1 of 23

Astronomical Spectroscopy

Fall 2013

Gerald Buxton,Justin Stockwell, and Annika Gustafsson

Dr. Scott Fisher

Professor Gregory Bothun

2 of 23

Astronomical Spectroscopy� Applications

chemical composition
temperature
density
mass
distance
luminosity
relative motion

Hertzsprung-Russell Diagram

GERALD PRESENT.

Spectroscopy is the fundamental tool with which an astrophysicist observes the universe. Everything we know about the universe, we know because of our analysis of incoming light. Primarily, the doppler shift can be utilized to unlock most of the universe’s secrets. For example, a spinning star emits light that is red shifted from one side (the side spinning away from the observer), and blue shifted on the other side, where the source is moving towards the observer. By treating this source as two separate sources, one can use spectral techniques to determine how fast the star in question must be spinning.

Along with rotational broadening, many other applications come into play in astronomical spectroscopy. If one would like to know the chemical composition of a particular star, the spectrum of that star is a great place to start in deducing the make-up. In addition, temperature broadening occurs because of the speed at which particles within the star are traveling. In hotter stars, they will obviously travel faster than in stars that are cooler. This difference shows up in the spectrum of the star, as the speed of the particles doppler shifts in accordance with physical law.

For the performer of astronomical spectroscopy, the key can be determining what effects are due to which phenomena, and this can be difficult when looking at stellar spectra. Really, this dives into the true long term goal of this project; that is, to be able to become “spectrally literate,” and understand subtle differences in a simple plot of intensity vs. wavelength. For all the many possible causes of a spectral feature, it would be nice to one day be able to look at a plot and know right away what we are looking at. To date, we have had a few high noise, low signal captures of Vega and Polaris, as well as Jupiter. So far, we have not been able to collect relevant enough data to observe any true spectral features. As we move forward, that is the immediate goal: a system that can deliver data good enough to analyze, repeat, and explore!

3 of 23

Classical Spectrometers

Slit on which the light from source is focused
Collimator parallelizes the light
Diffraction Grating splits the light
Camera to Focus

http://en.wikipedia.org/wiki/Spectrometer

4 of 23

Ocean Optics Spectrometer

http://www.oceanoptics.com/products/usb2000+.asp

GERALD PRESENT.

Here, we have a basic physical overview of our USB 2000+ Spectrometer and its relevant specifications. First, I would like to go over the physical mechanisms that go into making the spectrometer functional. First, we have an SMA connector. This is where the fiber inserts, and light enters the optical bench through this connection.

2. Next, we have the fixed entrance slit. This is dark piece of material that contains a rectangular aperture. The size of aperture can very, and this restricts the amount of light that enters the optical bench. If desired, the spectrometer can actually be used without a slit at all. In this scenario, the diameter of the input fiber regulates the amount of light incident on the optical bench.

3. Here, we have a filter. This filter restricts the light to a pre-determined wavelength region. By doing this, we are able to control another variable before light ever enters our optical bench. When we observe spectra on our interface (spectrasuite), the wavelength of incoming light is observed along this range. This filter also acts to block second and third order effects.

4. This is the collimating mirror. this mirror focuses light in a beam onto the diffraction grating. It is appropriately alligned with the 0.22 numerical aperture of the supplied optical fiber.

5. Next, we have our diffraction grating. This is one of the most elementary working parts of the spectrometer, and utilizes the most basic (and perhaps most important) of the scientific principles involved. The collimated light is diffracted, at specified wavelengths, onto the focusing mirror, and we are well on our way.

6. The mirror focuses only 1st order spectra onto the CCD. This mirror is high reflectance to promote the capture of light and ensure that as little light is lost as possible.

7. L2 Detector Lens. This lens focuses the incoming light (from the focusing mirror) onto the much smaller pieces of the CCD.

8, 9, 10. All elements of the CCD. Element 10 is actually an upgrade offered by ocean optics.

5 of 23

Astronomical Spectrograph

Standard Astronomical Spectrograph

Schematic of Spectrograph

http://tdc-www.harvard.edu/instruments/fast/fastpic.html

6 of 23

Useful Astronomical Terms

Vocabulary

Apparent magnitude is the apparent brightness of an object in the sky as it appears to an observer on Earth. Bright objects have a low apparent magnitude while dim objects will have a higher apparent magnitude.

