The 2007 RBSE Journal · The visible part of the Sun is referred to as the photosphere. Here, large...

The 2007 RBSE

Journal

The RBSE Journal i 2007 V1.9

The RBSE Journal 2007 The RBSE Journal is an annual publication that presents the research of students and teachers who have participated in the Teacher Enhancement program TLRBSE, “Teacher Leaders in Research Based Science” at the National Optical Astronomy Observatory in Tucson. This program, funded by the NSF, consists of a distance learning course and a summer workshop for high school teachers interested in incorporating research and leadership mentoring within their class and school. TLRBSE brings the research experience to the classroom with datasets, materials, support and mentors during the academic year. The journal publishes papers that make use of data from the TLRBSE program, or from its related programs such as New Mexico Skies and the SPITZER teacher observing program. These papers represent a select set of those submitted for publication by many students. All papers are reviewed both by the Senior Editor and the Astronomer responsible for the particular research project. In addition to papers representing classroom work, a number of these papers are based on observing done at part of the Teacher Observing Program (TOP) at Kitt Peak during the fall and winter. More information about both the TLRBSE and the TOP program can be found on the website, www.noao.edu/outreach/tlrbse I want to thank Dr. Travis Rector, Dr. Connie Walker, Dr. Steve Howell, and Dr. Jeff Lockwood for their generous help in reviewing these papers and working with the young scientists. Special thanks are due to Kathie Coil for her efficient editing of the final copy. Dr. Katy Garmany Senior Editor, RBSE Journal

The RBSE Journal ii 2007 V1.9

Table of Contents

SOLAR Magnetic Felid Fluctuations Before, During & After Flaring in NOAA AR 10930 ........................ 1 Caitlin DeWolf, Rebecca Mortensen, and Shelia Strong Chippewa High School, Remus, MI Teacher: Cris DeWolf, TLRBSE 2006 Characterization of NOAA AR 10930 .................................................................................................. 14 John Kucharczyk and Shuo Qiu Manhasset High School, Manhasset, NY Teacher: Peter Guastella, TLRBSE 2006 The Relationship Between the Area and the Magnetic Field Strength of Sunspot NOAO Active Region 10905 ........................................................................................................................................... 22 Brian Fisher, Isabel Garcia, Juan Vinagera South Mountain High School, Phoenix, AZ Teacher: Milton Johnson, TLRBSE 2006 Comparison of Magnetic Field Strengths in Circular and Irregular Sunspots................................. 28 David Young and Allison Osweiler St. Joseph Catholic High School, Pine Bluff, AR Teacher: Diedre Young, TLRBSE 2003 STARS Spectral and Photometric Analysis of Eclipsing Binary U Pegasi...................................................... 34 Olivia Claudio, Meredith Mead, and Sean Leahy Phillips Exeter Academy, Exeter, NH Teacher: John Blackwell, TLRBSE 2006 An Exploration for Extended Clusters in M33 TOP Run, October 2006......................................................................................................................... 43 Nathan Stano Grosse Pointe North High School, Grosse Pointe Woods, MI Teacher: Ardis Herrold, RBSE 1998, TLRBSE 2001 Spectral Analysis of Blazar S5 0716+714 using Spitzer Infrared Space Telescope and New Mexico ........................................................................................................ 49 Alekzandir Morton, Manutej Mulaveesala and Thomas Travagli Deer Valley High School, Antioch, CA Teacher: Jeff Adkins, TLRBSE 2002 The Age and Distance of the Open Cluster NGC 2345........................................................................ 57 Christina Clemens and Rachel Reece Sullivan South High School, Kingsport, TN Teacher: Thomas Rutherford, TLRBSE 2005

The RBSE Journal iii 2007 V1.9

GALAXIES Galaxy Clusters: The Local Effects on Star Synthesis ........................................................................ 66 Zackery Schroeder Grosse Pointe North High School, Grosse Pointe Woods, MI Teacher: Ardis Herrold, RBSE 1998, TLRBSE 2001 Star Formation Rate in Three High-Redshift Galaxy Clusters: A Contribution to the Study of Galactic Evolution .................................................................................................................................. 80 Vinay Patel and Matt Pellegrino Saint Joseph’s High School, South Bend, IN Teacher: Dr. Thomas Loughran, TLRBSE 2004 Deep Field Galaxy Classification........................................................................................................... 95 Daniel Kirpes Howenstine Magnet School, Tucson, AZ Teacher: Chris Martin, TLRBSE 2005

The RBSE Journal 1 2007 V1.9

Magnetic Field Fluctuations Before, During, & After Flaring in NOAA AR 10930

Caitlin Dewolf, Rebecca Mortensen, & Sheila Strong Chippewa Hills High School, Remus, MI Teacher: Cris L. DeWolf TLRBSE 2006

ABSTRACT This study analyzes the fluctuations in the magnetic fields of a strong, complex sunspot – NOAA AR 10930. During the time of the study, this active region was flaring, with X, C, B, and M class flares. Scans of the region were taken using the near-infrared array camera (NAC) of the McMath-Pierce National Solar Telescope. This data was reduced using the TLRBSE Solar Spectra Data Analysis Package – a script written for the IDL Virtual Machine. Comparisons of the total magnetic field power before, during, and after flaring of the sunspot were made by plotting the reduced data using Graphical Analysis and finding the integral for each resulting plot. INTRODUCTION The Sun was often worshiped by primitive societies as a glorious life-giving deity, and it’s easy to see why. As the source of all the Earth’s energy, the Sun is essential to life on our planet. Sometimes, however, the Sun can play mischievous games on the human race in the form of sunspots and solar flares, which can wreak havoc on radio communications and cause mass blackouts here on Earth. The Sun is estimated to be 5 billion years old, and is a gigantic mass of hot, glowing gas. Its gravitational pull keeps the Earth in a steady orbit and its heat and light allow for life on Earth. With a radius of 695,508 kilometers, the Sun is approximately 109 times larger than Earth, but has a density that is only about 25% that of Earth’s. The Sun is fueled by nuclear reactions that are triggered by its composition of hot glowing gas (Levine, 2004). The visible part of the Sun is referred to as the photosphere. Here, large and dark in composition, sunspots are often visible. One such solar blemish emerged from the back of the Sun in early December 2006. It was part of NOAA AR 10930.

These spots rotate with the Sun and can change in size or shape. A sunspot’s lifetime ranges from hours to months, and they increase and decrease in number on an eleven-year cycle. A sunspot’s magnetic field is thousands of times stronger than the Earth’s magnetic field, and is responsible for keeping sunspots cooler than their surroundings, resulting in the spot’s dark appearance in comparison to the sun. The magnetic field of the sunspot prevents heat, light, and energy from flowing out of the solar interior. This keeps the sunspot at a temperature of around 3230 degrees Celsius (Levine, 2004).

Following the cycle of the sunspot are areas on the Sun with exceptionally strong magnetic fields known as active regions. The rapid churning of these active regions sometimes causes magnetic fields that point in opposite directions to interact with each


other. The violent releases of energy associated with these interactions are known as solar flares. The explosion associated with solar flares heats the plasma around it to tens of millions of degrees Kelvin, creating protons, electrons, and heavier ions. These speed off into space at nearly millions of kilometers per hour, and produce electromagnetic radiation (Wikipedia, 2007). Ultraviolet and x-ray radiation reaches the earth eight minutes after a solar flare occurs on the sun. The GOES satellites record this incoming onslaught of radiation as it arrives to near-Earth space.

Atmospheric particles are raised to excited states and electrons are knocked free as the high-energy radiation is absorbed. This results in atmospheric layers increasing in density, in a process known as photoionization. The increased density causes short frequency radio waves to be absorbed, causing radio communications to be blacked out. These blackouts can last for minutes or hours, with the longest time durations being in the equatorial regions (Phillips,2007). The first recorded blackouts occurred on May 28, 1877 when telegraph lines in Boston, Baltimore, Philadelphia, and Washington D.C began to experience disturbances (New York Time, 1877). One of the places where solar flares are studied is the Kitt Peak National Observatory, which was dedicated on March 15, 1960 atop Kitt Peak in Arizona. With a telescope tower rising approximately one hundred feet above the ground, and a tunnel shaft running from the top of the tower 200 feet toward the ground at an angle of 32 degrees and continuing another 300 feet below ground, the McMath-Pierce National Solar Telescope is the world’s largest aperture solar telescope (Keller, 2002). The telescope’s heliostat mirror measures eight inches in diameter and rests one hundred feet above the ground. The mirror is mounted equatorially and rotates with the celestial sphere, once per day (Plymate, 2001). Sunlight reflects from the mirror down the shaft before reflecting off of a sixty-inch diameter mirror. The light is sent to a forty-eight inch diameter mirror before reaching its final destination in an observing room. Once in the observing room, an image of nearly one yard in diameter is available for the astronomers to examine (Keller, 2002). McMath-Pierce’s main instrument is the spectrograph. A cylinder resting on a bearing rotates at a twenty-four hour rotation to match the sun’s movements as it travels through the sky above the Earth. Inside the 1.8-m diameter, 21-m long vertical steel cylinder is the spectrograph. Slits ranging in width from 60-microns to 300-microns can be used by the spectrograph (Plymate, 2001). The light passes through the slit and down to a collimating mirror 13.7-m below, which reflects the beam up to the diffraction grating (Pierce 1969). A second mirror focuses the spectrum at the top of the spectrograph tank where it is accessed by detectors (Plymate, 2001).


The hypothesis in this study is that the magnetic field of an active region will gather strength before a solar flare, decrease in strength during the flare, and then return to a baseline when the flare is over. This is because flares require an interaction between magnetic fields to occur, and expend large amounts of energy. To test this hypothesis field strength versus relative frequency will be graphed for each scan by saving the data to a text file and importing it into the software Graphical Analysis by Vernier. The integral of each graph will be determined, with the area of beneath the line in each graph being used to compare the overall power of the magnetic fields over time, as well as their polarity trends. OBSERVATIONS On December 6, data was collected from a total of eight scans, due to the excellent visibility. Only five of those eight scans address the sunspot AR 10930 in the 1.565um wavelength, the rest are either scans of the smaller sunspot, AR 10927, or in the wavelength of magnesium lines (1.57um). The first three scans of the day were taken before the solar flare that occurred at about noon on that day. The fourth scan also occurred before the flare but most of the files in that scan are not useable because of the power outage that occurred shortly after 11:20 am. The X-class flare used in this study occurred around 12:00 noon and is represented by scan five in the data, it was followed by an M-class flare. On December 7, a total of five scans were taken. Only three of these scans were useable for observing Zeeman splitting. The C-class flare occurred in scan three of this set, at around 11:33 A.M. On December 8, a total of five scans were taken, and all five were of AR 10930. The visibility was poor at the beginning of the day but improved by scan three. A flare occurred at around 11:08 A.M. and was B-class, represented by scan four in the data. DATA REDUCTION During the first day of the study the Sun was quite active. A large sunspot in National Oceanic and Atmospheric Administration AR 10930 was still flaring. A visible light image of the solar disk projected from the East Auxiliary heliostat was projected on a light table to visually confirm the position of the region of interest and center the telescope on it. This active region was chosen as the region of interest that was then scanned. Staff members of the facility locked the telescope on this sunspot explaining how to use the telescope’s control system The telescope was then calibrated by using the adaptive optics array. Once the telescope was locked on the region of interest (in this case AR 10930) this apparatus made constant adjustments in a small mirror to correct for the distortions that our atmosphere makes in the Sun’s image. This allows for much greater resolution in the data collected, so that smaller scale structures in the magnetic fields of sunspots may be studied.


Light from the sunspot passed through the slit in the vertical spectroscope. The spectroscope scanned the sunspot by stepping the slit across the region of interest in seventy-five micron intervals providing a total of 160 data points in each scan. The scans started at the edge of the penumbra, moved across the umbra, and exited the sunspot on the opposite side. The wavelength of the scan was 1.565 um.

The data was taken in two polarization states, delineated as 0.8 and 1.9 in the nomenclature for the files. The numbers 1.9 and 0.8 refer to the voltage that was applied to the liquid crystal modulator during that observation. The 1.9 polarization represents right circularly polarized light, while the 0.8 polarization represents left circularly polarized light. The 0.8 voltage files are all one polarization state (left circularly polarized, or LCP) while all 1.9 v files refer to the other state (right circularly polarized, or RCP).

When this scan was complete, darks and flats were taken. Flats are taken with the shutter open and the telescope continuously and sporadically moving about the quiet part of the Sun. At the same time the bandpass is being continuously and randomly sampled in areas less populated by spectral lines. They were taken in order to remove the “background noise” in the images that is created by the instrumentation. Darks are taken with the same process as a normal scan would be, but with the shutter of the telescope closed. Flats are taken with the shutter open, and the telescope continuously stepping across the sun. The data was transferred from the telescope to the computer that held the archival data. Then using their FTP server, the data was transferred to the computer where the data analysis was to be performed. The first step was to reduce the data. The program used for data reduction was Teacher Leaders in Research Based Science Education (TLRBSE) Solar Spectra Data Analysis Package (Version 1.0 March 2006). This program runs in IDL Virtual Machine. The process of data reduction took the darks and flats, along with the images from the original data set, to produce the final images that were used for data analysis. Data reduction removed the sea turtle “ghost image” from the newly created images by subtracting the darks, and dividing by the flats to result in workable data. Upon completion of the data reduction, the data was resized by choosing new dimensions for the images that would allow easier analysis. The data was resized in order to make the files for the movies and images needed in the next steps a more manageable size. Without resizing, the amount of memory per file would have exceeded the processing capacity of the computer being used for data analysis. The reduced and resized data was then compiled into a movie using the TLRBSE Solar Spectra Data Analysis Package – an IDL virtual machine script written to work with solar spectral data. The movie was closely inspected for signs of Zeeman line splitting on the iron line. Once Zeeman line splitting became evident in the movie, the movie was paused and the frame number written down. The spectrum in this frame was used to make the magnetograms, or the magnetic field strength images.


DATA ANALYSIS As a result of using the analysis portion of the software, magnetograms (images of the magnetic field strength), Dopplergrams (images of the velocity flow of the plasma) and LineI images (images of intensity) were created. Our main interest revolves around the magnetograms. Specifically, inputting the LCP and RCP spectral line data allowed the software to create magnetograms. A region of interest, or ROI, around the Zeeman split portion of the spectral line at 1.565 μm is selected in both the LCP and RCP data. The software then creates the magnetograms from the formula B = 2.13 x 1012[Δλ/(Δλ2 x g], where λ is the wavelength of observation (15,650 Angstroms), g is the Lande factor which depends on quantum mechanics and for this spectral line is equal to 3, Δλ is the Zeeman split of the spectral line. Note that the larger the split, the larger the magnetic field strength, B.

A visual representation of this data was needed. The IDL script was limited in terms of being able to re-scale the axes of plots generated to show how field strength varied in frequency within the ROI so that they all had a common scaling. This type of plot would provide a method to visualize how field strength and polarity varied within the time frame that included pre-flare, flare, and post flare measurements of the field within the active region. A method using Excel to show this and allow for common scaling on all plots was developed. Data from the B-Gram images generated by IDL were written to text, opened with Excel and then copied into Graphical Analysis. Magnetic field strength was plotted versus relative intensity (how often a particular value occurred) to create integral graphs. Integrals measure the area under curves of graphs. These graphs were looked at in conjunction with the time that their original scans were taken to see fluctuations before, during, and after the flares. The more area that was under the curves of the graphs and the bigger the integral, the more power the fields held. On the first day of data collection there were two solar flares: an X-Class, and an M-Class. In the scan before the X-Class flare, the power contained in the magnetic fields of the active region decreased and the polarity was strongly positive. (Figure 1) The power continued to decrease during the flare itself, while the overall field polarity increased in negative polarity. (Figure 2) As soon as the X-Class flare was over, an M-Class flare occurred. During this flare, the power of the magnetic fields increased and the overall polarity became more positive. (Figure 3) Once the flare was over, the field power decreased and the negative polarity became predominant (Figure 4).


Figure 1 (Before X-Class Flare)

Figure 2 (During X-Class Flare)


Figure 3 (During M-Class Flare)

Figure 4 (After M-Class Flare)


There was one C-Class flare on the second day of data collection. Before the flare the field strength increased and negative polarities were more common. (Figure 5) The magnetic field strength decreased during the flare and positive polarities took over. (Figure 6) When the flare stopped the field strength of the active region was seen to increase, while negative polarities became dominant. (Figure 7)

Figure 5

(Before C-Class Flare)


Figure 6 (During C-Class

Figure 7 (After C-Class Flare)


On the last day of data collection there was one B-Class solar flare. In the scan before the B-Class flare, the field strength increased, while the main polarity was negative. (Figure 8) During the flare the strength decreased while the predominant polarity remained negative (Figure 9), and after the flare the magnetic field strength increased while the main polarity switched to positive (Figure 10).

Figure 8 (Before B-Class Flare)


Figure 9

(During B-Class Flare)

Figure 10

(After B-Class Flare)


DISCUSSION The hypothesis “the magnetic field strength of an active region will increase before a solar flare, decrease during the flare itself, and increase again once the flare is over” was refuted by this research. Every flare that was studied, with the exception of the B-Class flare, occurred when the field strength of the active region was increasing. The results for this last B-Class flare may be different due to human error. The data for the scan may not have been accurately transcribed, as far as time is concerned, because if the scans were not taken at the same relative time after the flare, then there would have been differences in the results. As previous research, such as that of Wang et.al. suggests, solar flares should occur on an increase of magnetic field strength because of the vast amount of energy required for a flare. Energy should gather before a flare, be given off during the flare, and only actually decrease when the flare is over. However, the majority of our data does not show the magnetic field strength decreasing after the flare. The only place where this is recorded as happening is after the combined X and M class flares. This is probably because of the difference in the time interval between the scan taken during the flare and the scan taken after the flare. The fifth scan on our second day of research, the scan after the C class flare, was taken at 1:18 P.M., an hour and forty five minutes after the flare was over, giving the field strength ample time to regain energy and increase. The same is true for the fifth scan on the third day of research, the scan after the B class flare, which was taken at 12:56 P.M, an hour and forty eight minutes after the flare was over. Unlike these scans, the seventh scan on the first day of research, the scan after the M and X class flares, was taken at 2:04 P.M.; only nineteen minutes after the flares were over. This did not give the active region time to regain strength, so it decreased, as previous research would suggest. This research was largely inconclusive due to the fact that only four solar flares were studied. More flares would have to be examined to see if any patterns emerge. If the opportunity for future research arose, the students would compare the data from more solar flares. It would also be constructive if the flares were of the same class. No two solar flares are exactly alike, and this fact could lead to some slight variations in the data. If permission could be obtained to use previous data from the archives, then this research could be done without actually returning to the McMath-Pierce Solar Telescope. SUMMARY AND ACKNOWLEDGEMENTS The hypothesis that the magnetic field strength of an active region will increase before a solar flare, decrease during the flare itself, and increase again once the flare is over was not supported by this data. Instead, the results of this research did show that solar flares tend to occur on an upward trend, or in other words, that the strength increases from one solar flare scan to the next, power building in the fields and then being released by the flare. This is useful to know because it may help to predict solar flares, and in doing so making this research valuable by allowing scientists to see upward trends in magnetic field strength and warn the population that a solar flare may occur. Then preparations


could be made for potential power outages, satellites could be sent into a “storm warning” stand-by mode, and astronauts in the ISS could take shelter.

Our warmest thanks are extended to:

• Connie Walker and the workforce of the National Optical Astronomy Observatories who made our research experience possible.

• Clyde Plymate and Eric Galayda who taught us to use the McMath-Pierce telescope and generously took time from their day to help us collect data.

• The Weidman Lions Club, Barryton Lions Club, and Chippewa Hills Education Association for the financial assistance that made this project possible.

• Cris DeWolf, for the knowledge and patience as our mentor that helped us put it all together.

REFERENCES Keller, Christoph. "History of the MP Facility." National Solar Observatory/Kitt Peak. 31 Oct. 2002. 9 Jan 2007 <http://nsokp.nso.edu/mp/history.html>. Plymate, Claude. "A History of the McMath-Pierce Solar Telescope." 01 June 2001. 9 Jan 2007 <http://nsokp.nso.edu/mp/cphistory.html>. "Short Wave Fadeouts (SWFs) and Solar Flares." Windows to the Universe. 2006. University Corporation for Atmospheric Research. 25 Nov 2007 <http://www.windows.ucar.edu/spaceweather/swf_flares.html>. “The Aurora Borealis.” New York Times 29 May 1877: 5. Levine, Randolph H. “Sun” Microsoft Encarta Encyclopedia CD-ROM.2003 ed. Microsoft Corporation, 2004. "Solar Flares." Astronomy 162: Stars, Galaxies, and Cosmology. 20 Dec 2006 <http://csep10.phys.utk.edu/astr162/lect/sun/flares.html>. "Solar flare." Wikipedia. 13 Jan 2007. Wikimedia Foundation. 20 Jan 2007 <http://en.wikipedia.org/wiki/Solar_flare>. Phillips, Tony. "The Classification of X-ray Solar Flares." Spaceweather.com. 3 Mar 2007 <http://www.spaceweather.com/glossary/flareclasses.html>. Phillips, Kenneth. Guide To the Sun. 2. Cambridge: University Press, 1992. Zirker, Jack. Journey from the Center of the Sun. 1. Princeton: Princeton University Press, 2002.


