MOPITT Mission Description Document (30 October 1993)

A.1 An Introduction to Correlation Spectroscopy

A.1.1 Introduction

There is however a fundamental problem since the above implies high dispersion to separate out the fine details of the spectrum. With high dispersion comes low sensitivity and high precision requirements which are difficult to implement in a space-based instrument.

Correlation Spectroscopy (CR) offers the opportunity for high selectivity without the attendant low sensitivity and high precision requirements.

A.1.2 Correlation Spectroscopy Techniques

a) The "equivalent filter profile" is zero between the spectral lines of the gas in the cell, eliminating signals from most of the spectrum.

b) The filter profile has a maximum at each spectral line and therefore the energy from each spectral line in a broadband emission is seen simultaneously. The system is therefore very sensitive to radiation with a spectrum identical or similar to that of the gas in the cell. Evidently the spectrum of the gas itself is the best correlated with the filter profile.

c) The apparatus contains no high precision optical adjustments. Quantum mechanics keeps the spectra aligned. In fact the only phenomena which affects the alignment is Doppler shifting of the cell and the emission spectrum.

The effect of the filter is shown in Figure A.2(c) where it can be seen that the spectral emission from lines co-incident with spectral lines of the gas in the cell (even if they originate from another gas) is detected and other emission lines are suppressed.

We can express the operation of a simple correlation spectrometer by the following equation

(1)

where and refer to the monochromatic transmissions of the gas in the two CR states and R is the monochromatic input radiation. The detector is not spectrally sensitive and therefore accepts all energy at its input although in practice some filtration is applied to restrict the range of the input. The terms and represent the physical reality that the two states of the CR system do not, in general, just change the gas but also change the optical system. They represent the transmissions (broadband) of the optical system in the two states. Equation (1) may be re-written as:

(2)

where the first term now represents the ideal CR system and the second or "balance term" represents the imperfections. The first term is the result of the differencing of two large numbers, the second has a large multiplier in the integral and therefore requires the change in transmission to be as small as possible. Much of the practical design of a CR system is concerned with the minimisation or elimination of the balance term.

The Nimbus 4 Selective Chopper Radiometer (SCR) instrument used the correlation spectroscopy technique to sense the temperature of the stratosphere using the 15µm band of CO2 as the source of radiation. The CR system is shown in Figure A.3. The design attempted to balance the optical signal in both arms of the system whilst the mirror vibrated between two extremes corresponding to the two states. Any variations in transmission between the two paths appears to first order as a signal indistinguishable from that due to a "true" CR signal. Evidently this was not easy to achieve or to monitor when in orbit and other means were sought to minimise the balance term.

The Nimbus 5 SCR instrument used a similar technique but elected to move the cells instead as shown in Figure A.4. This has the advantage of leaving much of the optics common to both channels. However it can now clearly be seen that the cell windows are performing two functions - vacuum sealing and optical transmission. Materials which are ideal for one are not necessarily ideal for the other, particularly when the requirement of identical transmissions is included. For high amounts of gas in the cells, with correspondingly high signals, this technique is successful.

The Nimbus 6 Pressure Modulator Radiometer (PMR) instrument was a much more advanced CR system. The use of a pressure modulator as depicted in Figure A.5 leaves the optics completely static and moves only the gas leading to a near-perfect balance condition. At this time an additional technique was employed to overcome another deficiency and that was the pressure monitoring and stabilisation within the cell.

Molecular sieves have been known for many years as absorbers of gases. The application of the technique to CR systems consists of attaching a cell of molecular sieve material to the CR cell. The molecular sieve materials used have been of the zeolite and silicalite families. For a given overall filling pressure at room temperature, the equilibrium vapour pressure over the molecular sieve obeys a non-linear, but known, temperature/pressure relationship. The effect of the molecular sieve in its thermostatted container is therefore to stabilise and control the pressure of the gas in the CR system. This gives confidence that the pressure of gas in the CR cell can be maintained over a long time, but requires that the cell be stationary.

An additional feature of the PMR system is that as implemented in the resonant system due to Houghton and Taylor (see for example Taylor (1983)), the resonant frequency is pressure dependent and therefore the cell pressure can be monitored via the operating frequency.

The use of PMR systems was advanced with the Nimbus 7 Stratospheric And Mesospheric Sounder (SAMS) instrument to composition sounding and at present the Improved Stratospheric And Mesospheric Sounder (ISAMS) instrument for stratospheric sounding is being tested for launch on the Upper Atmosphere Research Satellite (UARS). This instrument adds an additional feature to the correlation spectroscopy system which is shown in Figure A.6. The addition of the "conventional" chopper adds to the frequency content of the signal at the detector and provides essentially three signals:

1) The "baseband" signal is essentially the same as the CR signals discussed above. It is separated out using the same electronics as discussed above

2) The "average" signal is the overall signal through the system including the CR cell at an "average" pressure. It is separated out using a synchronous detector at the chopper frequency. The "effective average transmission function" is depicted in Figure A.2(b) and shows suppression at the spectral frequencies of absorption lines of the cell gas.

