Once upon a time spectroscopy was the realm of professionals, with spectroscopes costing several thousand pounds. However its now possible to buy diffraction gratings like the SA100 or SA200 which can be fitted to ordinary astronomical cameras or DSLRs and which are great for starting out in the field.
Why do Spectroscopy of Meteors?
The spectrum of an astronomical object can provide clues as to its composition. If you look closely at the spectrum you will see areas of brightness – spectral lines. The positions of these allow you to identify their wavelengths and thus you can identify the chemicals or elements responsible. The strength of the lines also give an indication as to the relative amounts (or absences) of each.
How do you create a Spectrum?
To obtain a spectrum, you pass the light from the object through a slit, prism or diffraction grating. A Slit is ideal but expensive to set up. A prism is cheaper, but a diffraction grating will produce a more linear dispersion of the spectrum and thus will make it easier to identify the individual spectral lines.
A diffraction grating produces whats called a first order spectrum to each side of the image of the meteor itself (known as the zero order). When imaging meteors – or indeed stars – your aim is to capture the zero and first order spectrum. To make this easier, most diffraction gratings are ‘blazed’ which means they direct most light into the first order spectrum on one side of the image which is consquently brighter.
However capturing meteor spectra always involves a compromise. The more that you disperse the lines, the easier it is to differentiate between lines of different wavelengths but the light is spread out more and fainter. Hence the dispersion from a 600 lines/mm grating will be greater than that from a 300 lines/mm grating meaning that the spectrum will be fainter and longer and part or all of the spectrum may fall outside of the camera’s field of view. As we said earlier its just luck whether the spectrum is all on the image!
Back to gratings: There are several gratings available on the market but its worth mentioning two : the SA100 and SA200, used by many amateurs, are blazed gratings mounted in a standard 1.25″ eyepiece filter which have 100 or 200 lines/mm. With a suitable adapter (like this one, or you can make your own with an old lens cap) they can also be fitted on the front of the lens of DSLRs. My own tests with the SA200 have shown that a spectrum fits nicely across the field of view of a basic DSLR with a 50mm lens.
Capturing Meteor Spectra
For meteor spectra, there is an obvious question. You don’t know where (or when) the meteor is going to appear, so in which direction should you point your camera?
All that you can do is to follow the same guidelines as for imaging of meteors in general – that is to say, point the camera about 40-45 degrees away from the shower radiant and about half way up from the horizon. If no meteor shower is active, then you would still point the camera of around 50 degrees altitude but in the direction of best darkness. Then set the camera to take 20-30s exposures, set it on continuous – and hope!
What to Expect
A typical meteor spectrum image is shown here. The meteor itself (ie zero order) is outside of the field of view and was moving roughly horizontally. So the individual spectral lines are aligned parallel to that path, with the spectrum stretching from centre top to bottom left.
Meteor spectra will typically consist of a number of bright emission lines as you can see here. Although spectra sometimes show an indication of a continuous background spectrum, there is debate as to whether this is a genuine feature or merely the combined result of many low intensity unresolved spectral lines.
In the above case, most of the spectrum was captured within the field of view. However as mentioned earlier this is often not the case, as the next image shows. The bright line across the centre of the image is the zero order image of the fireball. You can see one bright flare and one broader flare in brightness in the path of the fireball (ignore the bright diagonal line across the bottom right which is a moonlight artefact).
The first order spectra can be seen to the top right and bottom left – but as can be seen, large parts have fallen outside of the camera’s field of view. This is frustrating !
Processing a Meteor Spectrum
Generally once you capture the spectrum the meteor will be at one angle, the spectrum at another, and possibly incomplete. To process it you must first rotate the image so that the spectrum lies horizontally across the screen with the meteor/zero order towards the left. This is shown in the image below which has been rotated nearly upside down to allow the largest part of the spectrum to be chosen.
Next you should crop the image to remove everything except the zero order and spectrum. This is to make it easier to process in spectral analysis software like RSpec or VSpec. With a meteor spectrum you may want to go further and crop a set of long thin stripes from the spectrum at different points. This will allow a sharper spectrum to be created with less overlap between lines. The images below show the sort of thing you’re aiming for.
Spectral Analysis
Before you can analyse a spectrum you must obtain something to calibrate it with. Calibration means working out what wavelength each point on the spectrum corresponds to. The simplest way to do this is to take the spectrum of a bright star whose spectral class we know (and hence whose spectral lines are at known wavelengths). It is important that the star is imaged using the same camera with the same settings. This is so that the calibration is accurate. To be even more accurate, choose a star in the same part of the sky as you are looking for meteors and image it on the same night. Its therefore best to capture this calibration spectrum before starting the meteor hunt. You can also calibrate using a lamp or other source with known wavelengths such as an HPS street lamp, or even using common features of spectra such as infrared lines from water vapour in the atmosphere.
You can now load the thin strip you created earlier into suitable analysis software such as RSpec or VSpec (see links below), and calibrate it with the calibration data. A full explanation of how to use these packages is beyond the scope of this article, but the authors of both have provided extensive documentation that applies as much to meteors as to stars or even fireworks. There are also active spectroscopy groups on Groups.io which are full of advice.
An uncalibrated spectrum is interesting but doesn’t tell us much scientifically. Here’s the spectrum of the thin strip shown above. The problem is that we don’t know what wavelength each spike corresponds to. So we really want that calibration !