Characterization of NOAA AR 10930 Shuo Qiu and John Kucharczyk

Manhasset High School Teacher: Mr. Peter Guastella, TLRBSE 2006

INTRODUCTION Perhaps the most important of all stars, the sun, still has many mysteries to be solved. The most important of these is the prediction of solar storms, known as coronal mass ejections. Solar flares can cause radiation problems in astronauts that are in space. Strong flares like the one on Bastille Day can cause disruption in radio transmission and power lines. Some can even knock out entire grids. Dr Bhatnagar suggests that large active regions of the sun called sunspots that have a diameters of around 14000km or 8,700miles may be directly linked to the release of Coronal Mass Ejections. These ejections release massive amounts of X-rays. Sunspot NOAA AR 10930 first appeared on the east limb of the sun on December 4, 2006. The region evolved very quickly and by December 5 it emitted an X class flare. By the time it disappeared on the western limb on December 18, it had released 5 x class 13 M class and 60 class flares.

http://www.noao.edu/image_gallery/images/d1/02245a.jpg Figure 1- This is the side view of a coronal mass ejection. It can be seen coming from the sunspot.

Garcia (2000) reported that during solar cycles 21 and 22 there was an average of one M class(moderate) solar flare every two days and one x class every month. Therefore it is extrodianry to observe a single sunspot of with the activity of NOAA AR 10930. Schunker in 2000 showed that there was a sharp increase in magnetic field strength during the main portion of a solar flare. However, Lin in 1993 suggested that longitudinal magnetic fields remain the same during a solar flare. Similarly Rust in 2004 stated that magnetic field strength changes were insignificant during solar flares, despite the large amount of emissions.


To obtain absorption spectra from the NOAA AR 10930 and to interpret that data to determine the magnetic field strength of different areas of the sun spot, and to calculate the direction of plasma flow on the surface of the sun in the active region. A second goal is to obtain GOES satellite x-ray data and to compare solar activity to magnetic field strength and direction of flow.

METHODOLOGY Absorption spectra from the AR will be obtained using the McMath-Pierce Solar Telescope atop Kitt Peak. The lab technicians will be handling the complex equipment of the McMath Pierce, and so participation by the students will be minimal. For each scan there will be about 161 images of the spectra that will be received to interpret in raw Fitz file format. This will be done using the IDLVM TLRBSE software. Using the absorption images magnetic field strength will be calculated by measuring the size (lambda) of the Zeeman split in pixels. The software can also be used to create, when possible, magnetograms. The magnetograms will be used so that magnetic field strength can be calculated for the umbra and penumbra of the sunspot, in addition to the polarization of the active region. In addition the absorption spectra can be used to acquire dopplergrams using the IDLVM TLRBSE software. Dopplergrams show the direction of plasma flow on the surface of the sun. In addition, archival data from the GOES 12 x-ray satellite will be acquired to show solar activity during the times of the flare. X-ray data indicates when a solar flare or Coronal Mass Ejection occurs. The GOES x-ray data and data acquired by the McMath-Pierce will be compared to determine the effects of solar flares or lack of solar activity on a sunspot. ANALYSIS AND RESULTS On December 6, there was a decrease in magnetic field strength during the X-class flare. The flare occurred at about 6:40, and so the fifth scan was taken during the decline of the flare.


Magnetic Field Strength of AR 10930 on 12-6

0

500

1000

1500

2000

2500

4:13 4:40 5:30 6:20 7:00 8:45 9:00

Time

Mag

netii

c Fi

eld

Stre

ngth

(Gau

ss)

Graph 1- Shows the magnetism of AR 10930 on December 6. The scans were taken over a span of about 9 hours. There was a sharp decrease in magnetic field strength during the X-class flare.


Magnetic Field Strenght of AR 10930 on 12-7

0

500

1000

1500

2000

2500

3000

3500

5:25 6:33 8:18

Time

Mag

netic

Fie

ld S

treng

th (G

auss

)

Graph 2- Shows the magnetism of AR 10930 on December 7 over a span of about 7 hours. During this time you can see a sharp increase in magnetic field strength during the M-class flare. On December 7, during the M-class flare, there was a slight inc. in magnetic field strength. This flare occurred at about 7:30 and so the third scan was taken during or soon after the decline of the m-class flare.


Magnetic Field Strength of AR 10930 on 12-8

0

500

1000

1500

2000

2500

3000

3500

4000

3:41 4:34 5:06 6:08 7:56

Time

Mag

netic

Fie

ld S

treng

th (G

auss

)

Graph 3- Shows magnetism of AR 10930 on December 8 over a span of about 5 hours. There was a sharp spike in magnetic field strength even though there weren’t any significant flare action. Finally on December 8, there was a dramatic rise of 71% in terms of magnetic field

strength even though there were no significant flares during the period atop Kitt Peak.

Figures 1, 2 & 3- magnetograms AR 10930 before during and after the x-class flare on 12-6. The image to the left is a magnetogram taken before the x-class flare on December 6, the middle was taken during and the right shortly after.


The magnetograms formed before, during, and after flare showed that there were significant changes in magnetic field orientation. Before flare, there is one polarity. During the flare, two polarities appear, and then in the after flare phase the opposite pole appears.

Figures 4, 5 & 6- Dopplergrams of AR 10930 before, during and after the x-class flare on 12-6. Each dopplergram is accompanied with graphs of the flow velocities which were created using a horizontal region of interest across the colored area. The top dopplergram was taken before the x-class flare on December 6, the second was taken during, and the bottom dopplergram was taken after the flare.

The dopplergrams show the AR 10930 before, during and after the x-class flare on 12-6. Each dopplergram is accompanied with graphs of the flow velocities, which were created by drawing a horizontal region of interest across the colored area. As can be shown in the before flare flow velocity graph, you can see an increase in flow velocity in the umbra region of the sunspot. In the during flare graphs, you have one that has an increase in

Dopplergram scan 5 II

-2500

-2000

-1500

-1000

-500

0

500

90 92 94 96 98 100

102

104

106

108

110

112

114

116

118

120

122

124

126

128

130

132

134

136

138

140

142

144

146

148

150

152

154

156

158

160

162

164

166

168

Pixels

Velo

city

Dopplergram of scan 5

-1000

0

1000

2000

3000

4000

5000

85 87 89 91 93 95 97 99 101 103 105 107 109 111 113 115 117 119 121 123 125 127 129 131 133 135 137 139 141

Pixel

Velo

city

Dopplergram of scan 2

-1400

-1200

-1000

-800

-600

-400

-200

0

200

400

136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197

P ixel

Dopplergram Scan 6

-1000

0

1000

2000

3000

4000

5000

6000

87 9 0 93 9 6 99 1 02 10 5 10 8 11 1 1 14 11 7 1 20 1 23 12 6 1 29 13 2 1 3 5 1 3 8 1 41 14 4 1 47 15 0 15 3 1 56 15 9 16 2 1 65 1 68 17 1 1 74 17 7 1 80 1 83

Pixels

Velo

city


flow velocity and another with a decrease in flow velocity. In the after flare graph, you have a decrease in flow velocity as it goes through the umbra region. DISCUSSION - The X-class flare from sunspot 10930 on December 6 showed many glaring characteristics. First off, you can observe the decrease in magnetic field strength. This is shown in the bar graph of day 12-6. During the flare there was a 25.0% drop in magnetic field strength using the Magnetic Field Strength Formula. In addition the magnetic polarities were disfigured as can be shown by Figure 1,2,3. Lastly there was a change in the velocity of plasma flow as shown in Figure 4,5,6. Despite the decrease in magnetic field strength during the X-class flare, there was a 22.2% rise during an M-class flare on 12-7. In addition there was an abnormal increase in magnetic field strength of 71.4% and 57.2% during scans 17 and 18 on day 12-8, which was a non flare period. These findings refute the works of Lin and Schunker. CONCLUSION - Due to the findings of magnetic field strength during different intervals of activity and time, there does not appear to be any link between magnetic field strength and the magnitude of solar flares. This can be shown by the incongruities that occurred on all three days of scanned results. Without a link between magnetic field strength and magnitude of solar flares, prediction of solar flares remains unsolved. It has been found, however, that dipole characteristics of a sunspot are disrupted during flares as exposed by the results from the magnetograms. In addition, velocity in plasma flow fluctuates to the extremes before, during, and after solar flares. Future Studies - In the future, the development of better software will help to enhance the processing of the data that we obtained atop Kitt Peak. Also, incorporating Evershed flow into the Dopplergram data that gave flow velocities will help to elucidate the true motion on the sunspot during a flare. ACKNOWLEDGMENT - We would like to also thank NOAO, Dr. Connie Walker, and Mr. Claude Plymate for giving us an opportunity to conduct research atop Kitt Peak in the McMath Pierce Telescope.


BIBLIOGRAPHY Borner, P., High Resolution Observations of the Evershed Flow, 1992, Astronomy and Astrophysics, 307-312

Del Toro Iniesta, J., Cold, Supersonic Evershed Downflows in a Sunspot, 2002, The Astrophysical Journal 139-142

Hirzberger, J., 2D-Spectroscopy of the Evershed Flow in Sunspots, 2001, Astronomy and Astrophysics 1078-1086

Maltby, P., The Chromospheric Evershed Flow, 1975, Kluwer Academic Publishers 91-105

Penn, M., Weak Infrared Molecular Lines Reveal Rapid Outflow in Cool Magnetic Sunspot Penumbral Fibrils, 2003, The Astrophysical Journal, 119-122

Regnier, S., Evolution of Magnetic Fields and Energetics of Flares in Active Region 8210, 2006, Astronomy and Astrophysics, 319-330

Rust, D., An Active Role for Magnetic Fields in Solar Flares, 2004, Solar Physics, 21-40

Schlichenmaier, R., 1998, Astronomy and Astrophysics 897-910

Schunker, H., Variations of the Magnetic fields in Large Solar Flares, 2004, Space Science Reviews, 99-102

Sharmer, G., Dark Cores in Sunspot Penumbral Filaments, 2002, Nature Publishing Group 151-153

Sudol, J,. Longitudinal Magnetic Field changes Accompanying Solar Flares, 1986, The Astrophysical Journal, 647-658

Thomas, W., Solar Irradiance Variability During the October 2003 Solar Storm Period, 2004, Geophysical Research Letter Vol. 3

Thomas, J., Downward Pumping of Magnetic Flux as the Cause of Filamentary Structures in Sunspot Penumbrae, 2002, Nature Publishing Group, 398-401

Thomas, J., A Siphon-Flow Model of the Photospheric Evershed Flow in a Sunspot, 1993, Astrophysical Journal 398

Wang, H., Rapid Changes of Photosphere Magnetic Fields Around Flaring Magnetic Neutral Lines, 2006, Big Bear Solar Observatory

Uitenbroek, Evidence for a Siphon Flow Ending Near the Edge of a Pore, 2000, The Astrophysical Journal, 1-12

Yuanzhang, L., Variations of Magnetic Fields and Electric Currents Associated with a Solar Flare, 1993, Solar Physics, 133-138


The Relationship Between the Area and the Magnetic Field Strength of Sunspot NOAO Active Region 10905 Brian Fisher, Isabel Garcia, Juan Vinagera

South Mountain High School, Phoenix, AZ Teacher: Milton Johnson, TLRBSE 2006

ABSTRACT For three days during the week of August 21, 2006, sunspot data was collected for Sunspot AR 10905 using the McMath Solar Telescope at Kitt Peak National Observatory. It was thought that there would be a positive correlation between the area of the sunspot and the magnetic field strength. The data consisted of sunspot area and magnetic field strength. The objective was to compare and distinguish a relationship between these two characteristics. The data was scarce considering that only three days worth of measurements of the region were gathered due to cloudy weather. When the data was graphed comparing the area and the magnetic field strength, the representation showed an inconsistency in correlation. The area increased during the three days, but the magnetic field decreased. In the end, for this particular sunspot, a concrete relationship between the area and the magnetic field could not be established. INTRODUCTION A sunspot is a region on the Sun's surface (photosphere) that is lower in temperature, approximately 4000-4500 K, than its surroundings (5700 K). This is because it experiences intense magnetic field activity, caused by convective motion beneath the photosphere that slows down in areas where the magnetic field is passing through (the magnetic field loops out and back into the Sun’s surface). Although they are blindingly bright, in contrast to the surrounding material, they appear as dark spots. Sunspot cycles last roughly eleven years, and are marked by high and low activity. This particular sunspot, AR 10905, was formed at the end of the most recent eleven-year cycle, in which there was low activity. The strength of the magnetic field of a sunspot is measured with Zeeman shifts of the spectral lines. The particular line measured was the iron line found at 1.5648 microns. This line is found in the infrared spectrum and exhibits an unusually large Zeeman shift. Sample of this can be seen in Figure 1.

Figure 1. Sample of the Zeeman shift seen in


OBSERVATIONS At Kitt Peak National Observatory near Tucson, Arizona, data was taken of Sunspot AR 10905 activity for 3 days: Monday, August 21, 2006, Wednesday, August 23, 2006, and Friday, August 25, 2006. The intention was for data to be taken on Tuesday, the 22nd and Thursday the 24th, but weather conditions did not allow the possibility. The data was obtained using the National Solar Observatory McMath-Pierce Solar Telescope (M/P) infrared camera system operated by Claude Plymate and Eric Galayda. The scans were infrared spectra using the Fe line at 1.5648 microns, in ninety frames across the region. These were done for zero and ninety degree polarizations, along with flat and dark fields to minimize “noise.” Using this Fe line was preferable because it has the largest Zeeman shift, due to a Lande g factor of 3. This factor indicates the sensitivity of the spectral lines to magnetic fields. REDUCTION Using software, created by Dr. Frank Hill of the National Solar Observatory, which runs on an IDL Program platform, Scans 1, 2, and 3 were resized and a magnetogram was created. The images of the spectra were too large to reduce as they were, and required a reduction in size. They were then reduced by subtracting the flat-fields and dark-fields from the images. This removes blemishes from the spectra caused by defects in the ccd camera. Next, an animation of the ninety frames of spectra was produced with the software, which was used to find the frame with the largest Zeeman shifts in the sunspot region. This frame was needed to make a magnetogram of the sunspot. With the chosen frame, a magnetogram was generated. In order to generate the magnetogram, the particular frame was opened and a “region of interest” was selected. (Figure 2) This area of the selected spectrum contained the Fe line at 1.5648 microns. As mentioned before, this line is the one most affected by the magnetic field of the sunspot region, as noted by its large Lande g factor. The magnetic field causes the Fe line to spread, or shift. A stronger magnetic field, means a greater shift in wavelength.

Figure 2. The “region of interested” used to


The IDL program measures this shift in wavelength by converting the shift (measured in pixels) to wavelength. The strength of the magnetic field can be measured with the Zeeman shift using the equation: B= 2.13x10-12[Δλ/(λ2 x g)], where B (measured in Gauss) is the magnetic field strength, λ is the wavelength, Δλ shift in wavelength, and g is the Lande factor. The positive magnetic region of the sunspot was encircled (‘positive’ refers to the pixel value) in order to generate a histogram of the magnetic area using Image J software. This provided the mean and maximum magnetic field strengths used to graph against the area of the total sunspot. The mode was graphed on its own. These images are in Figure 3.

8/21 8/23 8/25

Figure 3 Top row: Optical Images from 8/21,8/23,8/25, respectively, taken from www.solarmonitor.org Middle row: Magnetograms taken from IDL software Last row: Histograms of the magnetograms using ImageJ software


The area of the sunspot was measured using optical photos from www.solarmonitor.org. To gather the area, the photos were imported into Image J software. First, the scale was set, drawing a horizontal line segment from one side of the sun to the other, and scaling it at 1,392,000 km, which is the average equatorial width of the sun. It was then zoomed in to 600% for better viewing. The area of the sunspot was measured by using the free-hand selection and drawing around each pixel that depicted the sunspot. The area was divided by cosine of the longitude angle (shown on www.solarmonitor.org) to get the corrected area. This resulted in two parts to each graph: Umbra and Umbra_adj. This process was repeated to insure accuracy. ANALYSIS Area was graphed versus both the mean and the maximum values of the entire sunspot’s magnetic field region. Both graphs show an inconsistency in correlation between the area and the magnetic field. The graphs comparing the area to the mean value, and the area with the maximum value of the magnetic region, did present an increase in strength the first two days, and then a dramatic decrease on the third day. The mode proves a disparity, showing a dramatic decrease between the first two days, and then a slight increase on the last day. This noticeable difference caused the mode to be graphed on its own, without the area data inputted. DATA

Magnetic Field Data Date Mean Mode Max Area

8/21/2006 1136.207 1237.183 2304.033 745,341,441.00

8/23/2006 1555.08 1360.896 2875.746 842,126,887.00

8/25/2006 1198.791 761.737 2716.894 1,082,000,000.00

Sunspot Area Data Umbra PenUmbra Longitude Radians Corrected Area

Date Area (km^2) (km^2)

8/21 78,187,194.00 255,355,536 70 1.221111 228215753 745341441

8/23 116,709,384 628,933,950 46 0.802444 167938989 905004621

8/25 234,561,584 1,022,949,120 19 0.331444 248062826 1.082E+09

Table 2

Table 1


GRAPHS

Area of AR10905

0.00

50,000,000.00

100,000,000.00

150,000,000.00

200,000,000.00

250,000,000.00

300,000,000.00

8/20/06 8/21/06 8/22/06 8/23/06 8/24/06 8/25/06 8/26/06

Date

Area

(km

^2)

UmbraUmbra_adj

Adjusted Area of AR 905

0

200,000,000

400,000,000

600,000,000

800,000,000

1,000,000,000

1,200,000,000

8/20/06 8/21/06 8/22/06 8/23/06 8/24/06 8/25/06 8/26/06

Date

Are

a (k

m^2

)

penumbrapenumbra_adj

Mean Magnetic Field vs. Area

y = 81409x + 8E+08R2 = 0.0119

0

200000000

400000000

600000000

800000000

1000000000

1200000000

0 500 1000 1500 2000Magnetic Field (G)

Are

a(km

^2) Magnetic Field Vs. Area

Linear (Magnetic Field Vs.Area)

Max Magnetic Field vs. Area

-

200,000,000.00

400,000,000.00

600,000,000.00

800,000,000.00

1,000,000,000.00

1,200,000,000.00

0 500 1000 1500 2000 2500 3000 3500

Magnetic Field (G)

Are

a (K

m^2

)

Magnetic Field Max vs. Area

Mode of Magnetic Field vs Area

-

200,000,000.00

400,000,000.00

600,000,000.00

800,000,000.00

1,000,000,000.00

1,200,000,000.00

0 500 1000 1500

Magnetic Field (G)

Are

a (k

m^2

)

Magnetic Field Mode

Graph 1

Graph 3 Graph 4

Graph 5

Graph 2

Graphs 1 &2 show the area of the sunspot vs. the date. Graph 3-5 compare magnetic field to area.


DISCUSSION From reading previous articles that experimented with the area of a sunspot with its magnetic field strength, the motive to discover a relationship between these two qualities was initially commenced. However, the data taken over the three-day period for this particular sunspot showed no correlation between the area and the magnetic field strength. The data for the area of the sunspot was the first to be graphed, and this provided an indication that the sunspot was growing. Although there was a slight variation in correlation for the un-corrected graph, noticed on day 8/22, the corrected data gave more confirming evidence that the size was increasing. The graphs for the magnetic field strength contradicted the graphs for the area, showing a decrease in activity. Three days of data was not enough to show a strong correlation. Also, the fact that the sunspot was caught coming off the limb increased hindrances for data observation. On top of that, by Friday, the sunspot’s magnetic field had already begun to dissipate. Unfortunately, the idea was apprehended that the area of the sunspot is not necessarily related to the area of the magnetic field strength. Based on the data is in not possible to provide a factual claim, such as, when the area increases, the magnetic field strength increases, or when the area increases, the magnetic field strength decreases. There is still the possibility that either is true, but more data would be needed to establish this. ACKNOWLEDGEMENTS We would like to thank the National Optical Astronomical Observatories (NOAO) for providing observation time through their Teacher Observation Program (TOP). We would also like to thank Dr. Connie Walker of NOAO for assisting and mentoring us through the year as we worked on this project. Also, we would like to recognize the telescope operators Claude Plymate and Eric Galayda for their help in collecting the data.

REFERENCES McMath Pierce Solar Telescope <http://nsokp.nso.edu/mp/mp.html>

Kitt Peak National Observatory <http://www.noao.edu/kpno/>

NOAO Research Based Science Education <www.noao.edu/education/arbse>

Corpuz, Gannon, Williams: The Association of Solar Magnetic Field Strength Variation with Sunspot Area: RBSE Journal 2005 V1.4 <www.noao.edu>

Mullins, Stanley: The Effect of Sunspot Size and Latitude on Magnetic Field Strength in Sunspots: RBSE Journal 2005 V1.4 <www.noao.edu>

Solar Monitor: <www.solarmonitor.org>

Sunspot: <http://en.wikipedia.org/wiki/Sunspot>


Comparison of Magnetic Field Strengths in Circular and Irregular Sunspots

David Young and Allison Osweiler St. Joseph Catholic High School, Pine Bluff, AR

Teacher: Diedre Young, TLRBSE 2003

ABSTRACT This research project investigates the comparative magnetic field strengths in circular and irregular sunspots by measuring the Zeeman splitting of the iron spectral line at 151650 Angstroms. Archived data from TLRBSE solar data disc July 2003 and February 2004 was used for research and analysis. It was concluded that circular sunspots, in general, have a greater magnetic field strength than non-circular sunspots. This information is useful in predicting the likelihood of solar flares. INTRODUCTION Sunspots have been seen and observed for centuries by astronomers. Galileo was the first to record sunspots, and believed that they were blotches on the sun. It was in the 1800’s when people started to believe that these blotches were caused by the extreme magnetism on the sun.