3) The "difference" signal is essentially the same as the baseband and CR signals discussed above. However it has one or two significant differences. It is a conventional am sideband of the chopper signal as shown in Figure A.7. where C is the chopper frequency and P1-3 are the sideband signal. The signals C1-2 are imperfections due to the chopping system. The sideband signal is at a high frequency (assuming that the chopper frequency is much greater than the modulation frequency of the CR cell) which is advantageous in some detector and electronics issues. It is formed by radiation which has been modulated by both the conventional chopper and the CR cell. It is not affected by synchronous signals emitted from the CR cell itself either mechanically or thermally. Unhappily the balance condition still applies.

A.1.3 MOPITT CR Design

The PMR suffers from two problems as the pressure is raised. Firstly the modulation ceases to be isothermal and tends towards the adiabatic condition in the limit. This is not an overall problem although it complicates the modelling of the cell conditions considerably. Secondly the mass flow required becomes more difficult to realise and the compression ratio falls. These problems led us to consider an alternative device for the MOPITT instrument which requires high cell pressures.

The Length Modulated Radiometer (LMR) is depicted in Figure A.8. It takes the advantages of the pressure modulator cell of a static cell with static vacuum windows, but adds a rotary component in the form of a "filler" which has two thicknesses ("thick" and "thin") which provide a varying residual path length for the gas in the CR cell. The optical system however is sensitive to the varying amount of the filler and therefore outside the cell is a "compensator" which is rotated synchronously with the cell rotor and has an exactly complementary filler thickness - i.e. the sum of the filler thicknesses in the optical path is constant. In this manner a CR system is constructed which has most of the advantages of a PMR but is capable of operation at high pressure. The balance condition in an LMR can be extremely good because the filler can be chosen without regard to vacuum sealing or other requirements (bi-refringent materials are however not acceptable). The first laboratory LMR system constructed showed a balance condition of 0.0008±0.0001 which is acceptable for the lower channels of MOPITT. The upper channels utilise PMRs for which no discernable balance error has ever been detected.

A.1.4 Application of Double-Chopping CR to Nadir Sounding

Figure A.9 shows a schematic of a nadir-sounding instrument for atmospheric work. In summary the instrument receives energy from the ground which may have originated as thermal emission or solar scattering (or both). This energy is modified by the atmosphere as it travels upwards and it is this modification which we are interested in.

The problem with nadir sounding for tropospheric work is that the surface radiance is intercepted by the instrument and this must be measured or deduced before a retrieval can be performed. The transfer equations for a double-chopping CR instrument in this situation for emission are:

Average:

(3)

Difference:

(4)

In the above equations I(0) is the surface radiance, is the overall optical filter transmission (the filter is used to isolate the region of the spectrum where the molecular absorption band is located). B(z,T(z)) is the Planck function at altitude z and is a function of the local temperature T(z). is the transmission from level z to and and are the monochromatic gas transmissions of the CR cell (idealised here to only two states).

Equations (3) and (4) can be re-arranged to isolate the "surface term" and the "atmosphere term" as follows:

Average:

(5)

Difference:

(6)

In the case of a solar channel where the atmospheric emission can be neglected:

Average:

(7)

Difference:

(8)

The quantity represents the surface reflectivity and is the solar zenith angle at the time of the measurement. is the solar emission and the vertical transmission has been approximated using a plane parallel atmosphere approximation and a height-dependent absorption coefficient k(z) and density .

Further analysis of the equation with attention to the sizes of the relative terms indicates that the average terms tend to be representative of the "surface term" in the thermal case and give most information about the solar intensity and the surface reflectivity in the thermal case. When applied to the difference signals the surface term can be eliminated leaving the atmospheric modification which is indicative of the gas concentrations.

A.2 Historical Reference List for Correlation Radiometers

Most of the projects involving correlation spectroscopy have been initiated by the Oxford University group under Prof. J.T. Houghton and, more recently, Prof. F.W. Taylor. Major contributions to the hardware have been made by Dr. E.J. Williamson and Dr. G.D. Peskett and to the theory of retrievals by Dr. C.D. Rodgers.

This list places in general historical order the development of correlation radiometers of various styles. It is not exhaustive and concentrates on those elements that are relevant to this proposal. A more comprehensive review of the development of the pressure modulator is given in:

     Taylor, F.W., (1983), Pressure Modulator Radiometry Spectrometric 
     Techniques Vol. III, Academic Press, 137-197.

