DSS-7 Deep Space Spectrograph

SBIG’s new Deep Space Spectrograph (DSS-7) is a spectrograph optimized for the types of spectral observations that an amateur has always been interested in, from stellar classification to nebular analysis to galactic red shifts. It is a more general purpose instrument than our Self Guided Spectrograph (SGS), which is optimized for stellar work, and is much less expensive. It is optimized for the ST-7XME or the low cost ST-402, and will work well with ST-8/9/10/2000 cameras and ST-237s. It will not work with the STL series due to their deeper backfocus required by the built in filter wheel. This memo describes the DSS-7 in detail, and present examples of observations that can be made by the amateur.

Spectroscopy Fundamentals: a spectrograph is a device that can produce a graph of the intensity of light as a function of color, or wavelength. A spectrometer is a device that measures only one selectable color, and a monochromator is a device that transmits only one color. The DSS-7 spectrograph is designed to separate and focus wavelengths from 4000 to 8000 angstroms across the width of an ST-7 CCD. The human eye is sensitive from about 4500 (deep blue) to 7000 (deep red) angstroms, with its peak sensitivity at 5550 angstroms. The silicon CCDs used in SBIGs cameras have a larger range of sensitivity than the eye. Most stars put out a continuum of wavelengths with a number of absorption lines superimposed on it. Most emission nebula like the Orion Nebula produce a spectrum this is composed of a few bright emission lines, such as H-alpha (a hydrogen line at 6563 angstroms), H-beta (a hydrogen line at 4861 angstroms), and O-III (a triply ionized oxygen line at 5007 angstroms). An angstrom is one ten billionth of a meter. You will also quite often see wavelengths written in nanometers, which is one billionth of a meter. 6563 angstroms (A) is 656.3 nanometers (nm). Galaxies have a spectrum that is an aggregate of many stars, and have a similar spectrum. Most galaxies only have a few obvious features – the cores tend to show a sodium absorption line due to the older stars there. Seyfert galaxies and other active galaxies show an excess of H-alpha, which is great since it makes a red shift much easier to determine. Quasars, nova and supernova in general exhibit strong 6563 emission. In the case of quasars it can be red shifted quite a bit, hundreds of angstroms, so it may actually appear at a different wavelength. For a nova, the line will only be shifted slightly since the star is in our own galaxy, but it may be greatly broadened. The individual hydrogen atoms are moving very fast due to the tremendous temperatures involved, producing Doppler broadening that smears out the line.

Stars can be classified spectrally into the well know OBAFGKM groups. The very hot stars have few features in their spectrum, perhaps only a few hydrogen lines. The spectrum of Vega shown later illustrates this. The cool stars tend to be old, with many metallic lines producing a very complex and structured spectrum. There are also several types of peculiar stars, which show strong emission lines or other structure. The DSS-7 can reveal these features.

Optical Design: the optical design of the DSS-7 is illustrated in Figure One. Light enters the spectrograph through an entrance slit and is folded and then collimated (made parallel) by the collimation lens. The light then impinges upon a diffraction grating, which causes different colors to be reflected at different angles. You can see a similar effect in the light reflected from a CD or DVD. The light diffracted from the grating is then collected by a focusing lens, and imaged onto the CCD. Light of a discrete wavelength through the slit will be imaged into a vertical line. If the light does not fill the slit (such as is the case with a star) the discrete wavelength will produce a starlike point on the CCD, with different wavelengths spread out along a line. This is illustrated by the next few figures.

Figure Two shows the DSS-7 entrance slit. The narrow (50 micron) slit in the center is flanked by a wider slit above (100 micron), and an even wider slit below (200 micron). 400 micron slits lie at the extreme top and bottom of the pattern. 50 microns is about 0.002 inch, so the slits are quite narrow.

Figure Three shows the spectrum collected when this slit is illuminated by hydrogen light – the two major wavelengths, 6563 and 4861 angstroms, produce two images of the slit displaced horizontally.