A calibrated and analysed spectrum is shown in the next image. The shorter wavelength (blue) part of the spectrum fell outside of the field of view and this shows the spectrum from the green to the near infrared. There is a bright green emission line from magnesium, along with a yellow emission line from sodium. Many meteors also show an emission line in the red from calcium, but this is not clearly seen here. The lines in the infrared are mostly related to the atmospheric gases with which the meteoric particle collided – the presence of these atmospheric lines in all spectra helps calibrate the wavelengths in the remainder of the spectrum.
People sometimes express surprise that some bright meteors appear green to the naked eye. This is probably because they don’t see green stars. However, whereas stars emit a continuous spectrum of light, the sum of which adds up to being close to white, meteors only emit light at specific wavelengths and thus if the spectrum is dominated by a strong green emission line from magnesium, this can cause the meteor show a strong green colour.
Interpreting the results
In principle, we could determine the composition of comets by inspecting the spectra of their associated meteor showers. In practice though, much of the volatile material from the comet debris will have boiled away by the time we detect it. However, do differences in spectra between meteor showers give clues as to differences in the non-volatile composition of their parent comets?
The answer is maybe! Although a number of articles have been written about the interpretation of meteor spectra, it is often not clear as to whether the statements included are based on theory, based on a general inspection of meteor spectra, or are based on detailed analyses of the spectra. This leaves us with a number of factors, listed below, that may affect the relative strengths of the different emission lines seen in the spectra and it is not immediately clear as to which are the most significant.
– the composition of the meteoric particle itself
– the size of the meteoric particle
– the speed at which it hit the Earth’s upper atmosphere
– the altitude at which the particle was vapourised
– local variations in the composition of the parent comet or asteroid
The above factors are not all independent of each other – faster, larger particles are likely to reach lower altitudes than are slower small particles.
Here are three examples, secured by Bill Ward during 2017
These colourised versions were generated by taking the intensity measurements collected and using the VSpec software to produce a colourised version of the spectrum.
How much do spectra differ between meteor showers?
From spectra captured during 2014, Bill Ward was able to compare the Perseids and Geminids,with three non-shower meteors.
As can be seen the relative strengths of the green (magnesium), yellow (sodium) and red (calcium) emission lines vary between the first four spectra.
The first and last spectra are from sporadic meteors. Whereas meteor showers are mostly related to (“icy”) comets (the Geminids being an exception), sporadic meteors are sometimes related to defunct meteor showers and sometimes asteroidal (“rocky”) in origin.
How much does the spectrum vary between meteors within a given meteor shower?
This can be tricky to answer as it requires the capture of a good number of quality spectra … and there is always the risk that the data may be “contaminated” by a sporadic meteor whose path just happened to line up with the shower’s radiant.
However, the accompanying intensity plot compares the spectra of seven bright Perseid captured by Bill at Perseid maximum in 2015 and it does suggest strong similarities.
The two bright emission lines to the left are related to magnesium and sodium.
The broader band near the centre is related to oxygen and nitrogen.
The bright emission line to the right is due to oxygen and is actually located in the near infrared rather than in the visible spectrum.
How much does the spectrum change during the passage of a meteor?
DSLR images of Perseids, such as the one shown here that was captured by Richard Fleet, often appear green at the start, whiter in the middle and red at the end. This effect is not seen in most other meteor showers however.
The reason that Perseids are different is that they are fast meteors (about 62km/s) and so carry more kinetic energy and therefore are likely to start emitting light at higher altitudes than do most other showers.
The colour changes can be understood by measuring that relative intensities of the emission lines at each of the stages.
This accompanying spectrum by Bill Ward (for a different Perseid) shows how the intensities of various emission lines changed during the passage (from top to bottom) of the meteor.
The oxygen line to the far right is in the infrared and can be ignored as it doesn’t affect the visible spectrum.
Note that the first line to appear in the visible spectrum is the green 557.7nm line from oxygen. This emission line can only be produced at high altitudes (above 110km). Hence the trail in the DSLR image is initially green.
Soon afterwards, more intense emission lines from magnesium and sodium appear and these swamp the oxygen line (while still above 110km). The combination of the yellow Na and green Mg produce a whiter colour.
Finally, if the particle is large enough to survive lower into the atmosphere, the broad red band from oxygen and nitrogen can dominate, causing the end of the meteor to appear reddish.
Note however, that the colours seen with the naked eye may be different as the eye is only sensitive to colour above certain intensities. Hence the initial green colour will usually not be visible to the naked eye.
Historically, the vast majority of spectra captured were from the Perseid and Geminid meteor showers – largely because people concentrated their imaging at the times when the chances of capturing a spectrum were higher. In pre-digital days, observers were reluctant to devote large amounts of photographic film and associated film developing time when capture rates were likely to be very low. In recent years, however, Bill and others are increasing our knowledge by expanding the imaging to other times of the year and thus capturing spectra of meteors from other showers, such as the Taurids and the Quadrantids, and by capturing spectra of a variety of sporadic meteors.
Links to Software
IRIS: http://www.astrosurf.com/buil/us/iris/iris.htm
VSpec: http://www.astrosurf.com/vdesnoux/download.html
Rspec: https://www.rspec-astro.com/
Further Reading:
A comparison of eight Geminid 2017 meteor spectra by Bill Ward
Bill Ward’s comparison of spectra for four 2017 meteor showers