This dissertation is available at Iowa Research Online:

Transcription

1 University of Iowa Iowa Research Online Theses and Dissertations Spring 2017 Radio wavelength studies of the Galactic Center source N3, spectroscopic instrumentation for robotic telescope systems, and developing active learning activities for astronomy laboratory courses Dominic Alesio Ludovici University of Iowa Copyright 2017 Dominic Alesio Ludovici This dissertation is available at Iowa Research Online: Recommended Citation Ludovici, Dominic Alesio. "Radio wavelength studies of the Galactic Center source N3, spectroscopic instrumentation for robotic telescope systems, and developing active learning activities for astronomy laboratory courses." PhD (Doctor of Philosophy) thesis, University of Iowa, Follow this and additional works at: Part of the Physics Commons

2 RADIO WAVELENGTH STUDIES OF THE GALACTIC CENTER SOURCE N3, SPECTROSCOPIC INSTRUMENTATION FOR ROBOTIC TELESCOPE SYSTEMS, AND DEVELOPING ACTIVE LEARNING ACTIVITIES FOR ASTRONOMY LABORATORY COURSES by Dominic Alesio Ludovici A thesis submitted in partial fulfillment of the requirements for the Doctor of Philosophy degree in Physics in the Graduate College of The University of Iowa May 2017 Thesis Supervisor: Associate Professor Cornelia Lang

3 Copyright by DOMINIC ALESIO LUDOVICI 2017 All Rights Reserved

4 Graduate College The University of Iowa Iowa City, Iowa CERTIFICATE OF APPROVAL PH.D. THESIS This is to certify that the Ph.D. thesis of Dominic Alesio Ludovici has been approved by the Examining Committee for the thesis requirement for the Doctor of Philosophy degree in Physics at the May 2017 graduation. Thesis Committee: Cornelia Lang, Thesis Supervisor Robert Mutel Kenneth Gayley Hai Fu Renée Cole

5 To all of my family, friends, teachers, and research advisors who have provided support along my journey. Without your support I would not be where I am today. ii

6 ACKNOWLEDGMENTS I would like to thank my advisor, Cornelia Lang, for her guidance throughout both my teaching and research careers. Without her, I would not have discovered my passion for teaching and active learning. I would also like to thank Robert Mutel for his support and advice through out my instrumentation work for the Van Allen Observatory and Iowa Robotic Observatory. The Iowa Robotic Observatory would not be the excellent instrument that it is today without Mark and Pat Trueblood of the Winer Observatory, and I thank them for their hospitality and assistance while I worked to install the instruments I developed. I would also like to include the acknowledgements from the published and submitted papers included in this document. For Ludovici et al. (2016): We wish to thank the staff of the Karl G. Jansky Very Large Array. We also thank Philip Kaaret for his advice in examining the Chandra X-ray data and discussing the possibility of a micro-quasar as the origin of N3. And for Ludovici and Mutel (2017; AJP submitted): Funding for this research was partially funded by NSF grant , the Carver Trust grant , and the University of Iowa College of Liberal Arts. iii

7 ABSTRACT The mysterious radio source N3 appears to be located within the vicinity of the Radio Arc region of the Galactic Center. To investigate the nature of this source, we have conducted radio observations with the VLA and the VLBA. Continuum observations between 2 and 50 GHz reveal that N3 is an extremely compact and bright source with a non-thermal spectrum. Molecular line observations with the VLA reveal a compact molecular cloud adjacent to N3 in projection. The properties of this cloud are consistent with other galactic center clouds. We are able to rule out several hypotheses for the nature of N3, though a micro-blazar origin cannot be ruled out. Robotic Telescope systems are now seeing widespread deployment as both teaching and research instruments. While these systems have traditionally been able to produce high quality images, these systems have lacked the capability to conduct spectroscopic observations. To enable spectroscopic observations on the Iowa Robotic Observatory, we have developed a low cost ( $500), low resolution (R 300) spectrometer which mounts inside a modified filter wheel and a moderate cost ( $5000), medium resolution (R 8000) fiber-fed spectrometer. Software has been developed to operate both instruments robotically and calibration pipelines are being developed to automate calibration of the data. The University of Iowa offers several introductory astronomy laboratory courses taken by many hundreds of students each semester. To improve student learning in these laboratory courses, we have worked to integrate active learning into laboratory activities. We present the pedagogical approaches used to develop and update the laboratory activities and present an inventory of the current laboratory exercises. Using the inventory, we make observations of the strengths and weaknesses of the current exercises and provide suggestions for future refinement of the astronomy laboratory curriculum. iv

8 PUBLIC ABSTRACT The center of our Galaxy is home to many intruguing objects. One of these objects, called N3, lies ontop of the bright, non-thermal filaments in a region known as the Radio Arc. To investigate the nature of N3, we have conducted radio observations using some of the worlds most powerful radio telescopes. Broad-band observations of the region reveal that N3 is an extremely compact and bright source with a non-thermal spectrum. Narrow-band molecular line observations reveal a compact molecular cloud adjacent to N3. We show that the properties of this cloud are consistent with other galactic center clouds. We are able to rule out several hypotheses for the nature of N3, though a micro-blazar origin cannot be ruled out. Robotic Telescope systems are now seeing widespread deployment as both teaching and research instruments. While these systems have traditionally been able to produce high quality images, these systems have lacked the capability to conduct spectroscopic observations. To enable spectroscopic observations on the Iowa Robotic Observatory, we have developed a low cost, high sensitivity spectrometer which mounts inside a modified filter wheel as well as a moderate cost, medium resolution spectrometer which recieves light from a fiber optic cable. Software has been developed to operate both instruments robotically and calibration pipelines are being developed to automate calibration of the data. The University of Iowa offers several introductory astronomy laboratory courses taken by many hundreds of students each semester. To improve student learning in these laboratory courses, we have worked to design activities wich actively enguage and challenge students. We present the pedagogical approaches used to develop and update the laboratory activities and present an inventory of the current laboratory exercises. Using the inventory, we make observations of the strengths and weaknesses of the current exercises and provide sauggestions for future refinement of the astronomy laboratory curriculum. v

9 TABLE OF CONTENTS LIST OF TABLES ix LIST OF FIGURES x CHAPTER 1 Outline of This Thesis I Radio Wavelength Studies of the Galactic Center: an Extreme Galactic Environment Introduction The Galactic Center Central Molecular Zone Goals of this Work The Unusual Galactic Center Radio Source N Abstract Introduction Observations and Data Calibration Low Frequency Observing and Calibration (2 12 GHz) High Frequency Observing and Calibration (30 49 GHz) Results Properties of N Flux Density and Spectrum of N The Environment around N3: the Radio Arc NTFs and Wake The Environment around N3: Molecular Gas Discussion Nature of Continuum Emission from N Relation between N3, the N3 molecular cloud, NTFs, and the Wake What is the nature of N3? Conclusions Current and Future Work Very Long Baseline Array Observations of N VLBA Observations Data Reduction and Results VLA GC Point Source Survey Proposed Observations vi

10 II Spectroscopic Instrumentation for Robotic Telescope Systems Introduction Spectrometer Designs Grating Spectrometers Echelle Spectrometers Robotic Spectrograph Control Goals of This Work A Compact Grism Spectrometer for Small Optical Telescopes Abstract Introduction The Compact Grism Spectrometer Optical Design Wavelength and flux calibration Sample CGS spectra and lab projects Limitations Adapting the CGS to other observatories Conducting Spectroscopic Observations with Robotic Telescope Systems Introduction Optomechanical Design Compact Grism Spectrometer Fiber-Fed Echelle spectrometer Robotic Spectrometer Observations Robotic CGS Observations Robotic Fiber-fed Echelle Observations Spectrometer Calibration CGS Calibration Echelle Calibration Conclusions and Future Work Echelle Cross-Disperser Upgrade Fiber Feed Stability Upgrade Automated Echelle Calibration III Improving Astronomy Education Using Active Learning and Student Observatories Introduction Fostering Active Learning at the University of Iowa Goals of this Work vii

11 10 Active Learning Astronomy Laboratories at the University of Iowa Developing a Collaborative Learning Environment Encouraging an Understanding of Concepts Solidifying Concepts through Hands-on Observing Astronomy Laboratory Inventory Laboratory Activities Examined Groupings of Laboratory Activities Laboratory Equipment Teaching Techniques Primary Focus Inquiry Level Inventory Evaluation Laboratory Curriculum Strengths Identified Unmet Needs Challenges Suggestions for Improvement Conclusions APPENDIX A Properties of an Echelle Grating A.1 Dispersion of an Echelle Grating A.2 Central Wavelength of an Echelle Order B Original Angles and Paralax Laboratory Activity C New Angular Size Laboratory Activity BIBLIOGRAPHY viii

12 LIST OF TABLES 3.1 Observed Fields Spectral Line Parameters GHz CH 3 OH Masers GHz CH 3 OH Masers Comparison of original and updated laboratory formats Inventory of equipment components used in the astronomy laboratory activities. Abbreviations for equipment components are as follows: class demonstrations (CD), group experiments (GE), dark sky observing (DSO), live CCD imaging (LI), and robotic observing (RO). Components marked with a L are contained in the laboratory activity while components marked with a T are contained only in the TA guide for the activity. Activities with blank rows have no equipment components Inventory of teaching techniques used in the astronomy laboratory activities. Abbreviations for teaching techniques are as follows: brainstorming activities (BA), misconception checks (MC), Hypothesis Formation (HF), On-line research (OR), class discussions (CD), and Reporting Results (RR). Components marked with a L are contained in the laboratory activity while components marked with a T are contained only in the TA guide for the activity Inventory of the primary focus emphasized within the astronomy laboratory activities. Abbreviations for learning objectives are as follows: night sky (NS), telescope observation (TO), digital data analysis (DDA), the changing universe (CU), the nature of light (NL), the structure of the universe (SU), size scales (SS), and gravitation and orbits (GO) Table 3 from Fay et al. (2007) Inquiry level of each activity according to the (Fay et al., 2007) rubric ix

13 LIST OF FIGURES 2.1 Right: Image of the GC showing the CMZ. Purple represents 20 cm radio emission, which traces non-thermal emission as well as thermal H II regions. 1.1 mm emission is shown in orange and traces dust and associated molecular gas. Cyan shows Spitzer IRAC data, and highlights emission from stars and polycyclic aromatic hydrocarbons. (Image courtesy of NRAO/AUI.) Left: Figure 1 from Ludovici et al. (2016) showing a 5 GHz continuum image of the Radio Arc region, showing the NTFs, Sickle and Pistol H II regions, N3, and the wake GHz continuum image of the Radio Arc region, showing the NTFs, Sickle and Pistol H II regions, N3, and the wake Continuum images of N3 from 49 to 4.5 GHz. All images are on the same scale and centered on the position of N3. The synthesized beam for each image is as follows: 49 GHz: 0.26 x 0.11, 44 GHz: 0.3 x 0.10, 36 GHz: 0.43 x 0.13, 30 GHz: 0.52 x 0.15, 11.5 GHz: 1.01 x 0.38, 10.5 GHz: 1.10 x 0.43, 5.5 GHz: 1.48 x 0.62, 4.5 GHz: 1.80 x The synthesized beam is shown in the bottom left of each image as a black outline Continuum spectrum of N3 from 2 GHz to 49 GHz. The solid line represents a broken power law fit to the data. At low frequencies (2-6 GHz) the spectral index rises (α +0.5) while at high frequencies, (10-36 GHz) the spectrum is falling (α 0.8) Left: 5 GHz image of region surrounding N3. Individual filaments of the NTFs are labeled. Right: Brightness profile of the NTFs as extracted from the black rectangle in the left figure. The main filaments located near N3 are labeled using the convention of Yusef-Zadeh & Morris (1987) Comparison of our 5 GHz continuum observations with HST Pa-α observations (Wang et al., 2010). The wake is obvious in both the Pa-α observations and the 5 GHz continuum, indicating that the source is thermal in nature. Important features of the wake are indicated. The position of N3 is marked with the cross x

14 3.6 Integrated intensity maps of the molecular transitions observed. Contours are at 3, 6, 10, 20, 30, 40, and 50 times the rms noise in each image (CH 3 CN: 46 mjy beam 1 km s 1, CH 3 OH: 153 mjy beam 1 km s 1, CS: 188 mjy beam 1 km s 1, HC 3 N: 98 mjy beam 1 km s 1, HCNO: 42 mjy beam 1 km s 1, SiO: 94 mjy beam 1 km s 1, SO: 100 mjy beam 1 km s 1, NH 3 (3,3): 18 mjy beam 1 km s 1, NH 3 (6,6): 19 mjy beam 1 km s 1 ). The position of N3 is marked with a cross Left: Velocity map of NH 3 (3,3) emission in the N3 molecular cloud. The contours are 5, 9, 13, 17, 21, and 25 times the RMS surface brightness (40 mjy beam 1 km s 1 ) of the integrated NH 3 (3,3) image. N3 is marked by the cross, while three additional positions in the cloud are marked with dots. Right: Spectra of the NH 3 (3,3) transition observed in the N3 molecular cloud. The profiles in this figure correspond to the positions labeled in the left-hand figure GHz continuum image (greyscale) overlaid white contours showing NH 3 (3,3) emission integrated between velocities of -25 and 100 km s 1. Contour values are 5, 9, 13, 17, 21, and 25 times the RMS noise value of 40 mjy beam 1 km s 1. The dashed white lines are the outlines of the arcs shown in Figure 3.5. The cross denotes the position of N Finding charts for detected 36 GHz (Left) and 44 GHz (Right) CH 3 OH masers in the N3 cloud. Greyscale shows maximum intensity maps of each CH 3 OH transition while the green contours show the thermal CH 3 OH ( ) transition at 3, 6, and 10 times the rms (153 mjy beam 1 km s 1 ) of the integrated intensity map. N3 is marked by the black cross. The inset shows the 36 GHz continuum image of N3 (grey scale) with 11 km s 1 CH 3 OH ( ) contours overlain at 4 and 8 time the channel RMS (RMS = 4.5 mjy/beam, channel width = 1 km s 1 ) Spectra for each detected 36 GHz CH 3 OH ( ) Maser. The dashed line marks the velocity for the detected maser Spectra for each detected 44 GHz CH 3 OH ( ) Maser. The dashed line marks the velocity for the detected maser Schematic drawings of the spectrum produced by a simple grating (Left) and a grism (Right). Curved line represents the curved focal plane of the spectrometer. Colors represent red, green, and blue light xi

15 6.1 Optical design of the compact grism spectrometer, showing ray paths at several wavelengths from 400 nm (magenta) to 800 nm (brown). The incident rays from the telescope (f/6.8) are first collimated by an achromatic doublet lens, then dispersed by a 10 wedge prism, followed by a 600 lpmm transmission grating, then a second 10 prism, and finally refocussed into the detector plane by a second achromatic doublet. The focal lengths of the collimating and refocusing lenses are determined by the f-ratio, location in the converging optical beam, and the back-spacing of the imaging sensor [top] Compact grism system, with 25 mm optical components, enclosed in a 3-d printed housing. [bottom] Compact grism installed in a 50mm diameter slot in a filter wheel. Note that the filter wheel housing has a 3-d printed wall extension with height 37 mm Cross-sectional view of a filter wheel with the grism installed Dispersed grism spectrum zero-order with stellar image at left. The image is focussed at the center of the dispersed spectrum so that the stellar image is strongly defocussed (a) Uncalibrated spectrum of the B3V star HD (b) Calibrated spectrum, (c) Gain curve applied to raw fluxes. (d) Calibrated spectrum of HD from Jacoby et al. (1984) Grism spectra of stars from 380 nm to 750 nm, illustrating the spectral sequence from hot to cool stars: (a) B3V:19,000 K (b) A1V: 9,000 K, (c) F7V: 6,240 K, and (d) M5III: 3,400 K CGS spectrum of the emission-line star HD76868 (B5e). Note the prominent chromospheric Hα emission line, which arises from circumstellar material ejected by the rapid rotation of the star. The circumstellar gas is optically thick in the Hα line, but with increasing frequency it becomes optically thin and the higher level Balmer lines are seen in absorption. The Hβ line has both emission and absorption components. The width of the emission component is λ 0.9 nm, indicating a spectral resolution R CGS spectrum of the extremely luminous Wolf-Rayet star WR7 (HD56925) at the center of the emission nebula NGC2359. This star has an extended hot wind responsible for the broad emission lines of helium, carbon, and nitrogen xii

16 6.9 CGS spectra of two low-redshift quasars, each with prominent red-shifted Balmer emission lines, as well as a forbidden oxygen line. (a) 1E ,V = 14.4, z = 0.096, (b) 3C273, V = 12.9, z = Each exposure was 15 min Graphical user interface to the CGS calibration and plotting program designed for ease of student use Photograph of the filter wheel used on IRO showing the CGS in the center and the pick-off mirror for the fiber feed on the right Engineering CAD model of the fiber feed used on IRO. Light enters though the filter wheel adapter (grey) and light from the science target is incident on the optical fiber (orange). Light not falling on the fiber is reflected by the mirror (green), focused by the lens (red) and is collected by the camera (blue) Engineering drawing of the echelle spectrometer assembly. All distances are measured in mm. Light travels from the fiber feeds (yellow) to the collimating mirror (cyan), with the calibration fiber reflecting off of the motorized flip mirror (magenta) when in use. After collimation, light is dispersed out of the page by the echelle grating (orange) then cross dispersed in plane by the prism (red). Finally, the collimated light is focused by the Canon 100 mm EF lens (green) and recorded on the CCD of the camera (blue) Subset of the Hg spectrum showing three bright lines as seen using the echelle spectrometer. The spectrum from the science fiber is in red, while the spectrum from the calibration fiber is in blue. It is easy to see that the two fibers are well aligned Image of the astronomy laboratory after the cubicles were replaced with curved desks xiii

17 CHAPTER 1 OUTLINE OF THIS THESIS This thesis concerns three unrelated astrophysics projects and has been divided into three parts: I) Radio Wavelength Studies of the Galactic Center Source N3, II) Development of Spectroscopic Instrumentation for Robotic Telescopes, and III) Active Learning Activities for Astronomy Laboratory Courses. For each project, I will provide an introduction to the topic and provide motivation for the projects in the first chapter. The chapters following each introduction will presented in the format of manuscripts for publication in Parts I and II. Only Chapter 3 has been published thus far. Chapter 6 has been submitted to the American Journal of Physics while all remaining chapters have not yet been submitted for publication. In Part III, the chapters after the introduction are written in the form of an internal report for the Department of Physics and Astronomy. The final chapter in each part will outline the conclusions of the previous chapters and present unresolved questions. These chapters will also describe future work to be undertaken in each area. Associated with Part II is one appendix which describes some of the mathematics of echelle diffraction gratings. While this information is not necessary for the understanding of this dissertation, it may be informative for the interested reader. Associated with Part III are two appedicies which provide an example of an astronomy laboratory actvity before and after our work to incorperate active learning. 1