Figure Four shows a spectrum collected while examining P Cygni, a peculiar star with permanent emission lines. The broadband radiation from the star produces a horizontal line, while the emission lines show up as bright points, and the airglow lines (some natural, some light pollution) show up as copies of the slit pattern. For this image the airglow lines have been exaggerated to illustrate them better – P Cygni is bright enough that exposures are short and airglow is not so prominent.

An obvious question at this point is how does one put the star onto a 50 micron slit? Figure Five illustrates the mechanics of the DSS-7.

The orientation shown here matches that of the optical schematic. The light enters from the left. Both the slit and diffraction grating are motorized. The slit is motorized such that it can be flipped in or out under computer control. The grating is motorized such that it can be driven between the first order (produces a spectrum) and the zeroeth order (produces an image) position. In the zeroeth order position the grating acts like a mirror, producing an image of the slit, if it is in place, on the CCD. To put a star on the slit, the user commands the slit out of the way, and the grating to the zeroeth order, and sets the camera to FOCUS mode. He then uses his telescope controls to move the star to a software generated line marking the slit position, and then clicks the CAPTURE SPECTRUM button. The computer immediately inserts the slit, flips the grating to first order, and starts the exposure.

Figure Six illustrates the image quality in the zeroeth order position of the grating. The image is good.

The DSS-7 is designed to accept an F/10 cone of light, a value typical of popular commercial Schmidt-Cassegrain telescopes. In the imaging mode, it acts like a 2:1 focal reducer, increasing the field of view of the CCD. It also is effectively a 2:1 focal reducer in spectrograph mode, increasing the sensitivity to extended objects like nebulas or galaxies. It will accept the center portion of the cone of light from a faster telescope, but light is lost around the edges of the collimator lens.

The small DC motors in the DSS-7 are powered by a 9 volt battery. The motors are controlled by signals from the CCD camera’s relay port through a phone jack connector. There is no provision for guiding. The length of exposure one can take will be limited by your telescope’s ability to track unguided unless you have another camera set up to work as a guider. For stellar work, it is not easy to keep the star on the narrowest slit. For diffuse objects it is much easier since a little motion still usually leaves some nebulosity passing through the slit. Reasonable spectra of stars as faint as 9th magnitude can be achieved in 30 seconds with an eight inch (20 cm) aperture telescope. Putting the star in one of the wider slits helps, but will yield some blurring of the spectrum. The 100 and 200 micron slits are included mainly for diffuse object observations.

Comparison to Slitless Spectrographs: the inclusion of an entrance slit in this design allows the user to obtain good spectra of extended objects, a measurement that was impossible with low cost slitless spectrographs using transmission gratings. The other advantage of the slit is the sky background is both resolved spectrally and reduced considerably, improving the signal to noise ratio for faint objects. With slitless spectrographs, guiding errors blur the spectrum. For the DSS-7, guiding errors cause the object to move away from the slit and light is lost.

Analysis Software: SBIG has modified the SPECTRA software originally developed for the SGS to make is simple to use with the DSS-7. The software allows the user to easily perform a wavelength calibration on collected data, and save the result as a text file that can be read by Microsoft Excel. Software features include the ability to subtract the sky background from stellar data, and display modes that smooth or colorize the data to aid in visualization of the spectra in a traditional manner.

Observations: We have used a prototype DSS-7 to measure a number of objects, which illustrate its capabilities. In Figure Seven we show the image of Vega’s spectrum, as well as a spectral curve. Vega is a good test star since its brightness is accurately calibrated. We used an E-finder to measure it (a very simple well defined optical system!), and also measured atmospheric transmission at the same time. This data enables the absolute sensitivity of the CCD-Spectrograph combination to be calculated. We have done this and the result is shown in Figure Eight. The accuracy of this data is probably +/-10%.

