THE CALIBRATION OF THE OPTICAL IMAGER FOR THE HOKU KEA TELESCOPE. Jamie L. H. Scharf Physics & Astronomy, University of Hawai i at Hilo Hilo, HI 96720

Transcription

1 THE CALIBRATION OF THE OPTICAL IMAGER FOR THE HOKU KEA TELESCOPE Jamie L. H. Scharf Physics & Astronomy, University of Hawai i at Hilo Hilo, HI ABSTRACT I have been calibrating the science CCD camera on the Hōkū Ke`a telescope, and to create an exposure time calculator for the use of future astronomers and students to accurately plan research projects. Our new Hōkū Ke`a telescope will revolutionize the teaching of physics and astronomy at UH Hilo, and will be instrumental for teaching physics and astronomy Majors (and Minors) hands-on observational astronomy, stellar and extragalactic astrophysics, and telescope instrumentation. These activities directly answer the goals of NASA to inspire, encourage, and develop the next generation of our nation s astronomers and space scientists. The ability to directly observe, and collect data for oneself, on a real, professional-grade telescope can only serve to inspire students to explore the Universe as dry textbooks never could. INTRODUCTION The basis of the project is to calibrate the Apogee Alta-U230 CCD (charged-coupled device) that was purchased on National Science Foundation grant [PI: David James] for use on the Hōkū Ke`a 0.9 meter telescope that is operated by University of Hawaii at Hilo (UHH). Accompanying this CCD in our project is an Apogee AFW50-7S filter wheel and an Apogee U47 Digital Imaging System (a smaller CCD), and a MegaMOAG off-axis guider. These will be assembled on a stainless steel interface plate that will be subsequently mated to the telescope. HOW A CCD OPERATES What is referred to as a CCD (charged-coupled device) is actually comprised of the CCD chip and its housing unit. The CCD is square silicon semiconductor. This is divided into pixels (i.e. potential wells storing photo-electrons), in the case of our Alta-U230, 2048 pixels by 2048 pixels. When a photon from a light source hits a pixel, it liberates an electron. The electrons are trapped in the pixel well until the charge is swept to the readout register, where the charges are counted and then transmitted via voltage to the output register (James 2010). The purpose of calibrating a CCD is to: Establish what the background level of electrons is (the so-called bias level) and what the average accumulation of thermal electrons -- generated by the thermal motion of electrons in the CCD -- is over time (the so-called dark current ). Doing this calibration process, we will also detect cosmic rays from our images. The purpose of cooling a CCD to as low a temperature as possible (say, -30ºC in our type of CCD) is to reduce the thermal noise, or dark current, to its lowest possible level. Higher

2 operating temperatures increase the dark current in our images, which makes them noisier and can act to drown out the detection of fainter stars. Cosmic rays are high energy particles (mostly protons) that bombard us constantly from outer space. They are unnoticeable on the macroscopic level but when they hit the CCD chip, they can liberate collisional-electrons, and often look like tiny-stars on the CCD. Many of our calibrations require us to know what the CCD chip looks like without the spikes left by cosmic rays, and we can achieve this by median -ing multiple calibration images (which acts to reject outlying data points). DESCRIPTION OF INSTRUMENTS Alta-U230 CCD: 2048 x 2048 pixels, triple air cooled camera head, E2V 230-BI MB CCD, USB2 interface. AFW50-7S filter wheel: 7-position, takes 50mm square filters, mechanical position system, USB2 interface. Apogee U47 Digital Imaging System: 1024 x 1024 pixels, single cooled, E2V CCD47-10 CCD, USB2 interface. Astrodon Photometrics Sloan filter set: 50mm square filters, consisting of u, g, r, i, z transmission filters. Kron-Johnson filter set: 50mm square filters, consisting of U, B, V, R, I transmission filters. Figure 1: Components of the instrument assembly. Clockwise from bottom left: blue U47 CCD, blue Alta-U230, black filter wheel housing. Figure 2: A closer look at the filter wheel. In this image, I was installing a green Johnson V transmission filter, which is 50 mm square. The V-filter has with a central wavelength of 5500 Angstroms, and a transmission bandpass of +/- 450 Angstroms. CALIBRATING THE CCD We need to calibrate and measure the bias of the CCD. A bias is an artificially induced electronic charge offset, specified by the manufacturer, to ensure that the CCD is capable of

