A Fabry-Perot spectrometer for auroral observations
Abstract
Spectrometers utilizing prisms or gratings have been in use
for a number of years. Interferometers such as the Fabry-Perot
and Michelson are often considered as being extremely delicate
and specialized instruments, not the kind of instrument one would
choose for routine observation of spectra. Recently, however,
special properties of the Fabry-Perot have been recognized and a
very useful and versatile spectrometer designed, which puts the
Fabry-Perot ahead of other spectrometers in many respects.
The Fabry-Perot was invented in the nineteenth century, and
used almost immediately in wavelength measurements and standardization
of the meter. The Fabry-Perot interferometer has also
been used a great deal for hyperfine structure studies. It has
always been recognized as being a useful instrument because of
its properties of being able to convert directly from wavelength
of light to a standard length and the fact that its theoretical
resolution has no limit. The Fabry-Perot is a multiple beam
apparatus that splits the beam up by successive reflections
between two plates, at each reflection allowing some of the beams
to pass through, thus making many beams interfere. The result is
that the interference fringes are very sharp.
The method by which a Fabry-Perot has been and still is often
used to measure wavelengths and study fine structure is what will
be called here the photographic method. In this method an objective
lens focuses the Fabry-Perot fringes, which are at infinity,
onto a photographic plate. When looking at a monochromatic source
the image consists of a number of concentric circles which become
closer together as one proceeds from the center. The Fabry-Perot
is often crossed with a prism or grating spectrometer, because the
free spectral range is small, when something other than a nearly
monochromatic source is being used. These fringe patterns are
quite difficult to analyse but for many years it was the best
method to get such high resolution, and it is still used.
The photoelectric method of recording Fabry-Perot fringes was
developed by Jacquinot and Dufour (1949). The method used is to
replace the photographic emulsion at the focus of the objective
lens with a diaphragm that allows a portion of the pattern to go
through, be collected by a lens and be focused on a photomultiplier
tube. If operated in this manner it is possible, by changing the
optical path length between the plates, to make the instrument
scan its pass bands linearly with time over a wavelength or wavenumber
interval; thus it is comparable to other spectrometers.
The recorded output is the spectrum, unlike the photographic
method where considerable analysis is required to get a spectrum.
Other advantages of a Fabry-Perot spectrometer compared to a
photographic Fabry-Perot are discussed below. A spectrum. can be
obtained in a shorter period due to higher sensitivity of the
photomultiplier which is especially" important in aurora because
of the short life of some forms that have interesting spectral
characteristics. The developing of plates and subsequent reduction
to intensities by microphotometer tracings is eliminated,
which is very important. This means the non linearities in
emulsions are overcome, the time required to get data is reduced,
and there is an increase in accuracy due to direct photoelectric
methods. The objective lens does not have to be of high quality
since only near-axial rays are used, and a high f/number to get
high speed is not required (in fact at high resolution a lens of
long focal length is advantageous).
An important advantage that any spectrometer has over a
spectrograph is that it is possible to see the results as the
observations are being taken, and thus observing time can be used
more efficiently. If rapidly moving sources are being studied
perhaps a photographic method has some advantages, but if the
spectrum is changing rapidly this advantage is lost.
The great advantage in light gathering power for a given
resolving power for a Fabry-Perot spectrometer, over prism and
grating spectrometers, is clearly pointed out by Jacquinot (1954).
He compares instruments of the same effective area and resolving
power and finds the grating spectrometer always will have greater
light gathering power than a prism spectrometer, and a Fabry-Perot
spectrometer will have as much as 30 to 400 times the light gathering
power as a grating spectrometer. This shows why the Fabry-Perot
can be superior even for low resolution where good light
gathering power may be important. A Fabry-Perot has further
advantages over a grating spectrograph in that one set of plates
can be used for any wavelength region in which they will transmit,
and the resolution can be adjusted to any value. This is not
true for a grating spectrograph since a particular grating is
designed to be most efficient in one spectral region and at one
resolution. This is why the Fabry-Perot spectrometer is a very
versatile instrument.
In upper atmospheric observations until very recently, the
Fabry-Perot has been limited to measuring wavelengths. The
classic examples using the photographic method are the exact
measurement of wavelength of the auroral green line by Babcock
(1923), and the confirmation of the sodium D lines in the twilight
by Bernard (1938). Others were the study of both 5577A and 6300A
by Vegard (1937) and a study of all of these by Dufay, Cabannes
and Gauzit (1942).