This idea was finally confirmed by Pieter Zeeman, who discovered the so called “Zeeman Effect”. This theory states that magnetism can affect the splitting of the ‘f’ electron in the Fe 1 atom. This theory was further expounded upon by George Ellery Hale, with the development of the spectroheliogram. This device enabled astronomers to actually measure the magnetic field strengths in sunspots. Sunspots are the product of the Sun’s magnetic fields located in the photosphere. The two fields travel at different speeds, and coil up to trap plasma in between the twisted fields. This plasma cools down and becomes less bright. The sunspots themselves are not actually ‘dark’, but because the sun is so bright, they appear dark as compared to the rest of the sun. Occasionally the sunspot, after accumulating enough mass, will burst releasing the plasma into space. Typically the plasma will be drawn back into the sun forming a Coronal Loop. Occasionally, the sunspot will burst releasing its plasma at such a speed that it cannot be retracted back into the sun. Sometimes, this plasma is pointed towards earth. Fortunately, the Earth is protected by a magnetic field created by the spinning of the iron core of the Earth. This magnetic field deflects the ‘solar radiation’ and protects the Earth from otherwise deadly radiation. The problem lies with man-made satellites that are orbiting the earth. When the magnetic field is hit, it is pushed back against the Earth. This exposes satellites to the solar radiation, which can disrupt communications and disable the satellite. Both communications and navigational satellites are vulnerable to these storms and even electrical arrays on earth can be affected. Information from this paper can be used to more accurately determine if a sunspot will break and to even predict the possibility of a sunspot breaking during its lifetime.


RESEARCH PROBLEM Based on past research in ‘Magnetic Field Strength Comparisons across Solar Active Region AR 396’and ‘Magnetic Field Strength Comparisons Between Lone and Clustered Sunspots of Similar Size’, it is reasonable to believe that the magnetic field strength of the umbra of a sunspot may vary based on its shape. This project investigates whether or not the magnetic field strengths of the umbra of a sunspot varies if that sunspot’s umbrae is circular or non-circular. OBSERVATIONS AND ANALYSIS This study was done by scanning Active Regions 375, 400 and 564 in infra-red light, and measuring the Zeeman splitting of the iron spectral line at 151,650 Angstroms. Images were taken from the McMath-Pierce solar telescope at Kitt Peak between July 3-7, 2003, and during February 23-24, 2004. The magnetic fields were determined by using the following formula (Klaveren et al. 10)

B=e/4πcme[Δλ/(λ2 x g)] Where B is the magnetic field strength in Gauss, e/4πcme is a value from quantum mechanics that equals 2.13 x 1012 at the wavelength λ, which is the wavelength in Angstroms. G is the lande factor which is determined from quantum mechanics, which is in this instance 3. Δλ is the separation in Angstroms between the pi and sigma components of the spectral line. The formula can thus be simplified to: B= 2.13 x 1012(Δλ/λ2g) To solve for B, the two variables, λ2 and Δλ, had to be determined. The variable λ has been determined to be 15,650 Angstroms. The sunspots were scanned to find Δλ. This was done by reducing and analyzing the images using a program designed by Dr. Marcel Bergman. After the images were taken at Kitt Peak, the raw data was reduced (eliminating errors and background noise), which is done by subtracting the “dark” images (showing hot pixels from the instrument) and dividing (normalizing) the data by the “flat” images. After the data was reduced, the spectra showed the Zeeman split lines clearly. The data set could then be ‘rearranged” into actual pictures of the sunspot using a ‘spectroheliogram’ program. Spectroheliograms are like snapshots of the sunspot at one wavelength, making use of the fact that the sunspot was scanned across in one spatial direction or dimension and the slit on the instrument providing the second spatial dimension. After reducing and running the images through the spectroheliogram software, they are ready to be analyzed to determine Δλ. After the spectroheliogram files were opened, one spectroheliogram file was chosen and the resulting image was displayed in a window. A region was then selected to measure the magnetic field, typically located in the center of the sunspot, and the resulting file number was opened to display the magnetic field lines. A horizontal box was drawn on each image at the appropriate y-axis for the region, and then using the “New Analysis Window” from the “Window” menu, a graphical spectrum was displayed, corresponding to the center of the sunspot. The set of three Zeeman lines were located in the spectrum.


An x-axis pixel reading was taken for left and center Zeeman lines. The pixel separation of the Zeeman split lines was therefore obtained. This measurement was converted into Angstroms from pixels and plugged into the formula for the magnetic field strength.

(Young, Mallet, Hayes, RBSE Journal 2004) DATA

0

500

1000

1500

2000

2500

3000

3500

Series1 2556.8 1988.6 3267 2343.7 1988.6 2566.8 2556.8 2059.7

regular 1

irregular 1

regular 2

irregular 2

regular 3

irregular 3

regular 4

irregular 4

SUNSPOTS DELTA LAMDA MAGNETIC STRENGTH 1 Reduced 5 regular (left 72; right 133; top and bottom 97) 8 pixels 2,556.81 gauss

1 Reduced 9 irregular (left 72; right 114; top and bottom 99) 13 pixels 1,988.63 gauss 2 AR 400 regular (left 254; right 266; top and bottom 145) 18 pixels 3,267.03 gauss 2 AR 400 irregular (left 191; right 203; top and bottom 47) 14 pixels 2,343.74 gauss 3 AR 400 regular (left 256; right 268; top and bottom 166) 16.5 pixels 1,988.63 gauss 3 AR 375 irregular (left 72; right 89; top and bottom 57) 23 pixels 2,566.81 gauss 4 AR 375 regular (left 22; right 32; top and bottom 106) 18 pixels 2,566.81 gauss

4 AR 375 irregular (left 190; right 200; top and bottom 218) 14.5 pixels 2,059.651 gauss Using ROI positioning


Reduced 5 Regular Sunspot


Reduced 9 Irregular Sunspot


DISCUSSION In general, circular sunspots have a greater magnetic propensity than irregular sunspots. The hypothesis that circular sunspots had a greater magnetic propensity than irregular was confirmed. Possible sources of error are having only used four sets of sunspots, calculations could have been off, and having used two different archived data sets. NEXT STEP The fact that one set of sunspots did not follow the trend could be expounded upon to see if this was an anomaly or if it is actually how a great number of sunspots exist, thus disproving the hypothesis. In the future, the information can be used to evaluate the strength of upcoming sunspots on their propensity for creating solar flares. Future research can be conducted on the reasons behind the difference in field strengths, and to confirm the data already received. ACKNOWLEDGEMENTS Daniel Young, Diedre Young, Kelly Railsback, TLRBSE REFERENCES Hathaway, David H. NASA/Marshall Solar Physics. Hathaway, Dr. David H., “Solar Flares” http://solarscience.msfc.nasa.gov/ January 18, 2007 Hathaway, Dr. David H., “Coronal Mass Ejections” http://solarscience.msfce.nasa.gov/ January 18, 2007 Hathaway, Dr. David H., “The Big Questions” http://solarscience.msfce.nasa.gov/ January 18, 2007 Hathway, Dr. David H., “The Key to Understanding the Sun” http://solarscience.msfce.nasa.gov/ January 18, 2007 Myers, Young, Stuart, Merriott, ‘Magnetic Field Strength Comparisons across Solar Active Region AR 396’ RBSE Journal (2004): 1-13 Young, Mallet, Hayes, ‘Magnetic Field Strength Comparisons Between Lone and Clustered Sunspots of Similar Size’ RBSE Journal (2004): 18-25 Young, Daniel. Personal Interview, March 6th, 2007 Young, Diedre. Personal Interviews, October 2006 – March 2007


Spectral and Photometric Analysis of Eclipsing Binary U Pegasi Olivia Claudio, Meredith Mead, and Sean Leahy

Phillips Exeter Academy, Exeter, NH Teacher: John A. Blackwell, TLRBSE 2006

ABSTRACT The results of a photometric and spectral study of the eclipsing binary star U Pegasi are presented in which confirmation was made that this star is a contact eclipsing binary, but no correlation between the spectral changes and photometric data were found. The Coudé Feed at Kitt Peak National Observatory was used to take spectral data and the Phillips Exeter Observatory 125mm telescope and the North Field Mount Hermon 152mm telescope were used to take photometric data. Light curves of integral flux over time were made to show the change in magnitude and luminosity of the star during its eclipse, and it showed no correlation to the photometric light curves. INTRODUCTION Eclipsing binary stars appear as a single point of light to an observer but actually consist of two stars in close orbit around one another. The variations in light intensity are caused by the eclipse, which is when one star passes in front of the other. Graphs can be composed depicting how brightness varies as a function of time for eclipsing binary stars. Light curve analyses of U Peg taken from Birouni Observatory in Iran and the National Observatory of Athens indicate that the star is an over-contact W Ursae Majoris system (Edalati, 2001) composed of two G2V spectral type stars (SIMBAD 2000). Of the many types of eclipsing binary stars, W UMa types are the most common to have a continuously changing light curve. They do not reach a stable level of brightness like many eclipsing binaries. U Pegasi became our target star after seeking stars with a high probability of eclipsing during our observation time at Kitt Peak National Observatory. Basic data about U Peg is given in Table 1 below. Since this is an over-contact binary system, the light curve never levels off at a peak brightness, which can be seen in the plot in Figure 1 below (Allegheny, 2001). In our study of U Pegasi, we analyzed the change in magnitude and luminosity using spectral data taken from Kitt Peak National Observatory and photometric data taken from the Phillips Exeter Academy Observatory and Northfield Mount Hermon School Observatory.

Name SAO 108933 FK5 2000 Coordinates 23 57 58.48 +15 57 10.1 B magn / V magn 10.27, 9.66 Spectral Type G2V Parallaxes 7.18 milliarcseconds

Table 1: Basic Stellar Data


This paper will present the results of an analysis of the spectroscopic and photometric data taken over the course of two evenings (18 September 2006 and 25 November 2006). The data has been analyzed in an attempt to determine correlations between spectral flux changes in the star system and photometric changes throughout eclipses. Along with the results of this experiment, this document will also explain the methodology used for the analysis, as well as showing the results from other similar studies.

Figure 1: Light curve of a full eclipse period of U Peg (Allegheny, 2001).

PROCEDURE The spectral data was obtained using the Coudé Feed at the Kitt Peak National Observatory, a .09m three-mirror telescope and spectrograph. Spectral data was taken between 4400 and 4650 Angstroms. Ten twenty-minute exposures of U Pegasi were taken while U Peg was eclipsing, with the first image taken at J.D. 2453997.645 and the last taken at J.D. 2453997.805. MaxImDL software was used to calibrate the spectral images with flat and bias images to remove instrumental noise and unwanted signal. Maxim DL was then used to convert spectral data into numerical data. The numerical data was transferred into Microsoft Excel and wavelength calibrated using VisualSpec software and data taken from an Iron Argon lamp while at the observatory.


The photometric data was taken by two observers, John Blackwell and Hughes Pack, using the Grainger Observatory's 125mm f/5.5 telescope and the Northfield Mount Hermon Observatory 152mm telescope respectively. The photometric data were taken on the night of November 25, 2006. Both telescopes were equipped with cooled CCD imagers and V photometric filters. Data was collected each minute throughout the night with breaks to allow for pier flipping as the star passed through the meridian. The data was processed with flat fields, dark and bias frames before being analyzed photometrically using the American Association of Variable Star Observers charts and comparison stars as standards. ANALYSIS AND RESULTS The spectra were examined initially using Excel graphs. Subsequently, a light curve was constructed to more easily view the change in flux throughout the eclipse. The data was transferred into Vernier Graphical Analysis, which was used to find the integral for each spectra. Taking an integral essentially returns the area beneath the curve between two wavelengths and is a measure of the flux being received from the object being studied as is shown in Figure 2 below. The units of the flux are normally in terms of energy per unit time per square area, however these spectra were not calibrated using a standard star throughout the course of the night. The units have been left in terms of pixel intensity from the FITS image over a calibrated wavelength in Angstroms.

Figure 2: Taking the Integral of a Spectrum to Obtain Flux


The integral was calculated from 4400 Angstroms to 4600 Angstroms in order to exclude excessive noise on the outer edges of the spectra. When calibrating instrumental magnitudes, the atmospheric extinction, or the dimming of starlight by the terrestrial atmosphere, must be corrected. Another source of variance in measuring flux and magnitudes is the seeing conditions during data collection process. Times of good seeing result in more starlight reaching the spectrometer's slit, while during times of poor seeing and high scintillation, less starlight reaches the slit. While measurements of the seeing conditions throughout the course of the observing run were not possible, the variance in atmospheric extinction was something that could be calculated. The longer the path length the starlight travels through the atmosphere, the more the starlight is dimmed. Thus, a star close to the horizon will be dimmer than a star closer to the zenith, and the observed brightness of a given star will change through the night as the zenith distance varies. Unless the data is corrected for atmospheric extinction in our data, the star will appear dimmer at times than it actually is, which could potentially misconstrue the data. Using MBK Team’s extinction calculator, the atmospheric extinction (in magnitudes) was calculated for each image. In the calculator, 2095.5 meters was entered as the elevation of Kitt Peak National Observatory, and the times entered were the midpoints of each image for greater accuracy. The atmospheric extinction in magnitudes was recorded in Excel. To obtain flux corrected for atmospheric extinction, the difference in flux, or extinction coefficient, was multiplied by measured integral flux. In equation 1 below, Mb -Ma is equal to the atmospheric extinction in magnitudes. Thus, 2.512(Ma-Mb) is the extinction coefficient, which is then multiplied by measured flux in units of Pixel Intensity * Angstroms. Pixel intensity is also referred to as Data Numbers (DN). See Table 2 for numerical values.

)(512.2 ab MM

b

a

b

a

II

FF −≅≅

Equation 1

Then the flux corrected for atmospheric extinction and Julian dates were graphed to form a light curve of the eclipse (Figure 3), using spectral data from 4400 to 4600 Angstroms.

Table 2


Relative Integral Flux vs. Time

0.00E+00

5.00E+04

1.00E+05

1.50E+05

2.00E+05

2.50E+05

3.00E+05

3.50E+05

2453997.6 2453997.65 2453997.7 2453997.75 2453997.8 2453997.85

Julian Date

Rel

ativ

e In

tegr

al F

lux

Figure 3: Relative Integral Flux of U Peg Spectra (4400 – 4600 angstroms)

The AAVSO 2006 Eclipsing Binary Ephemeris predicts a peak of one of U Pegasi’s eclipses to be at 8:30 UT on the night of September 19 -20, or 2453997.85417 in Julian Date. Minimum intensity of the star system should occur exactly in the predicted peak of the eclipse, when one star is most blocking the light of the other. Using photometric data from John Blackwell and Hughes Pack (Figure 4), the time between peaks in magnitude was calculated (using Vernier Graphical Analysis) to be 3 hours and 49 minutes. This value will remain constant for every eclipse of U Pegasi, and thus this value can be used to find the time at which the observed eclipse peaked. 3 hours and 49 minutes are subtracted from the listed eclipse peak time of 8:30 UT to get the time of the prior eclipse peak time, 4:41 UT or 2453997.69514 in Julian date. The starting time of the observed eclipse can be determined by subtraction of half the period (1:54:30) from the eclipse peak time, which is 2:46:30 UT or 2453997.61562 J. D. The first image was taken after the eclipse had already started, which is apparent, because the curve shows no peak but goes straight into a decline. The curve dips and begins to rise well before the eclipse peak time. The curve rises slightly and plateaus for a while before another sharp rise to peak just around the expected time when the eclipse starts over again at 2453997.77465. The decline at the start of the graph and the rise at the end concur with what is expected in a photometric curve. However, the fact that the curve rises before the eclipse peak time and the curve’s plateau are features unlike those demonstrated in photometric observations of U Pegasi. This graph shows that the relative integral flux of U Pegasi in the blue range (specifically 4400 to 4600 Angstroms) does not follow what would be expected as related to a photometric light curve. Though the photometric data from John A. Blackwell and Hughes Pack was taken using a V (green) filter, this still does not explain the discrepancies because the photometric light curve is very much the same (if not identical) in both the blue and green spectral ranges. Photometric light curves of U Peg taken with B and V filters from Stara Lesna and Skalnate Pleso observatories demonstrate this similarity. (See Figure 5)


U Peg: V Photometric Magnitude versus Julian Date

9.40

9.50

9.60

9.70

9.80

9.90

10.00

10.10

10.202454065.400 2454065.450 2454065.500 2454065.550 2454065.600 2454065.650 2454065.700 2454065.750

NMHPEA

Figure 4: Photometric Light Curve of U Peg (V filter) from PEA and NMH

Figure 5: Photometric Light Curves of U Peg (B and V filters) from Stara Lesna and

Skalnate Pleso Since both stars in U Pegasi are G2V type stars, they are almost identical in size. If this is true, the peak luminosity should be nearly twice the minimum. In a binary system where the stars were exactly the same size the sum of the luminosities is equal to twice the luminosity of one star. (See equation 2) Thus in the U Peg system, the peak luminosity


will is expected to be roughly, but not exactly, twice the minimum luminosity. Geometrically, the peak luminosity represents the point in the eclipse where both stars are visible and the minimum when one star is completely eclipsed. The peak of the relative integral flux was 3.23 and the minimum was 1.49. 3.23 divided by 1.49 is equal to 2.17, so this data validates the fact that the two stars of U Peg are, in fact, nearly the same size. In order to validate the observed data from this study, distance to U Peg was calculated using luminosity and then compared to the distance calculated using parallax from SIMBAD. The apparent magnitude of U Peg, 10.1, and its absolute magnitude, -5, were plugged into equation 3, which was then solved for distance. The distance calculated was 104.71 parsecs. A parallax of 7.18 milli-arcseconds was input into equation 4, which produced a value of 139.27 parsecs. The distance calculated using the data from this study and the distance calculated using data from SIMBAD are fairly close, which assures a fair level of accuracy in the observed data.

∑ L = 2L Equation 2

m – M = -5 +5logD

Equation 3

Distance = 1/ parallax/1000 Equation 4

The prominent absorption lines in the spectra are present throughout the eclipse, which validates that the two stars in U Pegasi are the same spectral type. If the two stars were different spectral types, a change in the spectra would be expected. The spectra taken in this study are consistent throughout the eclipse, as seen in Figure 6. Occasional abnormalities are disregarded as noise, particularly hot pixels which caused spikes that appear as strong emission lines. Figure 6: Wavelength Calibrated Spectra of U Peg

2453997.6453419

0

200

400

600

800

1000

1200

1400

4400 4450 4500 4550 4600

Wavelength (angstroms)

Inte

nsity

2453997.66356574

0200400600800

10001200140016001800

4400 4450 4500 4550 4600


Inte

nsity


2453997.67761828

0

200

400

600

800

1000

1200

4400 4450 4500 4550 4600


Inte

nsity

2453997.69927835

0

200

400

600

800

1000

1200

4400 4450 4500 4550 4600


Inte

nsity

2453997.72310661

0

200

400

600

800

1000

1200

1400

4400 4450 4500 4550 4600


Inte

nsity

2453997.73889051

0

200

400

600

800

1000

1200

1400

4400 4450 4500 4550 4600

Wavelength (angstroms)In

tens

ity

2453997.75447151

0

500

1000

1500

2000

2500

3000

4400 4450 4500 4550 4600


Inte

nsity

2453997.76913819

0

500

1000

1500

2000

2500

3000

3500

4400 4450 4500 4550 4600


Inte

nsity

2453997.78387523

0

500

1000

1500

2000

2500

3000

3500

4400 4450 4500 4550 4600


Inte

nsity

2453997.8048265

0

500

1000

1500

2000

2500

3000

4400 4450 4500 4550 4600


Inte

nsity

CONCLUSION The presented light curve of U Pegasi’s relative integral flux over time constructed using spectral data from 4400 to 4600 Angstroms does not correlate to photometric light curves of U Pegasi. Though the initial decline and finishing rise of the eclipse in our light curve appears similar to photometric light curves, the middle section where the curve plateaus is quite different. The accuracy of this study is supported by the validation of the ratio of maximum luminosity to minimum luminosity and the distance calculated from observed apparent magnitude. To learn more about this star system and the relationship between


change in luminosity as measured by spectral and photometric data, all spectral ranges of U Pegasi should be studied and compared to photometric light curves. Also, spectral data of U Peg in the blue range should be taken again to test whether the observed behavior is consistent. ACKNOWLEDGEMENTS We would like to thank those specific individuals who assisted in this project either directly or indirectly: John Blackwell (Phillips Exeter Academy), Dr. Katy Garmany (NOAO), Hughes Pack (Northfield Mount Hermon) and all the staff, technicians and other astronomers at Kitt Peak National Observatory. REFERENCES Allegheny Observatory, Publications of the Allegheny Observatory of the University of Pittsburgh. 2001 <http://digital.library.pitt.edu/parallax/aboutparallax.html> (12/08/06). American Association of Variable Star Observers, Prediction of Minima for Eclipsing Binary Stars. 2006. <http://www.aavso.org/observing/programs/eclipser/ebmono.shtml> (12/8/06). Astronomical Institute of the Slovak Academy of sciences, Photoelectric photometry at the Stara Lesna and Skalnate Pleso observatories. 2002. <http://www.astro.sk/~pribulla/lc.html> (12/12/06). Edalati, M.T. and Taheri, M. “The Photometric Observations and the Light Curve Analysis of U Pegasi,” Astrophysics and Space Science v. 278. 2001 Issue 4: p. 373-382. Smithsonian/NASA ADS. Online (12/10/06). Djurasevic, G and Rovithis-Livaniou, H. “A photometric study of the W UMa-type system U Pegasi,” Astronomy and Astrophysics v.367. 2001: p.840-847. Smithsonian/NASA ADS. Online (12/12/06).

Zhai, D.-S;Leung, K.C.; Zhang, R.-X. “A New Photometric Study of the Binary U Pegasi,” Chinese Astronomy and Astrophysics v. 9, June 1985, p. 98-105. Smithsonian/NASA ADS. Online (12/10/06).