Absolute Magnitude measures how bright an object would appear if it were exactly 10 parsecs away from Earth. On this scale, the Sun has an absolute magnitude of +4.8 while it has an apparent magnitude of -26.7 because it is so close.

Quantum Efficiency is the amount of counts the CCD camera picks up per pixel. This changes depending on the wavelength of light picked up by the detectors. When analyzing data we will have to take this into account.

HIDE

Apparent and absolute magnitude are useful in a variety of ways. Most importantly, they aid astronomical spectroscopy in calculating the distance of stars using a method called Spectroscopic Parallax. The word parallax does not indicate a trigonometric relationship, rather it is used as a reminder that this method is meant for approximating distances. For this method work, an observer must know the apparent magnitude of the star, as well as taken an accurate spectral measurement. The astronomer must have a Hertzsprung-Russel Diagram nearby, which plots the temperature of a star versus the luminosity of that star. This diagram exists because of the basic assumption that stars far from earth behave similarly to stars close to earth (in terms of the relationship described by the Hertzsprung-Russell Diagram. Once you have plotted a star on this diagram, it allows an assumed absolute magnitude to be read off. There are seven main types of spectral classifications that astrophysicists designate, and each classification corresponds to an approximate surface temperature range for the object.

Now, the larger story of this method starts way back when in the early stages of the 20th century at Harvard. A pioneering woman, named Annie Cannon, did a ton of early work on Stellar Spectra. She was an assistant at the Harvard Observatory, when she became famous for meticulously classifying hundreds of stars per hour. She completed a large catalog of spectral types for hundreds of thousands of stars.

The non-absorbed, continuous spectra plays the largest role in determining the temperature of a star. The temperature increases corresponding to an overall increase in light energy sent from the source.

Luminosity is the total energy produced by a star in one second. There are two main factors that come into play here: the radius of a star and its temperature. You can solve for luminosity by multiplying together the energy each section of a star produces in accordance with stefan-boltzman law, and the surface area of the star. L=4piR^2sigmaT^4

Now, back to the Hertzsprung-Russell Diagram. Observational comparisons led to the discovery of the H-R diagram. They realized that when stars are organized according to properties of temperature and luminosity, they almost always form a smooth curve. The stars along the main diagonal are known as “main sequence” stars, which makes up about 85% of the universe’s stars. Temperature, as we have seen, can be replaced with spectral class. Looking at the diagram, we can see that bright cool stars must be very large. Constant radius lines are plotted on the diagram. There is a correlation between main sequence stars mass and luminosity. L is proportional to M^3.5. This proportion exists because stars that are much more massive would collapse under their own gravity, while less massive stars would be blown apart by radiation pressure from the luminosity.

So, spectral class is closely related to temperature. What about luminosity? Instead of looking at which lines are present, we now want to look at the width of the lines. More luminous stars have narrower spectral lines, this we know due to the classification done by Annie Cannon.

Thus, using the HR diagram, we can use spectral type to place the star on the horizontal axis of the HR diagram. For vertical placement, we use the thickness of absorption lines. Given luminosity and temperature, we can now calculate absolute magnitude. This is then compared to the observed apparent magnitude, and the distance is calculated.

7 of 23

Useful Astronomical Terms

Stellarium

Altitude/Azimuth

Slew Speed

1 = .5x*	6 = 64x*
2 = 1x*	7 = .5°
3 = 4x*	8 = 2°
4 = 8x*	9 = 3°
5 = 16x*

* = (sidereal)

° = 60 arcmin

8 of 23

What have we been doing?

Vega with Night Sky subtracted

Vega from Vega Spectrum Atlas by the spectrometer MERIS

Vega

Perfectly Calibrated Star

GERALD PRESENT

This slide is meant to illustrate what the project has been about, and what we see it becoming in the future. In many ways, this term is a transitional term for the project. The fall means that nights out on the roof, and observations, are not very feasible. As a result, this time must be spent in the lab laying the table for future observations.

Last spring, we were able to take some interesting data, with one example in the accompanying slide. For astrophysicists, Vega is a meaningful reference point from which you can check your own data and observations. It is such, as it is a “perfectly calibrated star.” Looking at our data, we were encouraged at the beginning of some spectral features that would show we are pointing in the correct direction. However, we needed a high “Scan-to-average,” to eliminate random noise, while non-random noise still ran rampant throughout our data collection. In order to appropriately analyze spectral features, we knew that we must perfect our set up even further in order to move forward. As the spring wrapped up, we made it our groups goal to analyze and improve the optical set up we had, and, at the end of this term, determine if we would need to try a completely new approach moving forward. We hoped, however, that we would be able to just fine tune the one that we had.