18 Part I Radio Wavelength Studies of the Galactic Center: an Extreme Galactic Environment 2

19 CHAPTER 2 INTRODUCTION 2.1 The Galactic Center Central Molecular Zone The Central Molecular Zone (CMZ), which lies within the central 200 pc of the Galactic Center (GC) (Figure 2.1), is unlike any other region within our galaxy. While the CMZ contains less than 10 percent of the molecular gas within the Galaxy, over 80 percent of the dense molecular gas found in the Galaxy is found in this region (Morris & Serabyn, 1996). The dense molecular gas of the CMZ is mostly distributed within a population of giant molecular clouds with masses of M and sizes between 15 and 50 pc. These molecular clouds are vastly different than their counterparts in the Galactic Disk. The molecular clouds of the CMZ are characterized by large line widths (15 50 km s 1, Bally et al., 1987), high gas temperatures ( K, Hüttemeister et al., 1993; Mauersberger et al., 1986; Mills & Morris, 2013), and large gas densities (n > 10 4 cm 3, Zylka et al., 1992). Most CMZ clouds are also unlike clouds in the disk in that they show very little evidence for recent or ongoing star formation. When examining the molecular gas of the CMZ, many intriguing and unique regions, such as Sgr A, Sgr B2, Sgr C, and the Radio Arc, are apparent (See Figure 2.1). Of these, the Radio Arc region may be the most interesting. The Radio Arc region is noteworthy for its collection of molecular gas, H II regions, and non-thermal sources (Yusef-Zadeh et al., 1984; Yusef-Zadeh & Morris, 1987). Two prominent H II regions are found within the Radio Arc: The Arched Filaments (G ) and the Sickle (G ). Two stellar clusters have also been found in this region (Nagata et al., 1995; Cotera et al., 1996; Figer et al., 1999). The Arches cluster lies close (in projection) to the Arched Filaments while the Quintuplet cluster lies close in projection to the Sickle. Both of these H II regions are thought to be formed by the 3

20 ionization of a molecular cloud by the bright UV flux from their respective clusters. In addition to the thermal gas in the region, the Radio Arc is also home to a number of bright synchrotron powered non-thermal filaments (NTF). These NTFs are located 30 pc from the dynamical center of the Galaxy, Sgr A. The Radio Arc NTFs are long ( 40 pc), narrow ( 0.1 pc) and aligned roughly perpendicular to the Galactic plane (Yusef-Zadeh et al., 1984). Both the Sickle and pistol H II regions appear to be interacting with the NTFs (Yusef-Zadeh & Morris, 1987; Lang et al., 1997). While the Radio Arc NTFs are not the only such filaments in the GC, they are the brightest such filaments. Located along the line of sight to the brightest of the Radio Arc NTFs is the radio source N3. N3 is a compact radio source first described by Yusef-Zadeh & Morris (1987) and lies close to the both the Sickle and Pistol H II regions. Although N3 lies along the brightest of the NTFs, very few studies have examined this source. Yusef-Zadeh & Morris (1987); Lang et al. (1997) studied the continuum emission from N3, though they were unable to resolve N3 or determine its spectrum. Later, Radio Arc Sickle H II Region Sgr B2 Sgr A Sgr C Declination 48: :50:00.0 Wake N3 Pistol H II Region 52:00.0 Radio Arc NTFs :46: Right Ascension Figure 2.1: Right: Image of the GC showing the CMZ. Purple represents 20 cm radio emission, which traces non-thermal emission as well as thermal H II regions. 1.1 mm emission is shown in orange and traces dust and associated molecular gas. Cyan shows Spitzer IRAC data, and highlights emission from stars and polycyclic aromatic hydrocarbons. (Image courtesy of NRAO/AUI.) Left: Figure 1 from Ludovici et al. (2016) showing a 5 GHz continuum image of the Radio Arc region, showing the NTFs, Sickle and Pistol H II regions, N3, and the wake. 4

21 the molecular emission in the region around N3 was observed by Tsuboi et al. (2011). The molecular line observations revealed the presence of a compact clump of molecular gas along the line of sight to N3. Tsuboi et al. (2011) suggested that the properties of this molecular emission was suggestive of a physical interaction between N3 and the molecular gas. 2.2 Goals of this Work The intriguing position of N3, along with the extreme lack of information in the literature concerning the source make N3 an appealing target for an in-depth study. Not only is the emission from N3 not well understood, the resolution of the Nobeyama Radio Observatory used by Tsuboi et al. (2011) is insufficient to examine the possible connection between N3 and the molecular gas in the region. In this work, I will address the following outstanding science questions: What is the physical nature of N3? Is the molecular gas observed by Tsuboi et al. (2011) located in the GC? Does N3 show signs of interaction with the molecular gas in the region? Could N3 be a source of the relativistic particles needed to power the Radio Arc NTFs? To answer these questions, we carried out multi-frequency continuum and spectral line observations with the VLA in 2013 and The results are now published in Ludovici et al. (2016) and are described in Chapter 3. Additionally, we have conducted very high resolution continuum observations using the Very Long Baseline Array (VLBA) which are described in Chapter 4. 5

22 CHAPTER 3 THE UNUSUAL GALACTIC CENTER RADIO SOURCE N3 This chapter is taken directly from Ludovici et al. (2016) which was published in the Astrophysical Journal. 3.1 Abstract Here we report on new, multi-wavelength radio observations of the unusual point source N3 that appears to be located in the vicinity of the Galactic Center. VLA observations between 2 and 50 GHz reveal that N3 is a compact and bright source (56 mjy at 10 GHz) with a non-thermal spectrum superimposed upon the non-thermal radio filaments (NTFs) of the Radio Arc. Our highest frequency observations place a strict upper limit of mas on the size of N3. We compare our observations to those of Yusef-Zadeh & Morris (1987) and Lang et al. (1997) and conclude that N3 is variable over long time scales. Additionally, we present the detection of a compact molecular cloud located adjacent to N3 in projection. CH 3 CN, CH 3 OH, CS, HC 3 N, HNCO, SiO, SO, and NH 3 are detected in the cloud and most transitions have FWHM line widths of 20 km s 1. The rotational temperature determined from the metastable NH 3 transitions ranges from 79 K to 183 K depending on the transitions used. We present evidence that this molecular cloud is interacting with N3. After exploring the relationship between the NTFs, molecular cloud, and N3, we conclude that N3 likely lies within the Galactic Center. We are able to rule out the H II region, young supernova, active star, AGN, and micro-quasar hypotheses for N3. While a micro-blazar may provide a viable explanation for N3, additional observations are needed to determine the physical counterpart of this mysterious source. 6

23 3.2 Introduction The center of our Galaxy hosts a number of unique features not observed elsewhere in the Galactic disk. One of the most unusual regions in the Galactic Center (GC) is the Radio Arc region, lying 30 pc in projection from the dynamical center of the Galaxy, coincident with the radio counterpart of the Galactic black hole, SgrA. The Radio Arc contains both thermal sources (H II regions, molecular clouds) as well as a collection of long ( 40 pc), narrow ( 0.1 pc) non-thermal filaments Sickle H II Region 48:00.0 Wake Declination -28:50:00.0 N3 Pistol H II Region 52:00.0 Radio Arc NTFs :46: Right Ascension Figure 3.1: 5 GHz continuum image of the Radio Arc region, showing the NTFs, Sickle and Pistol H II regions, N3, and the wake. 7

24 (NTFs). The Radio Arc NTFs are aligned roughly perpendicular to the Galactic plane (Yusef-Zadeh et al., 1984) and appear to be interacting with the Sickle and Pistol H II regions (Yusef-Zadeh & Morris, 1987; Lang et al., 1997). Located (in projection) near the middle of the Radio Arc NTFs is an unusual, bright radio point source designated as N3. First observed by Yusef-Zadeh & Morris (1987) at 4.9 GHz, N3 appears to be located in the brightest filament of the Radio Arc NTFs and is well separated from the Sickle and Pistol H II regions. With the exception of the NTFs, no extended continuum emission is observed immediately surrounding N3. Figure 1 shows a finding chart of the various sources in this region. Despite the intriguing location of N3, the source has received little attention since it was first observed. Several radio studies of the GC in subsequent years have detected N3, but have not focused on the nature of this source (Lang et al., 1997; LaRosa et al., 2000; Yusef-Zadeh et al., 2004). Yusef-Zadeh & Morris (1987) consider that N3 could be a background source, except for its suggestive position on the brightest NTF and a faint wake of emission lying between N3 and the northern half of the Sickle HII region (see Figure 1). More recently, H 13 CO + and SiO emission lines were surveyed in this region by Tsuboi et al. (2011), who found a compact SiO emission component near the location of N3. In this component the brightness temperature ratio, R SiO/H 13 CO +, is high ( 4), indicating that the molecular gas is shocked. Tsuboi et al. (2011) conclude that N3 might be interacting physically with the molecular gas. Using the Karl J. Jansky Very Large Array (VLA) of the National Radio Astronomy Observatory, 1 we have conducted a large-scale study of the Radio Arc NTFs and surrounding regions. These observations consist of a multi-frequency, multiconfiguration study that includes spectral lines and is sensitive to polarization. Since the field of view of these observations included N3, we are able to utilize the data to conduct the first in-depth study of N3. In this work we present continuum observations of N3 and the surrounding regions, as well as detailed observations of molecular and hydrogen recombination lines. Detailed discussion of the study as a whole will 8

25 be presented in future papers. In this paper, we focus on the properties of N3 and examine possible interactions between this source and its surroundings. 3.3 Observations and Data Calibration We carried out multi-frequency, multi-configuration observations of the GC Radio Arc region with the VLA. The WIDAR correlator on the VLA enables the simultaneous observation of a wide spectral bandwidth for sensitive continuum measurements and large number of spectral lines. The observations spanned frequencies from 2 GHz to 49 GHz and used multiple array configurations, thus allowing high resolution observations of the region while still maintaining sensitivity to extended structures (see Table 1). The hybrid arrays of the VLA were used when possible to compensate for the low elevation of the GC from the VLA site. In addition, we utilized data (at GHz) from a survey of GC molecular clouds (Mills et al. 2015; Butterfield et al. in prep.). Below we describe the observing and calibration strategy for the low frequency (2 12 GHz) and high frequency (30 50 GHz) data separately. Table 3.1: Observed Fields Field RA Declination Integration Time 1 Name (J2000) (J2000) DnC 2 CnB 3 B 4 BnA 5 S-1 17 h 46 m s C-1 17 h 46 m s C-2 17 h 46 m s X-1 17 h 46 m s X-2 17 h 46 m s X-3 17 h 46 m s X-4 17 h 46 m s Ka-1 17 h 46 m s Ka-2 17 h 46 m s Ka-3 17 h 46 m s Q-1 17 h 46 m s Q-2 17 h 46 m s Q-3 17 h 46 m s Q-4 17 h 46 m s

26 3.3.1 Low Frequency Observing and Calibration (2 12 GHz) The low frequency observations were conducted using the DnC, CnB, B, and BnA array configurations over a period of several months from May 2013 to February Multiple pointings were used to cover the brightest portions of the Radio Arc NTFs (see Table 1). The WIDAR correlator was used in 8-bit mode for all observations. All observations were divided into two 1 GHz bands of 8 spectral windows, each consisting of 64 channels. Each set of observations used the same calibrators: 3C286 was used as both the absolute flux and bandpass calibrator, while J served as the phase calibrator. In addition, we observed J as a polarization leakage calibrator as it is a known unpolarized source. The data were then calibrated with the Common Astronomy Software Application (CASA) analysis package (McMullin et al., 2007) as provided by the NRAO using the standard processing technique for continuum data High Frequency Observing and Calibration (30 49 GHz) The high frequency observations were made between May 2013 and January 2014 (see Table 1). The observations were taken using separate 1 GHz sub-bands and consisted of a mixed setup: low resolution spectral windows for continuum observations and high resolution windows for spectral line observations. Continuum spectral windows consisted of 64 channels each, while spectral line observations had varying spectral resolution. Table 3.2 lists the observed spectral transitions, rest frequencies, velocity resolution, and whether the line was detected in our observations. The standard calibration procedures for high frequency continuum and spectral line observations were followed, including corrections for atmospheric opacity. We utilized 3C286 as an absolute flux calibrator and J as a phase calibrator. J served as the bandpass calibrator. We again observed J as a polarization leakage calibrator, however at these frequencies the calibrator was too 10

27 weak to detect. Initially using the DnC configuration, we mosaicked a larger field to image the Radio Arc NTFs, but ultimately we were not sensitive to the majority of these structures due to their large angular size. In subsequent observations, we focused only on the field containing N3 to improve our signal to noise on this interesting source. Table 3.2: Spectral Line Parameters Species+ Rest Channel Detection Transition Frequency Width (GHz) (km s 1 ) H56-α No H60-α No SO ( ) Yes CH 3 OH ( ) Yes HC 3 N (4 3) Yes CH 3 CN ( ) Yes SiO (1 0) Yes HNCO ( ) Yes H 2 CO ( ) No CS (1 0) Yes CH 3 OH ( ) Yes CH 3 OH ( ) Yes Continuum Imaging We created images using the clean task in CASA: a wide-field image at 5 GHz sensitive to the NTFs and other extended structures in the vicinity of N3 and a set of images focusing on only the compact source N3 at each band of our observations. To create the wide-field image (Figure 3.1), we concatenated data from all four array configurations, then created a two-field mosaic using Briggs weighting with a robust parameter of 0.5. In all other images of N3 (see Figure 3.2), we only used the data from B and BnA arrays with uniform weighting (robust= 2). 11

28 Spectral Line Imaging Table 3.2 summarizes the spectral lines observed in our spectral line setup. Twelve lines were observed, nine of which were detected in and around N3. Each spectral line was imaged individually and continuum subtracted using the CASA task imcontsub. All spectral lines, with the exception of CH 3 OH ( ) and CH 3 OH ( ), were imaged using only the DnC array data at their intrinsic spatial resolution using natural weighting. The images were subsequently smoothed to a spatial resolution of three or four arcseconds in order to improve the image sensitivity. The CH 3 OH ( ) and CH 3 OH ( ) transitions at 36 and 44 GHz are well known Class 1 (collisionally excited; Morimoto et al., 1985; Slysh et al., 1994; Sjouwerman et al., 2010) maser transitions. Masers are characterized by their point-like spatial distribution and narrow spectral profiles. The CH 3 OH ( ) and CH 3 OH ( ) transitions were imaged with uniform weighting using only the B array data. In addition, self-calibration was performed on a bright maser with minimal neighboring emission in order to increase the signal-to-noise ratio of the observations. The CASA task gaincal was used to perform phase self-calibration using a model built using clean. The calibration was applied to the data and the process repeated until the signal-to-noise improvement was no longer significant (2 3 iterations). The phase solutions for this single channel were then applied to all channels in the dataset, and the data imaged with clean to form the final image. 12

29 Figure 3.2: Continuum images of N3 from 49 to 4.5 GHz. All images are on the same scale and centered on the position of N3. The synthesized beam for each image is as follows: 49 GHz: 0.26 x 0.11, 44 GHz: 0.3 x 0.10, 36 GHz: 0.43 x 0.13, 30 GHz: 0.52 x 0.15, 11.5 GHz: 1.01 x 0.38, 10.5 GHz: 1.10 x 0.43, 5.5 GHz: 1.48 x 0.62, 4.5 GHz: 1.80 x The synthesized beam is shown in the bottom left of each image as a black outline. 3.4 Results Properties of N Continuum structure and size of N3 As seen in Figure 3.1, the radio continuum emission from N3 makes it the brightest source observed in the region of the Radio Arc at most frequencies. To examine the structure and size of N3, we analyze images made at the following frequencies: 49.0, 44.0, 36.0, 30.0, 11.5, 10.5, 5.5 and 4.5 GHz. Figure 3.2 shows continuum images of N3 created using only data from the B and BnA array configurations with the synthesized beam size in the lower left corner of each panel. The beam sizes range from (49.0 GHz) to (4.5 GHz). By utilizing only the B and BnA data (see Table 3.1), we effectively apply a 13

30 spatial frequency filter that removes the extended emission of the NTFs from the image, which allows us to more easily detect N3. N3 appears to be unresolved at all frequencies (Figure 3.2). The 49.0 GHz observations (Figure 3.2, panel 1) provide the highest resolution measurement of this source with a beam size of milliarcseconds (mas). To place a stronger constraint on the size of N3, we conducted a Monte Carlo simulation that examined our ability to deconvolve a source from the synthesized beam. In our simulation, we assumed a Gaussian source and SNR of the source, then varied the source size with respect to the synthesized beam. We next attempted to deconvolve the source from the convolution of the source and synthesized beam and determined to what lower limit we could reliably determine a source size. At the SNR of N3 ( 25), we found that a source can be reliably deconvolved from the synthesized beam if the source is 1/4 the size of the synthesized beam size. However, N3 cannot be deconvolved from the beam, so we place an upper limit of mas on the size of N Flux Density and Spectrum of N3 The intensity of N3 is brightest in our 10.5 GHz (62 mjy beam 1 ) and dimmest at 49 GHz (7 mjy beam 1 ); this intensity is nearly an order of magnitude greater than the extended emission of the NTFs and the Sickle and Pistol H II regions in Figure 3.1. Because N3 is bright and detected with a high signal-to-noise ratio in all images, we can study the spectrum of the source across the wide frequency range ( 2 36 GHz). Brightness measurements at 44 GHz and 49 GHz were not used due to uncertainties in the flux calibration scale. In order to make an accurate comparison between images, all images were smoothed to a common resolution of 3.5, which was chosen to match the lowest resolution images (23 GHz). In order to reduce extended structure from the NTFs that might confuse the measurement of N3, we utilized a UV cutoff when making the images. Forty-eight frequencies were imaged using a

31 MHz bandwidth across our frequency range: 2 4 GHz (5 images, UV cutoff > 40 kλ), 4 6 GHz (12 images, UV cutoff > 40 kλ), GHz (9 images, UV cutoff > 40 kλ), GHz (12 images, UV cutoff > 20 kλ), 30 GHz (6 images, UV cutoff > 10 kλ), and 36 GHz (4 images, UV cutoff > 10 kλ). For each of these 48 images, we used the maximum intensity of N3 to construct its spectrum. Figure 3.3 shows the spectrum of N3 across these frequencies. Figure 3.3: Continuum spectrum of N3 from 2 GHz to 49 GHz. The solid line represents a broken power law fit to the data. At low frequencies (2-6 GHz) the spectral index rises (α +0.5) while at high frequencies, (10-36 GHz) the spectrum is falling (α 0.8). In order to determine the nature of the spectrum, we compute the spectral index, α, defined as S ν ν α where S ν is the intensity in mjy and ν is the observed frequency. To calculate the spectral index, we fit a linear function to the logarithm of intensity versus the logarithm of frequency. Due to the high signal-to-noise ratio of the N3 detection, we assume that errors in the flux of N3 are dominated by the wavelength 15

32 dependant variation in the background due to the NTFs as opposed to noise in the images. Since each image represents data from a range of frequencies and not a single frequency, we assume a frequency error for each image of the bandwidth of the image (128 MHz). Due to the turnover of the spectrum of N3 at 8 GHz, we modeled the spectrum of N3 as a broken power law and fitted a spectral index separately above and below the turnover (Figure 3.3). At low frequencies (2 6 GHz), we assumed a 10% error in our flux measurements to account for the variable background and assume no variability. Our fit to the low frequency spectrum of N3 results in α =+0.56 ± 0.13 with a χ-squared probability of As will be discussed in Section , N3 appears to be variable over long time scale. This variability makes determining an accurate spectral index for the high frequency observations difficult since, unlike the low frequency data, our high frequency data were not taken on the same date. In order to produce a good fit the spectral index of N3 at high frequencies, we must assume a 10% variation in flux in addition to the 10% error in our flux measurements. Our best fit for the spectral index of N3 is α 0.86 ± 0.11 with a χ-squared probability of The fits for the high and low frequency spectra indicate that the spectrum turns over at 8.68 GHz The Environment around N3: the Radio Arc NTFs and Wake The two most prominent sources near (in projection) to N3 are the Radio Arc NTFs and the wake structure (Yusef-Zadeh & Morris, 1987) (see Figure 3.1). Our high-sensitivity images reveal greater detail than those of Yusef-Zadeh & Morris (1987) and allow us to examine these features and their properties N3 and the Radio Arc NTFs As first pointed out by Yusef-Zadeh & Morris (1987), N3 appears to be located within the brightest filaments of the Radio Arc NTFs. Though the NTFs are greater than 30 pc in length, they have widths of 0.1 pc (Figure 3.4, right) (assuming a 16