Babcock (1923) realized that the Fabry-Perot could be used
for temperature measurement even though it was not known at the
time what was producing the oxygen 5577A auroral and night airglow
green line. The method of obtaining temperature is called
the Doppler method. The profile of the spectral line is a pure
Gaussian and results from the Doppler shift arising from the
the atoms random motion in thermal equilibrium. The half-width
s, and absolute temperature T are related by s = 7.16 X 10-7 σ √T/M
where σ is the wavenumber or the line in cm-1, and M is the
molecular weight. To give good results when using the Doppler
method a line should have narrow natural half-width, due to
internal structure and transition probability. To fill the second
need a forbidden transition is best because of its long time in
the excited state. Another important consequence of the forbidden
transition is that the atoms spend a sufficiently long time in the
excited state to ensure that they will reach thermal equilibrium..
The oxygen 5577A [OI] fills these requirements since it is forbidden,
having life time of about 0.7 seconds in the upper state,
and is not known to have any internal structure that would broaden
the line compared to the Doppler width. Another important requirement
is that it be bright enough so that enough luminosity can be
obtained to use high resolution. The oxygen green line is one of
the most prominent features in the aurora. Babcock assigned an
upper limit to the half-width of the green line of 0.035 A.
Vegard (1937) tried to get a temperature for the oxygen red line
(6300A), but could not obtain enough resolution with the luminosity
required.
Because of this there was a time lapse of almost twenty years
where the idea of getting temperatures by this method was given
up. With the advent of dielectric multilayers, the absorption
in the reflecting layers was greatly reduced and interest was
again aroused. Wark and Stone (1955), Wark (1956), (1960), and
Cabannes and Duray (1955), (1956a), (1956b), resumed photographic
work on the aurora and night airglow using a Fabry-Perot, and
found it possible to obtain temperatures in this way.
When the Fabry-Perot interferometer was investigated by
Jacquinot (1954) interest was aroused in this method and some
Doppler temperatures were tried again. The reason for the difficulty
experienced by these many observers is the high resolution
required. Using the half-width obtained by Babcock (1923), and
assuming the instrument would need a passband at least as narrow,
the minimum resolution required becomes 5577A/0.023A = 2.4 X 10⁵.
However, with the increase in sensitivity of photomultipliers and of
course dielectric layers, it was thought worth while to try to
get Doppler temperatures. Armstrong (1956), (1959) used a Fabry-
Perot spectrograph to try and measure the Doppler temperature of
the 5577A line in night airglow and a bit of aurora and he obtained
some preliminary results. Karandikar (1956), (1959) has an instrument
capable of measuring Doppler temperatures, but at the time of
writing his results are not known since he had not the opportunity
to observe aurora until recently.
It is considered that Doppler half-width measurements of the
auroral green line, 5577A, would. give the most reliable spectroscopic
temperatures of those available, since the interpretation
is beyond question. The other methods of obtaining temperatures
are vibration and rotation. Vibrational temperatures are not
very reliable because, for temperatures below 1000°K the excitation
process dominates and largely determines the population of
the upper states, rather than temperature. Rotational temperatures
give reasonable values but the interpretation depends upon
the excitation process. Thus Doppler temperatures should be the
best and the oxygen green line should give very reliable results.
The reason that more temperatures have not been obtained by the
Doppler method is that high resolution is required that cuts down
the intensity since the product of luminosity and resolution is a
constant for spectrometers.
Because of the obvious suitability for auroral studies, it
seemed desirable to build a Fabry-Perot spectrometer at the
University of Saskatchewan. Of the possible studies, the most
worth while appeared to be a high resolution instrument that would
be capable of observing aurora for the oxygen green line with
Doppler temperatures in mind and Which would be easily convertable
to other studies. Other observers have had little chance to
observe aurora Where the gain in intensity over night airglow is
great. A pair of 4 inch diameter quartz plates were ordered from
Hilger and Watts before the author undertook this thesis. They
arrived in the winter of 1959 and this study was begun in the
spring. A Fabry-Perot interferometer was to be built with
the first purpose in mind to study the auroral green line for
temperatures.
The next chapter consists of same theory required for subsequent
chapters but it is not intended to be complete. Next in
Chapter III a full description of the apparatus is given. This
is described in detail in most parts because the design is thought
to be original and appears very satisfactory. In Chapter IV the
contours of the plates are determined. This is important because
of the great effect the defects of the plates have on the way in
which the plates are to be used and on the end result. The next
chapter contains some results on auroral observations which are of
a preliminary nature, but thought to be worth while. The final
chapter gives an estimate of the value of
the results and a few ideas on what may be done next with the
four-inch Fabry-Perot interferometer.