9 of 23

Progress for This Term

1. Inventor Drawings

2. Efficiency of Spectrometer

3. Optical Alignment

4. Efficiency of Setup

5. Determine Feasibility of Current Set-Up

10 of 23

1. Inventor Drawings

Old T-adapter System

Beam Splitter

Eye Piece

ANNIKA PRESENT

The initial setup at the beginning of Fall term consisted of the telescope with a beam splitter and lens tube system attached to the back. The beam splitter attached to an eyepiece on top and a lens tube at the back. The lens tube had a single 10cm converging lens which focused light directly on a fiber attached at the end of the lens tube. The fiber was attached using a fiber adapter with thumb screws which gave control over the three degrees of freedom of the fiber: 1 along the z axis where the length of the lens tube can be changed on the outside, and 2 and 3 along the x and y axis where the position of the fiber can be adjusted using the two thumb screws.

The “T-adapter” was one of our biggest concerns heading into this term. After taking stellar spectra in the spring, we realized the unreliability of our old setup made it difficult to get meaningful data. As a result, it was a priority this term to deliver an instrument with less variables. Firstly, we wanted one that did not allow for easy flexure of both system and fiber. The mass of our adaptor meant we did not have a perfectly parallel optical axis, the adapter would flex towards the ground, and this would affect our results.

In addition, we were not optimizing the light our system would receive by using a beam splitter. Instead, we installed a flip mirror, which allows for both visual and spectral observation, while not compromising 50 % of the incoming light. The flip mirror served two purposes. The first has already been stated, to optimize the amount of light our spectrometer sees. The second purpose is to make the adaptor more sturdy in general. The beam splitter prevents flexure between parts, due to better connections and newly machined parts. Whereas before we machined black plastic when making adaptor pieces, we now used aluminum. This meant far less flexure, and a more controlled optical environment.

Basically, we wanted to use this term to see if our setup, without any drastic changes, is viable at all. By optimizing all the little details, we ought to be able to get a maximum efficiency rating in the lab, ultimately proving whether it is possible to observe meaningful stellar spectra or not. As a baseline, we need an efficiency rating of around 50% to perform stellar observation.

11 of 23

1. Inventor Drawings

Flip Mirror

12 of 23

1. Inventor Drawings� Machined Adapter

13 of 23

1. Inventor Drawings

Lens Tube w/Fiber Adaptor

14 of 23

1. Inventor Drawings

New T-adapter System

15 of 23

2. Efficiency of Spectrometer

586 nm laser
iris
converging lens
ND filters
fiber adaptor
2 square mirrors
power meter

The mirrors added the necessary flexibility for fine alignment

16 of 23

3. Align Telescope

Align Finderscope to Eyepiece

17 of 23

3. Align Telescope

Align Fiber to Eyepiece

Willamette Basement Hallway

Pros:

Distance greater than focal distance of telescope

Cons:

Too much ambient light

JUSTIN PRESENT

Then, we needed to align the fiber to the eyepiece. We began with a similar setup in the Willamette basement hallway. We placed the telescope at one end of the hallway and a hydrogen spectral lamp at the other end.. With the fiber connected to the spectrometer, we had a nice hydrogen spectra to view in SpectraSuite.

We used this spectra to check the alignment of the fiber. By changing the x, y, and z-axes we were able to maximize the number of counts observed in the main hydrogen spectral peak. When our system is perfectly aligned we should be recording the maximum number of counts. We first adjusted the z-axis. The z axis is adjusted by changing the distance of the fiber, which we do by altering the length of the lens tube. We started with the lens tube completely screwed in and then slowly unscrewed the lens tube, increasing the distance to the fiber. We noticed that with each full turn we would get a max count value at approximately the same point in our rotation. We decided to settle on a distance that was close to the focal distance of our converging lens where we had observed a max count value, and also consistent with the numerical aperture of the fiber. Then, we proceeded to adjust the x and y-axes by slowly rotating the thumb screws on the fiber adapter. We observed maximum counts when both thumb screws were completely screwed in. However, we noticed that we had a difficult time returning to a specific max counts number. No matter how careful and precise we were, there was variance in the max counts number. We noticed that this count number would change just by moving the fiber. We also noticed how flexure in our fiber system affected the total number of observed counts.

