Department of Physics, Middlebury College 1992-93
Modern Physics Laboratory

XV. Spectroscopy of the Solar Photosphere and the Earth's Atmosphere


In previous labs you used a 1 m spectrometer to study various types of emission spectra, including the hydrogen and deuterium Balmer lines, laser-induced fluorescence lines of minerals, and the Zeeman effect splitting of mercury lines . In this lab, an absorption spectrum (the solar spectrum) will be the subject of your attention as you gain more experience in the operation of our spectroscopic hardware and software.

Electromagnetic spectra are broadly classified into three categories: emission line spectra, continuum spectra, and absorption line spectra. It is possible, and common, for a single physical source to show two, or possibly all three, of these characteristics. In the visible region, the sun shows an absorption line spectrum, that is, absorption features superimposed on a continuum. The continuum is formed in the solar interior and obeys approximately a black-body spectrum; the absorption lines are formed as the continuum radiation passes through the cooler outer layers of the sun and through the earth's atmosphere.1

A portion of the solar spectrum is illustrated in Fig. 1. The smooth, localized dips in intensity are the absorption features or "lines"; the continuum is the higher, slowly varying background which is visible between absorption lines. Each line corresponds to one (or more, in the case of several close, blended features) absorption line, which results from a photon-absorbing transition of an atom, ion, or molecule.

The location, strength, and shape of the absorption lines depend on the nature of the absorbing atom and on the physical conditions of the gas; therefore, it is possible to learn a great deal about the absorbing material through the careful analysis of the spectrum. Physical parameters of interest include temperature, pressure, abundances of various elements and compounds, magnetic field strength, and bulk velocity distribution. Some of these analyses are complex and beyond the scope of this lab, and some are rendered more difficult by the fact that the measurement process and conditions may alter some of the absorption lines' characteristics.

Your first investigation of absorption features will concentrate on their wavelengths, their strengths, and their widths. In the case of the solar spectrum, the absorbing agent may be either in the sun's photosphere, or in the earth's atmosphere. You will have the opportunity to observe examples of both in this laboratory exercise.


This lab must be performed in the daytime; furthermore, procedure steps (3)-(8) will be difficult or impossible in the presence of clouds!

Turn on the SPEX 1704 spectrometer and initialize the CD2A. Start up the McSpex data acquisition and analysis program on the IBM PC/XT system.2,3 Turn on the photon counting electronics and configure it as you did for previous labs, except that the lock-in amplifier should be bypassed. In this experiment photon counting rates will be measured by an ORTEC 449 Log/Lin Ratemeter.

(1) Arrange a mirror along the optical bench so that light reflected from the flat mirror mounted just outside the southwest window of the Laser Spectroscopy Lab will be reflected into the entrance slits of the spectrometer. For now, we will use whatever light from the sky is reflected into the lab.

(2) With the room lights out, so that the entrance slit is illuminated only by light reflected into the room via the flat mirror, take a spectrum of 5 - 10 Å width, centered somewhere between 4500 Å and 6000 Å. Use the default scan configuration in McSpex, but choose the scan rate intelligently.3 Repeat the scan of this same wavelength region until you are convinced that the variations in intensity in the scan are significant , rather than just random noise, as indicated by reproducibility in two consecutive scans. You may find it necessary to open the entrance and exit slits in order to achieve a sufficient signal-to-noise ratio. If you find it necessary to open them beyond 50 Ám, however, you should look for an alternative way to increase the signal strength. One possibility is to introduce a focussing lens between the final relay mirror and the entrance slits; another possibility is to dispense, for the moment, with the mirrors and use instead a large optical fiber bundle to introduce light into the spectrometer.

Save two scans which show the same spectral features.

(3) For the rest of the experiment, you will use light directly from the sun to achieve high signal strength. Go to the roof and set up the heliostat, mounted to the parapet four floors directly above the Laser Spectroscopy Laboratory, so that sunlight is reflected off of the tracking mirror and relayed down to the spectroscopy lab via the series of fixed mirrors, as shown in Fig. 2. As the alignment is not likely to be perfect, be prepared to used the remote hand paddle in the spectroscopy lab to make adjustments to the tracking mirror position in order to maintain a strong signal at the spectrometer. Your lab instructor will provide further instruction in operating the heliostat.

(4) Set the entrance and exit slits to 10-15 Ám and scan the solar spectrum in the vicinity of the sodium D lines at 5890 Å and 5896 Å. Make sure you can identify these two lines unambiguously. As you can see in Fig. 1, the sodium D lines are far stronger than any nearby lines. Now take another scan over a range just wide enough to show both D lines and clear continuum both above and below in wavelength. A total range of 10-12 Å should be sufficient. Finally, take a scan over the same range, but cover the entrance slit with a dark card at the beginning and at the end of the scan so that you will have a measure of the "dark" counts that are due to the instrumentation and which will allow you to normalize the intensity scales of your spectra. Many of the lines in this region of the spectrum are due to H2O vapor in the earth's atmosphere.