An Exploration for Extended Clusters in M33 TOP Run, October 2006

Nathan Stano Teacher: Ardis Herrold, RBSE 1998, TLRBSE 2001

ABSTRACT The goal of the study was to locate and characterize extended clusters in the Triangulum Galaxy (M33). Images of the sky field around M33 as well as a control cluster in the Andromeda Galaxy (M31) were obtained using the .9 Meter Telescope at Kitt Peak National Observatory (KPNO) near Tucson, Arizona. Due to errors with the Kitt Peak images, other images were obtained from the Hubble Space Telescope Archives. The images were examined to separate extended clusters from other phenomena, such as globular clusters or planetary nebulae. V-I data was used to correlate the clusters to those in M31. The presence of these clusters in M33 may relate the histories of the two galaxies or give insights into the formation of extended clusters. Results turned up no extended clusters, meaning that M33 has either already absorbed the clusters through tidal stripping, or that none of these clusters have existed. INTRODUCTION In April 2006, astronomers of the Royal Astronomical Society (RAS) were doing a high-resolution survey of the Andromeda Galaxy (M31) when they discovered a new form of star cluster, the extended cluster (Huxor, et. al, 2006) These clusters have a similar density to traditional globular clusters (GCs), but have a larger half-light radius, Rh, meaning that they are more extended than GCs. Their V-I colors are consistent with the metal poor stars in the globular cluster populations, at least in M31, without evidence of young main-sequence stars. These clusters help to fill a gap between traditional GCs and dwarf spheroidal galaxies (dSp), the first having almost no dark matter, the other being highly dominated by dark matter. This gap is currently being filled by a few recently-discovered cluster types: extended clusters and faint fuzzies. There have been relatively few postulations on the formation of extended clusters. There are two dominating theories: that they are the remnant of dwarf spheroidals that collided with the galaxies, or they are the remnants of tidally stripped dwarf spheroidals. Since only a limited number of these clusters have been found only in M31, no solid conclusions on their formation have yet been made. DESIGNING THE STUDY Since these clusters are very dim, they can only be effectively observed from Earth in the closest of galaxies. M33 is the next farthest large spiral galaxy from Earth, after M31. Other close galaxies were eliminated from the study based on magnitude and position in the sky. Also, constraints of the .9-Meter Telescope had to be taken into account. The .9-Meter Telescope’s limitations are that it can only observe objects with a declination greater than -20° and a right ascension within 4 hours of the Meridian. Since


M33 was well within these constraints during the period offered for observation, it became a perfect target. Other galaxies, such as M101 and M51 were eliminated because of these constraints. Using star cluster catalogues of proposed globulars in M33, a list of candidates was made. In Huxor et. al., it was found that such catalogues misclassified extended clusters in M31 as globulars. To ensure the best chance of getting an extended cluster, clusters described as “diffuse” were collected for study. The problem inherent in this method is that open clusters are the ones that are often described as diffuse. Several catalogues were used: (Mochejska+, 1998), (Chandar+, 2001), (Bedin+, 2005), and (Chandar+, 1999), to compile a list of possible targets. Some catalogues had an uncertainty factor, which was used to compile the best list of targets, especially those with large radii. Finder charts were printed using the VizieR and Simbad Astronomical Databases. Images from the Hubble Space Telescope Catalogues, by location query from the candidate cluster list, were obtained in FITS format and were opened in Image J as well as DS9. These images were used to help find clusters which may have been dimmer than the .9 Meter can “see”. They also have a higher resolution. Problems arose in the analysis of the Kitt Peak images, as coordinates of objects could not be directly obtained from the images in a matter within the scope of the study. Instrumentation and Setup NOAO accepted the proposal for the nights of October 21 and 22, 2006 and my teacher, Mrs. Ardis Herrold observed at Kitt Peak National Observatory on the .9-Meter Telescope since I was unable to make the trip. Over these nights, an area containing and slightly around M33 was observed, as well as a known extended cluster in M31. The .9-Meter Telescope is located on Kitt Peak as part of the Kitt Peak National Observatory. A consortium of universities, including the University of Wisconsin, Indiana University and Yale, as well as NOAO, operates the telescope. It has a .9-meter diameter mirror, which is about 36 inches. The telescope has the filters to take images in the U (Near UV), B (Blue), V (Green/Visual), R (Red), and I (Near IR) in H-Alpha. For this study, I have chosen to use the V band and I band filters, because these were used in the previous study of M31, as a comparison to GCs. The CCD camera, which is 2048x2048 pixels at .6 arc seconds per pixel, is cooled by a dewar to -105° C. The container holds the CCD camera as well as liquid nitrogen, which is refilled ever 12 hours. This is done to keep noise to a minimum. Data were reduced at the observatory by Dr. Katy Garmany and sent to me.

Calibration Techniques Certain procedures are taken to assess the reliability of the instrumentation and to standardize it. In order for this to occur, a series of calibrations are completed. The first calibration is taking an image of a white panel inside the dome. These images show where there are imperfections on the CCD chip, or possibly dust particles on the window of the CCD or the filter(s, which show up as circles on the image. These images allow the observer to account for such imperfections. Next a series of flat frames were taken. These are frames with the CCD shutter closed, to identify the sensitivity of each pixel.


Lastly, the telescope is pointed at the sky, and different focus settings are tested. This shows many images of the same star field as the telescope cycles through focus settings. Data Collection The first computer controlled the telescope. Coordinates for the star to be observed were entered, as well as the name and additional notes (which included the magnitude). The user would then select the star field to be observed from the newly made list, and then the telescope would slew to the area. All information was also recorded on an observer’s log sheet. Unfortunately, the guide motors, which allow the telescope to move with the sky and “track” the target, were not operational, decreasing image quality and causing stars to look like ellipses as opposed to circles. Another computer did the imaging/calibrations. Data is sent to it directly to and it displays the unprocessed image after it is done on a monitor for the observers to see. A third computer was used to record a log of all observations. OBSERVATIONS AND DATA REDUCTION The images taken at Kitt Peak were excluded from the study due to the poor quality images produced without the guide motor. Images from the Hubble catalog were used instead. These images were scrutinized in several ways. Photometric data was obtained by using the program Image J. This is done by comparing the mean relative brightness of a cluster to a ratio of a known cluster’s relative brightness to published magnitudes. This would give a mean relative magnitude for the cluster. By subtracting the I magnitude from the V magnitude, you would get the V-I magnitude, and a relative measure of a cluster’s redness. Methods of Differentiating Clusters The first method was to take a three dimensional intensity profile of the proposed cluster and a baseline of empty space around it. Should any of the clusters be extended, they would show a profile with many individual peaks clustered together, if they were globulars, they would show one central raised area. This is due to the fact that extended clusters do not typically have the same light distribution as globulars. While most of the light in globular clusters is concentrated in the center, making it the noticeably brightest part, extended clusters do not exhibit this feature, and tend to have relatively the same brightness throughout. The ease in picking out single stars in extended clusters is much greater than in globulars, due to the differences in density. V-I magnitudes were also compared to those of an average for globulars in M33 to see if the clusters had similar V-I values, one of the necessary components of being an extended cluster.


DATA Image RA Dec

Exposure Time Filter/Grating Candidates Candidate RA Candidate Dec Identifiers Mag V-I

No. (h:m:s) (d:m:s) (s) (h:m:s) (d:m:s) 1 01 34 41.77 +30 44 36.12 60 F555W (L2, 82) 01 34 40.39 +30 46 00.7 CBF 142 15.65 0.63 2 01 34 38.89 +30 45 00.29 300 F814W (L2, 82) 01 34 40.39 +30 46 00.7 CBF 142 15.02 3 01 34 00.79 +30 41 03.22 1300 F555W (L2, 39) 01 34 00.45 +30 41 21.8 CBF 99 18.21 0.78 4 01 34 00.80 +30 41 03.43 1300 F814W (L2, 39) 01 34 00.45 +30 41 21.8 CBF 122 17.43 5 01 34 05.60 +30 38 39.87 1300 F555W (L2, 62) 01 34 08.56 +30 39 22.4 CBF 122 17.4 6 01 34 00.80 +30 41 03.43 1300 F814W (L2, 62) 01 34 08.56 +30 39 22.4 d 7 01 33 33.39 +30 37 46.16 600 F555W (L2, 91) 01 33 28.40 +30 36 23.9 d 8 01 33 33.39 +30 37 46.16 600 F814W (L2, 91) 01 33 28.40 +30 36 23.9 d 9 01 33 32.67 +30 48 58.55 1300 F814W (L2, 51) 01 33 30.18 +30 49 29.0 CBF 111 19.22 10 01 33 32.67 +30 48 58.23 1300 F555W (L2, 51) 01 33 30.18 +30 49 29.0 CBF 111 20.1 0.88 11 01 33 53.05 +30 33 51.23 160 F555W (L2, 77) 01 33 49.56 +30 34 25.6 CBF 137 18.02 0.92 12 01 33 53.05 +30 33 51.23 160 F814W (L2, 77) 01 33 49.56 +30 34 25.6 CBF 137 17.1 13 01 34 00.79 +30 41 03.22 1300 F555W (L2, 42) 01 33 56.91 +30 41 37.3 CBF 102 18.64 0.71 14 01 34 00.80 +30 41 03.43 1300 F814W (L2, 42) 01 33 56.91 +30 41 37.3 CBF 102 17.93 15 01 34 32.42 +30 38 24.40 1300 F814W L2,16 01 34 33.05 +30 38 14.1 CBF 76 18.55 16 01 34 32.41 +30 38 24.70 1300 F555W L2,16 01 34 33.05 +30 38 14.1 CBF 76 19.55 1 17 01 34 42.56 +30 52 41.39 1300 F814W 24 01 34 45.84 +30 53 03.8 CBF 84 18.74 18 01 34 42.52 +30 52 41.42 1300 F555W 24 01 34 45.84 +30 53 03.8 CBF 84 19.7 0.96 19 01 34 22.33 +30 43 36.66 160 F555W 95 01 34 14.99 +30 41 18.9 d 18.3* 0.81 20 01 34 11.39 +30 42 19.00 100 F814W 95 01 34 14.99 +30 41 18.9 CBF 155 17.49 21 01 33 57.80 +30 43 40.32 70 F555W 92 01 34 08.70 +30 42 55.1 d 18.5* 0.68 22 01 34 11.39 +30 42 19.00 100 F814W 92 01 34 08.70 +30 42 55.1 CBF 152 17.82 23 01 34 42.56 +30 52 41.39 1300 F814W 22 01 34 43.14 +30 52 18.9 CBF 82 18.95 24 01 34 42.52 +30 52 41.42 1300 F555W 22 01 34 43.14 +30 52 18.9 CBF 82 19.72 0.77 25 01 34 32.42 +30 38 24.40 1300 F814W 14 01 34 35.26 +30 38 29.8 CBF 74 17.91 26 01 34 32.40 +30 38 24.50 1100 F555W 14 01 34 35.26 +30 38 29.8 CBF 74 18.74 0.83 27 01 34 16.11 +30 39 38.56 350 F555W 97 01 34 16.48 +30 40 27.3 CBF 157 18.9 0.97 28 01 34 05.60 +30 38 40.18 1300 F814W 97 01 34 16.48 +30 40 27.3 CBF 157d 17.93 29 01 34 04.14 +30 21 07.44 100 F814W 87 01 34 09.60 +30 21 30.5 CBF 147 17.89 30 01 34 04.14 +30 21 07.44 200 F555W 87 01 34 09.60 +30 21 30.5 CBF 147 18.45 0.56 F555W(V) F814W(I) *= published mag used d=image does not contain target, a "dud"

See Caption for Figure 1: This is a 3D profile plot of the light intensity of cluster CBF 142. In the data spreadsheet, the meanings of the column headings are thus: “Image #” refers to the number of the image, which was given for organizational purposes, “Instrument” refers to the camera that took the image, “RA” and “Dec” refer to the coordinates of the center of the image, “Exp. time” refers to the length of the exposure in seconds, “Filter/Grating” refers to the filters used in the image F555W refers to V and F814W is I, “Candidates” refers to the identification number of the cluster in the catalogue used, “Candidate RA and Dec” refer to the coordinates of the candidate, “Identifiers” refers to any identifiers from the SIMBAD database, “Radius” refers to the length of the radius in


pixels, “Mag. ” refers to the mean brightness obtained from the image, and “V-I” refers to the value of V-I.

ANALYSIS AND DISCUSSION At their most basic, these clusters have the same relative shape as globular clusters, though they are visibly more extended. The control cluster in M31 will be used as a visual reference as to how the clusters should appear. This was the first criterium for catagorization of the clusters. Next, the diameter of the cluster was measured, as these clusters should have a larger diameter than globular clusters. Finally, V-I photometry was analyzed, as these clusters should have high V-I values, however, this would not differentiate them from globular clusters. A galactic average for M33 was compiled by averaging the V-I values of 102 known globulars in M33, which was .713. The average of my candidates was .822. As three dimensional profiles were analyzed, they all mached profiles of globulars in both M31 and M33. They displayed a singular raised promentory in the center of the profile. This would indicate that the clusters were not extended, but rather were globulars. V-I values were consistant with those of globulars in M33, indicating further that the clusters were not extended but rather globular. CONCLUSION No extended clusters were detected by this study. While Huxor et. al. was able to explore for clusters in greater depth, with the use of specialized computer programs, the methods used by this study in both the attempted location of and analysis of possible clusters should theoretically work. The intensity plots, while a very simplified test, should show the large separation of stars that is indicative of extended clusters. Although the conclusions of this section are based off the assumption that these methods are valid, I do acknowledge that it was not the same procedure used in the original study. As previously stated, no clusters were detected by the study, this has repercussions for the recent past of M33. All leading theories about how these extended clusters are formed involve the interaction between large galaxies and smaller dwarfs. The lack of such clusters around M33 would suggest that its recent past has been calm, at least compared to M31. Two explanations can account for this. Either M33 has already absorbed its neighbors long ago, or it has had no neighbors to absorb. If it has absorbed its neighbors long ago, it would have either absorbed all the matter off the dwarfs as they collided, or been pulled off over time by tidal stripping. It is also possible that M33 has had no neighbors, or at least none that it could absorb. Today, scientists believe that M33 has only one orbiting dwarf galaxy, the Pisces Dwarf, which is only a proposed dwarf. If M33 had no neigbors to absorb, it would likely not show these new clusters. This gives us a better idea of how galaxies evolve. These clusters are another component of the evolution of galaxies. It would seem to give creedence to the theory that galaxies are the end result of the merger of many smaller objects, a leading theory on the formation of galaxies, as opposed to a simple crystalization of existing matter.


This also shows some of the results of the mergers of galaxies and smaller objects. We can then better come to understand exactly how galaxies form, as well as how they come together and merge. Finally, since these clusters are very old, they give insight into the structures of matter in the early universe. If, as is currently thought, these clusters are as old as some scientists think, they would be a valuable insight into how stars came together and how these may have come together to make galaxies. ACKNOWLEDGEMENTS I would like to thank my teacher, Ardis Herrold, for her help and support, as well as the RBSE program for the opportunity to be published, as well as the opportunity to take data at Kitt Peak. BIBLIOGRAPHY Aladin Sky Atlas. Centre de Données Astronomiques de Strasbourg. <http:// aladin.u-strasbg.fr/aladin.gml>. Huxor et. al. 2006, “A new population of extended, luminous star clusters in the halo of M31 SIMBAD Astronomical Database. Centre de Données Astronomiques de Strasbourg. <http://simbad.u-strasbg.fr/Simbad>. VizieR. Centre de Données Astronomiques de Strasbourg. <http://vizier.u-strasbg.fr/viz-bin/VizieR>.

Sparke, Linda S. and John S. Gallagher. Galaxies in the Universe. Cambridge, New York: Cambridge University Press, 2000.

Elmegreen, Debra Meloy. Galaxies & Galactic Structure. Upper Saddle River, New Jersey: Prentice Hall Inc., 1998.

Hubble Space Telescope Catalogues , MAST HST Multimission Archive at STScI. <http://archive.stsci.edu/hst/>. Globular Clusters in M31 and M33 (Mochejska+, 1998) VI photometry of M33 star clusters (Chandar+, 2001) New star clusters in M33 (Bedin+, 2005) Star Clusters in M33 (Chandar+, 1999)


Spectral Analysis of Blazar S5 0716+714 using Spitzer Infrared Space Telescope and New Mexico

Alekzandir Morton, Manutej Mulaveesala and Thomas Travagli Deer Valley High School

Teacher: Jeff Adkins, TLRBSE 2002

BACKGROUND An Active Galactic Nucleus (AGN) is the center of an active galaxy that is believed to be powered by a super-massive black hole. They emit large amounts of radiation, spanning all across the electromagnetic spectrum. The most prominent emissions are of Optical Light, Infrared, Radio and Gamma rays. Our project examined two aspects of a particular AGN and recorded a Spectral Energy Distribution (SED) diagram and light curves. An SED is based on all of the typical forms of electromagnetic radiation. Both the Spitzer Infrared Telescope and the New Mexico Skies ground-based telescopes were used. The ground-based telescopes observed the target using luminous filters while the infrared telescope observed the infrared radiation from the target. The term “blazar” is a more detailed classification of AGN. The aforesaid emissions come out in a large relativistic jet that is of intense energy and magnitude, which includes all types of radiation from the entire range of the electromagnetic spectrum. The jet is typically perpendicular to the plane of the accretion disk, which is the main collection of galactic matter as it falls into the AGN. There have been previous observations of the synchrotron radiation emitted from the Active Galactic Nucleus of S5 0716+714, however it is a rare occurrence for a flare to increase the synchrotron peak and create a ‘bump’ in the SED graph. A collection of observations of blazars showed that “No strong shifts in the synchrotron peak frequency are reported.” (BeppoSax ToO article). Thus, the observation of a shift in the synchrotron peak in the Spitzer Project data was highly unusual and markedly interesting. The similarity between the time-scales and variability is crucial to understanding the light curves of the target. The range of time-scales of the variability is broad. “On inter-night time scales, a bluer when brighter correlation was found when the object was in an active or flaring state, but this trend was absent during the quiescent time.” (C.S. Stalin et al). When a break in the spectrum occurs, it may be an indication of a temporary flare that happened in the blazar. These can be detected by the variability in the spectrum of the synchrotron radiation. The same BeppoSax observatory, on a different instance, noticed, “In the case of S5 0716+714 we detected again fast variability only in the synchrotron part of the spectrum.” (BeppoSax ToO article).


OBSERVATIONS For this research, two different telescopes were used. Two cameras on the Spitzer Infrared Space Telescope were used to take infrared images of the target. The Multiband Imaging Photometer for Spitzer, or MIPS, was used on April 8, 2006 and the Infrared Array Camera, or IRAC, was used on April 28, 2007. Because these two cameras could not operate simultaneously, a time period of three weeks was required between the two observations. During this time, and also before the MIPS and after the IRAC observations, ground based optical telescopes were used to take images of the target. A New Mexico Skies telescope was used by the authors to take luminous, red, and blue filtered images. Also, amateur astronomers contributed by sending in images. 91 images were taken using New Mexico Skies, and 61 were contributed by amateur astronomers. Reduction and Analysis of New Mexico Skies Data To Reduce the New Mexico Skies images, Image J was used to find brightness counts for the target and 8 standards in each image. Fathom was used to graph and the model the Magnitudes of the Standards against their brightness counts. The model was used to output a magnitude for the target along with an error bar. Once all the images were reduced, light curves were constructed. A light curve is a graph of Magnitude over time. They were used to determine whether the target had significantly changed during the two Spitzer Observations. As seen below, all the error bars overlapped in the light curves, indicating that there was no significant change. Light Curves

Fig. 1. Light curve using Red filters. Magnitude v. Time (Julian Days)


Fig. 2. Light curve using Luminous filters. Magnitude v. Time (Julian Days)

Fig. 3. Light curve using Visible filters. Magnitude v. Time (Julian Days) Spitzer Data Reduction The data from Spitzer, 7 images taken by the IRAC and MIPS cameras was reduced at the Spitzer Space Science Center in Pasadena, CA. With the help of a researcher, the exact keystrokes of the complicated data reduction were executed. The program necessary to reduce Spitzer images is called IRAF. The main purpose is to receive the magnitudes from the program. When operating IRAF, it is recommended to use a manual or written reference or a researcher who knows the program. After the operations are completed, three magnitudes will be displayed, pick the middle one as it represents the photometry measurement of the object. Derive the error bars by examining the range and dividing the difference by two, making the magnitude of the object with a plus or minus approximation of possible error. The final reduction is a data point of flux with an error estimate.


SED Data Flux (Jy) Lum Avg Flux�3.61474E-06

Lum Stdev 4.38503E-07 Error of Mean

6.02331E-08

Red Avg Flux�0.012262948

Red Stdev 0.000853732 Error of Mean

0.000220433

Vis Avg Flux�0.013075479

Vis Stdev 0.002796592 Error of Mean

0.001141704

Blu Avg Flux�0.012383625

Blu Stdev 0.002468409 Error of Mean

0.001425137

Ifr Avg Flux�0.011084003

Ifr Stdev 0.000203555 Error of Mean

0.000143935

� � Wavelength

(microns) Flux�(microJy)

Red�0.7 1.23E+04 Blu�0.556 1.31E+04 Ifr�0.8 1.24E+04 Irac Ch. 1�3.6 5.31E+04 Irac Ch. 2�4.5 6.10E+04 Irac Ch. 3�5.8 7.23E+04 Irac Ch. 4�8 8.69E+04 Mips Ch. 1�24 2.37E+05 Mips Ch. 2�70 3.63E+05 Mips Ch. 3�160 2.24E+05 Radio�210000 1.73E+06

Fig. 4. The data used to create the SED


SED

Fig. 5. The SED of the target. Flux (microJy) v. Wavelength (microns). Dr. Steve Rapp in Green Bank, West Virginia, submitted the radio data point

SED Analysis The Spectral Energy Distribution shows the flux of the target at several different wavelengths along the electromagnetic spectrum. This particular SED contains data points that lie in the visible light, infrared, and radio spectrum. In the infrared range of the SED, there was a slight irregularity, comprised of a rise and fall between three points0 (a bump) of flux reduced from MIPS camera images. This bump is peculiar and quite interesting, because it has never been observed for this target. Normally, the graph would be a rising logarithmic curve, caused by the synchrotron radiation induced by the jet of the blazar. When infrared flux is added to the synchrotron radiation in the SED, the curve typically remains unchanged. However, when there is a large source of infrared, there will be a change in the shape of the curve, illustrated by the bump seen in this SED. Modeling The SED yielded results that were uncommon for this AGN, and they were explored in further detail with the creation of a mathematical model. The model was built using Fathom, a software program for the plotting and analysis of data, and was based on Planck’s Law of Blackbody radiation.


