Spectroscopy – "looking at the spectrum"
Spectroscopy literally means, "looking at the spectrum". The spectrum we talk about here arises in the interaction of electromagnetic radiation with the substance under investigation. It describes the strength of the interaction as a function of the wavelength, frequency, wavenumber or photon energy of the radiation.
There are a few simple principles that govern all types of interaction processes
These principles make electromagnetic radiation (EMR for short) a powerful probe for investigating the properties of the system at hand. It has special advantages, well known but still worth repeating:
- The energy distribution of the probe radiation is a mirror image of the energy states of the sample. This makes it possible to perform analysis of the sample by analysing the radiation emitted, absorbed or scattered by it
- EMR does not affect the composition or chemical state of the sample
- EMR can be guided by various optical or other types of components, making remote sensing possible. Concentrating the radiation makes it possible to analyse samples even down to the size of the wavelength of the radiation
- The interaction process is instantaneous. Identification of components in the sample can be performed continually and in real time
Spectroscopy is considered here as recording the effect of the interaction on the radiation as a function of wavelength or wavenumber. The main component of the spectrometer is some kind of dispersive element for separating the different wavelengths in the radiation from each other. The radiation is let into the spectrometer through a slit or circular aperture, depending on the principle of operation.
Electronic energy states are usually related to radiation in the ultraviolet and visible region and the energy states of vibration and rotation of molecules to radiation in the infrared region. The corresponding branches of spectroscopy are called UV-VIS spectroscopy and IR spectroscopy, for short. These are, to be sure, if important, just two areas of spectroscopy among a proliferating multitude. They are chosen here for their growing importance for environmental analysis and for their applicability for real-time investigations.
Developments in UV-VIS spectroscopy
UV-VIS spectrometers of the 50s and 60s were bulky, ruling out the possibility for field use. Earlier designs used prisms for dispersing the radiation according to wavelength, but in the 1960s the grating manufacturing techniques were developed so that commercially competitive gratings became available and, eventually, replaced the prism. The radiation was directed with separate mirrors onto a narrow exit slit. Like the prism, the plane grating had to be combined with a separate mirror or separate mirrors to accomplish the necessary optical imaging of the entrance slit onto the exit slit. The gratings of those days were produced by mechanically ruling a glass substrate, which then was covered with a reflecting layer. As an alternative, the master grating was used as a mould, from which the final gratings were made.
An important step of developments in UV-VIS spectroscopy was the introduction of holographically manufactured concave gratings in the 1970s. The concave backing performs the imaging directly from the entrance slit to the exit slit. The dispersive function is performed by the grating grooves as in a plane grating. The grooves are made by letting the two components of a split laser beam interfere on a spherical substrate, which is covered with a thin layer of photoresist. On developing the photoresist grooves are formed in the layer.
The principle of the holographic grating was put forward already in the 1890s by the French physicist Alfred Cornu. He proposed essentially the presently used technique, but it could not be implemented for the lack of sufficiently intense monochromatic light. The first holographic gratings were produced on a laboratory scale in the 1960s, and the first commercial ones appeared shortly after 1970. Although quite some time has gone by since then, there still is a lot of elbowroom in the marketplace; there are only a few serious manufacturers around in the field.
Ruled gratings are not out!
Although the holographic grating revolutionised UV-VIS spectroscopy, conventionally ruled plane gratings still find their applications. In the 70s and 80s, the Finnish physicist Dr. Peter Lindblom developed a new type of grating spectrometer using very coarse ruled plane gratings. These gratings were so-called echelle gratings, looking like a staircase with plane steps. Such a grating made it possible to divide the wavelength range into several smaller ranges that could be spread out with a prism across a large matrix detector, much like the bandspread scale on a shortwave radio receiver. The new design concept was called MEGA, Multi Echelle Grating Arrangement. It employed four 160-mm wide gratings with 31,6 grooves/mm arranged so that each grating dispersed the radiation directly onto the next grating. The idea behind the MEGA was that the same resolving power is obtained with a spectrometer a quarter of the size of a corresponding spectrometer using one grating. Further, the width of the gratings was one fourth of that required if only one grating was used.
The MEGA did not enjoy commercial success, but the principles enunciated by Dr Lindblom set the path for a whole new generation of small high-resolution grating spectrometers.
Large detectors set the pace
Large detectors arranged in the shape of matrices were known to spectroscopists already from television. Numerous research groups developed spectroscopic instruments employing matrix or line detectors. Matrix detectors were used for building imaging spectrometers, where each horizontal row of pixels contains a complete spectrum.
Aberration-corrected spherical holographic gratings made it possible to develop high-speed spectrometers with only one optical component, i.e. a concave grating. These spectrometers are now abundant in the field. Their optical capabilities are excellent and the mechanical stability adequate for fieldwork as there are no moving parts.
Infrared spectroscopy revived
Infrared spectroscopy is used alongside with UV-VIS spectroscopy for detecting organic compounds because of the high sensitivity and specificity of identification.