(5) In this step you will investigate how the widths of the entrance and exit slits affect the widths of spectral features by taking several more scans of this same wavelength region, including the "dark" times at the start and end of each scan, with a variety of entrance/exit slit widths. For this set of scans, the ratio between the entrance slit and exit slit widths should be constant, and in the range 0.5-1.0. Take, and save to disk, a series of scans with a wide range of slit widths, including at least 5 Ám and 50 Ám, and several in between. You should notice that for some range of slit widths, the shape of the absorption lines depends on slit width.

(6) Take a scan in the wavelength range 6850 Å - 6950 Å. The very strong absorption lines here are due to the O2 molecule which is absent in the sun but quite abundant in the earth's atmosphere. Take a scan or, better, a series of scans that show, with the highest possible resolution, the band head near 6865 Å and the pairs of lines which lead away from the band head toward longer wavelengths, as shown in Fig. 3. Be sure to record the date and time accurately for these scans.

(7) In the vicinity of 6300 Å - 6303 Å there are four lines of special interest.3 Two are due to iron in the sun, and two are due to O2 in the earth's atmosphere. Take a scan in this region and adjust the spectrometer scan interval so that all four lines are visible, as well as some continuum on each end, as in Fig. 4. This is the region that is used for the solar rotation rate and the Zeeman effect measurements of another laboratory exercise.

It is sometimes desirable to configure the SPEX 1704 and the McSpex software so that the overall scan time and total number of data points may be chosen independently. For example, we may decide that we would like our scan of these four lines to cover a range of 3 Å, last 100 s, and have a resolution of 0.02 Å. In McSpex, change the acquisition configuration to "A/D 2" and adjust the parameters so that you may obtain a spectrum of the four lines in each of the following configurations:

(a) 3 Å interval, 40 s scan duration, 0.01 Å resolution;
(b) 3 Å interval, 100 s scan duration, 0.04 Å resolution.
Take, and save to disk, three scans of configurations (a) and (b), choosing the RC time constant on the ratemeter to be, respectively, (i) shorter than, (ii) approximately equal to, and (iii) larger than, the single-data point integration time.

(8) If scheduling permits, repeat single scans of each of the three wavelength regions already studied, but with the sun as low in the sky as possible. This will lead to an enhancement of the strengths of the O2 and H2O absorption lines because the sun's light is forced to encounter a much larger number of potential absorbers as it passes obliquely through the earth's atmosphere.


(1) Produce hardcopy plots of each scan you made. Label each with pertinent information concerning the experimental apparatus, such as slit widths, RC constants, and meteorological conditions.

(2) Identify the strongest lines in the spectrum of step (2) of the Procedure section by comparing with Ref. 4. To what accuracy does the SPEX 1704 seem to have been calibrated in wavelength?

(3) Identify the strongest eight or ten lines in your best spectrum of the sodium D lines. What fraction of them originate in the earth's atmosphere? If you have classmates who have completed this lab, or if spectra of this region are available from previous years, compare qualitatively the strengths of the H2O lines, using the prominent nickel line as a reference, between your scans and the others. Was your scan taken on a particularly humid day, or at a particularly low solar altitude?

(4) Using methods learned in previous labs, find the centroid wavelengths of the Na lines, and calculate their separation. Compare your value of the D line separation to the value given in Ref. 4.

(5) Discuss semi-quantitatively the effect of varying the spectrometer slit widths on the appearance of the spectrum. Does there appear to be an optimum setting, considering (a) signal-to-noise ratio only, (b) resolution of lines, and (c) some appropriate tradeoff between signal-to-noise ratio and resolution?

(6) Study your O2 spectra and compare with those offered by your instructor and/or those made by your classmates or by students from previous years. Using the date and time of each scan, calculate the solar altitude at the time of each scan (this is accomplished most easily by means of a "desk planetarium" computer program such as Voyager  by Carina Software) and discuss how this might lead to the differences in the appearances of the spectra. Using Ref. 4, see if you can identify lines due to 16O-18O molecular absorption in the earth's atmosphere.

(7) Using your best spectrum for the region near 6302 Å, calculate for the Fe lines and for the O2 lines, and compare to values given in Ref. 4. Repeat for the other five spectra and comment on how the choices of scan rate, wavelength resolution, and/or RC time constant might affect the resolution of the acquired data and any quantitative conclusions that might be derived from it.

(8) If you have a series of spectra taken over a significant range of solar altitudes, comment on the changes in the spectral features. Devise and execute a way to express quantitatively the correlation between solar altitude and some spectral feature.

1. J.C. Brandt, The Sun and Stars  (McGraw-Hill, New York, 1966), p. 14.
2. E.B. Anthony, The McSpex Project: Laser Spectroscopy Laboratory Interface Program, B. A. thesis, Middlebury College, 1987.
3. J.H. Cooley, Four Projects in Solar Spectroscopy, B.A. thesis, Middlebury College, 1991.
4. The Solar Spectrum: 2935 Å to 8770 Å, NBS Monograph No. 64, edited by C. E. Moore, M.G.J. Minnaert, and J. Houtgast (National Bureau of Standards,Washington, DC, 1966).