Derived by Planck in 1900, the expression refined earlier work by Wilhelm Wien, and is capable of accurately predicting the spectral intensity of blackbody objects. Planck’s Law is as follows:

v is the frequency in Hertz T is the temperature of the black body in Kelvins h is Planck’s constant = Joule per Hertz c is the speed of light = meters per second k is Boltzmann’s constant = Joule per Kelvin

The model correctly displayed the behaviors that characterize a blackbody curve, as described through the Stefan-Boltzmann Law and Wien’s Law. The former states that the intensity of a blackbody is proportional to its temperature to the fourth power, and that as temperature increases, intensity increases at all wavelengths. Wien’s Law states that as the temperature of a black body grows, the wavelength of the peak in the blackbody curve decreases. Both of these laws are clearly noticeable in the graph, which demonstrates its credibility as a functioning model and increases the likelihood that it is working properly. The model was superimposed on a graph, which also contained the SED from our measured data, and was vertically scaled for purposes of comparison (flux scaling is arbitrary, so this does not present a problem). The temperature, T, was used as a parameter and adjusted to fit the curve exhibited in the SED. It matched the general shape of the bump at approximately 100 Kelvin, which is often considered to be relatively cool for an object of this type. Images of the Target

Fig. 6. Image captured by Spitzer


Fig. 7. Image captured by New Mexico Skies (Visible) CONCLUSION The data reductions from the Spitzer images showed seven data points, which were linear on a logarithmic graph, with the exception of a slight “bump” in the curve between the MIPS data points. This slight difference changes the typical perception of a blazar having a complete logarithmic curve due to synchrotron radiation. The bump in the certain range of the infrared implies that there is an extra source of infrared that is significant enough to change the SED. The reason for this source of infrared could not be conclusively determined from the observations, because there are a few possibilities. One possibility is that the infrared is being generated in the torus. However, the blackbody curve somewhat contradicts this possibility, because it determines the temperature of the torus to be far lower than the amount needed to generate enough infrared. Another suggested, and perhaps more plausible possibility, is that the source is from the background stars of the host galaxy (the galaxy containing the blazar). A possible follow-up project could be to accurately find the redshift of the galaxy, so that the blackbody radiation curve can accurately find the temperature of the torus, because the number used for the redshift in this project was found from a different source. S5 0716+714 is a target worth scientific scrutiny, as it has proved to be a unique AGN.


REFERENCES Global Telescope Network Landessternwarte Heidelberg-Konigstuhl. <http://gtn.sonoma.edu/participants/catalog/details.php?object=3>. Credited to:

Hessman, Frederic V. . <http://www.astro.physik.uni-goettingen.de/~hessman/ImageJ/Astronomy/index.html>.

Hinckley, Brielle. "Micro-Variablity of 4C 29.45 Using the Spitzer Space Telescope, and Ground Based Telescopes." RBSE Journal 2006: 12-18. Krolik, Julian H. . Active Galactic Nuclei. Princeton University Press, 1999.

Tagliaferri, and Ghisellini. The BL Lac objects OQ 530 and S5 0716+714. Simultaneous observations in the X-rays, radio, optical and

TeV bands. 7 Jan 2003. <arXiv:astro-ph/0301117>


The Age and Distance of the Open Cluster NGC 2345 Christina Clemens and Rachel Reece

Sullivan South High School, Kingsport, TN Teacher: Thomas Rutherford, TLRBSE 2005

ABSTRACT The age and distance to the open cluster NGC2345 was determined. Using CCD photometry, images in B and V were obtained using a 14-inch telescope located in New Mexico. The open cluster was found to lie at a distance of 2280 parsecs (7430 light years) from the earth. The V and B-V magnitudes were plotted to form an H-R diagram. This plot was overlaid onto a ZAMS (Zero-Age Main Sequence) graph in order to determine the age of the cluster. Based on the MSTO (Main Sequence Turn-Off) from the ZAMS, the age of the cluster is approximately 320 million years. INTRODUCTION An open cluster consists of stars that are of approximately the same distance, age, and chemical composition (Frommert 2006). It is thought that our solar system was once part of an open star cluster that dispersed over time (MacRobert 2007). Because of this, the study of open star clusters is vital to astronomy. Studying them can help with a better understanding of the solar system. The cluster NGC2345 was selected because it had not been recently examined. Prior studies involved the use of photoelectric photometry (Moffat 1974). The current study involves the use of a Charged-Coupled Device (CCD) for data collection. When the V magnitudes are subtracted from the B magnitudes,the stars’ true colors may be determined. Because the light from the cluster must travel through interstellar dust, the stars appear both dimmer and redder than they actually are. The values of the stars’ magnitudes were adjusted to compensate for this. OBSERVATION AND DATA REDUCTION The images of the cluster were taken through two filters, B (blue) and V (green). The star’s true color (color index) was determined using the following equation:

B-V= color index of the star The data used in the project was collected with a remotely-operated Celestron C14 Schmidt-Cassegrain telescope provided by New Mexico Skies. Because of the presence of a nearly full moon, stars with magnitudes of 15 or fainter were not adequately recorded. Their true magnitudes and relationships to the cluster could not be determined. Six images were taken through each filter, with exposure times of 1 minute each. The images obtained were then processed using MaximDL’s kernel filter so that excess hot pixels were removed from the images. The program’s kernel filter setting was changed


from 20% to 10%, the default setting of 20% being too aggressive for the fainter stars in the image. The three best images were then chosen from each set and stacked, again using MaximDL, in order to provide deeper coverage. This gave the equivalent of a single three-minute exposure with each filter. Image 1. NGC 2345 was imaged through two different filters, B and V. Here an image of the cluster in V is shown. The image colors were reversed in order to make the stars easier to examine.

The program AutoStar Suite was then utilized to invert the images’ color, producing black stars on a white field. This made the star images more visible and made it easier to determine star brightness. The numbering system used by Moffat (1974) was used in this study (see Image 2).


Image 2. Below is the identification system used by Moffat (1974). Many of the fainter stars in the image are not part of the cluster.

NGC2345-2 was used as the reference star to determine all other star magnitudes in both B and V (SIMBAD) by comparison. The magnitudes were then entered into an Excel spreadsheet and the stars’ V magnitudes were subtracted from their B magnitudes to determine the their color indices. However, adjustments had to be made to the magnitudes because of interstellar dust. Interstellar dust causes the light from distant stars to be both reddened and dimmed (Strobel 2001). To compensate for the reddening, 0.62 (WEBDA 2006) was subtracted from each star’s B-V magnitude. In addition, a value of 1.9 (Garmany 2007) was subtracted from each star’s V magnitude to correct for interstellar extinction. ANALYSIS AND RESULTS Once the (B-V) values were determined for each star, the stars were classified by spectral type (O, B, A, F, G, K, and M) according to the criteria presented in Table 1.


Table 1. Stars may be classified into spectral types based upon their (B-V) values. These values also provide information on the life expectancies and masses of the stars (Washington 2005).

Spectral Type

Color B-V

Lifetime (years)

Mass (solar masses)

O -0.4 <106 40 Mo B -0.2 3 x 107 10 Mo A 0.2 4 x108 2.3 Mo F 0.5 4 x 109 1.4 Mo G 0.7 1 x 1010 1.0 Mo K 1.0 6 x 1010 0.7 Mo M 1.6 >1011 0.3 Mo

The numbers of each class of stars were then plotted as seen in Graph 2. The results revealed three O class stars and a large number of B and A class stars, suggesting the cluster was relatively young. The O-type stars 27 and 36 are thought to not be part of the cluster (Moffat 1974). They were not used in this study. An H-R diagram was then created by plotting V magnitudes verses B-V magnitudes. Onto this diagram, the Schmidt-Kaler ZAMS graph (1982) was overlaid. This allowed the turnoff point to be established, allowing an estimate to be made of the cluster’s age. In Graph 1, the star located farthest to the left is the youngest in the field, and the cluster can be no older than its youngest star. With NGC2345-27 and 36 not being part of the cluster, NGC2345-1 is the youngest star. However, star 1 is of dubious relationship to the cluster because it does not conform to the turnoff point in the graph. Assuming that the star cluster did collapse uniformly, NGC2345-35 would then be the youngest star in the cluster. Star 35 is an older class B star, making it between 30 and 400 million years old. According to interpolation of the data in Table 1, NGC 2345 is approximately 230 million years old.


Graph 1- When a ZAMS graph is overlaid upon an H-R diagram and the (B-V) axes aligned, the point where stars begin turning off of the main sequence can be established. The arrow indicates the star used as the ZAMS turnoff.

In order to determine the distance of the cluster from Earth, a cluster member’s absolute magnitude must be known. The absolute magnitude is the brightness that a star would have if it were 10 parsecs, 32.6 light years, from Earth (Comins 2003). Since the stars in the open cluster have unknown absolute magnitudes, a standard star was found which had a color index equal to one of the stars in the cluster, and had a known absolute magnitude. The star Regulus (Alpha Leonis) was chosen to be the standard star. Since both stars have similar B-V values, their absolute magnitudes are also similar. Star 42’s apparent V magnitude of 11.49 was used. Regulus’ absolute magnitude, 0.3, would be also be used. Using the distance modulus equation (Comins 2003):

d=10(m-M+5)/5 (equation 2)

where m=apparent V magnitude, M=absolute V magnitude, and d=distance in parsecs, it was calculated that the cluster is located about 2,300 parsecs or 7,400 light years away.


Class Distribution of Stars in NGC 2345

0

5

10

15

20

25

30

35

40

O B A F G K M

Graph 2- A frequency graph of the stars in NGC 2345. Most members of the cluster were relatively young A and B types. Very few fainter M types were detected, possibly due to the presence of a nearly full moon.

DISCUSSION In determining the age of the cluster the star deemed closest to the turn off point was NGC2345-35 after referring to Moffat (1974) and finding that two O stars were not included in the cluster. It was decided that the third O type star was also not a cluster member because it was positioned beyond the ZAMS turn off. This data provided an age for the cluster of approximately 230 million years. Even though M class stars are the most numerous type of star, this search detected few of them. This could be because the full moon prevented those stars from being detected by the telescope, thus not showing up in the final image. The distance calculated for the cluster is approximate since the position of a star within the cluster is unknown. As is evident in the results of both age and distance, NGC2345 is a relatively near and moderately young open cluster. It was also deduced from the studies of Moffat (1974) that this cluster was moderately young; so there is reasonable agreement between these results and the results of the previous study. SUMMARY In 1974, A. Moffat measured to distance to NGC 2345, determining it to be 1750 parsecs (5705 light years) away from the earth. Moffat’s analysis also placed the cluster’s age at 60 million years. By analyzing the current data obtained from the New Mexico Skies images it was calculated that the open star cluster NGC2345 is 230 million years old and located 7,430 light years away.


ACKNOWLEDGEMENTS Thanks to Dr. Katy Garmany, for advising us on our reddening problem, and to Mr. Thomas Rutherford for patiently teaching, guiding, and encouraging us through the entire project. Thanks to David Suit for his efforts in helping when he could. Also we appreciate David Clemens for his interest and support and thanks to New Mexico Skies for use of their equipment. REFERENCES University of Washington. "Cluster Color-Magnitude Diagrams and the Age of Stars." 16 May 2005. 19 Mar. 2007 < http://www.astro.washington.edu/labs/clearinghouse/labs/Clusterhr/cluster.html>. Comins, Neil F., and William J. Kaufmann III. Discovering the Universe. 6th ed. New York: W.H. Freeman and Company, 2003. 277. Dolan, Chris, ed. "Stellar Brightness." 31 Mar. 2007. <http://www.astro.wisc.edu/>. Frommert, Hartmut, and Christine Kronberg, eds. "Open Star Clusters." 20 July 2006. 22 Mar. 2007. <http://www.messier.obspm.fr/open.html>. Garmany, Katy. 2007 March 22. A Photometry Question [email]. Accessed 2007 March 23. MacRobert, Alan. "The Lives of Open Star Clusters." Sky and Telescope 113.3 (2007): 75. Moffat, A.F.J. "NGC 2345, a Moderately Young Open Cluster in Canis Major." Astronomy Astrophysics Supplement Series 16th ser. (1974): 33-42. NASA Astrophysics Data System. "Simbad Query Result: Simbad Search 07 08.3-13 11." SIMBAD. Donnees Center of Astronomics in Strasbourg. 8 Mar. 2007 <http://www.simbad3.u-strasbg.fr>. Schmidt-Kaler, Th. 1982, Landolt-Börnstein, Numerical data and Functional Relationships in Science and Technology, New Series, Group VI, vol. 2(b), ed. K. Schaifers, & H. H. Voigt (Berlin: Springer Verlag), 14 Strobel, Nick. “Interstellar Medium (ISM). 2001. <http://www.astronomynotes.com/ismnotes/s2.htm> WEBDA. Institute for Astronomy of the University of Vienna. 2006 September 24. <http://www.univie.ac.at/webda/cgi-bin/ocl_page.cgi?dirname=ngc2345>


Appendix 1- The locations and stellar magnitudes of all stars used in this study are shown in the following table. The stars marked with an asterisk at first appeared to be members of the cluster, but were later deemed to be either foreground or background stars.

Star Name X

Coordinate Y

Coordinate Magnitude

(B) Magnitude

(V) Color Index

(B-V) B-V r V r Spectral Class r

NGC 2345-1* 537.5 604.2 9.62 9.34 0.28 -0.34 7.44 O 2 476.4 569.3 12.25 11.86 0.39 -0.23 9.96 B 3 380.5 525.7 14.21 13.80 0.41 -0.21 11.90 B 4 341.4 441.6 12.67 12.17 0.50 -0.12 10.27 B 5 434.4 421.6 13.95 13.43 0.52 -0.10 11.53 B 6 509.4 450.6 13.92 13.38 0.54 -0.08 11.48 B 7 558.6 452.4 10.60 9.88 0.72 0.10 7.98 A 8 578.5 468.8 14.30 13.78 0.52 -0.10 11.88 B 9 636.5 463.2 14.27 13.81 0.46 -0.16 11.91 B

10 636.5 400.5 15.19 14.63 0.56 -0.06 12.73 B 11 614.4 374.6 14.44 13.85 0.59 -0.03 11.95 B 12 638.5 320.7 14.71 14.03 0.68 0.06 12.13 A 13 652.2 305.7 14.74 13.91 0.83 0.21 12.01 A 14 595.0 305.2 12.82 10.68 2.14 1.52 8.78 M 15 552.6 320.4 14.49 13.94 0.55 -0.08 12.04 B 16 537.3 356.4 14.47 14.01 0.46 -0.16 12.11 B 17 531.7 400.7 15.08 14.43 0.65 0.02 12.53 A 18 514.6 378.4 15.00 14.53 0.47 -0.15 12.63 B 19 562.5 259.8 13.58 12.85 0.73 0.11 10.95 A 20 562.0 240.0 14.54 13.87 0.67 0.05 11.97 A 21 598.9 241.6 14.59 13.89 0.70 0.08 11.99 A 22 665.6 272.5 12.75 12.06 0.69 0.07 10.16 A 23 648.5 175.0 14.19 13.41 0.78 0.16 11.51 A 24 574.0 163.0 14.96 13.98 0.98 0.36 12.08 F 25 532.5 213.0 14.16 13.49 0.67 0.05 11.59 A 26 491.0 219.1 13.90 13.23 0.67 0.05 11.33 A 27* 502.7 204.5 13.66 13.58 0.08 -0.54 11.68 O 28 493.0 198.2 12.94 12.39 0.55 -0.07 10.49 B 29 456.7 20.9 13.60 12.84 0.76 0.14 10.94 A 30 447.4 8.7 13.98 13.37 0.61 -0.01 11.47 B 31 427.0 164.0 14.93 14.11 0.82 0.20 12.21 A 32 448.0 184.2 14.43 13.74 0.69 0.07 11.84 A 33 438.3 235.9 14.49 13.82 0.67 0.05 11.92 A 34 413.5 229.5 11.44 9.89 1.55 0.93 7.99 K 35 397.4 225.1 11.40 10.91 0.49 -0.13 9.01 B 36* 193.8 208.7 10.33 10.16 0.17 -0.45 8.26 O 37 252.8 222.0 12.95 12.60 0.35 -0.27 10.70 B 38 276.5 250.5 13.72 13.05 0.67 0.05 11.15 A 39 226.8 273.0 13.21 12.83 0.38 -0.24 10.93 B 40 236.2 367.5 13.52 13.09 0.43 -0.19 11.19 B 41 264.0 323.9 13.49 13.06 0.43 -0.19 11.16 B 42 320.5 298.7 13.90 13.39 0.51 -0.11 11.49 B 43 344.3 287.2 12.51 10.63 1.88 1.26 8.73 K 44 358.8 339.0 13.70 13.27 0.43 -0.19 11.37 B


45 315.4 340.6 14.04 13.46 0.58 -0.04 11.56 B 46 310.0 358.8 13.76 13.17 0.59 -0.03 11.27 B 47 300.6 387.4 13.00 12.56 0.44 -0.18 10.66 B 48 320.8 388.0 13.16 12.68 0.48 -0.14 10.78 B 49 335.6 374.0 13.39 12.82 0.57 -0.05 10.92 B 50 336.3 372.9 12.44 10.40 2.04 1.42 8.50 M 51 338.0 373.0 12.82 12.37 0.45 -0.17 10.47 B 52 384.7 390.1 14.61 14.04 0.57 -0.05 12.14 B 53 355.0 438.0 14.17 13.62 0.55 -0.07 11.72 B 54 375.0 451.0 13.28 12.79 0.49 -0.13 10.89 B 55 379.9 479.7 14.59 14.01 0.58 -0.04 12.11 B 56 355.0 493.1 13.98 13.45 0.53 -0.09 11.55 B 57 405.0 518.0 14.62 14.03 0.59 -0.03 12.13 B 58 419.5 657.5 12.94 12.22 0.72 0.10 10.32 A 59 326.2 571.1 13.33 12.95 0.38 -0.24 11.05 B 60 285.8 461.0 12.37 10.45 1.92 1.30 8.55 K 61 291.2 420.0 13.74 13.32 0.42 -0.20 11.42 B 62 271.8 425.0 15.37 13.77 1.60 0.98 11.87 K


Galaxy Clusters: The Local Effects on Star Synthesis Zackery Schroeder

Grosse Pointe North High School, Grosse Pointe Woods, MI Teacher: Ardis Herrold, RBSE 1998, TLRBSE 2001

ABSTRACT Did galaxies in the early universe produce stars at the same rate as galaxies today? What environmental factors affect the star formation rate of galaxies? Data at 24-micron wavelengths for three intermediate-redshift galaxy clusters taken with the Spitzer Space Telescope was acquired to derive the star formation rates (SFR) of the cluster galaxies. Imaging data from the Hubble Space Telescope and the ESO Distant Cluster Survey (EDisCS) was used. The clusters differed in mass, degree of relaxation, and number of members. Physical properties such as morphology, VRIJK magnitudes, local density, and presence of merging galaxies were analyzed in order to correlate how factors in the internal and external environment of clusters affect SFR. The cluster galaxies were also compared to field galaxies. The SFRs of the clusters were compared to other redshift studies to provide a more complete picture of how the SFR of galaxies has changed over time. The SFRs of all galaxies correlate to their morphologies, color, and being merging galaxies. Also, galaxies within clusters tended to have lower SFRs overall compared to those in the field. Density within the cluster, dark matter, and ram-pressure stripping played a role in the difference between cluster and field galaxy SFR. When comparing the SFRs of clusters at 3 different redshifts, there is a large increase in SFR at intermediate redshifts. INTRODUCTION What different environmental or internal conditions in a galaxy cluster can affect the SFR, when compared against the SFRs of galaxies not in clusters? How has the average SFR of a cluster galaxy changed over the course of different redshifts? The Star Formations Rates of galaxy clusters have been extensively studied both in close and far redshifts, but those of intermediate redshifts have yet to be observed to the same extent. Three intermediate redshift galaxy clusters of differing masses and states of equilibrium were observed using the Spitzer Space Telescope as well as other data taken from the Hubble Space Telescope. By studying the 24-micron flux, the SFR was derived for each galaxy as well as for the entire cluster. The 24-micron flux will give a more accurate representation of the SFR because the light emitted by young, blue stars is absorbed by the dust and re-radiated at longer wavelengths (the infrared). In optical wavelengths, there is a large amount of light extinction from dust, thus making it difficult to get an accurate reading of SFR. OBSERVATIONS AND DATA REDUCTION 24-micron data was taken with the Spitzer Space Telescope’s “Multiband Imaging Photometer for Spitzer” (MIPS) for wide-field, broadband imaging, and the flux given from that was used for later deriving SFR.


