The measurement principle of several types of OPSIS gas analysers is called DOAS, which stands for differential optical absorption spectroscopy (you can read more about “DOAS” in this blog article).
A key component of the analyser is the “spectrometer”, a unit which separates incoming light into its individual wavelengths for subsequent recording of the light spectrum. But how does a spectrometer work?
The key is an optical component called a grating. At a distance, a grating may look like a small mirror which simply reflects light. If a light beam shines onto the mirror (or the grating) at a specific angle to its normal, it is reflected at the same angle on the other side of the normal. We are all familiar with that. However, the grating has something extra: a very fine, dense and carefully repeated pattern of parallel lines on the reflecting surface. There can be several thousands of lines per millimetre.
Wavelength, Wavefront, and Diffraction
To explain what happens now, we must think of light as electromagnetic waves. Each colour is represented by a specific length of the waves – the wavelength. The more towards the red and further to the infrared, the longer wavelength. The more towards the blue and the ultraviolet, the shorter wavelength. And a wave moves in a front – the wavefront. Now, when the wavefront of a specific colour hits the grating surface, it is disturbed by the lines on the surface. Much of the light is reflected as in a mirror, but there is also a different type of reflections. The wavefront reflections form constructive interferences, and some of the light exits from the grating also in other directions. The phenomenon is called diffraction. In-between these specific directions, there is no light emitted from the specific wavefront. And here is the beauty of the grating: the exit angle of the diffraction depends on the wavelength. A blue colour exits closest to the normal mirrored reflection (aka. zeroth order diffraction), then comes the green colour in a wider angle from the zeroth order, and the red colour exits even further away from the zeroth order.
The principle of a grating.
But why complicate things? We’ve all seen a rainbow, and that colour separation (occurring in tiny droplets of water) can be mimicked in a glass prism. Why not use a straight-forward prism instead of a grating? The answer is resolution. The grating produces a much wider spread of the colours than a prism, allowing a high-resolution spectrometer to be fairly compact.
Building a Spectrometer
So far so good, but what about the rest of the spectrometer? Well, it can be as simple as a plain black box, at least black on the inside. Somewhere along one of its sides, the light enters and is directed towards the grating mounted inside the box. In a different position along one of its sides, the diffracted light from the grating exists and is picked up by a light detector. That’s it. However, as a common feature, the grating can be rotated from the outside of the box, around an axis perpendicular to the axis of the incoming light. This allows selection of which wavelength region (range of colours, wavelength window) the detector is exposed to.
Some spectrometer types are as simple as this. It is then “just” a practical matter of matching light energy budget, f-numbers, grating resolution, geometry and detector sensitivity, and we have our spectrometer. To make it more compact, we can also add some mirrors to fold the light beams. Such design is used by OPSIS in its UV-DOAS gas analysers. It’s called a Czerny-Turner spectrometer, after its inventors.
The Czerny-Turner spectrometer design, applied in OPSIS UV-DOAS gas analysers.
Once we have our spectrometer, we only also need to know a bit about molecular absorption and computing, and we have our gas analyser. But that’s a different story!