33 distance of 8.3 kpc to the GC; Reid et al., 2014). Individual filaments in the NTFs exhibit gentle curvature and appear to intersect at a few locations, such as the intersection of two linear filaments 60 to the south-east of N3. They also display brightness variation along their length, with filaments brightening and fading independent of other nearby filaments. Using the nomenclature of Yusef-Zadeh & Morris (1987), N3 lies along the same line of sight as linear filament IV, which appears to connect with VS in the south and fades to the north of N3 (Figure 3.4, left). V S IV IV V S IV S IV Figure 3.4: Left: 5 GHz image of region surrounding N3. Individual filaments of the NTFs are labeled. Right: Brightness profile of the NTFs as extracted from the black rectangle in the left figure. The main filaments located near N3 are labeled using the convention of Yusef-Zadeh & Morris (1987) N3 and the Wake In addition to the NTFs near N3, the other prominent feature in this region is a wake feature first observed and labeled in Yusef-Zadeh & Morris (1987). The wake is a series of small, curved filamentary radio continuum features to the north of N3, indicated in Figure 3.1. Compared to the lower sensitivity images of Yusef- Zadeh & Morris (1987), our 5 GHz continuum images reveal a more complicated and 17

34 detailed structure. Figure 3.5 is a comparison between our 5 GHz image and the Paschen-α (Pa-α) image of Wang et al. (2010) with features of the wake labeled. These images reveal that the wake is composed of several separate wisps, filaments, and arcs, all located to the north of N3 (labeled in Figure 3.5). The east arc and west arc of the wake are the two closest Pa-α features to N3. Both arcs are 32 in length and are roughly oriented symmetrically around a line that would connect N3 to the Western edge of the Sickle HII region. There is no Pa-α emission at the position of N3 itself. To the north of the pair of arcs, a set of thermal linear threads can be seen in both the radio and Pa-α images. These threads appear to be oriented along the direction of the weak NTFs found in this region. While no bright Pa-α emission is found connecting the linear threads and the pair of arcs, there are numerous faint wisps and clumps found between these two sets of features in the surrounding environment The Environment around N3: Molecular Gas Surprisingly, strong molecular emission is detected in the region around N3 (see Table 3.2 for a listing of detected transitions). The molecular emission originates from a compact molecular cloud adjacent (in projection) to N3, hereafter referred to as the N3 molecular cloud. The two CH 3 OH maser transitions will be discussed in Section We use the observations of the NH 3 metastable transitions (NH 3 (3,3) through (6,6)) from Butterfield et al. (in prep.). As shown in Table 3.2, no detections of the H56α and H60α lines are made toward this molecular cloud. The lack of radio recombination line emission is also consistent with the lack of thermal radio continuum emission from this extended N3 molecular cloud region Morphology of the N3 Molecular Cloud Figure 3.6 presents integrated intensity maps of the observed transitions (the location of N3 is marked with a white cross). The emission from the N3 molecular 18

35 5 GHz Radio Continuum Pa-α 48:00.0 Sickle HII Region Sickle HII Region Declination 49: Linear Threads West Arc Linear Threads West Arc -28:50:00.0 East Arc East Arc 30.0 Pistol HII Region Pistol HII Region :46: Right ascension :46: Right ascension Figure 3.5: Comparison of our 5 GHz continuum observations with HST Pa-α observations (Wang et al., 2010). The wake is obvious in both the Pa-α observations and the 5 GHz continuum, indicating that the source is thermal in nature. Important features of the wake are indicated. The position of N3 is marked with the cross. cloud is roughly the same size in all observed transitions, and is elongated in the north to south direction. In nearly half of the observed transitions, the molecular emission exhibits curvature around the location of N3. Depending on the transition observed, the region of brightest emission is either to the north (HNCO, NH 3 (3,3), NH 3 (4,4), and NH 3 (6,6)) or to the west (CH 3 CN, CH 3 OH, CS, HC 3 N, SiO, SO, NH 3 (5,5)) of N3. The strongest and most widespread emission is from the NH 3 (3,3) transition (middle panel, bottom row in Figure 3.6). Figure 3.8 shows the 5 GHz continuum image of the region around N3 overlaid 19

36 49:41 CH 3 CN (4-3) 49:41 CH 3 OH ( ) 49:41 CS (1-0) Declination -28:50:10 Declination -28:50:10 Declination -28:50: :46: Right Ascension 17:46: :46: Right Ascension Right Ascension 49:41 HC 3 N (4-3) 49:41 HNCO (2-1) 49:41 SiO (1-0) 24 Declination -28:50: : Declination -28:50:10 Declination -28:50:10 Declination -28:50: :46: :46: :46: Right Ascension Right Ascension Right Ascension SO ( ) 49:41 NH 3 (6,6) 24 Declination -28:50: :41 NH 3 (3,3) 24 Declination -28:50: :46: :46: :46: Right Ascension Right Ascension Right Ascension Figure 3.6: Integrated intensity maps of the molecular transitions observed. Contours are at 3, 6, 10, 20, 30, 40, and 50 times the rms noise in each image (CH 3 CN: 46 mjy beam 1 km s 1, CH 3 OH: 153 mjy beam 1 km s 1, CS: 188 mjy beam 1 km s 1, HC 3 N: 98 mjy beam 1 km s 1, HCNO: 42 mjy beam 1 km s 1, SiO: 94 mjy beam 1 km s 1, SO: 100 mjy beam 1 km s 1, NH 3 (3,3): 18 mjy beam 1 km s 1, NH 3 (6,6): 19 mjy beam 1 km s 1 ). The position of N3 is marked with a cross. with intensity contours of the NH 3 (3,3) transition from integrated -25 to 100 km s 1. The south-western edge of the NH 3 (3,3) emission lies parallel to the NTFs (Figure 3.8). The eastern edge of the cloud is aligned with the Pa-α west arc. 20

37 3 N3 2 1 Figure 3.7: Left: Velocity map of NH 3 (3,3) emission in the N3 molecular cloud. The contours are 5, 9, 13, 17, 21, and 25 times the RMS surface brightness (40 mjy beam 1 km s 1 ) of the integrated NH 3 (3,3) image. N3 is marked by the cross, while three additional positions in the cloud are marked with dots. Right: Spectra of the NH 3 (3,3) transition observed in the N3 molecular cloud. The profiles in this figure correspond to the positions labeled in the left-hand figure Kinematics of the N3 Molecular Cloud Figure 3.7 shows the distribution of peak velocity derived from a single Gaussian profile of the NH 3 (3,3) emission within the N3 molecular cloud, as well as NH 3 (3,3) spectra towards N3 and three selected positions in the cloud. A velocity gradient exists in the cloud, with the velocity roughly increasing from east to west across the cloud. This gradient is strongest between positions 1 and 2 with a velocity gradient of 30 km s 1 pc 1. The gradient between position 1 and position 3 is 15 km s 1 pc 1 while the gradient between position 2 and position 3 is 21 km s 1 pc 1. These velocity gradients are significantly larger ( 2 times) than the velocity gradients measured in several other Galactic center clouds (i.e., Lang et al., 1997, 2001; Mills et al., 2015). The velocity profiles shown in Figure 3.7 show that the molecular lines in the N3 molecular cloud have broad line widths, particularly in the south-east region of the cloud. Near position 1 and N3, emission is detected between -20 to

38 km s 1. The profiles also show unusual structure; this wide structure may represent the superposition of multiple velocity components (multiple profiles in this region may also explain the unusually large velocity gradients, derived from a single broad profile at one of the locations). The lines in the western and northern portions of the cloud show narrower, single-peaked, Gaussian-like profiles with FWHM of 10 to 15 km s 1. We also note that no absorption is seen in the spectral profile toward N3. Figure 3.8: 5 GHz continuum image (greyscale) overlaid white contours showing NH 3 (3,3) emission integrated between velocities of -25 and 100 km s 1. Contour values are 5, 9, 13, 17, 21, and 25 times the RMS noise value of 40 mjy beam 1 km s 1. The dashed white lines are the outlines of the arcs shown in Figure 3.5. The cross denotes the position of N3. 22

39 Ammonia Temperature By observing multiple metastable (J=K) transitions of NH 3, we are able to measure the temperatures of the N3 molecular cloud. NH 3, due to its low dipole moment, is able to reach thermal equilibrium with H 2 relatively quickly, while the short lifetimes of the non-metastable (J K) states leave most molecules in the metastable states. By comparing the column densities of the metastable states, a rotational temperature can be determined that provides a good approximation to the kinetic temperature of the cloud (Morris et al., 1973; Hüttemeister et al., 1993). When calculating the temperature using the metastable NH 3 transitions, it is important to know whether the observed transitions are optically thick or thin. For an optically thick cloud, the hyperfine satellite lines of the transition will be enhanced relatively more than the main hyperfine line. While strong hyperfine features were observed in the NH 3 (1,1) and (2,2) transitions, no hyperfine structure is observed in any of the higher metastable transitions, thus we conclude that the N3 molecular cloud is optically thin for the NH 3 (3,3), (4,4), (5,5), and (6,6) transitions. To determine the temperature of a cloud, two transitions of either para (K 3n) or ortho (K=3n) NH 3 are needed. For optically thin clouds, the column density can be calculated from the velocity integrated brightness temperature by: N(J, K) = cm 2 ν J(J + 1) K 2 T mb dv, (3.1) as in Hüttemeister et al. (1993) and Mauersberger et al. (2003). We utilized the NH 3 (3,3), (4,4), (5,5), and (6,6) transitions to construct maps of column density from the integrated intensity images of each transition. These maps were then used in pairs to determine the average rotational temperature T rot (J, J ) of the cloud using the Boltzmann equation: 23

40 N(J, K = J) N(J, K = J ) = g ( ) op(k) (2J + 1) g op (K ) (2J + 1) exp ktrot (JJ ) E (3.2) where E is the energy relative to the ground state and g op (K ) is the statistical weight for para (g op (K = 3n) = 1) or ortho (g op (K 3n) = 2) transitions. To determine rotational temperatures within the N3 molecular cloud we examined both the NH 3 (3,3) and (6,6) ortho transitions and the NH 3 (4,4) and (5,5) para transitions. The average NH 3 (3,3) (6,6) temperature (T rot (3, 6)) over the N3 cloud is K. The NH 3 (4,4) (5,5) (T rot (4, 5)) temperature is much higher than T 36, with an average of K averaged over the N3 molecular cloud and 44 GHz Methanol Maser Lines We detect over a dozen compact, point-like sources in the 36 GHz ( ) and 44 GHz ( ) transitions of CH 3 OH (Figure 3.9). These transitions are well known to give rise to collisionally excited, Class I masers and to trace shocks (Morimoto et al., 1985; Slysh et al., 1994; Sjouwerman et al., 2010). To examine the possibility that the point sources in the N3 clouds are maser sources, we examined the properties of each source. In order to have uniform selection criteria for characterizing the point sources, we have utilized the source detection algorithm clumpfind (Williams et al., 1994). Clumpfind distinguishes sources that partially overlap in position or velocity. The algorithm identifies local maxima, then examines the emission surrounding the maxima both spatially and spectrally to determine the boundaries of the source. No assumptions about the clump geometry, either spatially or spectrally, are made during processing by the algorithm. Clumpfind produces a list of sources with uniform selection criteria. The output of clumpfind was used to construct a catalog of sources for both the 36 and 44 GHz transitions. The clumpfind analysis yields a total of sixteen 36 GHz CH 3 OH ( ) and twelve 44 GHz CH 3 OH ( ) point sources. 24

41 CH 3 OH ( ) [36 GHz] CH 3 OH ( ) [44 GHz] :50 49:50 M36-3 M44-9 M44-4 Declination -28:50: M36-9 M36-16 M36-4 M36-15 M36-1 M36-2 M36-11 M36-12 M36-8 M36-13 M36-7 M36-10 Declination -28:50: M44-2 M44-1 M44-8 M44-10 M44-11 M44-3 M44-7 M44-5 M M36-14 M M44-12 M :46:20 Right ascension :46:20 Right ascension Figure 3.9: Finding charts for detected 36 GHz (Left) and 44 GHz (Right) CH 3 OH masers in the N3 cloud. Greyscale shows maximum intensity maps of each CH 3 OH transition while the green contours show the thermal CH 3 OH ( ) transition at 3, 6, and 10 times the rms (153 mjy beam 1 km s 1 ) of the integrated intensity map. N3 is marked by the black cross. The inset shows the 36 GHz continuum image of N3 (grey scale) with 11 km s 1 CH 3 OH ( ) contours overlain at 4 and 8 time the channel RMS (RMS = 4.5 mjy/beam, channel width = 1 km s 1 ). Since the sources are not centered on the field, a primary beam correction was applied to the intensities of the detected sources. All maser candidates possess a brightness greater than six times the rms noise in their spectral channel. The point sources detected at 36 GHz and 44 GHz were then compared to determine any coincidences between the two transitions. Three point sources displayed both lines. The properties of all point sources can be seen in Tables 3.3 and 3.4 which correspond to the CH 3 OH ( ) and CH 3 OH ( ) transitions, respectively. These tables provide: (1) Catalog Number, (2) Galactic Name, (3) Right Ascension (HH:MM:SS.s), (4) Declination (DD:MM:SS.s), (5) velocity (km s 1 ), (6) FWHM (km s 1 ), (7) Peak Brightness (Jy beam 1 ), (8) Flux (Jy), (9) Brightness Temperature (K), and (10) any counterpart in the other transition. In addition to the source catalog, spectra are also presented for each source in Figures 3.10 and The spectra do not have velocity cutoffs for the edges of the line, thus faint sources may 25

42 not be the strongest peak in their spectra if a brighter source was located nearby. To aid in identification of weak sources near brighter sources, the central velocity of each source is indicated by the dashed line in each spectrum. Additionally, a finding chart of the detected sources can be found in Figure 3.9. The brightness temperatures of the detected sources are all in excess of 900 K. Since this temperature is nearly half an order of magnitude greater than the NH 3 temperature of the cloud, we conclude that these sources are not thermally excited and must be masers. 26

43 Table 3.3: 36 GHz CH3OH Masers ID Maser Name RA Dec Velocity FWHM Ipeak Flux Tb 44 GHz Counterpart HH:MM:SS.s DD:MM:SS.s km s 1 km s 1 Jy beam 1 Jy km s 1 K M36-1 G :46: :50: M36-2 G :46: :50: M36-3 G :46: :49: M36-4 G :46: :50: M44-2 M36-5 G :46: :50: M36-6 G :46: :50: M36-7 G :46: :50: M36-8 G :46: :50: M44-11 M36-9 G :46: :50: M36-10 G :46: :50: M36-11 G :46: :50: M36-12 G :46: :50: M36-13 G :46: :50: M36-14 G :46: :50: M44-12 M36-15 G :46: :49: < M36-16 G :46: :50:

44 Table 3.4: 44 GHz CH3OH Masers ID Maser Name RA Dec Velocity FWHM Ipeak Flux Tb 36 GHz Counterpart HH:MM:SS.s DD:MM:SS.s km s 1 km s 1 Jy beam 1 Jy km s 1 K M44-1 G :46: :50: M44-2 G :46: :50: M36-4 M44-3 G :46: :50: M44-4 G :46: :49: M44-5 G :46: :50: M44-6 G :46: :49: M44-7 G :46: :50: < M44-8 G :46: :50: M44-9 G :46: :49: M44-10 G :46: :50: M44-11 G :46: :50: M36-8 M44-12 G :46: :50: M

45 The majority of 36 GHz masers and all of the 44 GHz masers lie along an arc with a center offset by 2 to the north from the position of N3. This arc of maser emission also contains the three brightest regions of NH 3 (3,3) emission, though the masers are not clustered around the NH 3 (3,3) clumps. This arc is positionally coincident with the brightest regions of the other molecular transitions, though the intensity of the masers does not correlate with the intensity of the other observed transitions. The CH 3 OH masers have velocities ranging from 13 km s 1 to +25 km s 1, consistent with the full range of velocities observed in the other spectral lines. While there are many masers within the cloud, the weakest detected 36 GHz CH 3 OH maser in the cloud is perhaps the most interesting. M36-16 is positionally coincident with continuum emission from N3 to within the positional accuracy of our observations (Figure 3.9 inset). Apart from its interesting location, maser M36-16 resembles the other masers within the N3 molecular cloud in its velocity and velocity width. No 44 GHz counterpart to M36-16 is detected. 3.5 Discussion Nature of Continuum Emission from N Size As discussed in Section and Figure 3.2, N3 is unresolved at all observed frequencies. Our highest resolution data at 49 GHz has a resolution of mas for the angular size of N3. This angular size corresponds to pc ( AU) if N3 is located at the distance of the GC (8.3 kpc; Reid et al., 2014). However, our Monte Carlo simulation (Section ) shows that the upper limit to the size of N3 is 1/4 of the synthesized beam size at our SNR of 25, or mas, corresponding to a linear size of pc ( AU). The 29

46 Figure 3.10: Spectra for each detected 36 GHz CH 3 OH ( ) Maser. The dashed line marks the velocity for the detected maser. size of N3 itself constrains the nature of the sources by limiting physical counterparts to very compact sources as discussed in Section

47 Figure 3.11: Spectra for each detected 44 GHz CH 3 OH ( ) Maser. The dashed line marks the velocity for the detected maser Spectrum To understand the unusual source N3 we must understand the emission mechanisms at work within the source. As discussed in Section 3.4.2, the high frequency (10 49 GHz) spectrum of N3 falls with α = 0.86 ± 0.11 while the low frequency (2 6 GHz) spectrum rises with α = ± The high-frequency spectrum of N3 cannot be produced by thermal, free-free emission, which typically has a spectrum of α 0.1. Assuming a power law distribution of relativistic electrons, an optically thin synchrotron spectrum will have a 31

48 spectral index α = 0.5(1 δ) where δ is the power law index for the electron distribution. Consequently, at high frequencies, an optically thin synchrotron spectrum is the best interpretation of the observed spectrum of N3. Two possibilities exist to explain the spectrum at low frequencies: 1) free-free absorption of the synchrotron source and 2) synchrotron self-absorption (de Bruyn, 1976; Artyukh & Chernikov, 2001). Free-free absorption within a synchrotron source would act to flatten the spectral index of the synchrotron source over a narrow range of frequencies. However, the low frequency spectrum of N3 maintains a power law over a full dex in frequency. A free-free absorption from a uniform foreground source would not produce a power law spectrum over this large of a wavelength range. Additionally, a mixed synchrotron and thermal gas with internal free-free absorption cannot reproduce both the high and low frequency spectrum of N3. A non-uniform free-free absorbing medium could be designed to fit the observed spectrum, however the free-free absorbing material must be very compact (< 30 mas) in order for the thermal radio continuum emission to not be observed. Therefore, we conclude that free-free absorption of a optically thin synchrotron source not likely to produce the observed spectrum of N3. Self-absorbed synchrotron emission from a uniform source has a theoretical spectral index of +2.5, much steeper than the low frequency spectrum of N3. However, a self-absorbed synchrotron spectrum can be flattened if the source is non-uniform, with the magnetic field decreasing from the center to edge of the emission region (de Bruyn, 1976; Artyukh & Chernikov, 2001) Variability Although no significant short-term variability is observed from N3 within the timespan of our observations, N3 appears to be variable on long time scales. Yusef- Zadeh & Morris (1987) reported a peak intensity of 13.6 mjy beam 1 at 4.8 GHz with a bandwidth of 12.5 MHz and a beam of in observations taken between 32

