5. Instrument Line Shape Measurement Equipment

The equipment developed in this work, for determining the response function of Fourier transform spectrometers, consisted of a passively-stabilised infrared laser system (below), for direct measurement of the instrument line shape of high-resolution instruments, and a low-pressure gas cell with a portable filling rig (appendix A2), used for validation of these measurements and for direct measurements of the instrument line shape of lower resolution spectrometers.

Infrared Laser System

Figure 5-1. Schematic diagram of layout of IR laser system.

5.1 Infrared Laser

The infrared laser (figure 5-2) consisted of a ‘one off’ plasma tube (manufactured by Research Electro-Optics Inc., Boulder, Colorado) in a 30 cm resonant cavity (built in-house), with aluminium end-plates and Invar spacer bars.

Figure 5-2 Schematic diagram of infrared laser (section 5.1).

The high-reflectivity mirror was mounted in an acrylic holder screwed to one end-plate. The acrylic mount for the output mirror was mounted on a piezo-electric stack which facilitated remote, fine adjustment of the cavity length and this was, in turn, screwed to the other end-plate. The end-plates were fitted with micrometers to enable the mirrors to be tilted as required to align the cavity. Since the end plates are pivoted at their edges, the process of adjusting the micrometers to align the mirrors will result in significant axial movement of the mirrors for quite small changes in angle. This effect can be exploited in practice since it provides a mechanism for coarse tuning of the length of the laser cavity when it has been set up. One of the micrometers on an end plate may be adjusted so as to ensure that the oscillating mode is centred on the middle of the gain curve before the boxes are closed and the cavity length controlled from outside by means of the piezo electric stack. The amount by which the angle of the mirror concerned is altered in this process is negligible; altering the length of the cavity by one half wavelength results in a change in the orientation of the mirror of less than 30 mradians.

The two discharge-tube carriers (figure 5-3) each consisted of an aluminium beam spanning the gap between the lower two of the four spacer bars. One end of each was fitted with V-shaped saddles to ensure reliable location of the carriers with respect to the cavity.

Figure 5-3. Schematic diagram of discharge tube carrier assembly (section 5.1).

The beams were fitted with micrometer-driven translation stages which provided for alignment of the tube across the cavity. The discharge tube itself was held in insulated spring steel clips attached to the translation stages. Vertical adjustment was achieved by raising or lowering one end of each tube carrier by means of a micrometer and adjusting the translation stage as required to compensate for any lateral movement.

5.1.1 Laser Cavity Modes and Cavity Alignment

If the laser spectrum is to provide a direct and true measurement of the response of the spectrometer it is important to ensure that only a single longitudinal mode oscillates, as to obtain a line shape measurement from a multimode source would necessitate a detailed knowledge of the mode structure of the laser and for this to be deconvolved from the measured spectrum. This would not only add unwanted complexity but would also result in a measurement that was highly susceptible to noise.

5.1.1.1 Longitudinal Cavity Modes

Figure 5-4 shows an overlay plot of 429 laser spectra taken between 26th June and 13th November 1996. At various times during this period the laser was permitted to drift freely, or exhibited evidence of severe frequency drift. The spectra have been normalised, for ease of interpretation, such that their highest peaks have equal strength, although those peaks at the edges of the group will be much weaker, in practice because of their proximity to the edge of the usable part of the laser’s gain curve. It is, therefore, reasonable to assume that measured laser lines will cover the whole of that part of the gain curve that is above threshold for this laser, and hence the positions of the highest and lowest frequency laser lines may be taken as indicating the points where the gain curve rises above threshold.

The requirement, in terms of length, for resonance to occur in the cavity is that the length be equal to an integral number of half wavelengths. This may be written as

L = nl/2

(5.1)

where n is the number of half wavelengths that may be stacked in a cavity of length L, and s is the wavelength of the radiation. This may be written in terms of wavenumber as

s = n/2l

(5.2)

Since the adjacent modes result from the presence of n±1 half wavelengths in the cavity and the length L is unchanged the frequency difference between modes s₁ and s₂ (the free spectral range) will can be derived thus:

s1 - s2 = n/2L - (n+1)/2L = 1/2L.

(5.3)

In this case, where the cavity length is 30 cm, the free spectral range comes out as 0.017 cm^-1. The laser lines with the highest and lowest frequencies (figure 5-4) can be seen to have a separation of 0.015 cm^-1. Since this is narrower than the free spectral range of the cavity the laser will support only a single longitudinal mode.

5.1.1.2 Off-Axis Cavity Modes

Off-axis modes arise when the circulating photons are reflected by the wall of the discharge tube on one or more occasions during ‘a round trip’ in addition to the normal single reflection at each of the mirrors. It follows that such modes require some degree of misalignment of the cavity in order to exist and that they may, therefore, be prevented by careful alignment of the cavity. The procedure used to align the cavity is described below.

The laser tube is removed from the cavity and the red beam aligned using folding mirrors FM1 and FM2 (figure 5-1) such that it passes through the centres of both the high reflectivity mirror and the output mirror/coupler.

Figure 5-4. Overlay plot of 429 spectra used to determine the width of the usable part of the gain curve. Lower plot shows detail of upper one with ‘edges’ of gain curve marked (section 5.1.1.1).