3 recording very low photon input (the CCD controller floods the pixels with a few hundred up to a thousand counts or so, in order that the readout can process pixels having very small source counts in). Our procedure went as follows: 1. Set camera to cool to the desired temperature; 2. Take five bias frames a bias frame being a 0-second exposure -- the shutter does not open; 3. Take five dark frames, each lasting 60 seconds, with the shutter closed; 4. Go back to step 1, and repeat the process over a range of temperatures. After we took data for temperatures ranging from -35 to -10, with a 5 degree difference each time, we measure the charge on the bias frames, and, taking the median of each corresponding pixel from the different images, combined them into an image named Master Bias. Please see figure 3 for an example of a bias frame. Figure 3: One of our Bias images. A bias is a zero-second exposure, with the shutter remaining closed. The bias level can be measured from this frame. Figure 4: One of our 60 second exposure dark images, where the shutter remains closed for the entire exposure time. Please note that the small green rectangle in the image in the upper right corner is the area being displayed in the larger image. The white speckles are cosmic ray hits. We then did the same to each set of dark current frames, except that we subtracted off the masterbias frame from each dark frame (to subtract off the controller bias level). We then median-combined these bias-subtracted dark frames (to try to eliminate cosmic rays from our master dark frame). To measure the dark current, we used the following formula: master dark counts/time in seconds = dark current/pixel/sec Please see Figure 4 for an example of the dark frame, and please see Table 1 for our data summary.

4 RESULTS OF BIAS AND DARK CURRENT CALIBRATION Temperature Readout Master Master Noise Bias Dark Celsius Cts +/- 1 sig Mean StdDev Median Mean StdDev Median Table 1: Data from our comparative analysis of the master bias and master dark images for each temperature. For each cooling temperature, the data represent a series of five (5) bias and dark frames (60-second exposures) at each temperature. All numbers are CCD counts. READOUT Noise represents mean of the standard deviation on the individual bias frames, and represents the precision with how well we can readout data from an individual pixel. Master Bias = median of 5 individual bias frames. Master Dark = median of 5 dark frames, after each was Master Bias subtracted. As anticipated, we find that the dark current rises with increasing temperature. This is what we would expect, as the higher the temperature, the easier it is for electrons to move, increasing the noise we see on the CCD chip. As presented in Figure 5 below, we see that this curve is either exponential or polynomial based, with the polynomial presenting a better R 2 value, indicating it is a better fit to the data. Physically, this makes more sense, as the thermal noise component is not an exponential contributing factor to the detector being able to detect signal on the CCD. Figure 5: Temperature vs. Master Dark Median, showing best-fit lines for polynomial and exponential equations. Polynomial: y = x x R² = Exponential: y = e x R² = Figure 6: Temperature vs. Master Dark standard deviation, showing best-fit lines for an exponential equation. y = e 0.09x R² =

5 In Figure 6, we see that the standard deviation of the Master Dark plotted against temperature result in a nearly perfect exponential curve. It appears that, as the temperature rises, the variation in the dark current also varies widely, which is also to be expected, that the behavior of electrons becomes more unstable at higher temperatures. It should be noted that our CCD showed an increasing level of bias with decreasing temperature. The normal expectation is that the bias level remains constant, irregardless of temperature. We are unsure at this time why our results are different. COLLIMATION OF THE TELESCOPE OPTICS The process of collimating the telescope optics involves co-aligning the primary and secondary mirrors, and has been quite troublesome for us. The first laser collimator we had could not be Therefore, the optics would essentially have to be re-collimated every time someone needed to use the telescope. While another higher quality laser collimator was ordered, we still were able to take some images with the CCD and were able to focus the telescope to a limited degree (please see figures 7 and 8). Once our new laser collimator finally arrived, we were able to finish collimating the telescope. Figures 7 (left) & 8 (right): First Hōkū Ke`a images taken with the science CCD. The images are of the same star field, with the image on the right having been manually focused by adjusting the telescope. FINDING THE FOCAL LENGTH OF THE TELESCOPE It is important that we know where the focal point of the telescope is so that we may optimally set our CCD at that point. To this end, we set out to measure the focal length of the telescope.