49 1982 and To directly compare our data with the results of Yusef-Zadeh & Morris (1987), we created an image from our observations using the same center frequency, bandwidth, and synthesized beam. In our image, we find an intensity of 45.5 mjy beam 1, over a factor of three greater than the intensity measured by Yusef-Zadeh & Morris (1987). Contamination by the NTFs would affect both observations equally and thus cannot account for the difference. Lang et al. (1997) observed N3 at 8.3 GHz using data taken by the VLA during 1992 and In their image, the peak intensity of N3 is 15.4 mjy beam 1. While we do not have 8.3 GHz data to directly compare with the Lang et al. (1997) observations, if we assume that the spectral index of N3 has remained the same we can compare the intensity at 8.3 GHz to the intensities observed in this work and in Yusef-Zadeh & Morris (1987). Using the spectral index of determined in Section 3.4.2, we estimate that the 4.8 GHz intensity of N3 in 1993 was 11 mjy beam 1. While the inferred 4.8 GHz intensity of N3 in 1993 is comparable to the measured 4.8 GHz intensity in , N3 has undergone significant brightening between 1993 and We conclude that N3 is likely variable over decade-long time scales, however improved observations of radio variability are clearly needed Relation between N3, the N3 molecular cloud, NTFs, and the Wake Is the N3 molecular cloud located at the Galactic Center? In order to better constrain the location of the N3 molecular cloud, we can look for similarities in the physical properties between the N3 molecular cloud and molecular clouds known to reside in the GC. While clouds within the Galactic disk have narrow line widths (2-10 km s 1 ) and low temperatures, clouds in the Central Molecular Zone (CMZ) typically have larger linewidths (15 50 km s 1 ; Bally et al., 1987) and high temperatures ( K; Hüttemeister et al., 1993). Figure 3.7 shows that the molecular gas observed in the N3 molecular cloud has linewidths of

50 km s 1, consistent with most molecular clouds in the CMZ. The rotational temperatures of the N3 molecular cloud, as determined in section , are T rot (3, 6)=81 K in and T rot (4, 5)=165 K. These rotational temperatures are comparable to the temperatures typically observed in CMZ clouds (Hüttemeister et al., 1993) and are atypical of molecular clouds in the Galactic plane. Due to the broad line widths and high temperatures of the cloud, we conclude that the N3 molecular cloud likely lies within the CMZ. A location in the GC implies a physical size of the cloud of 0.8 pc 1.04 pc, assuming a distance of 8.3 kpc Origin and Nature of the N3 Molecular cloud? Assuming (as discussed above) that the N3 molecular cloud is located at the GC, then it has a relatively compact physical size ( 1 pc). Therefore, it is useful to explore its origin and possible connection to surrounding molecular material. As described earlier, both N3 and the N3 molecular cloud lie in projection near to the 25 km/s molecular cloud originally studied by Serabyn & Guesten (1991). In their Figure 3.2, emission is evident in the upper right panel (velocities of 5 to 10 km s 1 ) in the CS (J=3-2) emission line that is coincident with the position of N3 and relatively compact in nature. Faint emission at velocities near 25 km s 1 is also observed from this region in these early single-dish studies. It is therefore plausible that this compact cloud is physically related to a much larger molecular cloud structure in this region. The N3 cloud lies 4 pc in projection away from the massive Quintuplet stellar cluster, which is capable of ionizing the adjacent and extended Sickle HII region (Lang et al., 1997; Figer et al., 1999) at a large distance. Yet, there is no radio continuum arising from the edge of the N3 molecular cloud in our data. However, it is possible to explain the lack of ionization in the N3 molecular cloud if the ionizing photons from the Quintuplet were blocked by intervening gas. Several clumps of gas have been observed between the N3 molecular cloud and the Quintuplet (see Serabyn & Guesten (1991) and Butterfield et al. 2016, in preparation) which may be shielding 34

51 the N3 molecular cloud from ionizing radiation Physical arrangement of N3, the N3 molecular cloud, and the NTFs As Figure 3.6 shows, there is molecular emission present in all spectral lines at or near the location of N3. The N3 cloud clearly encompasses the position of N3. If N3 were located on the far side of the N3 molecular cloud, we would expect to see absorption in the spectral line profiles of molecular gas. However, as is shown in the spectral line profile in Figure 3.7, no absorption is seen at the position of N3. Since we do not see absorption and we do see line emission toward N3 from most observed transitions, it is likely that N3 lies on the near side of the N3 molecular cloud. As described earlier, maser source M36-16 is located at the same location as N3. We calculate the probability of a chance superposition of N3 with any of the masers within the cloud to be 0.3%. If the CH 3 OH ( ) transition were inverted in the foreground gas of N3, the continuum emission from N3 would be amplified at the local cloud velocity. The low probability of a chance superposition indicates that N3 is at least partially embedded in the cloud. If N3 is located just within the surface of the cloud, we could account for weak amplification of the CH 3 OH ( ) transition, while the background emission would dominate any absorption in the other molecular transitions. The morphology of the N3 molecular cloud is suggestive of a possible association between N3 and the N3 molecular cloud. The molecular line emission from the cloud appears to exhibit a curved morphology around the location of N3 (Figure 3.6). This curvature may indicate that N3 is interacting physically with the molecular cloud. Going further, in order to explore the physical arrangement between the NTFs and the N3 molecular cloud, one can examine the morphology of the observed spectral line emission. The southern edge of the N3 molecular cloud is aligned with the NTFs, while the western edge is roughly perpendicular to the NTFs. In the north, the molecular emission extends past the NTFs where the west arc of the wake 35

52 intersects the NTFs. (Figure 3.8). While the observed alignment of the southern edge of the molecular emission and the NTFs may be coincidental, the possibility exists that the N3 molecular cloud is bounded by the NTF at this location. Serabyn & Morris (1994) noted that molecular gas in the 25 km s 1 molecular cloud (to the north of the N3 molecular cloud) appears to be elongated along the NTFs, which suggests a possible interaction. The morphology of the N3 molecular cloud may indicate a similar interaction between the N3 molecular cloud and the NTFs Relationship between N3 and the Wake N3 is located to the south of the east arc and west arc of the wake. The east arc and west arc both appear to extend from near the position of N3 northwards away from the N3 molecular cloud (Figure 3.5). When considering both features, the curvature of both the arcs 22 (0.9 pc) north of N3 is reminiscent of a shell with a diameter of 12 (0.5 pc). The morphology of the arcs and shell like feature may be evidence of a past energetic event. The linear threads of the wake lie between the arcs and the Sickle H II region to the north and are aligned with a few of the NTFs that are located in this region. We conclude that they are likely associated with the NTFs, given their alignment along the filaments. We do not see evidence that the linear threads are associated with N What is the nature of N3? With the wealth of new knowledge concerning N3, we revisit existing hypotheses for its physical nature. These include: (1) UCH II Region: An Ultra Compact H II region typically possesses a flat radio continuum spectrum and would be detectable in the observed radio recombination line images and the Pa-α emission of Wang et al. (2010). The continuum spectrum of N3 is unambiguously non-thermal in nature, and no emission is detected 36

53 in radio recombination or Pa-α lines. This allows us to rule out an UCH II region as the physical counterpart to N3. (2) Young supernova: The radio emission expected from a young supernova (SN) could account for the spectrum of N3. However, a young SN would be expected to expand rapidly. N3 was first observed in 1982, while our observations were completed in If we assume a SN occurred in the GC shortly before the observations of Yusef-Zadeh & Morris (1987), the expansion velocity would be less than 38 km/s (assuming 250 AU) if the SN were to remain unresolved in our high resolution observations. Additionally, the shock generated by a SN within the N3 molecular cloud would likely generate X-ray emission. In Chandra observations of the region (discussed in detail below) no X-ray emission is seen. Due to the low upper limit on the expansion velocity and the lack of X-rays, can rule out a young SN within our galaxy. (3) Foreground Active Star: A foreground active star, such as an RS CVn star could account for the non-thermal radio emission from the continuum source. However, the emission from active stars is typically circularly polarized (Mutel et al., 1987) whereas we detect no polarization from N3. Also, such an object would have to be relatively nearby (< 100 pc) to account for the observed radio flux. An RS CVn star at that distance should be easily detectable in the optical, however, no optical counterpart is observed in the Second Digitized Sky Survey (McLean et al., 2000) nor in the 1.9 micron observations of Wang et al. (2010). (4) Active Galactic Nucleus: The bright self-absorbed synchrotron emission and very compact size are characteristic of AGN. While the physical properties of N3 itself seem well matched to the AGN hypothesis, the lack of absorption in our observations suggests that N3 lies in front of most of the molecular gas, and thus within the Galaxy. Also, if N3 were an AGN, the apparent morphological association of the molecular cloud with N3 would then have to be an accidental coincidence. (5) Micro-quasar: Micro-quasars are a type of X-ray binary in which material from a companion star is accreted by a stellar mass black hole. Micro-quasars can 37

54 produce many of the observable features of N3, including self-absorbed synchrotron emission and point-like emission (Rodriguez et al., 1995; Dhawan et al., 2000; Soria et al., 2010). A micro-quasar origin for N3 would also fit well with the observed radio flux. However, if N3 were a micro-quasar, the bright radio emission would indicate that bright X-ray emission should also be present (Remillard & McClintock, 2006). Chandra images of the region (Obs ID 14897) show no source at the position of N3. The column density along the line of sight to N3, as determined using the WebPIMMS utility 2, is N H cm 2. Using the non-detection in the Chandra observation and this value of N H, we place an upper limit on the unabsorbed X-ray flux from N3 of erg cm 2 s 1. The observations with the VLA and Chandra are separated by 28 days. The observation of strong radio emission with no detectable X-ray counterpart is unlike any observed micro-quasar, thus we conclude that the micro-quasar hypothesis is also unlikely. [NOTE: The X-ray luminosity of known microquasars, such as SS433 and S26, is erg cm 2 s 1 (Soria et al., 2010; Fabrika & Medvedev, 2011). These sources typically show a hard power law spectrum with a photon index γ 1.4. At the distance of the GC, this luminosity corresponds to an unabsorbed X-ray flux of erg cm 2 s 1. In order to reconcile this unabsorbed flux with the upper limit on the observed flux, a column density of N H cm 2 is needed. This is is a factor of 800 greater column density than is exected for this line of sight. Even if we assumed that the X-ray emission from N3 was 100 times less than SS433, an unreasonably high column density of N H cm 2 would be needed to explain the lack of detection in the Chandra observations.] (6) Other: Since all of the above hypothesis for N3 are essentially ruled out, we must consider that N3 represents a more exotic object. Could N3 possibly represent a black hole larger than the stellar mass black holes of micro-quasars? To explore this possibility, we attempted to constrain the possible mass of N3 using the fundamental plane of black hole activity (Equation 5, Merloni et al., 2003). Unfortunately, the 2 /w3pimms.pl 38

55 errors in this relation are too large to provide a meaningful constraint on the nature of N3. However, we also examined Table 1 of Merloni et al. (2003), which lists the X-ray and 5 GHz radio luminosities of Galactic black holes and AGN used to derive the fundamental plane of black hole activity. In this table, every black hole source possessed an X-ray luminosity greater than its radio luminosity. This situation is reversed for N3, which has a radio luminosity greater than the upper limit of its X- ray luminosity. Because of the low X-ray luminosity, we conclude that N3 is unlikely to be a typical hard state black hole. The sample of black holes examined by Merloni et al. (2003) excludes blazars. In blazars, the accretion powered radio jet is directed nearly along the line of sight to the observer. Relativistic beaming is then able to increase the intensity of the observed jet with respect to the X-ray emission of the hot corona. While the increase in brightness is heavily dependent upon the orientation of the beam to the observer and the speed of the relativistic particles, an amplification of more than two orders of magnitude is possible. If N3 represents a Galactic micro-blazar, the increase in the brightness of the radio jet could account for the lack of X-ray emission observed in N3. A micro-blazar origin for N3 predicts variability on short time scales due to large dependence of relativistic beaming on the particle velocity and beam orientation. Small changes in either would result in a large change in the observed intensity of N3. The micro-blazar hypothesis could be easily tested by examining N3 for short term radio variability. N3, despite our analysis, continues to be shrouded in mystery. Although a micro-blazar hypothesis cannot be ruled out by our data, additional observations will be necessary to distinguish between a micro-blazar and the possibility that N3 represents an object not considered in this work. Future radio observations will be essential for uncovering additional clues that might lead to a viable working hypothesis. For 39

56 example, the character and timescale of the variability, as well as precise measurement of the spectral energy distribution will be valuable for taking the next steps. Furthermore, high-resolution VLBI observations may reveal structure in N3 because of its very compact size, although because of scatter-broadening by the foreground interstellar medium, such observations would need to be carried out at high radio frequencies. [NOTE: VLBA observations of N3 at 43 GHz failed due to poor phase solutions due to interstellar scintillation of Sgr A*. See Section 4.1] Meanwhile, the N3 molecular cloud should also be observed to search for any weak atomic or molecular absorption of the continuum emission from N3, maser emission species other than CH 3 OH, or molecular line indicators that might reveal whether the molecular gas projected near N3 shows any excess heating. 3.6 Conclusions Using the VLA, we have conducted the first in-depth study of the GC radio source N3. Our observations reveal that: (1) N3 is an extremely compact source and is unresolved at all frequencies in our data. Using our 49 GHz data, we place an upper limit of pc or AU (at a distance of 8.3 kpc) on the size of N3. (2) The brightness of N3 is measured to be 62 mjy at 10.5 GHz, making N3 the brightest source in the radio arc region at these frequencies. Furthermore, the brightness of N3 has increased by more than a factor of three since the observations of Yusef-Zadeh & Morris (1987) and Lang et al. (1997). (3) The spectrum of N3 can be modeled as a broken power law peaking near 8.68 GHz. Our fit to the spectrum of N3 is consistent with self-absorbed synchrotron emission. (4) Adjacent to N3 in projection lies a compact molecular cloud, in which we detect 13 molecular transitions. The cloud morphology is suggestive of an interaction with N3. The high temperatures (T 45 = 165 K) and large linewidths (20-30 km s 1 ) 40

57 strongly suggest that this molecular cloud is located in the GC. (5) Weak molecular emission, and no absorption is detected along the line of sight to N3. Additionally, a weak 36 GHz collisionally excited methanol maser is found to be positionally coincident with the continuum emission from N3. Combined, these two observations suggest that N3 lies within the near side of the molecular cloud. The location of the molecular cloud within the GC then suggests that N3 also lies within the GC. (6) The arcs in the southern portion of a thermal wake structure seen in Pa-α images appear to extend northward from the position of N3 and may show evidence of a past energetic outburst from N3. (7) We are able to rule out a UCH II region, young supernova, nearby active star, AGN, and micro-quasar as possible physical counterparts for N3. While a microblazar explanation cannot be completely ruled out, more observations are needed to discern between a micro-blazar model and more exotic or unknown phenomena not considered in this work. 41

58 CHAPTER 4 CURRENT AND FUTURE WORK 4.1 Very Long Baseline Array Observations of N3 Although the VLA observations of Ludovici et al. (2016) have revealed a wealth of new information about N3, they are insufficient to determine the true nature of N3. As was discussed in Section 3.5.3, we concluded that the most likely physical counterpart to N3 would be a micro-blazar. A micro-blazar would be characterized by extremely compact, high brightness temperature synchrotron emission and thus would be a natural target for the Very Long Baseline Array (VLBA). VLBA observations have the potential provide answers to two science questions: 1. What is the angular size of N3? Our VLA observations discussed in Section 3 placed a strong upper limit of mas on the size of N3. The VLBA synthesized beam at 43 GHz can be as small as 0.2 mas ( pc or 1.7 AU at the distance to the Galactic Center.) This drastic increase in resolution has the potential to determine the true size of N3. The high resolution of these observations will also permit stronger constraints on the brightness temperature of N3. With the VLA, we have placed a limit of T b >6200 K at 43 GHz. At this same frequency, N3 will be detected by the VLBA if it is a uniform source and the brightness temperature is 10 6 K or greater. 2. What is the structure of N3? If N3 is indeed a micro-blazar, high resolution images with the VLA may reveal structure in the form of jets. These jets, if found would provide clear evidence of accretion onto a compact object. If N3 exhibits a bow shock structure as opposed to a jet, it may indicate that a high velocity compact object or flow is encountering the N3 molecular cloud. If N3 is revealed to show no structure on these size scales, we will conclude that the source is likely a background AGN, as the source would be prohibitively small 42

59 if located at GC distances. The possibility does exist that the VLBA will be unable to image N3 if the scale of the emission is larger than the largest angular scale of the instrument. However, a non-detection due to the source being over-resolved would still provide a strong lower limit on the size and an upper limit on the brightness temperature of N VLBA Observations In August 2016, we observed the GC radio source N3 using the VLBA using the 7 mm receiver in dual polarization mode. The observations were taken during a single observing session lasting 4.5 hours. The 7 mm receiver was chosen to minimize the effects of the strong interstellar scattering in the direction of the GC. Sgr A* served as the phase calibrator as it is nearby, bright, and well studied. NRAO 530 was also observed to serve as a flux calibrator. During our observations, the Pie Town (PT) antenna was unable to observe due to a broken axle. Additionally the St. Croix (SC) antenna was down for maintenance. The Owens Valley (OV) and Mauna Kea (MK) antennas both had recording errors at the beginning of the observation. The OV antenna was brought online 18 minutes after the start of observations, while the MK antenna began recording 34 minutes after the start of observations. Total time on source was 110 minutes Data Reduction and Results The VLBA observations of N3 were processed using the standard VLBA pipeline within the Astronomical Image Processing System (AIPS) software package. We attempted to phase calibrate using Sgr A*, however the phase solutions for Sgr A* suffer from significant fluctuations, likely due to interstellar scintillation. While we were able to use self-calibration to clean up the Sgr A* observations, there is not enough signal to noise on individual baselines to perform self calibration on N3. Because of this, the phase calibration on N3 is unreliable and we are unable to image 43

60 the field. 4.2 VLA GC Point Source Survey While our VLA observations of N3 have revealed a great deal of new information about N3, many questions about the source remain. While the details of N3 are still shrouded in mystery and difficult to study, a broad and easier to address question remains: Is N3 an unique source or is there a population of N3-like sources in the GC? The presence of massive star clusters in the CMZ and the large number of X- ray binaries (Skinner, 1993) imply that large numbers of compact objects should be present in the GC. If N3 represents a compact object within the Radio Arc region, many other similar sources would be expected to be found within the GC. Past observations of point sources were either low sensitivity (Lazio & Cordes, 2008) or covered only a small region of the GC (Toomey, 2014). A large scale point source survey of the GC using the upgraded VLA would allow us to obtain an order of magnitude greater sensitivity to these point sources than obtained in previous large scale surveys. Among the observed sources, we would also expect to detect UCH II regions, radio pulsars, and possibly additional N3-like sources Proposed Observations To search the GC for point sources, we plan to propose for VLA observations using the A array configuration during the 2018A semester. By utilizing the 8-bit receivers and 2 GHz of bandwidth, we should achieve a RMS noise of 10 µjy with a 30 minute observation per pointing at 1.5 GHz and a 10 minute observation per pointing at all higher frequencies. To separate thermal point sources such as ultra-compact H II regions from nonthermal sources, we will need to determine the spectral indices of the observed point sources. To do this, we will observe at several frequencies ranging from 1 15 GHz. 44

61 We will also observe a variety of spectral lines simultaneously with the continuum. These spectral lines will allow us to examine maser transitions in the GC as well as to look for absorption lines towards the brighter point sources. 45