The high-reflectivity mirror is then removed and the red alignment beam is steered so that, after reflection at the output coupler the beam falls on the aperture plate attached to the the front of the alignment laser. There will, of course be reflections from both faces of the output coupler; this is useful because only when the red beam is incident on the centre of the correctly aligned mirror will the two reflected spots form a single, symmetrical pattern on the front of the red laser. It is possible to obtain a reflected spot from either or both of the surfaces of the output coupler on the front of the alignment laser depending on the relative orientations of the alignment beam and the output coupler. Since the end plates of the infrared laser pivot about their edges and the piezo-electric stack holds the output coupler some 30 mm away from the plane of the plate, any attempt to realign the output coupler will cause its centre to move away from the red spot. As a result of this, the red beam and the orientation of the output coupler must be repeatedly adjusted, in an iterative manner, until the two conditions (that the red beam be centrally incident on the mirror, and that the reflected spot be centred on the output aperture of the alignment laser) are satisfied; this may involve replacing and removing the high reflectivity mirror to check that the red beam is still aligned along the axis of the cavity.

The high-reflectivity mirror is then replaced and the previous procedure is repeated for this mirror while ensuring that the red beam remains centred on the output coupler. As the mirror holder is screwed directly to the end plate, no significant sideways displacement of the mirror occurs as it is tilted. When both mirrors are properly aligned (this can be seen by the presence of a symmetrical pattern around the aperture of the alignment laser caused by reflections from all four surfaces of the infrared cavity mirrors) the plasma tube is replaced in the cavity and centred on the red beam using the micrometers on the tube carriers.

This procedure ensures that the principal axes of both laser mirrors and the axis of the plasma tube are co-linear and hence that the cavity will be optimally aligned for single longitudinal mode operation and the possibility of off axis modes negated.

A Judson LD10 indium antimonide detector was used in conjunction with a Hewlett Packard spectrum analyser (141T frame with 8554B and 8552A plug-ins) to visualise the output of the laser in real time. In this way it was possible to confirm that the output of the laser consisted of a single frequency as required.

5.2 Thermometry and Power Measurement

The emergent laser beam from the infrared cavity (figure 5-1) was incident on a zinc selenide beam splitter that diverts about ten percent of the radiation, via a chopper and a zinc selenide focussing lens, on to an indium antimonide (Judson LD10) detector, connected to an oscilloscope, which was used to monitor the relative power output of the infrared laser.

Three thermocouples were added to the laser system to record the degree to which fluctuations in ambient temperature affected the laser cavity: one was fitted to (but not in contact with) one of the Invar spacer bars, one outside the infrared laser box but still inside the infrared system box, and one was fitted outside the system box.

These thermocouples were not fitted until shortly before the intercomparison at Lauder (section 6.3). The output from these three thermocouples and the indium antimonide detector were recorded simultaneously on a Psion LZ64 data logger to give an indication of thermally induced laser drift. These data shall be considered in section 6.3.

5.3 Beam Expansion and Collimation

The remainder of the beam, passing through the beam splitter, is incident on a 90º off-axis paraboloid which expands the beam, via a folding mirror, onto a reflective diffuser (scatter plate). The diffuse reflected infrared radiation from the scatter plate is collected by a large (f3) 30º off-axis paraboloid identical to the entrance mirror of the 120M, which collimates the diffuse beam and directs it out of the outer enclosure.

5.3.1 Reflective Diffuser

The scatter plate was manufactured from flame-sprayed aluminium and was shown to scatter through a wide angle but, while this reduces the risk of ‘hot spots’ in the scattered infrared beam, it also caused much of the scattered infrared beam to be lost past the collimating off-axis paraboloid. This loss of output from the laser system was exacerbated when the plate was orientated so as to prevent any of the specular reflection from reaching the collimating off-axis paraboloid (figure 5-1). By aligning the scatter plate with its normal directed towards the collimating mirror one ensures that any specular reflection of the incident radiation must undergo at least one further diffuse reflection in order to reach the collimating paraboloid, as described below.

Figure 5-5. Schematic diagram of scatter plate and hemisphere used to expand, flatten and add divergence to infrared laser beam (section 5.3.1).

To improve the throughput of the system the scatter plate is covered with a gold hemisphere, drilled with 5 mm diameter holes for the incident and reflected beams. The hole drilled for the incident beam is aligned such that the infrared beam is incident on the plate at the centre of parent sphere (figure 5-5); the output hole is sized so as to prevent any radiation leaving the hemisphere/scatter plate combination at such an angle that it would miss the collimator; some loss through the input hole is unavoidable. Any reflected radiation that does not pass out through these holes is reflected back to the centre of the hemisphere whence it may be scattered out towards the collimator; this includes any primary specular reflection since the diffuser is orientated such that the axis of the output beam is normal to the plane of the plate. These multiple reflections within the scatter plate/hemisphere combination not only improve the throughput of the system but also increase the degree to which the intensity profile of the reflected beam is flattened.

Summary
Acknowlegements
Contents
Chapters: 1, 2, 3, 4, 5, 6, 7, 8
Appendices
References