6 Figure 9: The setup for calculating the focal point of the telescope. The light beam hits the mirror that is positioned at 45º relative to the base of the telescope, reflecting the light out so that the focus point will be seen on the side of the telescope rather than directly below it. Due to an infrared camera that had been installed on our telescope by JAXA (Japan Aerospace Exploration Agency) and Dr. David James last semester, we had to insert a mirror (see Figure 9) at a 45º in order to redirect the light beam so that we could find where the focal length was, using a piece of transparent plastic. In this case, we needed to measure the distance from the base of the telescope to the point of the mirror where the light from our focus object (Jupiter) struck it. From that point we measured to where Jupiter focused on our plastic sheet. The focal length came out to be 14.5 inches from the base of the telescope, so the CCD should be installed at that point. We will need to account for the thickness of the filter wheel and the offaxis guider that will direct light to the guide CCD. CALCULATING THE AMOUNT OF TILT An important component of attaching the CCD assembly to the telescope is knowing how much the CCD can be tilted relative to the incoming light beam while still retaining a focused picture across the CCD chip. Please refer to Diagram 2 for an illustration. Figure 10: The position of the CCD chip, the focal plane, t, and x. As Figure 10 shows, the CCD chip is tilted at an angle relative to the focal plane. There is a light beam shown coming in and striking the CCD chip at some distance t from the plane. Due to the tilt of the CCD chip, the light beam does not strike the chip evenly. We need to keep this unevenness restricted so that x, shown as the horizontal plane where the light beam strikes the chip, is restricted to 2 pixels, or 30 μm. Our starting equation (Heacox 2011) is: (Equation 1) = 10

7 Knowing this, we can easily solve for t, so (Equation 2) t = 10* x So that if x = 30 μm, t = 300 μm. As we see in Figure 11, the angle θ on the triangle of which t and x form the two sides of is the same angle formed between the focal plane and the CCD chip, so, that if we can find the θ for t and CCD chip (which is equal to 30,000 μm, or, we will have the θ and so be able to calculate what the amount of tilt the entire CCD must be in order to keep the CCD chip within the margin that will allow the entire chip to be in focus at one time. In that case, we know that (Equation 3) tanθ = In which case, tanθ = 30 μm/30,000 μm, or 1/100, giving us a θ = 0.573º. The top of the CCD measures 8 inches, or cm. Knowing our maximum angle allowance, we know that tan(0.573) = opp/20.32cm, or that opp = tan(0.573)*20.32cm, giving us a maximum tilt of cm, or mm. This means that our entire CCD case must have no more than a 2mm drop from one side to the other relative to the focal plane, which is perpendicular to the light coming in from the telescope. Figure 11: Showing the relative angles of t, x, focal plane, and CCD chip. CALIBRATING THE U47 With the Apogee U47, we took 100 bias frames and took the average of them to make a master bias. We then visually examined the master bias to see if there was evidence of structure, which there appeared to be (Fig. ). We then took a Fourier Transform using ImageJ software to see if there were abnormalities. The FFT indicated that there was, but when a line graph was taken across the master bias, it indicated that the counts varied from , which is a low variance and indicates a fairly clean chip. Please refer to figure 12 for an image of the U47 master bias.

8 CONCLUSION Although the U230 had to be shipped back to the manufacturer on account of software malfunction, we were able to do a great deal of work, both with the telescope and the CCDs. Future work to be done include making the steel interface for the CCDs and filter wheel and mating them to the telescope, as well as actual testing on the telescope. Figure 12: Master bias of the Apogee U47 ACKNOWLEDGEMENTS I would like to thank Dr. Bill Heacox for being my mentor for the Spring semester, Dr. Josh Walawender for helping me with the CCD, Dr. David James for being my mentor during the Fall semester, Dr. Ken Hon for helping me with the paperwork for this grant, and fellow Briana Hurley for helping me do calibration work on the telescope. Heacox, Bill. Personal interview. Feb James, David J. Personal interview. Aug REFERENCES