62 Part II Spectroscopic Instrumentation for Robotic Telescope Systems 46

63 CHAPTER 5 INTRODUCTION The traditional mode of operation for optical telescopes is for an observer or operator to directly control the telescope from the observatory site. While this method works well in many cases, it can severely limit the utilization of telescopes operated by a single observer or group as the telescope will sit idle if other obligations keep the observer from operating the instrument. Robotic telescope systems, where a computer controls all aspects of telescope operation without human intervention, eliminate this problem (Trueblood & Genet, 1985). To utilize a robotic telescope system, observers simply submit an observing request to a scheduler. This scheduler then constructs a command list that is sent to the telescope. At sunset, the observatories roof is opened, the telescope performs some basic calibration functions, and the observing program is executed. The telescope then observes until sunrise, when the observatory roof is closed and the telescope is parked until the next observing session. This system not only allows an observer to collect data without managing the telescope, but also decreases the downtime between observations, leading to an increased efficiency of observing. The University of Iowa has been a leader in robotic telescope development since the early 1990s. The first robotic telescope operated by the University of Iowa was the Automated Telescope Facility (ATF) (Mutel & Downey, 1994) which was installed on the roof of Van Allen Hall in Iowa City. The ATF was in service until 1997 when the Iowa Robotic Observatory (IRO) was installed at the Winer Observatory near Sonoita Arizona (Mutel, 1998). The Winer Observatory offers clear, dark skies that are unavailable to observers in Iowa City. The original 0.5 m IRO telescope was replaced with the 0.37 m Rigel telescope in 2002 Mutel (2000, 2002, 2003) and the Rigel telescope system was replaced with the current Planewave CDK m telescope in May

64 Since the 1990s, many universities have followed in the footsteps of the University of Iowa in developing robotic telescope systems for both research and teaching. The telescopes are often equipped with sensitive, cooled CCD cameras that allow observers to obtain impressive images of a wide variety of astronomical objects, including comets, planets, stars, nebulae, and galaxies. These systems enable a wide variety of photometric and astrometric investigations. However, with the exception of a few sophisticated instruments, such as the MINERVA telescopes Swift et al. (2015), spectroscopic observations are not possible with these systems. This is because the majority of these robotic observatories utilize a single instrument port. Traditional spectroscopic instruments often require the physical removal of the imaging system on a telescope, making them impractical for an unmanned robotic system. Additionally, these spectrometers often require positioning the target with sub-arcsecond accuracy on the slit of the spectrometer. In contrast, the blind pointing accuracy of most robotic telescopes is on the order of arc-minutes. The capabilities of these smaller, robotic observatories stands in contrast to large optical observatories. Nearly all of the currently operating large optical telescopes have spectroscopic instruments. These instruments dominate the observing schedules of these telescopes. This is because spectroscopic observations reveal a wealth of information not accessible in photometric observations such as temperature, composition, and radial velocity. Enabling spectroscopic observations on robotic telescope systems will not only help enable science, but student learning as well. Spectroscopy is at the heart of many of the subjects discussed in introductory astronomy courses (ex. Composition of the Sun, Wiens Law, Doppler shift, exoplanets, and Hubble expansion). However, astronomy laboratory exercises focus on imaging due to the difficulty of spectroscopic observations. This leads to students struggling with concepts such as the nature of light (Bardar, 2006; Brecher, 1991). 48

65 5.1 Spectrometer Designs While spectrometers take on many different forms, most spectrometers used in astronomy can be divided into four broad categories: prism, grating, echelle, and Fourier transform spectrometers. Prism spectrometers utilize wavelength dependent refraction in a glass prism to produce a spectrum. This type of spectrometer operates at very low spectral resolution (R < 100) and is typically only used for specialty instruments. Fourier transform spectrometers utilize interference from a modified optical interferometer to produce very high resolution spectra, however the increased resolution is accompanied by decreased sensitivity. Because of the small collecting area of most robotic telescopes, this type of instrument is not practical for most educational and scientific projects Grating Spectrometers A grating spectrometer utilizes a diffraction grating to separate incoming light into its component wavelengths. In these gratings, lines are placed in a periodic structure on the optical substrate using either holographic printing or surface etching. These lines generate a many slit diffraction pattern described by sin θ = nλ a ; (5.1) where λ is the wavelength of the incoming light, a is the spacing between lines of the grating and n is an integer called the order number. In blazed gratings, the shape of the ruled groves is modified to concentrate light into a particular spectral order, increasing the overall efficiency of the grating. These gratings can either be transparent transmission gratings or be coated with a reflective material to form a reflection grating. A variation on the transmission grating is a prism-grating or grism. A grism 49

66 consists of either a separate grating and wedge prism, or a single wedge prism with a grating etched onto one surface. Combining a wedge prism and diffraction grating deflects the diffracted spectrum back to the optical axis of the input light (Figure 5.1). When designing a grism, the grating dispersion and angular deflection of the wedge prism must be matched. For example, using equation 5.1 we see that for a central wavelength 550 nm, a 300 line mm 1 grating requires a prism with a 10 deflection when operating in the first order. Grating Grism Excess Focal Path Figure 5.1: Schematic drawings of the spectrum produced by a simple grating (Left) and a grism (Right). Curved line represents the curved focal plane of the spectrometer. Colors represent red, green, and blue light. A generic grating spectrometer design includes five primary components, a small slit, a collimating optic, the grating or grism, a focusing optic, and a detector. The slit controls the spectral resolution of the spectrometer by controlling the angular size of the source, though it can be optional if observed sources are of a small angular size. The collimating optic creates a parallel beam of light, thus preventing many aberrations in the final image. The focusing optic is then used to refocus the light onto the detector. In most modern instruments, this detector is a CCD camera Echelle Spectrometers An echelle grating is a special type of diffraction grating with a very large blaze angle. The large blaze angle concentrates light into very high spectral orders. As is easily seen in Equation 5.1, increasing the order number also increases the diffraction angle of the grating, leading to much higher resolution than a grating operating in 50

67 the first order. The downside to operating at high spectral orders is that neighboring orders overlap when operating at high order number. (For a more complete view of the properties of an echelle grating, see Appendix A). To correct for this problem, a second, dispersive element is placed perpendicular to the dispersion direction of the echelle grating. This cross-disperser is usually either a low resolution grating or prism. While gratings offer a linear dispersion with wavelength, prism cross-dispersers provide much greater throughput. The general design of an echelle spectrometer is similar to that of a standard grating with two exceptions. First, a slit is required for an echelle spectrometer because of the two dimensional nature of the spectra. Secondly, the grating in a grating spectrometer is replaced with the echelle grating and cross disperser in an echelle. Otherwise, the requirements for a collimator and camera optic are similar Robotic Spectrograph Control Robotic Telescopes must be able to operate all night without any human input. This means that all operations needed to observe a target must be performed automatically including pointing, exposing the instrument, collection of calibration frames, and instrument switching. Of primary concern is the ability to switch instruments if the robotic telescope system is to be used for spectroscopy and photometry. Traditionally, a telescope that makes use of both photometeric and spectroscopic instruments either utilizes a Nasmyth focus with multiple instrument ports (eg. the MINERVA Telescopes, Swift et al., 2015) or the instruments are physically removed when switching between observation modes. For smaller robotic telescope systems which do not have multiple instrument ports or are located at remote sites where instruments cannot be switched easily, spectroscopic observations have been impractical. 51

68 In order to enable spectroscopic observations with the majority of robotic telescope systems, the telescope and spectrometer must allow: A simple method for switching between photometric and spectroscopic instruments Automated calibration of spectroscopic observations Precision alignment of the spectrometer with target object 5.2 Goals of This Work The ability to conduct spectroscopic observations with robotic telescope systems would enable many educational and scientific research projects not only at the University of Iowa, but at universities and colleges around the world. New technologies, such as 3D printing and improved robotic telescope control systems offer exciting opportunities for novel instrument designs. In this work, I will address my instrumentation work that was focused on meeting the following instrumentation challenges: Constructing a low cost, high sensitivity, low resolution spectrometer that is easily adaptable to a large range of robotic observatories. Addressing the challenges inherent to conducting robotic observations with both in-line and fiber-fed spectrograph systems. Designing a pipeline for automatic calibration of robotic spectroscopic observations. To, provide solutions to these challenges, I have designed and built a compact grism spectrometer (see Chapter 6) and a fiber-fed echelle spectrometer (see Chapter 7). I will describe both of these spectroscopic instruments, which are now installed at the Iowa Robotic Observatory (IRO). I will also describe robotic observations utilizing the CGS and fiber-fed echelle spectrometer alongside the photometric camera on IRO in Chapter 7. 52

69 CHAPTER 6 A COMPACT GRISM SPECTROMETER FOR SMALL OPTICAL TELESCOPES This chapter is taken directly from Ludovici and Mutel (2017), which has been submitted to the American Journal of Physics. 6.1 Abstract We describe a low-cost compact grism spectrometer for use with small astronomical telescopes. The system can be used with existing CCD cameras and filter wheels. The optical design consists of two prisms, a transmission grating, a collimating lens, and a focusing lens, all enclosed in a 3-d printed housing. The system can be placed inline, typically in an unused filter wheel slot. Unlike conventional spectrometers, it does not require the target to be precisely positioned on a narrow slit. The mean spectral resolution (R 300) is sufficient to resolve the spectral lines of many astronomical objects discussed in undergraduate astronomy labs, such as stellar absorption lines along the main-sequence, emission lines of early-type hot stars and galactic novae, and redshifts of bright quasars and supernovae. 6.2 Introduction Many colleges and universities have on-campus observatories that are integrated into the undergraduate astronomy lab curriculum. The telescopes are often equipped with sensitive, cooled CCD cameras that allow students to obtain impressive images of a wide variety of astronomical objects, including comets, planets, stars, nebulae, and galaxies. However, spectroscopic observations, in which the light is separated into hundreds or thousands of spectral channels, dominate the observing schedules of most professional telescopes. This is because spectroscopy allows astronomers to determine a large number of physical conditions that are either difficult or impossible 53

70 to discern from an image. These include elemental composition, temperature, density, motion (both translational and rotational), and even magnetic field strength. Understanding spectroscopy and the nature of light is critical for understanding the results of astronomical research. However, evidence shows students often fail to fully understand the nature of light and spectra (Bardar, 2006; Brecher, 1991). Exposing students to spectroscopic data in undergraduate laboratories may help students gain a deeper understanding of light and the properties of astronomical objects. Despite the benefits of spectroscopic observations, relatively few small telescopes have the capability to conduct spectroscopic observations. This is because conventional spectrometers are expensive to build, very sensitive to vibration and temperature variations, and difficult to operate, since the target object must be placed on the spectrometer entrance with arcsecond accuracy. In addition, the spectrometer subsystem often interferes with CCD camera imaging or eyepiece observing, necessitating cumbersome, time-consuming equipment removal and replacement when the spectroscope is installed. Most spectroscopic observations in undergraduate astronomy labs have either made use of handheld spectroscopes for observations of terrestrial sources (Bardar, 2006), or have concentrated on collecting spectra of the Sun (Ratcliff et al., 1992). While systems that permit telescopic observations of stellar spectra exist, they require the removal of the imaging equipment, are limited to observations of the brightest stars (Ratcliff et al., 2011), or produce low quality spectra (Beaver & Conger, 2012). A transmission grating placed directly in the optical path provides a simple lowcost way (Beaver & Robert, 2011; Beaver & Conger, 2012) to obtain low-resolution spectra with minimal equipment changes. This can be done either by inserting the grating into an eyepiece for visual observing (Rainbow Optics, 2017) or in a filter wheel using the existing imaging camera as a detector (Rspec Star Analyze, 2017; Hood et al., 2012). However, the dispersed spectrum focuses on a curved surface rather than a flat plane (Petzval field curvature). This limits the effective resolution 54

71 of simple transmission gratings to R = λ λ 100 over the visible wavelength range. Grism spectrometers, which combine a transmission grating with a prism, minimize field curvature aberration by redirecting the dispersed rays closer to the optical axis. These systems can be used in large focal ratio ( slow ) optical systems without additional corrective optics since the incoming rays are close to paraxial (e.g., the Hubble Space Telescope WFC3 grism (HST WFC3, 2017)). However, most small observatory telescopes have fast optics (f/5 - f/10) so the incident rays are significantly non-paraxial. This requires corrective optics to collimate the incident rays and refocus the exit rays onto the image plane. Here, we describe a low-cost, compact grism system with corrective optics that can be constructed using commercial off-the-shelf optical components. It is small enough to fit into existing commercial filter wheels with the addition a small housing extender. The grism enclosure and filter wheel extension are easily fabricated using a 3-d printer. The spectral resolution (R 300) is sufficient to observe a wide range of astrophysically relevant targets, such as stellar spectral types, emission lines from hot star and novae, and red shifts of bright quasars and Type II supernovae. 6.3 The Compact Grism Spectrometer Optical Design The compact grism spectrometer (CGS) optical design consists of five elements: A high-efficiency transmission grating, two achromatic lenses for collimating and refocussing, and two wedge prisms. In principle, a single prism can be used, but a typical grism design requires a large (θ 20 ) prism refraction angle, which is not commercially available as a stock item. Fig. 6.1 shows a ray trace of the optical design. We now consider the design criteria for each element. Diameter of optical elements. 55 Achromatic lenses, transmission gratings

72 Collimator Grating Focuser Prisms Figure 6.1: Optical design of the compact grism spectrometer, showing ray paths at several wavelengths from 400 nm (magenta) to 800 nm (brown). The incident rays from the telescope (f/6.8) are first collimated by an achromatic doublet lens, then dispersed by a 10 wedge prism, followed by a 600 lpmm transmission grating, then a second 10 prism, and finally refocussed into the detector plane by a second achromatic doublet. The focal lengths of the collimating and refocusing lenses are determined by the f-ratio, location in the converging optical beam, and the backspacing of the imaging sensor. and wedge prisms are commonly available(thorlabs, 2017; Edmund Optics, 2017; Ross Optical, 2017) in both 25 mm and 50 mm sizes. For a compact design and to minimize weight, 25 mm diameter components are preferable. However, to avoid vignetting, the lens clear aperture must be larger than the beam size δx at the grism, δx = d/f < 25 mm, where d is distance from the grism to the focal plane and f is the focal ratio. Collimating lens. The collimating lens corrects for the converging telescope beam, creating a parallel beam. The lens is a negative (diverging) achromatic lens whose focal ratio is chosen to collimate the converging rays from the telescope optics. The exact focal length depends on where the lens is place in the optical path. For a fully illuminated lens, a f/6.8 telescope beam can be collimated with a 25 mm diameter achromatic lens with a focal length -170 mm. 56

73 However, for a partially illuminated lens (i.e. closer to the focal plane), the (negative) focal length increases. Wedge prism. The wedge prism compensates for the wavelength-dependent dispersion angle of the grating. The prism refraction angle, which is nearly wavelength-independent, is chosen to redirect the center wavelength of the observed spectrum back onto the optical axis. For example, for a 600 lpmm grating and a center wavelength 550 nm, the grating dispersion angle is 19. Commercially available stock wedge prismsross Optical (2017) range from 2 to 10 in 2 increments, so we chose to use two 10 prisms. Transmission grating. Stock transmission gratings range from 300 lines per millimeter (lpmm) to 1200 lpmm groove spacing, with larger values providing increased spectral resolution, but at lower efficiency. Also, higher resolution gratings require larger wedge prism refraction angles and larger sensor sizes. In order to cover a wavelength range λ, a sensor located a distance d from a grating with groove spacing a = 1/lpmm must have a linear dimension D s at least, D s > λ a d, For the full visible wavelength range λ = 300 nm, d 50 mm, and a 600 lpmm grating (a = 1.67µ), the sensor size must be at least 9.4 mm. Refocusing lens The refocusing lens is a positive achromat, with a focal length approximately equal to the physical distance from the grating to the focal plane (sensor). For a two prism system, such as described here, the distance is somewhat smaller, since the second prism lies between the grating and the re-focussing lens. We have found that both the collimating lens and refocussing lens focal lengths 57

74 do not need to be exactly matched to the telescope focal ratio and sensor backfocus exactly, since a small change in telescope focus can compensate for the optical path difference. Enclosure The optical elements are housed in a plastic enclosure fabricated using a 3-d printer. Fig.6.2 shows the CGS in its enclosure, and installed in the filter wheel. For most commercial filter wheels, an enclosure extension is required to allow adequate clearance for the grism, as illustrated in Fig.6.3. This can be inexpensively printed using a 3-d printer Wavelength and flux calibration An example of a raw dispersed spectrum recorded on the imaging camera is shown in Fig In order to obtain an astronomically useful spectrum, the raw image requires both CCD and spectral calibration. The CCD calibration, consisting of thermal and bias subtraction, and cosmic ray removal, is standard procedure for CCD imaging at most observatories and will not be discussed here. Note that a flat field correction, which is normally applied to images, is not needed for spectroscopic observations since the gain variations across the dispersed spectrum are corrected by gain calibration, as described below. Spectral calibration has two components: Wavelength and flux calibration. The wavelength calibration is done in two steps. First, the target star is observed with a broadband filter. Small pointing offsets are then applied to place the object exactly at the field center. This can be done either manually or by solving the for the center coordinates using an astrometric image solver (e.g., Pinpoint(DC3 Pinpoint, 2017)). After re-centering, the grism image is taken and major spectral features (e.g. Balmer sequence for an A-type star) are matched with corresponding pixel values. The wavelength-pixel data are fit with a low-order polynomial, and the corresponding coefficients stored in a configuration file. 58

75 The flux calibration, which corrects for wavelength-dependent efficiencies in the imaging sensor, transmission grating, and telescope optics, can be determined by comparing raw spectra with flux-calibrated spectra of standard stars taken a comparable spectral resolution and wavelength range. Fortunately, these are readily available in downloadable formjacoby et al. (1984); Silva & Cornell (1992) and were used to calculate a table of gain coefficients for each spectral channel. Fig.6.5 shows an example of a raw spectrum and wavelength-flux calibrated spectrum. 6.4 Sample CGS spectra and lab projects The grism spectrometer is installed at the Iowa Robotic TelescopeIowa Robotic Observatory (2017), a 0.5m diameter f/6.8 Cassegrain reflector located at the Winer Observatory(Winer Observatory, 2017), about 80 km SE of Tucson AZ, and operated from campus at the University of Iowa in Iowa City. The imaging system consists of a 2048x2048 format 13.5µ pixel back-illuminated CCD camera equipped and a 12-position filter wheel, of which one slot contains the CGS. This system is used by students and faculty at the University of Iowa for teaching and research. The telescope is typically operated robotically using a queued observing list but can also be operated in realtime using a high-speed Internet connection. A few examples of grism observing projects that have been done in undergraduate laboratories are shown below. They illustrate the range of astronomical objects that can be investigated spectroscopically. Fig.6.6 shows a sequence of four stellar spectra, with surface temperatures ranging from 19,000 K to 3,400 K. The spectra illustrate how the Balmer series of hydrogen lines become dominant near spectral type A, then rapidly decrease for solar-type stars, with molecular metal band dominating the coolest stars. Fig.6.7 shows the spectrum of an emission-line star, in which circumstellar gas is heated in the chromosphere or in a disk by rapid rotation. This hotter gas 59

76 produces emission lines at longer wavelengths where the gas is optically thick, but at shorter wavelengths one see absorption lines arising from the cooler photosphere. This is a nice illustration of Kirchhoff s laws. Fig. 6.8 show the spectrum of WR7, an extremely luminous (280,000 L sun ) star with a chemically-enriched high-speed wind whose emission lines dominate the spectrum. The line widths ( λ 2 nm) are spectrally resolved, indicating wind speeds of several thousand km/s. Fig.6.9 shows the spectrum of two bright (V=14.4, 12.9) quasars whose redshifts are readily obtained from the prominent Hα emission line that are displaced from the rest wavelength of nm. In addition to several other Balmer lines, both spectra also show prominent forbidden-line emission of doublyionized oxygen III [OIII] arising from the narrowband region of the quasar. 6.5 Limitations The grism system described here is slitless, so that only point sources can be observed with good spectral resolution. In principle one could add a slit in front of the grism, but since it is in the converging light cone of the telescope, there would be substantial light loss from vignetting. Also, centering the target is critical for accurate wavelength calibration, particularly in the direction parallel to the dispersion axis. For example, with a 600 lpmm grating and a 3m telescope focal length, the wavelength displacement is about 0.5 nm per arcsecond. Finally, modification of some filter wheels to accommodate the grism height is more difficult with some cases than others. For example, some filter wheels have electronics boards that interfere with a simple housing extension modification, so users should carefully inspect their filter wheel design before deciding to incorporate it into a grism system. 60

