2000: Star Counts: [#] Figure 4. These columns can be plotted to see variability in the values and summed to attain the total star counts and each column sum (except area) can then be used in the following equation to calculate the signal to noise ratio. 3. References Star Area: [pixels] Background counts: [#] Dark current: [#] [1] Zombeck, Handbook of Space Astronomy and Astrophysics, Cambridge University Press 2010 COSGC Space Research Symposium Page 5

6 Appendix A: Flux Calculations: 75mm units Flux per Unit area of a 0 Magnitude Star: Reference[1] E-08 [erg/(cm^2 *sec * Ang)] Energy Per Photon (5500 A) E-12 [erg]/photon Converting to Photons with average wavelength of 5500 angstroms [photons/(cm^2 *sec *Ang)] Across a 2000 Angstrom band E+06 [photons/(cm^2 *sec)] Aperture Light gathering power E+08 [photons/sec] 10% loss due to optics E+08 [photons/sec] 70% quantum efficiency of CCD E+08 [photons/sec] Exposure Time Calculations: QSI 504 at 70% QE, 100k well depth Total Time (with 1 sec readout) 10x10 defocus 5x5 defocus 10x10 5x5 Mag Mag Mag Mag Mag Number of frames we will need to get 10^5 SNR 10x10 defocus 5x5 defocus 100k Well Depth (80k used) Estimate mm Ideal light spread will be more like a doughnut Plate Scale ArcSeconds/mm Pixel Scale 2.32 ArcSeconds/pixel Field of View ArcSeconds 2010 COSGC Space Research Symposium Page 6