Additionally, imaging data taken by the Hubble Space Telescope’s Wide Field Planetary Camera (WFPC2) was acquired in order to categorize the galaxies with proper types according to the Hubble Sequence and view the shape of the cluster. Photometric redshifts were also necessary to determine which galaxies were cluster members, and were determined by fitting template Spectral Energy Distributions (hereafter SEDs) to the observed wavelengths of BVIRJK. Access to the ESO Distant Cluster Survey (EdisCS) provided multi-band optical and infrared photometry and spectroscopy. The freeware program DS9 was used to observe the cluster visually for shape and density. During July of 2006, I traveled with my teacher to the Spitzer Space Telescope Center on the Caltech campus in Pasadena, California to work with three astronomers, five additional teachers, and three other students from across the nation for the Spitzer Space Telescope Research Program for Teachers and Students1. While there, I was given access to MIPS and HST data along with the photometric and spectroscopic membership flags (from which I can determine which galaxies are members of the cluster), model SED curves, HST types, 24-micron flux, velocity dispersions of the individual galaxies, which tell how fast each galaxy is moving within the cluster (the faster the galaxies are moving, the more massive the cluster must be in order to retain them, so therefore velocity dispersion is proportional to the mass of the cluster, as well as the virial radius), access to the EdisCS database (for photometric magnitudes and irregularity data), and the routine on how to derive SFR from 24-micron flux, as described below. Selection of Cluster Candidates Three clusters of similar intermediate range redshift (.5414 < z < .6355) yet inherently different physical properties were selected based upon their differences in mass, degree of relaxation, and virial radius (also R200), which is the radius at which the mean density of the cluster is 200 times the critical density of the universe, meaning that the galaxies will be able to resist universal expansion and separation from the cluster. Since R200 needs to be a certain radius to contain the galaxies, it is proportional to the mass of the cluster.

CL 1037.9-1243 This cluster is the least massive of the three clusters, with a virial radius of .57 mega parsecs (Mpc) and the lowest velocity dispersion (the faster the galaxies are moving, the more massive the cluster must be in order to retain them, so therefore velocity dispersion is proportional to the cube root of the mass of the cluster, as well as directly proportional to the virial radius2. There are 206 star forming galaxies in the 24 micron field, but only 45 of them are identified as members of the cluster, determined by their spectroscopic and photometric membership flags, making this cluster the second least populated. This cluster is also the least “relaxed” due to its cigar-like shape, meaning that its galaxies 1 Rudnick, et al. “Star Formation in High Redshift Clusters with Spitzer.” <http://coolcosmos.ipac.caltech.edu/cosmic_classroom/teacher_research/7-StarForm/index.shtml.> 2 Finn, Rose A., et al. “Hα -DERIVED STAR FORMATION RATES FOR THREE z~0.75 EDisCS

GALAXY CLUSTERS.” The Astrophysical Journal. 1 Sept 2005. Equations 8 and 10.


have not yet entered equilibrium. This also means that the velocity dispersions are not quite as accurate in determining the mass of the cluster, since the velocities could be those of when the galaxies entered the cluster, not yet retaining a uniform velocity. CL 1227.9-1138 This cluster is the middle of the road candidate, with the medium mass, velocity dispersion, and virial radius of 1.00 Mpc. The MIPS image has 145 galaxies present, but only 26 of those are cluster members, making it less populated with star forming galaxies than CL1037, and yet more massive and spread out. This cluster is also in the middle state of relaxation. Its galaxies are traveling less erratically than those of CL1037, but faster. Since the galaxies are moving considerably faster than those of CL1037, it is true that CL1227 is much more massive, since it needs to be so to retain its galaxies. However, it is not nearly as “relaxed” or massive as CL1232. CL 1232.5-1250 This cluster is by far the most massive of the three and has the highest velocity dispersion and virial radius (1.99 Mpc, twice that of CL1227). It is by far the most populated with 162 HST galaxies in the image, 54 of those being identified as cluster members. This cluster is also the most “relaxed”, with the most uniform shape, but its galaxies have the highest velocity dispersion, and therefore this cluster is the most massive of all three. Deriving SFR from Flux and Ltir The photometric and spectroscopic redshift membership flags were used to determine which galaxies were cluster members. This narrowed the number of star forming galaxies in CL1037 down from 206 to 45. The observed wavelength 24-micron flux is equivalent to the luminosity at 15 microns, since the redshift of the galaxy results in a shift from 15 microns to 24 microns. This was done because light emitted by newly formed stars is nearly all extincted by dust absorption, since most star formation occurs in areas of high concentrations of dust and cool gas. However, the light absorbed by the dust is re-emitted at infrared wavelengths and was therefore an accurate reading of the amount of young starlight, and thus SFR could be derived from it. The luminosity at 15 microns was converted to the total infrared luminosity (Ltir) by using the 5 conversion values derived from the template SED curves. Thus, there were effectively 5 different possible Ltir for each galaxy. The reason 5 values are present is due to the 5 different SEDs: these curves estimate the type of emission from the dust in a typical galaxy, and since it is impossible to determine which type of dust emission each galaxy has, the 5 values must be eventually averaged to get one value of SFR.


The final step was to convert the Ltir to the actual SFRs for each galaxy, which was done by substituting the values of Ltir into the following equation:

SFR (Msol yr−1) = 4.5 x 10−44 LTIR (ergs s−1)3 Wherein the Ltir was multiplied by 4.55*10-44 to obtain SFR in Msol/year. Now 5 different SFRs for each were obtained, which could be used to get the average SFR for each galaxy. The average SFRs for each galaxy were used to calculate the average SFR for a typical star-forming galaxy in a specific cluster. By averaging the five different SFRs, it was possible to get the average total SFR for the cluster. Identifying Mergers Cluster members that were cited in the EDisCS database as possible mergers were visually inspected in the HST data. If the galaxies appeared visually close to one another and appeared to exchange gas, then they were classified as mergers. Merging galaxy SFR was compared against normal star-forming galaxy SFR to observe any changes.

Mass-Normalization of SFRs In order to compare the total SFR in each intermediate redshift cluster to those at high or low redshifts it was necessary to normalize the SFR to account for the clusters having different masses. In order to do so, the total SFR had to be calculated within ½ the virial radius because the number of star forming galaxies is related to the distance from the cluster center4. Also, this was the calculation made in the other redshift studies as well, so it had to be done to compare them all. The sum SFR within ½ the virial radius was then used to calculate the Sum SFR per cluster mass per 1e14 solar masses, or ∑SFR/Mcl/1*1014 Msol. Since each cluster has a different mass, their Sum SFR had to be divided by it to normalize them for comparing. DATA The spreadsheet labeled “Basic Cluster Data” outlines for each cluster member the center RA, center DEC, redshift (z), virial radius (R200, in degrees), ½ R200 (in degrees), velocity dispersion (in km/s), number of galaxies in the cluster (Galaxies in cluster), the cluster mass (in solar masses per year), the sum SFR in 1/2 the virial radius, and the normalized SFR per cluster mass. The other spreadsheets given the progression of average SFR as a function of each environmental effect: morphology, radius, density, and color. Additionally, they display the average SFR in cluster galaxies and non-cluster galaxies.

3 Kennicutt, Robert C. Jr. “STAR FORMATION IN GALAXIES ALONG THE HUBBLE SEQUENCE” 4 Finn, Rose A., et al. “Hα -DERIVED STAR FORMATION RATES FOR THREE z~0.75 EDisCS GALAXY CLUSTERS”


Basic Cluster Data

Cluster Name Center RA (degrees)

Center DEC (degrees) z

R200 (degrees)

1/2 R200 (degrees)

Velocity Dispersion (km/s)

Galaxies In Cluster

Cluster Mass (Msol)

Sum SFR/1/2 VR SFR/Mcl

Cl 1232.5-1250 188.1271 -12.8433 0.5414 0.08682 0.04341 1080 54 4.86E+14 2588.13 182.66 Cl 1227.9-1138 186.9746 -11.6381 0.6355 0.04045 0.02023 572 26 1.82E+14 269.47 50.78 Cl 1037.9-1243 159.4633 -12.7242 0.5789 0.02407 0.01204 315 45 4.06E+13 214.45 181.17

Avg SFR for each HST evolution group Avg SFR for each radius group Cluster Ellipticals Early Type Late Type Irregulars 1 2 3 4 5 CL1037 36.46 70.11 40.72 42.28 49.84 60.25 69.30 70.51 Cl1227 36.58 65.2 47.02 28.72 59.71 74.88 79.93 76.31 39.97 CL1232 32.26 55.35 41 45.31 50.30 58.75 54.18 54.95 34.46

Avg SFR for each density group Avg SFR per Photometry Group Cluster 1 2 3 4 6 2 3 4 5 6 CL1037 64.64 53.64 35.83 39.59 80 54 59 56 53 Cl1227 86.32 49.29 55.58 87 38 49 65 CL1232 43.65 46.83 46.83 50.53 40.88 50 59 58 51 47

SFR in Cluster Field Galaxy SFR 42.89 53.27

CL1037 – On the left is the MIPS Image, on the right is the HST Image. Blue circles represent cluster members


ANALYSIS AND RESULTS Analysis of Morphology Using the HST and EDisCS data, each galaxy in the cluster was labeled according to their HST morphological code. Having a visual representation of where which galaxies are located within the cluster where makes it possible to look for groups of certain types in certain areas, such as those surrounding the central galaxy. These areas of certain morphologies were correlated against the SFR for each galaxy as well as the average for each group versus other groups. Results proved inconclusive. Additionally, the HST types were broken into 4 groups corresponding to their evolution along the Hubble Sequence. Values of –7 to –1 were ranked as elliptical type, 1 to 4 as early type spiral galaxies, 5 to 8 as late type spiral galaxies, and 11 as irregular galaxies. These new groups were also correlated against SFR individually as well as on an averaged basis versus each other to reveal how the certain groups of HST types behave.




There were, however, some cluster members of which the HST type could not be identified, and they were therefore excluded from this particular aspect as well as correlations to radius and SFR. In CL1037, when taking the average for each HST grouping, there at first appears to be no obvious trend when advancing on the Hubble Tuning Fork, but when removing two abnormal merging galaxies (see page 8), a trend is apparent. The elliptical galaxies have a rather low average SFR, as would be expected, followed by a spike in early type spiral galaxy SFR and a gradual downward trend from late type spirals to irregular galaxies. Unlike the previous cluster, in CL1227 the trend of increasing then decreasing SFR in the different types of morphologies peaks in the late type spirals as opposed to the early type. However, when removing 3 somewhat outstanding outliers with high SFRs, the previously mentioned trend is reproduced in the data. In CL1232, a similar trend also exists as in CL1037, until reaching the irregular type galaxies. This could possibly be explained by an outlier and small sampling size for the group. Analysis of Relative Density and Radius from Center It is impossible to perceive the depth of the galaxies within the clusters due to viewing a 3-dimensional volume on a 2-dimensional area, so galaxies that appear close to one another on one plane may in reality be rather far apart. However, it is safe to assume that in simulating the cluster in 3 dimensions that the galaxies’ apparent distance from the center of the cluster can place the galaxies in sets of spheres, and from there a density can be determined among different ‘annuli’ of volume containing a certain number of galaxies.

SFR per Morphology Mergers Removed

0

10

20

30

40

50

60

70

80

Elliptical Early Spiral Late Spiral Irregular

SFR

(Mso

l/yea

r)

CL1037

CL1227

CL1232


The progression of SFR in cluster galaxies as a function of radius. Annulus 1 is the center of the cluster, while annulus 5 is the edge of the cluster.

Looking at the MIPS 24-micron data and the HST images, the galaxies in the clusters were categorized into different ‘annuli’ of volume in the cluster sphere by their distance from the cluster center in mega parsecs, a new annulus every .3 Mpc (to have an even spread of galaxies in each shell for comparison). The SFRs of the galaxies within each annulus were averaged and compared to the other annuli to see how SFR changes from the cluster center to the edge. Pertaining to density, an annulus of .4 Mpc was placed around each cluster member, and the number of galaxies within each annulus was counted. A larger number of members in each annulus means that the area is denser, and this was observed to see how a more or less dense area stimulates or stunts star formation.

In CL1227, an increasing trend in average SFR occurs from annuli 1 to 4 (with a dramatic peak in annulus 4), followed by a tremendous decrease in SFR in annulus 5, the edge of the cluster. Also, annulus 4 was the most densely populated, which may have contributed to the spike in SFR. Relating to density, there is a decrease in SFR as density increases, but the correlation is not strong due to only density sets being compared, one of them only containing one galaxy.

In CL1232, the average SFR amongst the different annuli remains rather stagnant, no matter how densely populated the galaxies are together. This could be attributed to CL1232’s being the most ‘relaxed’ cluster of the three, having a uniform velocity dispersion meaning that it’s galaxies have reached an equilibrium within the cluster.

SFR within Radius Annuli

0

10

20

30

40

50

60

70

80

90

1 2 3 4 5

Annulus

SFR

(Mso

l/yea

r)

CL1037CL1227CL1232


Therefore, the absence of changing the average SFR over distance and density is plausible. Relating to density, there was no apparent correlation between SFR and density, but rather it fluctuated irregularly, unlike the other two clusters. In CL1037, the SFR gradually increases as one goes farther from the core of the cluster, as does the population density until annulus 4, wherein it drops. However, the average SFR does not drop although it is less populated, and the increasing trend is only seen when 2 merging galaxies in annulus 2 are removed form the averages, because their average SFRs are nearly 10 times more than the ‘normal’ galaxies. Despite these discrepancies, this trend seems uniform within CL1037 and CL1227. Relating to density, there was a definite decrease in SFR with an increase in density. Analysis of Mergers Using the EDisCS survey, data on which of the galaxies in the cluster presented irregularities was analyzed along with the HST imaging data to determine which galaxies could possibly be merger galaxies. Once identified, the SFRs of the merger galaxies were averaged to compare to the average SFR of the cluster. Also, attention was paid to the correlation between merger activity in certain radii or density shells. Additionally, the merger SFRs were compared to other non-merger galaxies of the same HST morphological type. In all three clusters, merging galaxies had no apparent difference in SFR form normal, non-merging galaxies. However, there were a few cases of high SFR merging galaxies having SFRs of over 150 solar masses per year. These were inferred to be starburst galaxies. When two galaxies collide, sometimes their gas collapses into a small volume and begins to produce stars very quickly, and is dubbed a starburst galaxy. These galaxies skewed the data in the density and morphology studies, and were therefore removed. Analysis of Color The color magnitudes given from the EdiSCs database range from the green optical wavelengths to that of IJK infrared photometry. By subtracting the K magnitude from the V magnitude, the resulting numbers showed how ‘red’ the galaxy is, with a high value meaning very red, and a low value meaning less red. However, since there is dust absorption in some wavelengths, and not in others (namely infrared), the blackbody curve will be emitted or absorbed at certain wavelengths, throwing off the data. To make an accurate judge of how the galaxy actually differs in which wavelengths it emits in, one cannot take an X-K (X being another wavelength) value that is too far away from the K value, because then the dust absorption will throw off the data. However, one also cannot pick a wavelength too close the K wavelength, because it wouldn’t really show anything useful. So, the galaxies were binned into ranks according to their R-K and I-K values (because although R-K might take into account light extinction, I-K is also rather close to the K magnitude range, so both indices were rather intermediate), separated by differences proportionally to their range of values. Next, the average SFRs of each bin were taken and observed against on another to see if redder galaxies had any definite changes in SFR.


In CL1037, aside from the galaxies with R and I magnitudes very close to their K magnitudes, there is a definite trend for increasing SFR when going from redder to less red galaxies. In CL1227, as in CL1037, there is, although less noticeable, a similar trend for less red galaxies to have higher SFRs. In CL1232, there is again a trend wherein most of the bluer galaxies have higher SFRs. Analysis of Field Galaxies All non-cluster galaxies in the fields of all three clusters were also examined for the effects of their morphologies, density, and color versus their SFRs, and were cross-correlated to the results of the cluster galaxies to give insight into how galaxies in and out of cluster differ as a function of their environment. When comparing the average SFR of a galaxy in a cluster versus the average SFR of a galaxy not in a cluster, it was found that the latter had higher SFRs overall (50% more star formation was noticed outside the clusters in all three cluster fields). Comparison of Mid-z SFR to High and Low-z SFR After the mass normalizations of the average SFRs of each cluster was obtained, they were correlated against the mass-normalized SFRs of low and high redshift cluster, as presented in Finn et al 2005. Trends in average SFR of cluster were identified as the age of the universe increased.

Progression of SFR (Msol/yr/1e14 Msol) with z Redshift Low z Mid z High z

6.97 182.66 52.18 0.80 50.78 12.97 1.78 181.17 19.33

Sum SFR/Mcl/1e14 Msol in Clusters

72.01 71.60 Average 20.39 138.21 39.02

Conclusions: Changes in SFR due to Morphology I found that in both cluster and non-cluster members that the average SFR of certain HST types follows the trend of having elliptical type galaxies with mid-range SFRs, followed by a sudden spike in the SFR of early type spiral galaxies, ending with a gradual downward trend in SFR from late type spirals to irregular galaxies. The fall of SFR from early type spirals to irregular galaxies can possibly be explained due to the shape of the galaxies themselves. In early type spirals, the arms are numerous and tightly wound, as is the gas and dust, and stars form readily. However, later type spirals have fewer arms and are more tapered, which in some cases should have more

As shown in the data table, there is a large spike in SFR at the intermediate redshift range


young stars5. This can be explained by active galactic nuclei, strong sources of radiation in the centers of some galaxies. These emit over all different wavelengths, including the infrared, so the observed 24 micron flux yields a higher value of SFR than is actually present. AGN are more prominent in early type spiral galaxies due to their large central bulges. The large difference in the SFRs of elliptical galaxies and early type spiral galaxies makes sense due to elliptical galaxies containing low amounts of interstellar matter, which includes gas and dust to condense and form stars. Conclusions: Relating Color to SFR I found that overall bluer galaxies tended to have higher SFRs. However, since galaxies with high SFR have lots of dust, and dust makes the galaxy redder, this result seems to cast some doubt. It could be that the reddest galaxies are elliptical, and thus have redder stars (not much dust) and little cool gas to form stars. Even so, the correlation was not as strong as expected due to interpretation constraints: there was a small sampling size of galaxies in certain index groups, and I could not viably quantify the errors in the R and I bands. Conclusions: Changes in SFR due to Density and Radius form Cluster Center I found that SFR in a cluster increases in the intermediate radii of the cluster, and decreases at the edges and center of the cluster. This could be due to multiple factors. At the center of a cluster, there is a lower concentration of cool gas available to form stars, while at the edge of the cluster SFR is lower because of ram-pressure stripping, which occurs when an object (galaxy) enters a more fluid environment (cool gas in a cluster) and said environment strips away gas from the object, producing a lack of star formation. Difference in SFR of Cluster and Non-cluster Galaxies It was found that galaxies in clusters had a lower star formation rate overall than those who were not in clusters. Although the densest part of the clusters form more stars, they also exhaust their cool gas reservoirs faster, and the galaxies outside the cluster have a higher, constant SFR over a long period of time. DISCUSSION Environmental Factors that Relate to SFR Internally, meaning what characteristics of the galaxy itself can relate to SFR, morphology and color are the largest signifiers, where being bluer and a spiral type galaxy indicates a higher presence of star formation. Additionally, merging with another galaxy may in some cases produce violent starburst. Externally, meaning what characteristics of the cluster can affect galaxy SFR, density within the cluster and radius from the cluster center are most likely the cause of fluctuations in SFR, due to galaxy harassment and ram-pressure stripping.

5 Sparke, Linda S. and John S. Gallagher. Galaxies in the Universe. Cambridge, New York: Cambridge

University Press, 2000.


Evolution of SFR in Galaxy Clusters In previous studies there was a decrease in SFR from high redshift to low redshift clusters, and model of an aging universe was proposed wherein as the universe aged, less cool gas was available to form stars, and thus SFR decreases. However, in my study of intermediate redshift clusters, I found that the average SFR is considerably higher than both high and low redshift clusters. I propose I new model of the evolution of star formation in the universe: at the beginning of the universe, galaxies and galaxy clusters were very disorganized and did not create new stars very quickly. At the middle-aged universe (my study), galaxies and clusters were more organized and began to create stars much more quickly. Finally, at the present age of the universe, the cool gas in the galaxies was exhausted from the high star formation during the middle ages of the universe, and thus lower SFRs are found. There was a peculiar finding with CL1227 though: it’s Sum SFR/Mcl was considerably lower than that of the other two mid z clusters; This might be due to two things: while the other two clusters have very similar redshifts, CL1227’s redshift is the highest of them all, so perhaps there is a time in the universe where galaxies have violent bursts in star formation, or perhaps since CL1227 had no merging galaxies, their were no starburst galaxies present to add to the Sum SFR, as in the other clusters. The reason mergers may not be present is because although CL1227 is not as massive as CL1232, it is very massive yet has the least amount of star forming members, so it is less likely that merging and starburst would occur. SUMMARY AND ACKNOWLEDGEMENTS In order to examine how the star formation rates of galaxy cluster change as a function of their environment as well as a function of time, 24-micron data from the Spitzer Space Telescope as well as I-Band Hubble Space Telescope Data of three intermediate redshift galaxy clusters of differing mass, number of members, and degree of relaxation were observed in a number of environmental areas such as morphology of the galaxies in the cluster, color magnitudes, cluster density, and radius from the cluster center. This data was also correlated against galaxies not in clusters to examine how a cluster environment affects SFR, and against average SFRs of high and low redshift galaxy clusters from other studies to see how the SFRs have changed over the evolution of the universe. I found that SFRs of galaxies in and out of clusters were affected by their morphologies, with a small amount of SFR in elliptical type galaxies, then a large spike in SFR of early type spiral galaxies, and a gradual decrease in SFR as the morphologies advanced along the Hubble Sequence. Also, high-density galaxies in the cluster have high SFRs, due to plentiful gas and lack of dark matter ram-pressure stripping. However, field galaxies still had the highest SFRs of all due to their isolation to have a high, constant SFR. Finally, when correlating the SFRs of mid-range clusters to those of high and low redshift clusters, it was found that there is a large jump in SFR of mid redshift clusters.