77 6.6 Adapting the CGS to other observatories Details for design and construction of a compact grism system, including Winlens(Winlens3D, 2017) files for the optical design and OnShape(Onshape.com, 2017) files for the grism enclosure and housing extenders for several commercial filter wheels, can be found on the Iowa Robotic Telescope website. The Python scripts for calibration and display of CGS spectra are open-source and available on GitHub(Github.com, 2017) and include a user-friendly GUI interface (Fig.6.10). 61

78 Figure 6.2: [top] Compact grism system, with 25 mm optical components, enclosed in a 3-d printed housing. [bottom] Compact grism installed in a 50mm diameter slot in a filter wheel. Note that the filter wheel housing has a 3-d printed wall extension with height 37 mm. 62

79 FW8S-STXL Imaging Filter Filter Wheel Disk Filter Retainer Miniature Grism System SBIG STXL-6303E Camera CCD Figure 6.3: Cross-sectional view of a filter wheel with the grism installed. Zero-order star image Dispersed spectrum Figure 6.4: Dispersed grism spectrum zero-order with stellar image at left. The image is focussed at the center of the dispersed spectrum so that the stellar image is strongly defocussed. 63

80 (a) (c) (b) (d) Figure 6.5: (a) Uncalibrated spectrum of the B3V star HD (b) Calibrated spectrum, (c) Gain curve applied to raw fluxes. (d) Calibrated spectrum of HD from Jacoby et al. (1984). 64

81 (a) HD (B0V) (c) HD (F7V) (b) HD (A1V) (d) SAO (M5III) Figure 6.6: Grism spectra of stars from 380 nm to 750 nm, illustrating the spectral sequence from hot to cool stars: (a) B3V:19,000 K (b) A1V: 9,000 K, (c) F7V: 6,240 K, and (d) M5III: 3,400 K. 65

82 1 nm Hε Hδ Hγ Hβ Telluric Hα Figure 6.7: CGS spectrum of the emission-line star HD76868 (B5e). Note the prominent chromospheric Hα emission line, which arises from circumstellar material ejected by the rapid rotation of the star. The circumstellar gas is optically thick in the Hα line, but with increasing frequency it becomes optically thin and the higher level Balmer lines are seen in absorption. The Hβ line has both emission and absorption components. The width of the emission component is λ 0.9 nm, indicating a spectral resolution R

83 HeII NIII HeII HeII CIV+HeI HeII NV Figure 6.8: CGS spectrum of the extremely luminous Wolf-Rayet star WR7 (HD56925) at the center of the emission nebula NGC2359. This star has an extended hot wind responsible for the broad emission lines of helium, carbon, and nitrogen. 67

84 1E V=14.4, z = Hα +[ΝΙΙ] Hγ Hβ [ΟΙΙΙ] 3C273 V=12.9, z = Hγ Hβ+[ΟΙΙΙ] Hα +[ΝΙΙ] Figure 6.9: CGS spectra of two low-redshift quasars, each with prominent red-shifted Balmer emission lines, as well as a forbidden oxygen line. (a) 1E ,V = 14.4, z = 0.096, (b) 3C273, V = 12.9, z = Each exposure was 15 min. 68

85 Figure 6.10: Graphical user interface to the CGS calibration and plotting program designed for ease of student use. 69

86 CHAPTER 7 CONDUCTING SPECTROSCOPIC OBSERVATIONS WITH ROBOTIC TELESCOPE SYSTEMS 7.1 Introduction Robotic telescope systems are telescopes that utilize computer controlled mounts and instruments to make observations without real-time input by a human operator. The benefit of these systems is that an operator does not have to stay up all night to control the telescope. Users do not need to know how to operate the telescope hardware and data is available soon after observations are completed, usually the next morning. This makes robotic telescopes extremely valuable for providing undergraduate students with their own data for analysis or conducting long scientific observing projects. These telescopes can operate either using a pre-determined schedule for an entire night, or by using a dynamic scheduler that determines the best observing target in real time. In both cases, the scheduler receives an observation request from a user and uses the information in that request to assemble a series of commands to be issued to the telescope. These commands tell the control computer how to operate the hardware components of the telescope, such as the mount, cameras, and filter wheel, to perform the tasks needed for completion of the observation request. Robotic telescope systems have seen a large increase in popularity over the last twenty years due to improvements in hardware and computing capabilities. These telescopes, which are mostly operated by universities for research and teaching, are able to produce hundreds of high quality images a night for use in photometric and astrometric studies. However, despite the technological progress in imaging and control systems for robotic telescopes, the capability to make robotic spectroscopic observations has been limited. Traditional spectrometers often require sub-arcsecond 70

87 pointing precision to place a target on the instruments slit, where as even the best robotic systems have blind pointing errors on the order of tens of arcseconds. These systems are also prohibitive from a mechanical standpoint as well. Traditional spectroscopic instruments are mounted on an instrument port of the telescope. For the majority of robotic telescope systems, which only possess a single instrument port, this means that the telescope must be dedicated to either imaging or spectroscopy, greatly reducing the flexibility of the instrument. In this paper, we describe the optomechanical installation and software control of two spectroscopic instrument designs developed for the Iowa Robotic Observatory (IRO). Both designs are installed alongside a high sensitivity imaging camera on a single instrument port and operated using a Windows PC. Robotic control software has been developed for both instruments, and details on robotic control are provided. 7.2 Optomechanical Design To fully enable spectroscopic observations at IRO, both new spectrometers need to operate without input from a user. Thus, not only must the telescope be able to switch from imaging to spectroscopy using only computer inputs, all aspects of the spectrometer s operation must also be autonomous. We have enabled switching between imaging and spectroscopy by rotating a compact spectrometer into the optical path using the telescopes filter wheel and by using an off-axis feed Compact Grism Spectrometer The compact grism spectrometers (CGS) are slitless spectrometers of simple design that are designed to be installed in the FLI CFW3-12 filter wheel of the main imaging camera. The optical design of this instrument was described in detail in Ludovici & Mutel 2017 (Submitted.) and thus we keep the description of the system here to a minimum. The CGS uses a five optical element design consisting of a collimating lens, two wedge prisms, a diffraction grating, and a focusing lens. Two 71

88 different CGS spectrometers have been designed for the IRO. The first uses two 10 wedge prisms and a 600 lpmm grating and achieves a resolution R 300 while the second produces uses a 4 and 6 wedge prism and a 300 lpmm to produce spectra of R 150. The optics for both CGS are installed within the 3D printed enclosure, and a 3D printed extension is used on the filter wheel to accommodate the length of the CGSs. These instruments, when combined with the back illuminated Apogee Aspen GC-42 back illuminated CCD on IRO, can collect spectra of 15th magnitude continuum sources and up to 19th magnitude spectral line sources with a 15 minute exposure Fiber-Fed Echelle spectrometer The second spectrometer installed at IRO is a fiber-fed echelle spectrometer. There are two major pieces of the echelle spectrometer: The fiber feed and the echelle system. The fiber feed system is not specific to the echelle spectrometer and could be utilized on any fiber fed spectrometer, including many commercially available models. The spectrometer itself is of a standard echelle design, though some minor modifications were needed to fully enable robotic operation of the instrument Off-Axis Fiber Feed Since IRO must be able to switch between imaging and spectroscopy on demand, we opted to design the spectrograph to use an off-axis fiber optic feed. Placing the spectrometer feed off-axis using a mirror allows us to switch from imaging to spectroscopy by simply off-setting the telescope pointing. Using a fiber optic cable instead of mounting the spectrometer on the telescope allows much more flexibility in terms of spectrograph design as weight and balance concerns are eliminated. Since we were already designing a filter wheel extension to accommodate the CGS system, we opted to build in an off axis pickoff mirror into the filter wheel extension (Figure 7.1). To securely affix the fiber feed to the filter wheel and to reduce the back focus 72

89 Figure 7.1: Photograph of the filter wheel used on IRO showing the CGS in the center and the pick-off mirror for the fiber feed on the right. of the fiber feed, we opted to use a flat, bolt on interface between the filter wheel and the fiber assembly. This also means that the filter wheel extension is easily adaptable to other telescope systems. Traditional fiber feeds simply place a fiber optic cable at the focal point of the telescope. While this make the mechanical design of the system simple, it poses a challenge to focusing and aligning the fiber with the science target. To aid in aligning and focusing the fiber, we have opted to design an imaging fiber feed for IRO (Figure 7.2). In an imaging fiber feed, the fiber optic cable is mounted within a mirror which is positioned at 45 to the incoming beam. While light from the science target fall into the fiber and is directed to the spectrometer, light from the surrounding field is reflected at a right angle and refocused onto a CMOS camera. This allows an observer to finely position the target on the fiber optic cable and to ensure the fiber is located at the focal plane. In our system, light from the f/6.8 beam of the telescope is directed towards 73

90 Figure 7.2: Engineering CAD model of the fiber feed used on IRO. Light enters though the filter wheel adapter (grey) and light from the science target is incident on the optical fiber (orange). Light not falling on the fiber is reflected by the mirror (green), focused by the lens (red) and is collected by the camera (blue). the fiber feed using a 12.5 mm right angle prism. This light is then incident on a 1 inch aluminum substrate mirror with a 150 µm hole drilled in the center. Mounted in the 150 µm hole is a 105 µm core, high OH fiber optic cable. This fiber is 35 m long, and runs directly to the science fiber input of the spectrometer. The fiber is monolithic and not spliced to eliminate the large losses incurred in fiber connections. After reflecting off the mirror, the now diverging light is focused by a 35mm focal length lens onto a Levenhuk C310 CMOS camera. This camera provides a live video feed to the telescope control computer and can be used to manually focus and align 74

91 Figure 7.3: Engineering drawing of the echelle spectrometer assembly. All distances are measured in mm. Light travels from the fiber feeds (yellow) to the collimating mirror (cyan), with the calibration fiber reflecting off of the motorized flip mirror (magenta) when in use. After collimation, light is dispersed out of the page by the echelle grating (orange) then cross dispersed in plane by the prism (red). Finally, the collimated light is focused by the Canon 100 mm EF lens (green) and recorded on the CCD of the camera (blue). the fiber feed. While this fiber feed was designed for use with our fiber-fed echelle spectrometer, the fiber can be attached to any instrument which utilizes a SMA fiber connection. Thus this fiber feed could be used with any number of off the shelf or custom built spectrometers Echelle Spectrometer Figure 7.3 is an engineering drawing showing all optomechanical components of the spectrometer. Two 105 µm fiber optic cables are used to input light into the spectrometer. The science fiber is connected to the off-axis fiber feed assembly and carries light from the telescope into the spectrometer. The second calibration fiber is used to inject calibration sources into the optical path of the spectrometer. 75

92 The science fiber adapter is held onto an optical post using a fixed optical mount. This optical mount is positioned mm from the collimating mirror and defines the optical axis of the system. The calibration fiber is positioned 50.8 mm up and 50.8 mm right of the science fiber and is mounted within a 5-axis kinematic mount. This mount provides tip/tilt as well as X, Y, and Z axis translational adjustments for the mounted fiber. The calibration fiber is positioned perpendicular to the optical axis of the science fiber. During observations of the science target, the science fiber will directly illuminate an off-axis parabolic mirror. The mirror has an elliptical profile, giving a 50.8 mm diameter circular profile when viewed at a 45 angle. The focal length of the mirror is mm, giving the collimator a focal ratio of f. The mirror is positioned 3 at a distance equal to its focal length from the fiber optic inputs, and thus produces a collimated beam at a right angle from the input beam. A 3 mm hole in the mirror parallel to the input beam from the science fiber allows light from the input fibers to be sampled during testing. When using the calibration fiber, a motorized flip mount moves a flat mirror into the optical path of the calibration fiber. This mirror acts to steer the light from the calibration fiber into the optical path of the science fiber. In order for the calibration fiber to be useful, the reflected beam must be aligned with the science fiber. The alignment of the two fibers is accomplished using the five-axis kinematic mount that retains the calibration fiber. When using an iterative adjustment process using a line source imaged with the spectrometer, the two fibers can be aligned to within < 5 µm (Figure 7.4). After collimation, the input light is incident on an echelle reflection grating with a ruling of 52 lines mm 1 and a blaze angle of The grating disperses the incident light vertically and the blaze angle concentrates the light into high spectral orders (Orders ). The grating is positioned in a γ configuration, meaning that the elevation angle of the grating is zero (θ = 0) while the azimuthal angle is γ =

93 546 nm 577 nm 579 nm Figure 7.4: Subset of the Hg spectrum showing three bright lines as seen using the echelle spectrometer. The spectrum from the science fiber is in red, while the spectrum from the calibration fiber is in blue. It is easy to see that the two fibers are well aligned. Since the spectral orders from the echelle grating are overlapped after leaving the grating, a cross disperser rotated 90 from the echelle dispersion direction must be used to separate the orders. A prism was chosen as the cross disperser for our spectrometer as the prism is more efficient and provides more evenly spaced spectral orders than a diffraction grating. The prism is a anti-reflection coated 60mm equilateral dispersive prism. The two dimensional spectrum then enters a 100mm Canon EF f 2 camera lens. This camera lens focuses the spectrum onto an Apogee Alta F47 camera. The F47 uses a back illuminated CCD with µm pixels. To examine the performance of the spectrometer, we used the sodium D doublet at 589 nm. The combination of camera and lens results in a wavelength scale on the CCD of 0.17 Å pixel 1 at this 77

94 wavelength. The FWHM of the of the 105 µm fiber image at the CCD is 4.4 pixels, giving the system a spectral resolution of R = 7900 at the sodium doublet Calibration Sources Calibration of high resolution spectrometers like the echelle is more difficult than for their lower resolution cousins. Simultaneous wavelength calibration of several dozen spectral orders requires a spectral line source with multiple lines in every order to be calibrated. Additionally, a continuum spectrum is needed for order extraction and for flat fielding the spectrum. Echelle spectrometers produce many overlapping spectral orders, all of which must be wavelength calibrated. Because of the large number of orders, echelle spectrometers must be calibrated with spectral line sources with a large number of lines distributed nearly uniformly across the operational range of the spectrometer. In professional astrophysical observatories, wavelength calibration is accomplished using thorium-argon cold cathode lamps. These lamps are chosen because they possess hundreds of spectral lines with a semi-regular spacing (Palmer & Engleman, 1983). However, thorium argon cold cathode lamps are prohibitively expensive ( $900) for small observatories. A low cost alternative to the thorium-argon lamp is the RELCO SC480 glow starter. The SC480 is produced commercially and houses a bimetallic strip within a glass bulb containing helium, neon, argon, as well as traces of hydrogen and oxygen (Walker & Rifferswil, 2015). The combination of elements in the SC480 produces 240 identified emission lines, making it ideal for the calibration echelle spectra. Walker & Rifferswil (2015) have produced an atlas of lines for the SC480 from 389 to 814 nm. This atlas uses spectra taken at an R = 20000, well above our spectral resolution of 7900 and thus should provide sufficient resolved spectral lines for our calibration. A 5W halogen lamp is utilized as a continuum source for order extraction and 78

95 flat field images. Halogen lamps are able to burn at higher temperatures than standard incandescents. The higher temperature of the bulb produces a Plank function with more blue and UV emission, allowing orders to be extracted down to a wavelength of 400 nm. Additionally, a red 650 nm laser diode is used to identify spectral orders to aid in data processing. While the halogen lamp is powered by a 12 V electrical supply, the SC480 requires 240 V to light. To light the SC480, we use a DC voltage converter to produce 250 V DC electrical power. The high voltage output from the converter passes through a 24 kω resister before connecting to the SC480. All three calibration sources are powered by a common 12 V power source, and will be switched on and off using a computer controlled USB relay. 7.3 Robotic Spectrometer Observations Robotic CGS Observations In theory, robotic observations with the CGS are straightforward. Using the existing control software used for imaging, the telescope can be pointed at the science target and the filter wheel is rotated to place the CGS into position. The main imaging camera can then be instructed to take a standard image to capture the target spectrum. In practice however, capturing a CGS spectrum requires more than a simple blind pointing if calibration is needed. This is due to low level aberrations due to the GCS optical design. Spectra taken off-center have slightly different scaling and the positioning of the spectrum as a function of pointing off-set is not linear. To simplify calibration, we have implemented a re-centering routine when collecting CGS spectra. After the initial blind pointing, an imaging filter is selected, an astrometric image is taken, and an astrometric pointing solution is determined for the image. The pointing of the telescope is then corrected for the offset. The option exists to perform 79

96 this pointing correction iteratively, however a single correction is usually sufficient for CGS observations. After the pointing correction is applied, the CGS is rotated into position and a spectrum is collected using the main imaging camera Robotic Fiber-fed Echelle Observations Observations with the fiber-fed spectrometer require control of two systems: the telescope mount and the spectrometer itself. In the following sections, we will discuss each system independently Telescope pointing and Fiber Alignment The first step in automating the collection of spectra from a fiber fed spectrometer is to determine the location of the fiber optic cable within the focal plane. This is particularly challenging when using an off-axis fiber feed, as the position is offset from the nominal pointing center. Using the position angle of the fiber with respect to the imaging camera and the physical offset between the pick-off mirror and optical axis, it is possible to determine a rough offset. To determine the precise position of the fiber, the telescope is pointed at a bright star, and the pointing offset is applied. Using this technique, the star can be positioned where it is visible from the fiber-feed camera. The star is then centered on the hole in the mirror, noticeable as a dark spot on a flat field image. When centered, the star will disappear from the view of the camera, and the light will be directed down the fiber optic cable to the spectrometer. After centering the star, an image is taken with the primary imaging camera, and an astrometric solution is acquired. This is used to determine the precise offset between the main imaging camera and the fiber position. This offset is recorded and used for all future pointing alignments. To center the optical fiber on the star in robotic mode, the same re-centering script developed for the CGS is used to position the target on the fiber. Instead of the pointing center, the script is instructed to use an off-set pointing. When aligning 80

97 a target on the fiber, a precision of 1 is desired, necessitating up to five iterations of the re-centering algorithm Spectrometer Automation To automate the echelle spectrometer, three devices must be automated: the camera, calibration sources, and flip mirror. The camera is automated using the same commands as is used to take images with the telescope s imaging camera. To automate the calibration sources, we developed a python script to operate the USB relay according to the manufacturer s instructions. The flip mirror was supplied with C programs from Thorlabs, and python wrappers were written to integrate them into control scripts. The spectrometer is controlled by a separate Echelle Control Computer or ECC. The ECC listens for instructions from the telescope control computer (TCC). When a spectroscopic observations is requested, the TCC sends an image request to the ECC specifying the flip mirror position, calibration lamps to illuminate (if any), and the exposure time of the observation. When complete, the ECC sends the images back to the waiting TCC where they are saved and transferred back to the University of Iowa. After receiving the image, the telescope then proceeds to the next scheduled observation. 7.4 Spectrometer Calibration CGS Calibration As with the optical design of the CGS, calibration routines for this system have already been discussed in Ludovici & Mutel submitted. We refer any interested reader to this paper for details on CGS calibration. 81

98 7.4.2 Echelle Calibration To calibrate and process a raw spectrum from the echelle spectrometer, four images are used: 1) White light image provided by the halogen bulb, 2) line image from Relco SC480 lamp, 3) monochromatic image from 650 nm red diode laser, and 4) a spectrum of a known stellar source. The first step in processing the echelle spectra is to determine the location of the spectral orders. To do this, we use image erosion improve the definition of the orders and subtract the results of a 30th percentile filter to remove the background from the image. All pixels above the noise level are then identified and grouped with their neighbors. Groups consisting of less than 2000 pixels and greater than pixels are removed in order to eliminate blended orders and small scale features. The positions of the pixels in each order are fit with a quadratic that defines the extraction region for each order. After identifying the order positions in the echellogram, spectra of the target and calibrators are all extracted. Next, we examine the monochromatic light source to perform order identification. The laser light only appears in orders 51 and 52, allowing us to identify the spectral order number easily with an automated program. After identifying the order number, a rough wavelength calibration is performed using the echelle parameters: λ c (n) = 2a cos γ sin Θ B n = K n (7.1) dβ dλ = 2 tan Θ B. (7.2) λ where λ c (n) is the central wavelength of the order, n is the order number, a is the line spacing of the grating, γ is the off-axis angle, and Θ B is the grating blaze angle. We assume the central wavelength of each order is located at the maximum 82