Infrared spectrometers have been built since 1908, but the first commercial spectrometers appeared only after 1945. They used a dispersing prism. Soon after that, coarse diffraction gratings were introduced. Recording a spectrum was done by turning the prism or grating. In 1887 Albert Michelson in his famous investigation together with Edward Morley built an interferometer for investigating the dependence of the speed of light on the direction of motion of the Earth. Michelson discovered that his instrument could be used for high-accuracy analysis of the structure of spectral lines. The light beam in the Michelson interferometer is collimated into a parallel beam and led through a beamsplitter, which divides it into two beams. One beam propagates towards a fixed mirror and the other towards a movable mirror. Both beams are reflected back towards the beamsplitter again. One of these beams is partly reflected and the other is partly transmitted, and these beams combine again.
According to Fourier's theorem any finite mathematical function can be expressed as a sum of sine and cosine functions. The interferogram is such a function, and the amplitudes of the cosine waves of which it is made up constitute the spectrum of the radiation. As the Fourier transform is the basis for this type of infrared spectroscopy, it is called Fourier Transform Infrared Spectroscopy, or FTIR for short.
The first working FTIR spectrometers were built in the 1950s. They used plane mirrors. The mirror alignment was extremely critical in order to ensure interference, and implementing the stable motion of the movable mirror proved very difficult.
The FTIR spectrometer has revolutionised infrared spectroscopy and the interferometer has replaced the diffraction grating almost completely in that field. Further, the price level has come down to the point where acquiring a reliable analytical instrument no longer is a major economic setback.
The main advantages of an FT instrument run as follows:
1) The Fellgett advantage. All wavelengths are recorded simultaneously, increasing the signal-to-noise ratio. This advantage pertains explicitly to grating instruments where parallel recording like with a line detector is not used. This is in general the case in the IR region, where decently priced line detectors are yet to appear, let alone matrices
2) The Jacquinot advantage. In a grating instrument, the incident radiation enters through a narrow entrance slit. That limits the available energy on the grating, because the linewidth is governed by the slit width. In an interferometer, the entrance aperture can be of a much larger area than in a grating spectrometer entrance slit for the same linewidth. This increases the sensitivity of the instrument considerably over that of a grating instrument
3) Calibration of the interferometer is done alongside with measuring the spectrum using a He-Ne laser as reference. Calibration is as accurate as the frequency of the laser, i. e. better than 0,001 cm-1. The calibration is transferable between spectra measured with different instruments as the calibration is performed for each spectrum
The smallest linewidth ?? obtainable with a grating spectrometer with grating width L is of the order of 1/L expressed in units of wavenumber. With an interferometer, the linewidth obtained with moving mirror path D is of the order of 1/2D, i. e. one half the linewidth obtainable with a grating of equal width. To be precise, the quantity determining the linewidth is the optical path difference OPD between the two beams. Expressed using OPD the linewidth is of the order of 1/OPD. For one moving mirror OPD is equal to twice the amplitude of the motion. Mounting the mirrors back to back on the moving mechanism increases the OPD by an additional factor of two as compared to the motion amplitude. Adding further mirrors to the optical path additionally increases the OPD. The large FTIR spectrometer in the University of Turku in Finland has an OPD of 16 m using a mirror movement of +/- 2 m. This corresponds to a linewidth of 0,0004 cm-1.
FTIR goes field-proven
Stringent emission regulations and the advantages of the FTIR technique made it a natural goal during the 90s to build field-proven FTIR instruments capable of performing environmental and process analyses in real time.
Another obstacle was that the necessary computer resources were not readily packed into an instrument of portable size and weight. After 1990, processor speeds have increased to the point where both the control program for the interferometer and the calculation program for evaluating spectra could be combined in a computer of portable size and weight.
Here it is in order to point out the enormous advantage of replacing the interferometer plane mirrors with cube corner retroreflectors. A cube corner reflects incident light back exactly in the opposite direction. Needless to say, this opens up completely new possibilities for the design of the mirror movement mechanism – as most contemporary manufacturers have noticed.
From present emission regulations, it is obvious that the sensitivity of the instrument had to be as high as reasonably possible. The whole optics setup from the radiation source to the detector mirror has been carefully optimised over a period of several years. A few percent here and a few there finally amounted to more than a twofold increase in signal amplitude from an already pretty efficient original design.
In a story like this, the author is supposed to deliver his or her views on future trends. This author believes, in a very subjective way, that important development will take place along these lines:
- Line and matrix IR detectors will be available. This could mean a renaissance for IR grating spectroscopy. With a matrix, spectral imaging in the IR region would be possible
- The NIR region will be accessed profitably using FTIR techniques
- Detectors will develop towards lower wavenumbers. Much of the fingerprint region down to 600 cm-1 is now inaccessible with semiconductor detectors without LN2 cooling
- Data analysis software is going to be much more powerful. Higher data throughput makes it possible to extract concentrations in much more detail than now, making the analysis more flexible
- Developments in mechanical and optical design make it possible to access aggressive and difficult new environments
Published: 10th Dec 2005 in AWE International