The first paper on the selective chopper technique for remote sensing was:

     Houghton, J.T. and Smith, S.D. (1970), Remote Sounding of Atmospheric
     Temperature from Satellites, I. Introduction. Proc. Roy. Soc. Lond.
     A320, 23-32.

and initiated a series of papers under that general title in Proceedings of the Royal Society some of which are referred to below.

The first selective chopper instrument was built for the NIMBUS 4 satellite. This instrument was designed to measure temperatures in the upper troposphere and lower stratosphere using the differential absorption of cells of carbon dioxide. Since carbon dioxide is evenly distributed through the atmosphere up to at least mesospheric levels, the upwelling radiation in the 15 µm band of CO2 is indicative of the temperature profile. Sensitivity to various atmospheric levels may be obtained by varying the cell pressures. The instrument is described in:

     Abel, P.G. et al. (1970), Remote Sensing of Atmospheric Temperature from
     Satellites, II.  The Selective Chopper Radiometer for Nimbus D. Proc. 
     Roy. Soc. Lond., A320, 35-55.

The instrument operated by switching the optical beam between two different cells thus obtaining a chopping effect due to the differing cell pressures.

The next instrument was an advanced selective chopper radiometer for the NIMBUS 5 spacecraft.

     Ellis, et al. (1973), Remote Sensing of Atmospheric Temperature from 
     Satellites, IV. The Selective Chopper Radiometer for NIMBUS 5.  Proc. 
     Roy. Lond., A334, 149-170.

In this instrument the optical beam remained static and the cells were rotated into the beam by a "filter wheel" system. Channels were also included for looking at some other atmospheric effects and to view the surface radiance to assist with the retrieval of the temperature in the lower atmosphere.

The first paper on the pressure modulator was:

     Taylor, F.W. et al., (1970), Radiometer for Remote Sounding of the Upper
     Atmosphere. Appl. Opt., 11, 135-141.

The principle was first tested in a balloon instrument and was followed by the PMR instrument on the NIMBUS 6 spacecraft.

     Curtis, P.D., et al. (1974).  Remote Sounding of Atmospheric Temperature 
     from Satellites V.  The Pressure Modulator Radiometer for NIMBUS F. 
     Proc. Roy. Soc. Lond., A377, 135-150.

The instrument successfully measured temperatures in the stratosphere and mesosphere.

The first use of the pressure modulator for composition sounding rather than temperature sounding was in:

     Chaloner, C.P., et al. (1978), Infrared Measurements of Stratospheric 
     Composition I. The Balloon Instrument and Water Vapour Measurements, 
     Proc. Roy. Soc. Lond. A364, 145-149.

     Drummond, J.R., et al. (1978), Infrared Measurements of Stratospheric 
     Composition II.  Simultaneous NO and NO2 Measurements, 
     Proc. Roy. Soc. Lond.  A364, 237-254.

The next satellite instrument was the SAMS Instrument on NIMBUS 7.

     Drummond, J.R., et al. (1980), The Stratospheric and Mesospheric Sounder
     on NIMBUS 7.  Phil. Trans. Roy. Soc. Lond., A296, 219-241.

This composition and temperature sounder possesses a number of channels for viewing concentrations of various components such as methane and water vapour as well as temperature profiles. It differs from the instruments mentioned above in being a limb- scanner rather than a nadir sounder.

Several parallel projects were now in progress. A notable one was a pressure modulator radiometer on the Venus orbiter spacecraft:

     Taylor, F.W., et al. (1979), Infrared Radiometer for the Venus Orbiter. 
     I. Instrument Description. Appl. Opt., 18, 3893-3900.

Current developments include the ISAMS instrument for the UARS (Upper Atmosphere Research Satellite) and several other planetary projects.

Developments outside of the Oxford group included a "GASPEC" instrument

     Ward, T.V. and Zwich, H.H. (1975), Gas Cell Correlation Spectrometer:  
     GASPEC, Appl. Opt., 14, 2896-2904.

This instrument is a selective chopper style instrument in which the differencing of the signal from passing through two cells is performed electronically, rather than optically. A second signal at a different frequency is superimposed on the main radiometric signal and used to operate an automatic gain control circuit to maintain system balance.

The GASPEC instrument was the forerunner of the MAPS (Measurements of Atmospheric Pollution from Satellites) instrument which flew on the shuttle (STS-2) in 1981:

     Reichle, H.G., et l. Middle and Upper Tropospheric Carbon Monoxide 
     Mixing Ratios as Measured by a Satellite-Borne Remote Sensor During
     November 1981.  J. Geophys. Res., 91, d10, 10865-10887.