Special thanks to my instructor, Ardis Herrold, Julie Krugler, Dr. Gregory Rudnick, Dr. Vandana Desai, Dr. Rose Finn, John Blackwell, and all else who participated in/ran the Spitzer Teachers Program. REFERENCES Finn, Rose A., et al. “Hα -Derived Star Formation Rates For Three Z~0.75 Ediscs Galaxy Clusters.” The Astrophysical Journal. 1 Sept 2005. Equations 8 and 10. Kennicutt, Robert C. Jr. “Star Formation In Galaxies Along The Hubble Sequence” Rudnick, et al. “Star Formation in High Redshift Clusters with Spitzer.” <http://coolcosmos.ipac.caltech.edu/cosmic_classroom/teacher_research/7-StarForm/index.shtml.> Sparke, Linda S. and John S. Gallagher. Galaxies in the Universe. Cambridge, New York: Cambridge University Press, 2000 BIBLIOGRAPHY Aladin Sky Atlas. Centre de Données Astronomiques de Strasbourg. <http:// aladin.u-strasbg.fr/aladin.gml>.

Brand, K., et al. “The AGN Contribution to the Mid-IR Emission of Luminous Infrared Galaxies” Astro-ph. 08 Feb 2006.

Dale, D. A. and George Helou. “The Infrared Spectral Energy Distribution of Normal Star-Forming Galaxies: Calibration at Far-Infrared and Submillimeter Wavelengths.” The Astrophysical Journal. 8 May 2002.

Dale, D. A., et al. “Infrared Spectral Energy Distributions of Nearby Galaxies.” The Astrophysical Journal. 26 Jul 2005.

Dale, D. A., et al. “The Infrared Spectral Energy Distributions of Normal Star-Forming Galaxies.” The Astrophysical Journal. 1 Nov 2000.

ESO Distant Clusters Survey. Max-Planck-Institut fur Astrophysik. < http://www.mpa-garching.mpg.de/galform/ediscs/index.shtml.>

Finn, Rose A., et al. “Hα -DERIVED STAR FORMATION RATES FOR THREE z~0.75 EDisCS GALAXY CLUSTERS.” The Astrophysical Journal. 1 Sept 2005.

Halliday, C., et al. “Spectroscopy of clusters in the ESO Distant Cluster Survey.” Astronomy and Astrophysics. 28 Jul 2004.

Kennicutt, Robert C. Jr. “Star Formation of Galaxies Along the Hubble Sequence.” Astro-ph. 17 Jul 1998.

Poggianti, Bianca M. “Color, spectral and morphological transformations of galaxies in clusters.” The Astrophysical Journal. 10 Oct 2005.

Poggianti, Bianca M. “Emission Line Galaxies in Clusters.” Proceedings of Science. 7 Jan 2005.


Poggianti, Bianca M. “Evolution of galaxies in clusters.” Proceedings of Science. 9 Oct 2004.

Poggianti, Bianca M. “Evolution of Galaxies in Clusters.” The Astrophysical Journal. 9 Oct 2004.

Poggianti, Bianca M. “The Evolution of the Star Formation Activity in Galaxies and Its Dependence on Environment.” The Astrophysical Journal. 16 Dec 2005.

Rudnick, Gregory, et al. “A K-Band Selected Photometric Redshift Catalog in the Hubble Deep Field South: Sampling the Rest Frame V Band to z = 3.” The Astronomical Journal. 25 Jul 2001.

Rudnick, Gregory, et al. “Star Formation in High Redshift Clusters with Spitzer.” <http://coolcosmos.ipac.caltech.edu/cosmic_classroom/teacher_research/7-StarForm/index.shtml.>

SIMBAD Astronomical Database. Centre de Données Astronomiques de Strasbourg. <http://simbad.u-strasbg.fr/Simbad>.

Sparke, Linda S. and John S. Gallagher. Galaxies in the Universe. Cambridge, New York: Cambridge University Press, 2000.

Thompson, Andrea. “Galactic Birth Control: Unknown Factor Prevents Star Formation.” <www.Space.com>. 02 October 2006.

VizieR. Centre de Données Astronomiques de Strasbourg. <http://vizier.u-strasbg.fr/viz-bin/VizieR>.

White, D. M. “The ESO distant cluster survey.” Astronomy and Astrophysics. 28 Jul 2005.


Star Formation Rate in Three High-Redshift Galaxy Clusters: A Contribution to the Study of Galactic Evolution

Vinay Patel and Matt Pellegrino Saint Joseph’s High School, South Bend, IN

Teacher: Dr. Thomas Loughran, TLRBSE 2004 ABSTRACT In our universe, where there are ten to the twentieth stars and an average of a million solar systems being formed every hour, the human mind has pondered the uncertain future of our Milky Way. How will our galaxy evolve in response to the surrounding galaxies of the Local Group? What is the ultimate fate of our own galaxy? If we knew the evolutionary trajectory for galaxies such as our own, we might know something of its future. This evolutionary trajectory might be determined from the astronomical analogue of the fossil record, which is the collection of galaxies whose light is only now reaching us after billions of years. Because the universe is expanding, these older galaxies are redshifted. By analyzing data taken from the Spitzer Space Telescope, star formation rates (SFRs)—a key indicator that varies greatly over the lifecycles of galaxies—have been calculated for each of the 127 galaxies of three high-redshift clusters (cl1037, cl1227, cl1232, z = 0.54 to 0.64). (These three galaxy clusters are part of a much larger EDisCS catalogue, which consists of thirty such clusters.) SFR estimates are arrived at by averaging across five theoretical models for determining total infrared luminosity from 15 micron luminosity, which is in turn derived from Spitzer’s 24 micron flux measurements. These calculated SFRs have been compared with other data for the cluster members, such as distance from the center of the cluster and Hubble type. Furthermore, these SFRs have also been compared to the morphological distribution of other clusters of similar redshift, and there is substantial agreement. By characterizing these high-redshift clusters in these ways, we contribute to an understanding of the evolution of SFR and thus improve our understanding of galactic evolution and ultimately of our history and future. INTRODUCTION The study of galaxy clusters has greatly increased recently, with a focus on the evolution of the cluster itself and its galactic members. Observation has largely dominated this study of evolution, and within the last forty years some of the first evolutionary trends have been noticed. The first trend noticed was the Butcher-Oemler effect, which described the tendency of galaxy colors to be bluer at higher redshifts. As a result, a number of theoretical frameworks have been developed in an effort to explain these and other observational results (Rudnick et al. 2003). The most popular theory among astronomers is composed of several different physical mechanisms that cause cluster galaxy evolution by acting as environmental forces. The first of these mechanisms are mergers and strong galaxy interactions. These directly observable types of mechanisms are “most efficient when the relative velocities between the galaxies are low” and affect the galaxy structure and morphology (Poggianti 2004). On the other hand, harassment, a type of tidal force, describes the cumulative gravitational effect from many cluster


galaxies on a galaxy (Poggianti 2004). Harassment is especially influential in the lower end of the galaxy mass spectrum and is not directly observable and thus has a more theoretical backing. Another influence on galaxies is gas stripping—the interaction between galaxies and the intergalactic medium (IGM). Using viscous stripping, thermal evaporation, and ram pressure stripping (which is most efficient when IGM gas density is high and the relative velocity between the galaxy and IGM is high), this mechanism removes the interstellar medium (Poggianti 2004). Ram pressure stripping is considered to be highly influential in the galaxy cluster cores and does have empirical evidence. Both ram pressure stripping and another phenomenon known as strangulation indirectly affect the appearance of galaxy morphology, because they inhibit star formation, which in turns alters the observed color and morphology of the galaxy. Strangulation is defined as the removal of the galactic envelope of hot gas that is responsible for cooling and fueling star formation (Poggianti 2004). Strangulation, along with tidal force stripping, has a much more theoretical background, and neither has been directly observed. Theoretical frameworks have also made large advancements lately because of the ability to create models and run simulations on computers (Rudnick et al. 2003). Direct observation helps to constrain theory and provide more accurate predictions and explanations. Current results suggest that although no single physical mechanism dominates galactic evolution, evolution is determined largely on a case-by-case basis. Because of this, galaxy clusters are seen as extremely powerful models for studying galactic evolution as a function of environment and redshift, since each cluster provides a large number of galaxies at nearly the same redshift and in a cluster many galaxies are exposed to relatively similar environment conditions (Rudnick et al. 2003). Plus, galaxy clusters can be studied in detail from a redshift of z = 0 to z ≈ 0.8, where evolutionary effects are the greatest. However, constraint on theoretical models for clusters at z > 0.3 has been dominated by x-ray detected samples, creating a bias towards the most massive and dense systems (Rudnick et al. 2003). Because of this bias, it is difficult to make accurate and informative comparisons between samples of clusters at lower redshifts and samples at higher redshifts. In addition, there have been few detailed observations of clusters at redshifts greater than 0.5, the site of dramatic evolutionary changes (Rudnick et al. 2003). The EDisCS Sample The ESO Distant Cluster Survey (EDisCS) is a European Southern Observatory Large Program aimed at studying galaxy and galaxy cluster evolution. It is a survey composed of 30 galaxy clusters, with 15 clusters at a redshift of z ≈ 0.5 and 15 clusters at z ≈ 0.8. The survey combines quite a lot of different data from the Very Large Telescope (VLT), New Technology Telescope (NTT), and Focal Reducer Spectrograph (FORS2) on the VLT for spectroscopy (Halliday et al. 2004). The goals of this program are first to complete ultraviolet to near infrared Spectral Energy Distributions (SED), study galaxy morphology as a function of SED, study cluster luminosity functions (LF) as a function of redshift and cluster properties, and estimate cluster masses through gravitational


lensing. The second set of goals is to study the masses of the individual cluster galaxies, star formation rates, and star formation histories, and to compare cluster galaxies with galaxies in the field (Rudnick et al. 2003).

The EDisCS catalog can provide the necessary information for this, and this is the reason it is being used in this research. We are interested in studying the evolution of galaxies, and it will begin with analysis of three high-redshift galaxy clusters—CL1037, CL1227, and CL1232. OUR PROJECT This project began with a focus on studying the evolution of star formation in galaxy cluster environments by looking back to the earliest time in which large samples of clusters exist, at z ≈ 0.8. This was proposed because this has not yet been accomplished in any great degree of detail or accuracy (Rudnick et al. 2003). Previously, Hα emission lines were used to derive star formation rates, and early results showed that star formation depended both on cluster mass and redshift. Because of this, it was obvious that a broad range of cluster mass had to be studied. Our research of star formation within clusters of redshift 0.54, 0.58, and 0.63 will help to fill in the missing gaps of a study of star formation activity in galaxy clusters from z = 0 to z = 1. The original justification for studying star formation rates came from a difference in observations of galaxies in dense environments that have low star formation and galaxies in less dense environments that have significant star formation. The question then became, “Why? What drives environmental variations in galaxy star formation properties?” Theoretical models have made predictions, and the total star formation rates per cluster mass have helped to constrain this theory, but significant errors still exist because of uncertainty in the mass of the cluster and even more so from the effect of dust on H-alpha derived star formation rates. The data we are using comes mostly from the Spitzer Space Telescope, an infrared telescope in orbit around the sun. Even though Hα has been used for measuring SFR in lower redshifts, it is limited severely by extinction. Infrared, however, can detect around 10 to 100 times more star formation than Hα can, which led Spitzer to be chosen specifically for this measurement of flux collection because of its infrared capability (Finn et al. 2006). Due to the huge amount of dust and gas which obscures our view of the cosmos, optical telescopes such as the Hubble Space Telescope are unable to view visible light in large sections of the universe. This dust and gas is even more of a problem when studying star formation rates: young stars are surrounded by the excess hydrogen gas that created them, so they are even harder to see in visible wavelengths. The light, however, is absorbed by the dust and gas surrounding the stars, and then re-emitted as heat. Spitzer allows us to see this re-emitted infrared light, and then infer the presence of stars and galaxies which surround and heat the dust. Thanks to the Spitzer and its infrared technology, astronomers are able to probe optically-blocked galaxies that were previously invisible and then measure their flux. This is our rationale for using Spitzer data, rather than ground-based telescopes or the Hubble Space Telescope.


PRIOR WORK Our research on this SFR project consisted of three basic pieces—the five month period of preparation in which we became familiar with basic astronomy, terminology, and this project; the three day meeting at the Spitzer Space Center (SSC) in Pasadena, California, in which we completed our initial analysis; and the eight month period after the meeting in which we completed more analysis, as well as learned, in depth, the meaning of the analysis we began during the meeting. In this section, we will detail the prior work and contributions of others in our project. This description will necessarily parallel our methodology section below, but this section focuses on who did what work, not the work itself. Beginning in January 2006, as part of the Spitzer Teacher Research Program for Teachers and Students, six Spitzer research teachers (John Blackwell, Velvet Dowdy, Rosa Hemphill, Ardis Herrold, Thomas Loughran, and Dwight Taylor), with the help of three professional astronomers, prepared the Spitzer observing proposal. This proposal asked for 2.5 hours of observing time to examine three galaxy clusters, which are part of the EDisCS catalog. In February 2006, we joined the team and began reading preparatory studies, allowing us to participate in this project. This led to our visit to the Spitzer Space Center (SSC). From July 26, 2006, to July 28, 2006, a meeting of the Spitzer research teachers and the three supporting astronomers, Gregory Rudnick (NOAO), Rose Finn (Siena College), and Vandana Desai (California Institute of Technology), took place. Here, we split up into three teams, one which consisted of ourselves and Dr. Loughran; one which consisted of John Blackwell, Dwight Taylor, and Ardis Herrold and her student, Zak Schroder; and one which consisted of Velvet Dowdy and Rosa Hemphill and her student, Emily Petroff. With the help of these astronomers, these groups performed the initial analysis of the data from one cluster (CL1037) and learned how to compute SFR. With Gregory Rudnick’s instruction, one of the first milestones we completed was learning to navigate the astronomical imaging and data visualization tool, DS9. Vandana Desai provided the Hubble images of CL1037, and using DS9, we compared Hubble and Spitzer images. (Prior to the SSC meeting, we had already obtained the Spitzer image of CL1037 using the Spitzer Pride software, Leopard.) Rose Finn also provided several domain files for the cluster members, so we could (using DS9) locate CL1037 and its cluster members. DS9 also allowed us to visually display cluster members as colored circles, and using the blink tool (which “blinks” from one image to another), the comparison of the Hubble and Spitzer images reinforced our views of the difference in resolution between Hubble and Spitzer images as well as the difference in what is visible and not visible in different wavelengths of light. In addition to a visual analysis, we also performed numerical analysis using data files given to us by Greg Rudnick. These files contained necessary information about every one of the 176 galaxies in the Hubble image, such as the EDisCS catalog number, right ascension, declination, and flux. Rose Finn added to this data by contributing galaxy morphologies, which was entered into the spreadsheet and visually verified. Furthermore,


Gregory Rudnick’s determinations of the redshifts of each galaxy as well as the cluster itself, through spectroscopic and photometric redshifts, enabled us to determine which of these galaxies belonged to CL1037. From this basic data, we calculated the luminosity of each member galaxy at the rest-frame wavelength. With the luminosity in hand, we were able to generate spectral energy distributions (SEDs), thanks to a program written in the Python language. This program applied a combination of detected emissions from both obscured and unobscured stars to produce SEDs for each of the cluster members. Gregory Rudnick then provided us with c values for each of the five models (which were calculated from the area under each SED and served as a ratio between the total infrared luminosity and the luminosity at one specific wavelength), letting us convert from the luminosity at the rest-frame wavelength (15 microns) to the total infrared luminosity according to the equation

LTIR = c x L15. After this, Gregory Rudnick introduced us to the equation to convert from the total infrared luminosity to SFR, allowing our group to be the first to arrive at a SFR for each galaxy (i.e. each cluster member) as well as the SFR for CL1037 as a whole. The only problem with this, however, is that there are five different SED models (and therefore also five different c values), which remain to be narrowed down via Infrared Array Camera (IRAC) data analysis. This led to five different SFR values per galaxy, which were simply averaged as a first-order approximation. Additional IRAC data at five additional wavelengths has become available and will be applied at some future date to help select among these five models. Figure 1 below is a sample SED depicting each of the five models for one cluster member of CL1037.

Figure 3 A sample SED for one member of CL1037. SEDs are basically separated into two parts—the left blue section, representing optical emissions from stars not obscured by dust, and the multicolored right section, representing infrared emissions from stars that are obscured by dust. The six points on the left from the dotted blue line are part of the stellar model. The seventh point, at a rest-frame wavelength of 15 microns, is the only one we have in the infrared range, which leaves us with some uncertainty of the exact model. Instead of one exact model, we have five approximate ones, each of which is represented in the SED above.


Since our initial calculation of SFR for each cluster member and for the entire CL1037 cluster, we have also calculated the SFR values for clusters CL1227 and CL1232, thanks to the arrival of additional data in November. Along with this, we have attempted to find a correlation between SFR and distance from the center of the cluster, SFR and morphology (i.e. galaxy type), and morphology and distance from the center of the cluster for each of our three clusters, and of course, SFR and redshift. (See Figures 3 and 4 for these graphs.) In addition to this analysis, we have also attempted to understand every bit of what went on during the SSC meeting—particularly the methodology. METHODOLOGY The procedure to calculate the star formation rate of CL1037, CL1227, and CL1232 includes the identification of cluster members; the conversion of 24 micron flux (measured using the MIPS instrument on the Spitzer Space Telescope) of those cluster members to luminosity at a specific, rest-frame wavelength; model-guided extrapolation from luminosity at the rest-frame wavelength to total infrared luminosity (LTIR); and the conversion of the total infrared luminosity to SFR. The first step to calculating SFR after attaining the data is to identify the cluster membership: we need to know which of the galaxies in the observed field belong in the cluster so we can accurately measure the SFR of the entire cluster. This is done by calculating the redshift of each galaxy and then comparing this to the redshift of the cluster. The redshift of each galaxy was calculated in two ways: spectroscopically and photometrically. The spectroscopic redshifts were calculated using the spectra of the cluster members. The shift of each galaxy’s emission lines is directly related to the redshift of the galaxy, and by calculating this shift in emission lines, we can accurately estimate the redshift of a galaxy. This is the most precise method of determining the redshift, but it did not help much, in our case. Many of the individual membership galaxies are relatively faint objects, and because of this, obtaining spectral data requires much telescope time. As a result, very few of the galaxies being examined had spectra to analyze. For the remaining galaxies, photometric redshifts were used. Using two photometric codes, the publicly available HyperZ package and Gregory Rudnick’s self-created code, the redshift was approximated for each galaxy. The goal of both of these codes is to estimate the galaxy’s redshift based on its observed fluxes by using a standard SED fitting procedure. These codes contain a set of galaxy SEDs, or templates, for each type of galaxy (spirals, ellipticals, starbursts, etc.), that are shifted to all redshifts from 0 to 2. Then, all of these redshifted templates are compared to the SEDs of the galaxies, and depending on which template at which certain wavelength most closely matches the galaxy SED, the most likely redshift is indicated. Photometric redshifts are not as accurate as the spectroscopic ones; however, they are much easier to obtain. After finding all of the redshifts for each galaxy, we determined cluster membership by seeing whether the redshifts of the galaxies were reasonably close to the redshift of the cluster. Once cluster membership is determined, the flux (the number of particles that flows through a unit area per unit time) detected by Spitzer must be converted to luminosity. In addition, since luminosity is a measure of the amount of energy per unit time that an object radiates at the source, whereas flux is energy detected by us, luminosity needs to


be given in rest-frame wavelength and not in the observed wavelength. Thus the luminosity at 24 micron wavelengths is converted to luminosity at the rest-frame wavelength, according to the equation

λrest = λobserved/(1 + z).

Using this formula, where CL1037’s redshift is 0.58, the rest-frame wavelength was calculated to be 15.3 microns. This information, however, is not enough to calculate SFR. The total energy emitted at wavelengths from 8 to 1000 microns, which is the total infrared luminosity (LTIR), of each galaxy is required. Using a set of five model spectral energy distributions (SEDs), we can estimate this total energy output. These models have been computed using properties of the dust in the galaxies, such as how the dust emission depends on the energy that the dust absorbs and the composition of the dust itself (i.e. density of the dust or size of grains of dust). Since galaxies are different, there are a range of models to span the entire range of possible infrared properties in the galaxies. (When compared to local galaxies, these infrared models are found to accurately describe infrared emission.) These SEDs give us the luminosity at every wavelength from 8 microns to 1000 microns, so if the luminosity at a certain wavelength is known, the entire energy output can be determined. A c value was then calculated which would convert from luminosity at a certain wavelength (in our first case, 15.3 microns) to the LTIR. The c value was calculated by integrating the SED model as a function of wavelength to find the total area under the SED curve, and then dividing this value by the luminosity at the rest-frame wavelength. By multiplying the observed luminosity at the rest-frame wavelength by the c value, the total infrared luminosity was calculated. The last step to calculating SFR is converting LTIR to SFR. Assuming most of the light emitted by the newly-formed stars is absorbed by dust, the amount of light that is re-emitted in the infrared is proportional to the total number of young stars and also the SFR. Kennicutt, by observing star formation rates of nearby galaxies in the Hubble sequence, derived a set of self-consistent SFR vs. LTIR conversions. The one used in our study is presented below:

SFR (M yr-1) = 4.5 x 10-44 x LTIR (ergs s-1). This equation, with others in Kennicutt’s paper, has been used many times as aids to workers in the field of astronomy. Using this equation, we were able to convert from LTIR to SFR, and thus calculate SFR values for each of the three clusters being examined.