18 of 23

3. Align Telescope

Align Fiber to Eyepiece

Willamette 100

Pros:

Distance greater than focal distance of telescope

Cons:

Dark enough to not allow for much ambient light

JUSTIN PRESENT

We used this spectra to check the alignment of the fiber. By changing the x, y, and z-axes we were able to maximize the number of counts observed in the main hydrogen spectral peak. When our system is perfectly aligned we should be recording the maximum number of counts. We first adjusted the z-axis. The z axis is adjusted by changing the distance of the fiber, which we do by altering the length of the lens tube. We started with the lens tube completely screwed in and then slowly unscrewed the lens tube, increasing the distance to the fiber. We noticed that with each full turn we would get a max count value at approximately the same point in our rotation. We decided to settle on a distance that was close to the focal distance of our converging lens where we had observed a max count value, and also consistent with the numerical aperture of the fiber. Then, we proceeded to adjust the x and y-axes by slowly rotating the thumb screws on the fiber adapter. We observed maximum counts when both thumb screws were completely screwed in. However, we noticed that we had a difficult time returning to a specific max counts number. No matter how careful and precise we were, there was variance in the max counts number. We noticed that this count number would change just by moving the fiber. We also noticed how flexure in our fiber system affected the total number of observed counts.

So, we went forward with the recommended procedure. We first ran into trouble when the power meter would not pick up our LED. After messing around with the power meter and calibrating it a couple of times, we were able to get some measurements. We measured a power of 4.8 microwatts through the back of the telescope and 17 nanowatts through the entire telescope system with the fiber as best aligned as possible. We calculated that we were losing 99.7% of the light through our telescope system. Of course we were surprised by this result as it was several orders of magnitude away from our expected value. At a separate time, we were able to successfully gather power readings at different stages of our telescope system to track power losses. We took the T-adapter off of the telescope and used a piece of paper to find the focal point of the light to try and determine where the problem was. We noticed that the back-focal distance of the telescope was 14.5cm. On our T-adapter, 14.5cm is just past the flip mirror, so the light is being focused far too fast and has already gone through the focal point by the time it reaches the fiber. This is consistent with the power readings we got where we had a max power at approximately the focal point of the light. We replaced the 12.5cm biconvex lens in our lens tube with a 100cm biconvex lens, but the light still focused far before our fiber, even with our fiber as close to the flip mirror as our setup allowed. Next, we tried to use a two lens system with a diverging lens and a converging lens to see if we could get the light to focus slower, but again we were unsuccessful. We determined that the best solution to this problem would be to use the Optica software and determine exactly what is happening to the light in our T-adapter.

19 of 23

4. Efficiency of Set-Up

Data

3.1 5.5 3.81 0.06

56.3 69.1% 1.1%

@ back of telescope @ back-focal distance @ start of fiber @ end of fiber

20 of 23

4. Improving Efficiency

The goal for Winter Term was to improve upon the efficiency of our current setup.

-The first major component of this is to determine precisely the focal distance of our telescope.

-We also needed to determine the spot size of our incoming light

-To do so we used a beam profiler which accurately determines both components of our light profile.

21 of 23

Going Forward...

T-Adapter

Optica

Improve Components on Adapter:

Fiber Adapter

Need thumb screws to move fiber past central point

Fiber

Need sheath to hold fiber stable

Adapter to Telescope

Need to machine a piece to screw flip mirror onto telescope to eliminate flexure
Currently the flip mirror slides in and attaches by screws. “Pins in”

22 of 23

Going Forward...

Optica

Telescope Dimensions:

Primary Mirror
Spherical Mirror
Radius = 101.6mm
Hole Aperture = 38.1mm
Focal Ratio = F/2
Secondary Mirror
Schmidt Lense: Corrects for Spherical Mirror
Radius = 63.5mm
Magnifies the Primary Mirrors focal ratio by 5
Combination of Mirrors Creates a Focal Length of 2032 mm
Optical Tube Length
This is about 431.8mm (17 in)

It changes as the telescope is focused

23 of 23

Going Forward...

Goals for Next Term:

Improve T-Adapter
**Optica Drawing of �entire telescope system**
Investigate New Astronomical Spectrometers, as opposed to current set up where spot size is a problem
Align entire telescope system
Characterize the spectrometer
Find an absolute sensitivity measurement using absolutely calibrated light source
Determine feasibility of using the Ocean Optics spectrometer for astronomical purposes