Figure Nine shows spectra of P Cygni and a Wolf Rayet star, WR135. Note that both have emission line structure completely unlike Vega, and that WR135 has more lines and broader lines. The wider spectral width is real, and is indicative of higher temperatures in its stellar atmosphere.

Figure Ten shows spectra of M57, the Ring Nebula. The slit lay vertically across the ring. Note the obvious H-alpha, H-beta and OIII emission lines. One can also tell from the spectral image that the center of the ring has different proportions of these elements than the edges.

Figure Eleven compares the spectrum of the zenith sky from my light polluted backyard with a much darker site. The natural airglow line at 5577 angstroms is pretty insignificant compared with the mercury and sodium vapor lines from my backyard. The broad hump across the spectrum is integrated starlight. Each airglow measurement is 15 minutes worth of exposure. We have been tempted to make an aurora detector by just monitoring this oxygen emission line – if enough users are interested we will write a program to do this. One advantage of some light pollution is one can use these lines as calibration lines for longer exposures!

Figure Twelve shows some data that is hard to come by. In this figure we have shown the absolute brightness of the dark site compared with a suburban site on a night with a full moon – CCD quantum efficiency and spectrograph throughput have been removed, so this graph is in absolute photon units. By the way, the sky is BLUE on a night with a full moon, even though it looks gray to the eye! Your eyes cannot perceive color under dim light conditions. Dark skies in light polluted areas are greenish yellow.

image030a.gif (8576 bytes)

Figure Twelve: Absolute Bright and Dark Sky Brightness

Figure Thirteen illustrates the red shift of NGC 7603, a 14^th magnitude Seyfert galaxy in the Virgo cluster. The red shift of the H-alpha line at 6563 angstroms is obvious, and is about 190 angstroms, an easily measurable amount (35 pixels). This required three 15 minute exposures using a Celestron 8 guided by an STV. Some residual artifacts from subtraction of the light pollution lines remain between 5400 and 6000 angstroms. This galaxy is interesting since the relative brightness of the 6563 emission relative to the continuum has increased three-fold since the 1970’s, a change an amateur can now track!

A final use of the DSS-7 spectrograph is illustrated in Figure Fourteen. By taking a spectrum of a stable light source or star with and without a filter in place and taking a ratio one can measure the spectral characteristics of the filters. This composite curve illustrates data we used to characterize narrow band filters as part of an absolute calibration of the spectrograph.

In summary, the DSS-7 will provide a powerful new capability to the amateur, revealing the complexity found in the third dimension of starlight, the spectra. We here at SBIG have been amazed at the complexity of structure found in many stars and objects. For example, note the zero order image of Nova Scorpii-2 shown in Figure Six. In that image it is a completely unremarkable star in the field. Its spectra, shown in Figure Fifteen, reveals a huge H-alpha line obviously broadened by the power and heat of the explosion. Visually, this star is probably a cherry red color!

	DSS	SGS
Input F/number	F/10	F/6.3 x F/10
Dispersion	5.4 Angstroms /pixel	High Res = 1.07 Angstroms / pixel Low Res = 9 Angstroms / pixel
Resolution with 9u pixels (ST-7)	15 Angstroms	Higth Res = 2.4 Angstroms Low Res = 9 Angstroms
Spectral Range (ST-7)	4130 Angstroms	High Res = 820 Angstroms Low Res= 3290 Angstroms
Projected width of narrowest slit on CCD	25 microns	18 microns
Blur perpendicular to slit	~ 25 microns	~ 100 microns
Lower resolution slit choices	50, 100 and 200 mcirons	72 microns
Ideal for measuring	Extended Objects	Stars
Relative Sensitivity for dim extended objects near H-alpha	5 - 10X	1X
Dimensions (excluding connectors)	2.2 x 4 x 4.3 in.	3 x 4 x 7 in
Weight (excluding camera)	1.5 lb.	1.5 lb.