Following these calculations, we examined SFR as a function of morphology. Before doing so, however, it was necessary to verify that our three clusters had typical morphologies (see Figure 2). Seeing regular morphology distributions, except for low elliptical concentrations, we proceeded to fairly compare SFR vs. morphology in our three clusters. We also examined the SFR-distance relation for each cluster (see Figure 3).


RESULTS6 Using the methodology described in the previous section, we calculated the SFR for each of the cluster members in CL1037, CL1227, and CL1232. The average SFR of the five models of SFR for each of the cluster members of these clusters (in no particular order) is located at the end of the paper in Table 1, pending the application of IRAC data at V, R, I, J, and K bands to adjudicate between the five models. In addition, Figure 3 compares graphically SFR to morphology type for each cluster. For CL1037, the average SFR per cluster member was 86.0574 M /yr. These forty five cluster members contributed to a total SFR of 3872.59 M /yr for the entire cluster. The five galaxies within half of the virial radius of the cluster totaled about 6% of the total SFR, 214.450 M /yr. We also categorized SFR according to the Hubble type for each cluster member. The elliptical galaxies detected have an average SFR of 28.4975 M /yr. The cluster’s two lenticular galaxies, one of which has a particularly high SFR, averaged 312.061 M /yr, and the SFR of spiral galaxies average 81.2230 M /yr. The irregular cluster members of this cluster have an average SFR of 42.2799 M /yr. The average SFR per cluster member of CL1227 was 74.9301 M /yr, slightly lower than that of CL1037’s. Each of these twenty six members contributed to a much lower total SFR of 1948.18 M /yr for the entire cluster. The four clusters within half of the virial radius of the cluster totaled 269.473 M /yr, about 14% of the total SFR. In this cluster, the elliptical galaxies have an average SFR of 34.1923 M /yr, lenticular galaxies have an average SFR of 30.1390 M /yr, spiral galaxies average 90.4489 M /yr, and the irregular cluster members of this cluster have an average SFR of 28.7295 M /yr. The third cluster, CL1232, had a total SFR of 3114.48 M /yr, an average SFR of 56.6268 M /yr for each of the fifty five cluster members. From these clusters, a staggering forty seven were within half of the virial radius, totaling 2625.03 M /yr, about 84% of the total. SFR, according to the Hubble type for each cluster member, was also calculated. The elliptical galaxies detected have an average SFR of 63.5693 M /yr. Lenticular galaxies have an average SFR of 30.3968 M /yr, and the SFR of spiral galaxies average 53.3709 M /yr. The irregular cluster members of this cluster have an average SFR of 90.1752 M /yr. DISCUSSION Initially we assumed there would be generally increasing star formation rate activity within clusters of increasing redshift. Logically, this seemed sound; galaxies are younger in the past (at higher redshifts) and would seem, on average, more densely packed, giving them more star formation material—gas—closer at hand to form more stars.

6 In this section, we present the “average” SFR for objects detected by MIPS. These values, however, are not necessarily accurate, because a number of elliptical galaxies, for example, may not be detected at 24 micron wavelengths. As a result, our averages are not an average over all of the elliptical galaxies, but only an average over the elliptical galaxies that were detected by MIPS.


We found, however, that the evolution of star formation rate within our three clusters at different redshifts did not exactly match our expectations. Instead of forming the direct relationship that was expected, we found a concave down curve for the SFR of the entire cluster, and in fact a sharp decrease in SFR as z increased for the fraction of members inside the virial radius. As a result, we need to determine whether our calculations are incorrect or whether there is a relationship more complicated than originally predicted. Our next steps will be to grapple with the reasons why our cluster’s SFRs produce such interesting behavior around z = .6 or so, and whether other observational studies verify the surprising star formation variation we observed in this region.

REFERENCES Finn, Rose A, et al. "Hα-Derived Star Formation Rates for Three z ≈ 0.75 EDisCS Galaxy Clusters." The Astrophysical Journal (2005): 206-277.

Finn, Rose A, et al. SFRs of 9 Intermediate-Z Clusters. Sienna College. 2006. 1-12.

Halliday, C, et al. "Spectroscopy of Clusters in the ESO Distant Cluster Survey (EDisCS)." Astronomy & Astrophysics (2004): 397-413.

Kennicutt, Robert C. "Star Formation in Galaxies along the Hubble Sequence." Annu. Rev. Astron. Astrophys. (1998): 189-231.

Poggianti, Bianca M. Color, Spectral and Morphological Transformations of Galaxies in Clusters. Padova Astronomical Observatory. The Netherlands: Kluwer Academic, 2006. 1-10.

Poggianti, Bianca M, et al. "The Evolution of the Star Formation Activity in Galaxies and Its Dependence on Environment." The Astrophysical Journal (2006): 188-215.

Poggianti, Bianca M. Emission Line Galaxies in Clusters. INAF-Padova Astronomy Observatory, Italy. 2005. 1-13.

Poggianti, Bianca M. Evolution of Galaxies in Clusters. INAF-Padova Astronomy Observatory, Italy. Novigrad: SISSA, 2004. 1-16.

Rudnick, Gregory. Agenda for Spitzer Teacher's Meeting in Pasadena CA in July 2006: 24 μm Observations of CL1037. Spitzer Teacher's Meeting, July 2006, Spitzer Science Center.

Rudnick, Gregory, et al. "Studying High Redshift Galaxy Clusters with the ESO Distant Cluster Survey." Reports From Observers: 19-24.

White, S D M, et al. "EDisCS — the ESO Distant Cluster Survey." Astronomy & Astrophysics (2005): 1-20.


CL1037 Average SFR CL1037 Average SFREDisCS ID Number (Mο/yr) EDisCS ID Number (Mο/yr)

EDCSNJ1037521-1243392 62.493 EDCSNJ1037529-1241349 32.265EDCSNJ1038022-1242560 39.521 EDCSNJ1037520-1241119 20.171EDCSNJ1038016-1241080 97.730 EDCSNJ1037514-1244162 32.653EDCSNJ1038012-1242150 36.951 EDCSNJ1037514-1243266 45.022EDCSNJ1037596-1243469 41.042 EDCSNJ1037509-1241063 135.87EDCSNJ1037590-1245212 63.713 EDCSNJ1037507-1241068 135.87EDCSNJ1037590-1242557 23.250 EDCSNJ1037504-1241447 61.310EDCSNJ1037589-1245578 35.582 EDCSNJ1037501-1243350 46.723EDCSNJ1037586-1244022 49.457 EDCSNJ1037498-1243240 24.622EDCSNJ1037582-1243466 50.018 EDCSNJ1037497-1242015 80.099EDCSNJ1037576-1243360 23.471 EDCSNJ1037496-1240584 26.800EDCSNJ1037574-1242454 18.868 EDCSNJ1037496-1241252 57.588EDCSNJ1037573-1245551 44.589 EDCSNJ1037492-1243403 35.590EDCSNJ1037573-1241263 57.402 EDCSNJ1037489-1245388 122.72EDCSNJ1037570-1243291 120.40 EDCSNJ1037487-1244461 28.929EDCSNJ1037563-1241204 45.733 EDCSNJ1037481-1241298 32.918EDCSNJ1037562-1241460 192.99 EDCSNJ1037474-1241072 205.79EDCSNJ1037561-1244252 102.12 EDCSNJ1037447-1243270 38.907EDCSNJ1037539-1244024 583.08 EDCSNJ1037440-1244409 29.412EDCSNJ1037538-1244010 583.08 EDCSNJ1037440-1242399 118.42EDCSNJ1037535-1242005 54.029 EDCSNJ1037427-1243470 102.53EDCSNJ1037534-1242193 23.385 EDCSNJ1037424-1243362 77.202EDCSNJ1037531-1241354 32.265


EDCSNJ1228001-1136095 44.754 EDCSNJ1227553-1136118 88.558EDCSNJ1227572-1135552 128.89 EDCSNJ1227539-1135589 28.446EDCSNJ1227554-1139178 27.940 EDCSNJ1227530-1138474 54.441EDCSNJ1227533-1136527 67.784 EDCSNJ1227530-1138085 145.49EDCSNJ1228031-1136039 34.192 EDCSNJ1227523-1135184 30.869EDCSNJ1228030-1139157 35.689 EDCSNJ1227510-1137559 41.603EDCSNJ1227586-1136295 55.581 EDCSNJ1227506-1135282 64.037EDCSNJ1227586-1139362 54.718 EDCSNJ1227505-1136072 103.26EDCSNJ1227585-1140265 304.18 EDCSNJ1227498-1139019 182.11EDCSNJ1227583-1139255 22.842 EDCSNJ1227486-1135342 28.730EDCSNJ1227570-1135193 30.139 EDCSNJ1227467-1136464 193.08EDCSNJ1227569-1136423 49.321 EDCSNJ1227444-1138305 56.771EDCSNJ1227565-1137141 44.180 EDCSNJ1227433-1139362 30.586



EDCSNJ1232373-1249247 100.42 EDCSNJ1232321-1249464 101.40EDCSNJ1232327-1249057 52.147 EDCSNJ1232320-1250423 41.969EDCSNJ1232323-1251267 27.947 EDCSNJ1232313-1248364 25.670EDCSNJ1232309-1249408 30.397 EDCSNJ1232311-1249451 58.834EDCSNJ1232307-1249573 75.514 EDCSNJ1232308-1252171 36.243EDCSNJ1232297-1250080 33.422 EDCSNJ1232306-1249579 75.514EDCSNJ1232294-1249028 36.994 EDCSNJ1232304-1249573 75.514EDCSNJ1232275-1248540 88.151 EDCSNJ1232302-1251229 86.742EDCSNJ1232273-1251080 19.184 EDCSNJ1232301-1253181 29.596EDCSNJ1232238-1252111 40.034 EDCSNJ1232290-1252600 37.396EDCSNJ1232365-1253082 54.743 EDCSNJ1232288-1251516 44.211EDCSNJ1232400-1248353 179.90 EDCSNJ1232288-1249386 35.410EDCSNJ1232394-1252042 28.889 EDCSNJ1232287-1249376 35.410EDCSNJ1232387-1248459 70.328 EDCSNJ1232272-1250593 21.879EDCSNJ1232376-1248420 77.469 EDCSNJ1232268-1247563 45.764EDCSNJ1232372-1249258 100.42 EDCSNJ1232266-1247569 45.764EDCSNJ1232369-1251280 56.889 EDCSNJ1232265-1249266 136.00EDCSNJ1232364-1250394 58.983 EDCSNJ1232264-1249254 136.00EDCSNJ1232355-1249470 26.298 EDCSNJ1232262-1250380 36.731EDCSNJ1232348-1250073 44.517 EDCSNJ1232261-1250355 36.731EDCSNJ1232347-1252164 33.537 EDCSNJ1232252-1248324 33.965EDCSNJ1232343-1253415 34.461 EDCSNJ1232236-1252123 40.034EDCSNJ1232340-1250133 118.37 EDCSNJ1232216-1249463 53.930EDCSNJ1232339-1252492 31.335 EDCSNJ1232212-1250468 39.051EDCSNJ1232334-1249030 31.793 EDCSNJ1232208-1251077 59.269EDCSNJ1232332-1251228 40.853 EDCSNJ1232207-1251106 59.269EDCSNJ1232324-1251464 79.182 EDCSNJ1232202-1250144 36.900EDCSNJ1232323-1252070 77.113

Table 1 Calculated SFRs (in M /yr) for each of the cluster members for each cluster are presented above.


Figure 4 From Fasano et al. (2000). The morphological classifications used in our study were given to us by Vandana Desai, who specializes in classifying galaxy morphologies. The fraction of elliptical, lenticular, and spiral galaxies is shown for clusters between z = 0 and z = 0.6, with our clusters’ galaxy types shown in color.


0

10

20

30

40

50

60

Perc

enta

ge o

f Tot

al S

FR

Elliptical Lenticular Spiral Irregular

Galaxy Morphology

CL1232: SFR Contribution vs. Morphology

0

10

20

30

40

50

60

Perc

enta

ge o

f Tot

al S

FR


Galaxy Morphology


0

10

20

30

40

50

60

70

80

Perc

enta

ge o

f Tot

al S

FR


Galaxy Morphology


z = 0.54 n = 55 Galaxies z = 0.58 n = 45 Galaxies z = 0.64 n = 26 Galaxies

Figure 5 SFR contributions classified by galaxy morphology for each of the three clusters described in this paper.


CL1232: SFR vs. Distance from Cluster Center

0

20

40

60

80

100

120

140

160

180

200

0 200 400 600 800 1000 1200 1400

Distance (kpc)

SFR

(Sol

ar M

asse

s/Ye

ar)

EllipticalLenticularSaSabSbSbcScSdmIrregularHalf Virial Radius


0

100

200

300

400

500

600

700

0 200 400 600 800 1000 1200 1400

Distance (kpc)

SFR

(Sol

ar M

asse

s/Ye

ar)

EllipticalLenticularSaSabSbSbcScScdSdIrregularHalf Virial RadiusVirial Radius



0

50

100

150

200

250

300

350

0 200 400 600 800 1000 1200 1400

Distance (kpc)

SFR

(Sol

ar M

asse

s/Ye

ar)

EllipticalLenticularSaSabSbSbcScIrregularHalf Virial RadiusVirial Radius

Figure 6 SFR is classified by distance from the cluster center and presented above for each of the three clusters.


Deep Field Galaxy Classification Daniel Kirpes

Howenstine Magnet School, Tucson, AZ Teacher: Chris Martin, TLRBSE 2005

ABSTRACT Over three months images of sixteen galaxies were taken, with varying morphologies, using the 0.359m telescope at the New Mexico Skies observatory. The images were studied using the program ImageJ. The brightness in pixels (source) was found and plotted on a chart. From the source the magnitude was calculated. Then the magnitude versus B-I, of each galaxy type, was graphed to more easily check for any patterns. INTRODUCTION Is there a relationship between the Blue magnitude minus Infrared magnitude (B-I) and the morphology of a galaxy? The first project of the year was studying galaxies in images of the Boötes field from NOAO and the Sloan Digital Sky Survey. This project was created by Mr. Martin and Dr. Garmany and also used a NASA grant (Understanding galaxies using electromagnetic radiation). Part of this project was the study of B-I which seemed to have some sort of correlation to the galaxies’ shape. The Deep Field is a part of the universe that is so far away that the light coming from the galaxies there is from when our universe was young. Images of Boötes Field were taken from the NOAO Deep Wide Field Survey (NDWFS) www.noao.edu/noao/noaodeep/. NDWFS is a deep optical near infra red survey that covers the Boötes Field. The images of the Boötes Field used in this study are from the 4m Mayall telescope at Kitt Peak. According to Hubble’s theory of galactic evolution, usually galaxies start off as an elliptical shape. As they age and get consumed by the black hole in their center, galaxies change to a ordinary spiral (SO) then either to a spiral or barred spiral shape. Most stages are represented in Edwin Hubbel’s tuning fork diagram (as in the Montana State University web site). However, this diagram does not include irregular galaxies. Neither did this experiment, because they come in a wide variety of shapes nor would likely have a wide range of B-I. In addition, they are fairly uncommon, so this shouldn’t matter too much in the end results. Using the equation magnitude=22.35-2.5[log10(source/exposure time)] the apparent magnitude of each galaxy was found. The source is the brightness in pixels, and the exposure time is the time period used to capture the light coming from the galaxy. The B-I was then calculated for each galaxy. B is the magnitude of the blue filtered image, while the I is the magnitude of the infra red-filtered image. Originally, the plan was to use the ultraviolet minus Infra red for this study. However, New Mexico Skies do not have an ultraviolet filter available on the particular telescope used for this study.


This is important because if it works well it could be a way to better classify distant galaxies that could not be identified by their shape (Hubble’s tuning fork). This would eventually further increase the size of the charted universe. It could also be used to check the existing classification of local galaxies. METHOD Over three months (Dec 6, 2006- Feb 18, 2007) the images of sixteen galaxies were taken, with varying morphologies, using the 0.359m (14’’) telescope at New Mexico Skies Observatory in Mayhill, New Mexico. The galaxies were chosen from the NGCIC website. Specifically, the galaxy needed to be viewable from the observatory from 8 PM to 10 PM. To use New Mexico Skies, first you have to make an appointment for using the telescope to allow the staff to set it up. Then you log on to the telescope, on your own home computer, during your appointment. From your home you can move the camera, change filters, and take the images. After collecting the images using the Johnson-Blue and Johnson-infrared filters, the images were studied using a free Java program called ImageJ. The brightness in pixels (source) was found and plotted on an Excel spreadsheet. Eventually, the source (the average brightness in pixels of the galaxy) was used to obtain the magnitude and the B-I. OBSERVATIONS AND DATA REDUCTION From December 6, 2006 to February 18, 2007, there were a total of 54 images taken of 18 different galaxies. All the images were taken using the 14” telescope at New Mexico Skies Observatory using the blue, red, and infrared filters. The images were reduced to biased, dark, and flat. This made the galaxy stand out much better and allowed for accurate measurements to be taken. The magnitudes were found, then graphed, and the graphs were examined for a pattern.

NGC shape Blue Source Exp. Time

Magnitude Infrared Source

Exp. Time Magnitude B-I

714 SO 30120.83 180 16.79 55463.34 60 14.94 1.85 860 E 10788.02 120 17.47 17705.83 60 16.18 1.29 938 E 9036.00 60 16.91 47364.43 60 15.11 1.8

1022 S 55423.62 120 15.69 209994.9 120 14.24 1.45 7514 S 23426.26 180 17.06 114915.5 180 15.34 1.72

596 E 8694.86 60 16.95 280534.8 60 13.18 3.77 1541 SO 7688.59 160 18.15 25424.04 60 15.78 2.37 1615 SO 11372.9 160 17.72 46556.95 60 15.13 2.59 1875 SO 4875.04 160 18.64 21049.81 60 15.99 2.65 1993 SO 19885.31 160 17.11 61994.16 60 14.81 2.3 2089 SO 30329.42 160 16.66 84238.89 60 14.48 2.18 1507 SB 4878.45 120 18.33 38886.07 60 15.32 3.01 1580 S 10362.5 120 17.51 30900.9 60 15.57 1.94 1635 SB 3025.69 120 18.85 48682.97 60 15.08 3.77 1819 SB 1020.76 120 20.03 21657.6 60 15.96 4.07

1961 S 13673.2 120 17.21 3898.2 60 17.82 -0.61 2300 E 38592.1 240 16.83 137841.7 60 13.95 2.88 2426 E 10567.7 120 17.49 20977.92 60 15.99 1.5


Abbreviation Explanation E Elliptical; spherical, not flattened SO Ordinary Spiral; flattened but the arms haven’t developed yet S Spiral; flattened and has arms that start curving at the galaxies center SB Barred Spiral; flattened with arms that start curving at the end of a bar shape extending from

the center on both sides

0

5

10

15

20

-1 0 1 2 3 4B-I

Mag

nitu

de

Series1


Eliptical

0

5

10

15

20

-1 0 1 2 3 4B-I

Mag

nitu

de

SO

0

5

10

15

20

-1 0 1 2 3 4B-I

Mag

nitu

de


Spiral

0

5

10

15

20

-1 0 1 2 3 4B-I

Mag

nitu

de

Bared Spiral

0

5

10

15

20

-1 0 1 2 3 4B-I

Mag

nitu

de

ANALYSIS AND RESULTS The magnitude was approximately the same for each galaxy. A correlation was observed between the B-I value and the morphology of the galaxies. The elliptical galaxies were between one and two B-I, the SO galaxies were between two and three B-I, spiral galaxies were also between one and two B-I, but the barred spiral galaxies were the highest with a B-I of three to four.


DISCUSSION The least reliable grouping is the elliptical galaxies. This was probably from picking the wrong galaxy in the exposures. This is easier to do with elliptical galaxies because they can look more like stars than other galaxies. This makes them hard to differentiate if they are small. All the rest are sufficiently accurate. The shape can be identified unless it’s B-I is right between the range of the shapes (2 or 3). In some images the galaxy that was being looked for was not obvious. Meaning, the measurements of the wrong galaxy might have been taken. This was the problem with NGC2300 and NGC2426. In the future, when selecting galaxies to image, larger galaxies will be chosen to prevent this problem from happening again. In addition, if the right galaxy can’t be discerned, that image/galaxy will be discarded so as to not upset the results. It would be beneficial to increase the number of local galaxies in the study. Ideally, there would be at least ten elliptical, SO, spiral, and barred spiral galaxies in this study. This would validate the classification system. Using the Boötes Field blue and infra red images from NWDFS, this classification system would allow a detailed examination of deep field galaxies. SUMMARY The results indicated a well established connection between the B-I and the shape of a galaxy. With this data one can tell if a galaxy is a barred spiral or an SO shape. It was not possible to tell the difference between a spiral and elliptical galaxy at the moment. If it was possible to get more data sets, it could eventually be possible to find a difference between the two. If a difference was found then almost any galaxy could be identified, and for the most part, they would have the right classification. ACKNOWLEDGMENTS NOAO National Optical Astronomy Observatory Dr. Garmany Dr. Croft REFERENCES <www.ngcic.org/dss/dss_ngc.asp>; used to chose the galaxies I would take images of. <http://en.wikipedia.org/wiki/Galaxy_morphological_classification> ; images and some information btc.montana.edu/CERES/html/Galaxy/galbackground.htm. ; Galaxies and Galactic Structure for Hubble’s theory of galactic evolution.

The 2007 RBSE Journal · The visible part of the Sun is referred to as the photosphere. Here, large...

Documents

Transcript of The 2007 RBSE Journal · The visible part of the Sun is referred to as the photosphere. Here, large...