In general, monochromators have a reputation for relatively poor efficiency, but this is not always deserved. However, it inevitably remains the case that a comparable arrangement using optical filters will give lower losses. Therefore, where it is important to conserve as much of the light as possible within a given wavelength range, a monochromator is NOT a good choice. As far as fluorescence measurements are concerned, this means that monochromators are not really suitable for emission wavelength selection, where reduced efficiency results in degraded signal-to-noise performance (although the relation is only as the square root). On the other hand, reduced efficiency on the excitation side may well be tolerable - and in practice it is - because other considerations limit the maximum permissible excitation intensity.

In our filter wheel fluorescence excitation systems, for example, it is typically necessary to reduce the excitation light intensity by around 75% in order reduce fluorescent dye photobleaching. Therefore monochromators CAN give excellent performance on the excitation side of a fluorescence system, but they must nevertheless be designed with care for this application. Many monochromators are designed primarily for very high WAVELENGTH resolution, of sometimes well under 1nm. Although useful, that is not a prime requirement for a fluorescene system (although ours can give you that if you want, as we'll explain). The problem is that high resolution and high efficiency tend to conflict. Rather than efficiency per se, one normally speaks of optical throughput, which simply relates to the total optical output within the selected wavelength range, although the two measures are of course very closely related. As with any optical system, the optical quality improves as the aperture is reduced. What this means in practice is that the light entering the instrument is more nearly parallel, but that implies lower throughput, for the following reason.

A monochromator consists of two slits, at the input and exit of the instrument, plus other optical components (usually mirrors) to focus an image of the input slit onto the exit slit. In between these is the diffraction grating, which disperses the light so that the lateral position of the image of the input slit is wavelength-dependent. By rotating the grating, the wavelength range of interest can be focussed onto the exit slit. The efficiency of the grating itself is quite high, typically in the range 50-70%, but the potential throughput problem arises because only a small proportion of the available light may be able to pass through the input slit. To get as much light as possible through the input slit requires the light source to be focussed to as small a spot size as possible. However, that light is converging and diverging more strongly on either side of the image, and a faster-aperture monochromator is needed to pass the more extreme rays. The only way to reduce the beam angle at the image is to focus the light source to a larger spot, but then relatively less light can pass through the input slit.

What this means is that for greatest throughput the aperture of the light source (which can be modified by accessory lenses if required) must be matched to the aperture of the monochromator, which itself must be as fast -i.e working over as wide a beam angle - as possible. Conventional monochromators tend to have apertures no greater than about f/4, as optical aberrations tend to increase drastically if the aperture is increased further. However, by using aspheric mirrors we have been able to obtain excellent optical performance at the faster aperture of f/2.5, which not only gives greater throughput but also matches directly to our standard light source.

Even so, there is another important constraint to monochromator throughtput, which our design has also addressed. The bandwidth of a monochromator depends on the widths of the two slits, but in a rather interesting and important way. Imagine that the monochromator is being illuminated with monochromatic light, and that the focussed spot size on the input slit is always larger than the slit width. When the grating is at the appropriate angle, the image of the input slit will fall exactly on the exit slit, giving greatest throughput. As the grating rotates either side of this position, the input slit image will scan across the exit slit, to give a bandpass characteristic that is triangular, as shown in the bandpass diagram below (trace a). Note that if one slit is wider than the other - and in this case let us imagine that the exit slit is made wider (as shown in the left hand diagram) - there will be a range of grating angles over which the throughput is a constant maximum value, giving a trapezoidal bandpass characteristic which is wider overall (trace b). But what happens if we now widen the input slit to match? The bandpass characteristic is again triangular, but since more light can now pass through the input slit, the throughput at the optimum grating angle is now proportionately higher, as shown in the dashed trace in the diagram. The situation with polychromatic light is effectively the same. The only difference is that instead of seeing a clear image of the input slit on the exit slit, there is effectively an entire array of images, at different positions according to wavelength, so we see a continuous spectrum there instead. However, we can conceptually analyse the situation by summing the responses for all illumination wavelengths that can pass through the exit slit to any extent, which gives the same overall picture.


Image of narrow input slit at exit slit





Bandpass profiles





The consequence of this is very important. As the diagram shows, the bandwidth of the monochromator increases linearly as both slits are widened, but the throughput - represented by the area under each graph - increases as the SQUARE. Therefore, if we can tolerate a wider bandpass - and in many fluorescence applications we certainly can - we can effectively make the throughput problem disappear altogether! For this reason we consider a facility to vary slit widths to be invaluable, and our fast system allows this adjustment to be made on a wavelength-by-wavelength basis for still greater benefits. And furthermore, one can close the slits right down to give high wavelength resolution if required, so that detailed spectral scans can be made too.