99 intensity of the order. This technique provides a wavelength calibration that is good to within 1 nm. To fine tune the wavelength calibration, the Relco SC480 spectrum is examined using an interactive calibration script. The use matches catalog lines with the observed lines in each order, and the wavelength calibration is automatically updated with every line selected. After all orders of the Relco spectrum are calibrated, the calibration is applied to the other observed spectra. After wavelength calibration is complete, a gain correction for each order is determined using observations of Rigel and a spectrophotometric spectrum. This gain calibration is then applied to each order. Since the lengths of the calibration and science fiber are drastically different, the absorption along the science fiber is significantly more than that of the calibration fiber. If a gain calibration is desired for the calibration sources, it is determined using the halogen bulb. 83

100 CHAPTER 8 CONCLUSIONS AND FUTURE WORK We have been successful in constructing the CGS, a low cost spectrometer which fits inside a modified filter wheel. The CGS is an unprecidented spectrometer design and serves as a valuable new teaching tool due to its abillity to collect professional quality spectra using a small telescope in a short amount of time. Since the CGS relies on hardware already installed on most robotic telescopes, the CGS is easily adapted to other observatories. The CGS is now in regular robotic observation at IRO. We have also developed a system for robotic control of a fiber-fed echelle spectrometer at IRO. While a few minor adjustments to the hardware and software are still needed to enable fully robotic observation and calibration of the echelle, these upgrades will be completed soon. Details on these upgrades are found in the following three sections. 8.1 Echelle Cross-Disperser Upgrade While the echelle spectrometer is performing well, there is significant cross-talk between the spectral orders redder than 6500 Å. Since a large portion of the CCD is currently unused, we have decided to add a second prism to the cross-disperser. By adding the second cross-disperser prism, we will double the separation between each spectral order. This upgrade will decrease cross-talk in the echelle and allow extraction of orders in the infrared. 8.2 Fiber Feed Stability Upgrade While the current fiber feed allows us to easily place a bright star on the optical fiber in live observing mode, the fiber can move independent of the main imaging camera. While the movement is small ( 10), it is significant enough to prevent robotic operation of the instrument. We have determined that this error is caused 84

101 by a wobble in the focusing mechanism. To correct this, I will install a fixed spacer in place of the focusing mechanism for the fiber feed. This will prevent movement of the fiber and enable regular robotic operation of the instrument. 8.3 Automated Echelle Calibration While the software to calibrate the echelle spectrometer is currently functioning, the software is not user friendly and wavelength calibration takes a substantial time investment. In order to make the echelle spectrometer available to laboratory students, we have decided to fully automate the echelle calibration software. To do this, we plan to have an initial calibration performed using the current software to produce a set of gain curves and wavelength calibrations. The wavelength calibrations will be applied to the new data, then the software will look for changes in the line position between the initial calibration and the current calibration. Any necessary corrections will be applied to the data, and then the spectrum will be written out in a CSV file. Students will then be able to examine this file in programs such as LoggerPro and Microsoft Excel. All observers will thus receive both the raw data, and the echelle pipeline calibrated spectra. 85

102 Part III Improving Astronomy Education Using Active Learning and Student Observatories 86

103 CHAPTER 9 INTRODUCTION The University of Iowa offers several astronomy laboratory courses to the undergraduate student population: 1) laboratories aimed at astronomy and other STEM majors and 2) non-majors courses that satisfy a general elective science credit. The non-major courses are taken by hundreds of students each semester. Since these courses fulfill a general science credit, they may often be the final science class of an individual student s academic career (Fraknoi, 2001; Lawrenz et al., 2005). As such, the effectiveness of these courses, both in teaching key astronomy concepts and developing science literacy, is especially important. Historically, all astronomy laboratory courses at the University of Iowa were taught in the traditional cookbook format where students followed step-by-step instructions to complete the activity. While the laboratory activities were well designed in terms of science content, students were not retaining the concepts and process of science emphasized in these activities. To improve student outcomes in the laboratory and associated lecture, we sought a new pedagogical approach to teaching the astronomy laboratory courses. 9.1 Fostering Active Learning at the University of Iowa Active Learning is a broad teaching approach that can be defined as any instructional technique that actively engages students in the learning process. In an active learning environment, students work on activities and reflect upon the process involved in completing those activities (Bonwell & Eison, 1991). These activities can be as simple as a single class discussion in a traditional lab, or transforming the entire laboratory exercise into the active format. Included under the broad topic of active learning are collaborative learning and problem-based learning (PBL). Collaborative learning can broadly be defined as any method of instruction where students work together in small groups (Prince, 2004). PBL involves presenting relevant problems 87

104 at the beginning of class that provide context and motivation to the learning objectives (Prince, 2004). Astronomy laboratory courses, where students often conduct experiments in groups, are well suited to integration of these teaching methods. In general, it has been shown that inquiry-guided approaches to student learning (i.e. group problem-solving during class) improves learning outcomes compared to traditional lectures (Singer et al., 2012). In Freeman et al. (2014), all well studied STEM disciplines show statistically significant effect sizes for the improvement in examination scores and failure rates when lectures are taught in an active format. Additionally, physics showed the largest improvement when implementing active learning. Since active learning improves learning outcomes in STEM fields and has the greatest effect on the closely related field of physics, it is likely that active learning will also be effective in astronomy labs. 9.2 Goals of this Work The astronomy laboratory activities have been undergoing a series of changes and additions since the fall of In this work we summarize the current status of the laboratory development by Describing the changes made to the astronomy laboratory courses at the University of Iowa as part of our effort to create active learning laboratory activities for these courses. Presenting an inventory of the astronomy laboratory activities and our observations of the current state of the activities. Offering suggestions aimed to guide future development of these activities. The changes made to the laboratory activities focus on three main areas: fostering collaborative learning, using active learning to promote an understanding of astronomy concepts, and creating hands-on laboratory exercises. The laboratory inventory will focus on examining the teaching techniques and equipment utilized within 88

105 each activity as well as the inquiry level of the activity and learning objectives. 89

106 CHAPTER 10 ACTIVE LEARNING ASTRONOMY LABORATORIES AT THE UNIVERSITY OF IOWA 10.1 Developing a Collaborative Learning Environment Prior to the introduction of active learning techniques, the astronomy laboratory setup did not foster student collaboration. Although students worked in groups of two, each student was responsible for their own laboratory sheet. This encouraged students to share the data required to complete the laboratory activity but not collaborative problem solving. The laboratory rooms themselves were not conducive to larger group discussions. Each group of two was placed at a cubicle with walls that partially blocked their view of their peers. This line of sight problem also extended to the teaching assistant (TA) in charge of the laboratory section. If students did not sit upright in their chairs, the TA may not be able to see the students. The physical barrier of the cubicle walls inhibited the TA from creating class discussions, as students could not always see their classmates speaking. Table 10.1: Comparison of original and updated laboratory formats Original Design Updated Design Desks Cubical Rounded Desks Group Size Two Students Three Students Reports Individual Reports Shared Group Report Pre-Lab None Collaborative Quiz and Discussion Class Discussions None Facilitated by TA Lab Procedure Step by Step Instruction Technique Instruction Telescopes Not used Local and Remote Observing To encourage students to collaborate within their own groups and between groups, we redesigned both the physical laboratory environment as well as the structure and content of the laboratory activities themselves (Table 10.1). Students are 90

107 now placed in groups of three, who submit a single laboratory worksheet for which all group members are equally responsible. Our aim was to encourage students to discuss problems and solutions with their group members. By giving one grade for the entire group, the grade of each student hinges on the success of their peers. If students engage with each other to discuss the laboratory activity, no one individual in a group will lag behind the others. To extend the collaborative environment beyond the student groups, the two seat cubicles were replaced with curved tables capable of comfortably seating three to four people (Figure 10.1), thus removing the physical barriers to a group discussion. The laboratory activities were also re-designed to encourage class discussions. In the original laboratory activities, students immediately began the activity after a brief introduction to the material by the TA, preventing the TA from evaluating the skill level of the students. In the new activities, class begins with a pre-lab quiz. This quiz is graded for completion, not correctness and consists of questions designed to stimulate discussion on the subject of the laboratory activity. After the quiz, the class discusses the questions and answers, with students explaining the reasoning behind their answers. These quizzes not only start a discussion on the activity topic, but also offer a valuable opportunity for the TA to check for student misconceptions. We have also worked to encourage students to discuss their answers and reasoning with their classmates. The combination of the pre-lab quiz and class discussion serves to have students discuss their answers and reasoning with their classmates. In addition to the pre-lab quizzes, TAs are encouraged to start group discussions when reaching portions of the laboratory activity where students typically have difficulty. In some activities, groups rely on data analyzed by other groups to compete the final questions of the worksheet. In others, the goal was for all groups to report on their findings and discuss sources of error at the end of the activity. 91

108 Figure 10.1: Image of the astronomy laboratory after the cubicles were replaced with curved desks Encouraging an Understanding of Concepts To switch the emphasis of the astronomy laboratories from obtaining correct answers to understanding the process, we restructured the exercises within each laboratory activity. Ideally, students are not given cookbook step-by-step instructions. Instead, students are provided with the techniques and methods needed to complete the activity. After being introduced to these tools, some simple questions are asked that allow the students to practice using the techniques introduced to them. After becoming acquainted with the techniques needed for the laboratory activity, students are given a PBL exercise. The PBL exercise requires the use of the previously practiced techniques, but does not include explicit instructions on completing the task. The goal is that students understand how to apply a skill to the assigned problem. For example, in the Angular Size activity, students are first presented with the definition of angular size and asked to determine the angular size of objects in the classroom by measuring their linear size and the distance to the 92

109 object. Students are then challenged to construct a device that can make angular size measurements of distant objects. They then use the device they built to measure the angular size of the Old Capitol dome from the roof of Van Allen Hall. This task requires them to understand important features of angular size and apply them in a different manner from when they practiced the technique at the beginning of lab. A before and after comparison of this laboratory activity is located in Appendices B and C for the interested reader Solidifying Concepts through Hands-on Observing Slater (2008) states that It is generally accepted that most undergraduate science majors taking an astronomy course spend little to no time actually looking through a telescope. This was true of the cookbook laboratory activities at the University of Iowa as well. Students completed nearly all laboratory exercises using only a computer. While students occasionally were able to utilize the Iowa Robotic Observatory (IRO) for obtaining images, the system is the epitome of a black box in the laboratory. The robotic nature and remote location of IRO prevented students from investigating or understanding the instrument used to complete their activity. Adding to the challenge of using IRO, students could not determine if their observations were successful before lab. When observations failed, students were simply given stock images to use with no opportunities for reflection or troubleshooting. Jacobi et al. (2008) suggests that overall achievement in an astronomy course may be correlated with the number of clear nights available for observing. To take advantage of the potential benefits of clear sky observing, we have begun incorporating the Van Allen Observatory (VAO), located on the roof of Van Allen Hall, into the astronomy laboratory exercises. Students are able to control the roof-top observatory to gather data for their laboratory exercises. While one group is using VAO, other groups are able to use small telescopes located on the roof for eyepiece observations. Currently, three laboratory activities have been designed around the use of VAO: 93

110 Introduction to Observing with VAO, Observing the Sun, and Introduction to Stellar Spectroscopy. Activities that take advantage of the smaller rooftop telescopes include Introduction to Telescopes and Observing the Night Sky. These activities aim to help students apply the basic knowledge gained in the lectures to real observing conditions. We are also currently experimenting with a new approach to astronomy majors laboratory courses. This new approach to the course begins with four to five introductory laboratory activities to teach the basics of astronomical observation. Students then choose from a list of astronomical research projects, and use all remaining laboratory sessions to complete the chosen project. All the potential research projects have been designed to be useful to current research in astronomy. Results from the projects will either be added to databases of observations (eg. monitoring observations of asteroids), be published in journals like the Journal of the American Association of Variable Star Observers, or contribute to the operation of IRO and VAO. This course structure is similar to other Course-based Undergraduate Research Experience (CURE) courses. Our hope is to improve student motivation and impression of astronomy as is seen in CURE courses in other disciplines (Kerr & Yan, 2016). 94

111 CHAPTER 11 ASTRONOMY LABORATORY INVENTORY The transition to active learning laboratory activities is currently a work in progress and is being undertaken by multiple individuals. During the continued development of these activities, it is important to understand the current state of the laboratory activities and assess where improvements are still needed. To help focus the continued development of these laboratory activities, we have completed an inventory of the astronomy laboratory activities and present observations on the state of the laboratory activities. We then present initial conclusions after looking at the inventory across several parameters. The astronomy laboratory inventory does not evaluate the educational value of the activities, but simply documents which techniques and learning objectives have been incorporated into the laboratory activities. The inventory is separated into four primary sections: 1) laboratory equipment, 2) teaching techniques, 3) broad learning objectives, and 4) inquiry level. In the following sections, we describe the inventory items in each section and present the results for each inventory section. We then discuss the results of this inventory and provide suggestions for improvement of the laboratory activities Laboratory Activities Examined For this inventory, we consider most of the laboratory activities designed for the Exploration of the Solar System (ESS) and Stars, Galaxies, and the Universe (SGU) laboratory courses. Those activities not considered were removed from consideration because they are either not completed or have been superseded by a new activity. Properties of Supernovae has been left out because no worksheet is available on the webpage, and thus we are unable to complete most of the inventory items. Astronomical Spectroscopy is a old version of the Introduction to Spectroscopy laboratory 95

112 and is not considered as it has been replaced and is no longer taught as part of the laboratory curriculum. Additionally, some activities are considered twice in our inventory. This is because these laboratory exercises utilize separate worksheets for daytime and nighttime laboratories. In these cases, the daytime and nighttime versions of these laboratory activities are considered separately, as significant differences may be present between these two versions Groupings of Laboratory Activities The laboratory activities present on the Imaging the Universe laboratory website 1 are divided into three categories: Foundational Labs, Observational Labs, and Advanced Labs. Typically, astronomy laboratory courses will begin with Foundational Labs and then progress on to the other categories after several weeks. The exact curriculum taught is different each semester, based on the professor and TA associated with the course. Foundational labs are designed primarily to introduce students to the tools and techniques they will need for observational and advanced laboratory activities. The difference between the observational labs and advanced labs is less clear. The observational labs are a mix of activities that utilize naked eye and telescopic observations using an eyepiece and digital analysis of CCD images. The advanced labs focus mainly on laboratory activities that utilize digital data analysis and CCD imaging from the VAO or IRO telescopes. In the astronomy laboratory activity inventory, labs are grouped for reference Laboratory Equipment In the astronomy laboratories, three main methods of equipment usage are employed. Class demonstrations utilize a single experimental setup operated by the TA. This setup is typically located in the front of the classroom and is viewed by all

113 Table 11.1: Inventory of equipment components used in the astronomy laboratory activities. Abbreviations for equipment components are as follows: class demonstrations (CD), group experiments (GE), dark sky observing (DSO), live CCD imaging (LI), and robotic observing (RO). Components marked with a L are contained in the laboratory activity while components marked with a T are contained only in the TA guide for the activity. Activities with blank rows have no equipment components. Lab Activity CD GE DSO LI RO Foundational Labs Intro. Active Learning (SGU) T Intro. Active Learning (ESS) T Obs. The Night Sky (Day) T Obs. The Night Sky (Night) T L Angular Size L Measuring the Sky L Intro. Spectroscopy L Intro. Telescopes (Night) L L Intro. Telescopes (Day) L Parallax L Intro. Obs. With Gemini L Intro. Image Analysis L Intro. Stellar Spectroscopy L L Image Analysis (Solar System) L Classifying Galaxies Observational Labs Observing the Sun L Observing the Sun with VAO L Observing the Giant Planets Observing Lunar Features L Observing Comets The Moon L Tracking Solar System Objects L Observing with VAO Advanced Labs Independent Research Projects L Spectra of Wolf-Rayet Stars L Photometry of a Globular Cluster Properties of Nebulae Astronomical Red-shift L Stellar Photometry The Mass of Jupiter Near-Earth Asteroids L L Main Belt Asteroids L L Exoplanet Discovery L Eclipsing Binary Stars L 97

114 students simultaneously. These demonstrations are meant to spur discussion amongst the students and demonstrate principles important to the laboratory activity. Group experiments in contrast are equipment setups utilized by each individual group. Each student group completes the portion of the lab utilizing the setup on their own. Finally, groups may utilize naked eye and telescopic observations of the night sky to complete their activities. We divide these telescopic observations into three categories: Dark sky observations which consist of naked eye observations and observations made using small telescopes utilizing an eyepiece, live CCD imaging where students use the cameras on VAO or IRO in real time, and robotic observing where students submit observation requests to the IRO telescope. In Table 11.1 the inventory of equipment is shown for all laboratory activities. Entries marked with an L are present within the laboratory worksheet or lab manual while those marked with a T are only found within the TA manual for the activity and is thus used at the discretion of the TA. The eight activities with no entries do not utilize equipment. In total, only seven of the thirty-six laboratory activities utilize in class demonstrations led by the TA. Labs utilizing class demonstrations are split almost evenly between foundational activities and advanced activities. Ten activities utilize group experiments. Like the class demonstrations, this is split almost evenly between foundational activities (6) and advanced laboratories (4). There are no class demonstrations or group experiments utilized within the observational labs. Only 12 out of the 36 activities utilize any dark sky observations or CCD telescopic imaging with VAO or IRO. With the exception of the Independent Research Projects activity, none of the advanced laboratory activities utilize telescopic observations. The activities utilizing dark sky observations, VAO imaging, or IRO imaging are split evenly between the foundational activities and observational activities. It should be noted that if a TA was dedicated to conducting telescopic observations, most of the observations used for the advanced lab activities could be easily conducted using IRO. However, after reviewing the IRO telescope observation logs, 98

115 it appears the current TAs are not making use of the instrument Teaching Techniques Table 11.2 shows the results of our inventory of teaching techniques in the astronomy laboratories. Our goal was to catalog some of the teaching techniques that are often employed in active learning environments. The teaching techniques considered in this inventory are brainstorming activities, misconception checks, hypothesis formation, conducting online research, class discussions, and reporting results. Brainstorming activities ask students to work in a group and write down facts or questions they have without researching the topic. Misconception checks ask the students to complete questions that probe their understanding of the activity and provide the TA an opportunity to intervene and address common misconceptions. Hypothesis formation questions ask the students to make a prediction based off of their current understanding of a phenomenon. Online research questions ask the students to conduct research on a topic of which they possess no prior knowledge, class discussions are guided by the TA and involve all students in the course, while reporting results is the aggregation and utilization of experimental results between laboratory groups. Students are only asked to complete brainstorming activities as part of four of the foundational labs. Hypothesis formation is more prevalent than brainstorming, though it is only present in seven of the thirty-four laboratory activities. In all of the laboratory activities, misconceptions are evaluated using a pre-lab quiz. While all laboratory activities with the exception of Spectra of Wolf-Rayet Stars have a pre-lab quiz, not all have questions that probe student misconceptions. Ten of the thirtythree labs with quizzes utilize questions that do not directly address the material contained in the lab or have obviously wrong answers. Despite the remodeling of the laboratory rooms to promote group discussions, only two laboratory activities (Introduction to Stellar Spectroscopy and Independent Research Projects) includes group discussions or the sharing of results between the 99

116 Table 11.2: Inventory of teaching techniques used in the astronomy laboratory activities. Abbreviations for teaching techniques are as follows: brainstorming activities (BA), misconception checks (MC), Hypothesis Formation (HF), On-line research (OR), class discussions (CD), and Reporting Results (RR). Components marked with a L are contained in the laboratory activity while components marked with a T are contained only in the TA guide for the activity. Lab Activity BA MC HF OR CD RR Foundational Labs Intro. Active Learning (SGU) L L L T Intro. Active Learning (ESS) L L L T Obs. The Night Sky (Day) L T Obs. The Night Sky (Night) L T Angular Size L L Measuring the Sky L Intro. Spectroscopy L L Intro. Telescopes (Night) L Intro. Telescopes (Day) Parallax L L Intro. Obs. With Gemini L L Intro. Image Analysis L Intro. Stellar Spectroscopy L L L L Image Analysis (Solar System) L Classifying Galaxies T T Observational Labs Observing the Sun L L Observing the Sun with VAO L L Observing the Giant Planets L Observing Lunar Features L Observing Comets L The Moon L Tracking Solar System Objects Observing with VAO L Advanced Labs Independent Research Projects L L L Spectra of Wolf-Rayet Stars Photometry of a Globular Cluster L T Properties of Nebulae L T Astronomical Red-shift L T Stellar Photometry L The Mass of Jupiter Near-Earth Asteroids Main Belt Asteroids L Exoplanet Discovery L L Eclipsing Binary Stars 100

117 student groups. If including activities where the TA manual mentions group discussions, this number is increased to ten total activities Primary Focus While most of the laboratory activities emphasize different specific learning goals, these topics addressed fall into several broad categories of big ideas: Night Sky Observation These activities promote an understanding of the night sky. Usually, this focuses on activities that teach skills necessary for naked eye observations such as celestial coordinate systems, navigating constellations, and the movement of stars and solar system objects in the night sky. Telescopic Observation Laboratory activities that concentrate on setting up and observing with small visual telescopes as well as CCD imaging with VAO and IRO. Digital Data Analysis Students learn various techniques to analyze digital images and spectra. The Changing Universe These activities focus on the common misconception that the night sky is static. This broad category deals with temporal variability in the form of both moving objects and variable objects. The Nature of Light A firm understanding of light is essential to a strong understanding of astronomy. These activities aim to teach students about light, colors, spectra, absorption, emission, and the Doppler Effect. The Structure of the Universe Laboratories in this category emphasize the large scale structure of the universe, including individual galaxies, groups, and clusters. Size Scales The size scales encountered in astronomy are often so large that students cannot relate to them. These activities either relate size scales to more 101

118 regularly encountered scales or to the size of other known astronomical systems, such as the solar system. Gravitation and Orbits This big idea focuses on applications of Kepler s Third Law and orbital motion. Table 11.3 shows which laboratory activities address each big idea. While most of these big ideas are covered by many laboratory activities, The Nature of Light, The Structure of the Universe, and Gravitation and Orbits are only addressed by a few activities Inquiry Level Fay et al. (2007) developed a rubric to examine the inquiry level of chemistry laboratory activities. This rubric, though designed for chemistry, is easily transferable to astronomy laboratory activities. This rubric can be easily understood from Table 11.4 (Table 3 of Fay et al. (2007)). The level of inquiry is defined by which of the problem, procedure, and solution is constructed by the student. The results of our inventory of inquiry level in the astronomy laboratory activities can be seen Table The inquiry levels listed in Table 11.5 represents the highest inquiry level found in the activity. It should be noted that these levels were determined by a single individual who was involved with the development of the laboratory activities and should be used only to guide future improvements to the laboratory activities and not as final evaluation of the activity s inquiry level. The only activity that had the problem constructed by the student is the Individual Research Projects activity. Very few activities utilize student developed procedures for completing the laboratory exercises, with only seven of the thirty-four requiring students to determine which method is appropriate for the assigned experiment. In many laboratory activities the students are provided with the solution to the question either before or immediately after examining their data. Exactly half of the laboratory activities do not provide the solutions to the students. 102

119 Table 11.3: Inventory of the primary focus emphasized within the astronomy laboratory activities. Abbreviations for learning objectives are as follows: night sky (NS), telescope observation (TO), digital data analysis (DDA), the changing universe (CU), the nature of light (NL), the structure of the universe (SU), size scales (SS), and gravitation and orbits (GO). Lab Activity NS TO DDA CU NL SU SS GO Foundational Labs Intro. Active Learning (SGU) X Intro. Active Learning (ESS) X Obs. The Night Sky (Day) X Obs. The Night Sky (Night) X Angular Size X Measuring the Sky X Intro. Spectroscopy X Intro. Telescopes (Night) X X Intro. Telescopes (Day) X Parallax X X Intro. Obs. With Gemini X Intro. Image Analysis X Intro. Stellar Spectroscopy X X Image Analysis (Solar System) X Classifying Galaxies X Observational Labs Observing the Sun X X X X Observing the Sun with VAO X X X X Observing the Giant Planets X X Observing Lunar Features X X Observing Comets X X X The Moon X X Tracking Solar System Objects X X X X Observing with VAO X Advanced Labs (Variable) Independent Research Projects Spectra of Wolf-Rayet Stars X X Photometry of a Globular Cluster X X Properties of Nebulae X Astronomical Redshift X X X X Stellar Photometry X The Mass of Jupiter X X X Near-Earth Asteroids X X Main Belt Asteroids X X Exoplanet Discovery X X X Eclipsing Binary Stars X X 103

120 Table 11.4: Table 3 from Fay et al. (2007). Level Problem Question Procedure Method Solution 0 Provided to student Provided to student Provided to student 1 Provided to student Provided to student Constructed by student 2 Provided to student Constructed by student Constructed by student 3 Constructed by student Constructed by student Constructed by student Table 11.5: Inquiry level of each activity according to the (Fay et al., 2007) rubric. Lab Activity Level Lab Activity Level Foundational Labs Observational Labs Intro. Active Learning (SGU) 2 Observing the Giant Planets 0 Intro. Active Learning (ESS) 0 Observing Lunar Features 2 Obs. The Night Sky (Day) 0 Observing Comets 1 Obs. The Night Sky (Night) 0 The Moon 0 Angular Size 2 Tracking Solar System Objects 0 Measuring the Sky 0 Observing with VAO 0 Intro. Spectroscopy 1 Advanced Labs Intro. Telescopes (Night) 1 Independent Research Projects 3 Intro. Telescopes (Day) 0 Spectra of Wolf-Rayet Stars 0 Parallax 2 Photometry of a Globular Cluster 0 Intro. Obs. With Gemini 0 Properties of Nebulae 0 Intro. Image Analysis 1 Astronomical Redshift 0 Intro. Stellar Spectroscopy 1 Stellar Photometry 0 Image Analysis (Solar System) 1 The Mass of Jupiter 0 Classifying Galaxies 0 Near-Earth Asteroids 1 Observational Labs Main Belt Asteroids 1 Observing the Sun 1 Exoplanet Discovery 2 Observing the Sun with VAO 1 Eclipsing Binary Stars 2 104

121 11.3 Inventory Evaluation Laboratory Curriculum Strengths The astronomy laboratory activities at the University of Iowa cover a wide range of important astrophysical concepts. Nineteen of the thirty-four activities include either class demonstrations, individual student experiments, or dark sky observing. A strong emphasis on digital data analysis with real telescope data allows students to utilize modern astronomical research techniques used by professional astronomers. By emphasizing size scales and changes in the night sky, students become familiar with the temporal and spatial scales of the universe. The foundational laboratory activities provide the necessary introduction needed to enable students with no prior astronomy experience to be successful in the laboratory course. The pre-lab quizzes in most activities enable the TA to assess and address misconceptions prior to students beginning work on the day s exercises Identified Unmet Needs Few of the current activities emphasize the nature of light. Additionally, the activities which do cover this topic are typically taught later in the semester. Emphasizing activities that discuss the nature of light early in the semester may help students to more easily grasp the material in later activities, such as understanding filters in the Introduction to Observing with VAO activity. Presently, most astronomy laboratory activities are limited to inquiry level 0 105

122 and 1 based off of the Fay et al. (2007) rubric, though this may be by design in some activities. The Van Allen Observatory and Iowa Robotic Observatory are severely underutilized. While most of the advanced and observational activities could make use of these instruments, no structure is in place to make observation planning easy for students. Additionally, students are not trained to check the quality of their images before class, increasing the chances that students will not obtain suitable observations before class. Some professors prefer eyepiece and dark sky observing to imaging with digital detectors. In this case, few activities exist which allow students to make scientific measurements without digital data analysis. The creation of dark sky observing activities associated with even simple measurements could be beneficial. The topics of galaxies and large scale structure receive little attention in the current activities despite the importance of this topic in current astrophysical research. Gravitation and orbits are only examined in one activity. Additional activities covering this topic should be added if this topic is to be emphasized in the laboratory or lecture course Challenges The climate of Iowa City limits the number of clear nights for dark sky or VAO observing. Predictability is also often a problem, as haze and thin clouds can greatly obscure astronomical objects when combined with the light pollution of Iowa City. While the Exploration of the Solar System laboratory courses are scheduled at 106

123 night, the Stars, Galaxies, and the Universe laboratory courses are scheduled during the day. This means that clear sky observing is limited to the Sun and Moon unless students are required to attend an observing session outside of the regular laboratory time. Technical problems with VAO and IRO can cause difficulties during activities where students rely on these instruments. TAs receive little training concerning VAO and IRO. This encourages TAs to forego observations with these instruments in favor of images that were collected earlier and are guaranteed to produce successful results. Laboratory students do not have access to classrooms outside of class, making it difficult for them to ensure their observations were successful before class. This means students cannot re-attempt failed observations and learn from their mistakes. The curriculum is highly variable from semester to semester as each TA and professor chooses the labs they prefer. In addition professors and TAs will make changes to the activities in order to make them match their particular teaching style. This is a challenge because some of the activities have been almost completly rewritten over the last three years due to professors having very different expectations for the same activity topic. The laboratories thus never reach a final state and new problems are found everytime the activity is taught. There is often little coordination between the TA and professor in charge of the course. Oftentimes, the TA chooses the laboratory activities to be taught. While this sometimes works out well, the laboratory often gets out of sync with the astronomy lecture course. 107

124 11.4 Suggestions for Improvement In this section, we present several suggestions that could improve the current laboratory activities as well as assist in development of new activities that could address the unmet needs of the laboratory courses. These suggestions are meant to guide future developers of the laboratory course materials or assist future evaluators of the material. To discriminate between the activities that are designed around dark sky observing and those utilizing digital data analysis, we suggest defining the observational laboratory activities as those that focus primarily on dark sky observations and the advanced laboratory activities as those focusing on digital data analysis. This would allow TAs and professors to easily discern between these two types of laboratory activities when selecting activities for their courses. With the current selection of laboratory activities, this would leave few activities in the observational laboratory category, though we recommend that this group of laboratory activities be a focus for the next round of activity development. Many of the digital data analysis activities could serve as models for future observational laboratory activities. If students were able to practice using setting circles in prior activities, new activities utilizing small telescopes to look at asteroids, nebulae, and galaxies could be implemented. We also suggest providing TAs with additional training for VAO and IRO. This additional training would encourage TAs to utilize these facilities as part of their courses. To further encourage TAs to utilize the observatories, a free and easy to use FITS image viewer should be available to the laboratory students to enable them to verify their observations prior to class. Additionally, observing guides should be provided in the advanced laboratory activities to guide students and TAs. Finally, we suggest that additional TA manuals for the laboratories be developed and provided to new TAs. These manuals contain suggestions for class discussions, illuminate common misconceptions, and provide TAs with a strong starting point 108

125 when preparing to teach an activity. While the TA manuals have been developed for the most commonly taught activities, most of the laboratories either do not have a manual or the current manuals need expansion. These manuals have the potential to increase TA skill in the laboratory courses, thus improving student learning as well. 109

126 CHAPTER 12 CONCLUSIONS The astronomy laboratory courses have undergone many changes in order to create the active learning activities used today. Some of these changes were physical changes to the classrooms, such as removing physical barriers to group discussions. Most involve the introduction of active learning teaching techniques such as class demonstrations, shared group responsibility, and class discussions. Our inventory of the laboratory activities have revealed that the current activities are succeeding in many areas, however some weaknesses are still present. We have made several suggestions for the improvement of existing activities. Guidance for the development of future activities has also been provided. Combined, these suggestions alongside the inventory offer a guide for the future of astronomy laboratory activity development at the University of Iowa. 110

127 APPENDIX A PROPERTIES OF AN ECHELLE GRATING A.1 Dispersion of an Echelle Grating To calculate the properties of an echelle grating, we will begin with the generic form of Equation 5.1: (sin α + sin β(λ)) cos γ = nλ a. (A.1) Where α is the angle of incidence, β is the diffraction angle, and γ is the off-axis angle. To calculate the dispersion of an echelle grating, we take the derivative with respect to λ d d sin β(λ) = dλ dλ nλ a cos γ sin α (A.2) dβ dλ = n a cos γ cos β(λ). (A.3) By substituting Equation A.1 into Equation A.3 we obtain dβ dλ sin α + sin β(λ) =. (A.4) λ cos β(λ) This leads us to a curious property of the echelle grating: the spectral resolution of the echelle grating is independent of the line spacing of the grating and order number. If we then choose to use the echelle grating in the Littrow configuration where the angle of incidence and angle of diffraction is equal to the blaze angle of the grating (α = β = Θ B ) we find that 111

128 dβ dλ = 2 tan Θ B. (A.5) λ Thus, in a Littrow configuration the resolution of a echelle grating depends only on the blaze angle. A.2 Central Wavelength of an Echelle Order Continuing with our assumption that the echelle grating will be operating in the Littrow configuration, it is now possible to solve for the central wavelength of each echelle order (λ c (n)) using Equation A.1. λ c (n) = 2a cos γ sin Θ B n where K is known as the echelle constant. = K n (A.6) 112

129 APPENDIX B ORIGINAL ANGLES AND PARALAX LABORATORY ACTIVITY The following is a copy of the astronomy laboratory activity, Angles and Paralax. This appendix is taken from the 2007 edition of the Imaging the Universe laboratory manual used at the University of Iowa before the update to active learning activities. 113

130 θ θ θ θ θ θ θ 114

131 θ θ π θ ¼ θ 115

132 116

133 θ φ φ φ φ φ φ θ 117

134 Δ Δ δ 118

135 θ 119

136 φ φ φ φ θ φ φ θ θ 120

137 Δ δ Δ θ 121

138 APPENDIX C NEW ANGULAR SIZE LABORATORY ACTIVITY The following is a copy of the astronomy laboratory activity, Angular Size. This appendix is taken directly from the University of Iowa s Imaging the Universe website during the spring of 2017 (see foundational-labs/angular-size/). 122

139 Angular Size Learning Goals: The goal of this lab is understand how an object's distance, physical size, and angular size relate to one another, and to learn how astronomers use this relationship to determine the sizes of distant objects. Challenge: Using only the provided materials and your understanding of angular size, your team will develop a method for estimating the diameter of the Old Capital building's dome. Tutorials: none Background: Resources: Worksheet, Google Maps, meter sticks, tape measure, rulers, calculators, paper, pens Terminology: small-angle formula, percent error formula Whenever you look at an object, you are measuring its angular size - the amount of space it takes up in your field of view in degrees, minutes and seconds (or radians if you're mathematically-inclined). You can't directly measure an object's size in centimeters or inches unless you walk up to it and use a ruler. You know that faraway objects look small and nearby objects look big, so your brain puts together an object's angular size with your guess as to its distance to give you an idea of its actual size. Humans have evolved binocular vision to help us make these distance guesses for things that might affect our survival, such as bears or lions. We also compare sizes of known objects, such as buildings and trees, when they are near the object in question. In astronomy, distances are more uncertain and objects that appear close in the sky may be many light-years apart. Thus, our basic measurement of size in astronomy is angular size. In order to learn the true physical size of an object, one must find the distance to the object by some independent method. Conversely, if the physical size of an object is known this can be combined with its apparent angular size to determine its distance. In both cases the desired quantity can be calculated from the others using the small-angle formula. Complex and precise instruments exist and can be constructed for measuring the angular size of objects, but a set of rough measurement tools can be found at the end of most people's arms. Because humans are built to mostly the same proportions, if you hold your arms outstretched with your palms facing forward, your hands will have about the same angular size in your field of vision regardless of whether you are tall, short, big or small. Your fingers and knuckles can be used to make rough measurements of angular sizes and distances on the sky as shown in the diagram to the right. Other useful angular size rulers exist as well. For example, the moon is almost exactly one-half degree in extent as viewed from the surface of the Earth. 123

140 Pre-Lab Quiz: Angular Assessment Use the image above to complete the quiz. Discuss each question with your team and be prepared to explain the reasoning behind your team's answers after the quiz. 1. Assuming the balloons are the same size, which is closest to the observer? 2. Assuming the balloons are the same size, which is farthest away? 3. Still assuming the balloons are the same size, where are B and E in relation to each other? B is four times closer than E B is eight times closer than E B is twice as close to the observer as E E is twice as close to the observer as B B and E are the same distance from the observer 4. Now suppose C is twice as big as D. If C is 2 miles away, how far is D? 2 miles 4 miles 1 mile 1.5 miles impossible to say 124

141 Part 1: Calculating an Angle As you are aware from the background section of this lab (and hopefully your own experience), an object's angular size depends on its physical size (in feet, meters, etc.) and its distance. Since you will make a device for measuring angular size in this lab, you need something of a known angular size to compare it with. Large Distances and Small Angles Using trigonometry, if you know the length of two sides of a right triangle, you can find the angles. Suppose the building in the picture to the right is 50 feet tall and 30 feet away from you (A=50', B=30'). What would its apparent angular size be in degrees? If the object is farther away (i.e. B is much bigger than A), as is often the case in astronomy, you get a long, skinny triangle and a small apparent angular size. This makes the math easier, because for small angles the tangent of an angle is approximately equal to the angle itself (measured in radians). This is the basis of the small angle formula. Pick a small angle (less than 10 degrees) and show that this is true. Use trigonometry or the small angle formula to find the angular size of something in the lab. Is the small angle formula precise enough to use for the object you are measuring? Does the angular size depend on where you stand in relation to the object? Part 2: Creating an Angular Measurement Tool In this part of the lab you will design, build and test a device for measuring the angular size of objects. As discussed in the introduction, you can estimate angular sizes using your fingers, the moon, etc. However, for the measurement you will make in this lab you want to create something that is capable of greater accuracy than these rough methods, although it need not be as fancy as the tool shown in the image here! Constructing Your Measurement Device You will have access to paper, pens, paper clips, rulers, tape, and other materials you can use to construct an angular measurement device. You want to make something that a person can hold up in their line of sight and compare to distant objects to find their angular size. Your device should have markings of a degree or less for accuracy. HINT: You found the angular size of an object in the lab in the last section. You can use this to calibrate and test your device. Estimate how precise your measuring device is. Does it measure angles within a half-degree? One degree? This will help you determine whether your result is consistent in the next part. 125

142 Part 3: Measuring the Old Capitol Dome Collecting Data The instructor will escort the class to the roof of Van Allen Hall, where the Old Capitol building dome can be seen. Use the devices you constructed in the last section to measure the angular size of the dome at its base. Find the diameter of the dome in meters. As you should understand at this point, you need to combine the angular diameter with the distance to the dome to find its physical size. How or where could you find the distance from the roof to the Old Capitol Dome? Analyzing Your Results At some point, your instructor will tell you the actual diameter of the dome. Use the precision you estimated for your instrument to find the maximum and minimum possible values for the physical size. Does the actual size fall within these values? What does this tell you about the precision of your instrument or the accuracy of your value of the distance to the dome? 126