5.4 Delivery Optics

When in use with the National Physical Laboratory’s Bruker, the infrared system is located alongside the Bruker’s scan arm with its output beam parallel to it. The laser beam is introduced to the spectrometer by means of two folding mirrors (figure 5-6). The folding mirrors are mounted on a purpose-built shelf that can be attached to the instrument. The second of these mirrors is fitted on a kinematic mount so that it can be replaced with a beam splitter (zinc selenide with 0.3º wedge) to enable measurements of the infrared laser to be taken simultaneously with solar/atmospheric measurements.

Figure 5-6. Schematic diagram showing layout of the delivery optics used to introduce the IR laser to the spectrometer for simultaneous solar and line shape measurements (section 5.4).

5.5 Alignment Procedure

Correct alignment of the infrared system internally is essential to ensure good output power and beam profile, while alignment to the instrument must be correct if accurate measurements are to be obtained. Due to the great scope for misalignment between the infrared system and the spectrometer and the nature of the low-power infrared beam, the alignment procedure is carried out in two stages.

5.5.1 Alignment of the Red Beam

The red laser, having been used to align the cavity and tube of the infrared laser, will be approximately coaxial with the infrared beam¹ and so can be used in the alignment of the infrared system both internally and to the spectrometer. When misaligned, two red spots are clearly seen on the scatter plate; they may be observed through the exit hole in the hemisphere with the aid of an inspection mirror. These are the incident beam from the laser and the reflection of that beam refocussed onto the scatter plate by the hemisphere. The two spots are brought together using folding mirror FM3 (figure 5-1).

The infrared system is aligned to the spectrometer by inspection initially; this should be sufficient to cause the focussed red laser beam to fall on the aperture wheel in the Bruker (the expanded red beam may only be seen in a fully darkened room). From this point the beam may be steered so as to produce a round spot centred over the field stop, using the two folding mirrors shown in figure 5-6.

5.5.2 Alignment of the Infrared Beam

When the red beam is aligned as described above the system will be misaligned for the infrared beam, due to the transmissive optics, but only minor adjustment of the folding mirror FM3 (figure 5-1) is required for a signal to be detected by the spectrometer. From this point the system is aligned so as to maximise the signal reaching the spectrometer’s detector.

The beam incident on the scatter plate should be readjusted first as this is both critical to the output of the laser system, and straightforward. Since this alignment has already been carried out for the red beam, only slight adjustments will be required and since the main cause of the displacement of the red and infrared beams is the beam splitter used for power measurements, then most of the displacement will be in the horizontal plane. It follows from this that this adjustment should be made first and that that should produce a noticeable increase in the signal with a pronounced maximum. The beam may then be aligned in the vertical plane. This will ensure that the maximum throughput is being achieved in the infrared system optics. From this point the two folding mirrors (shown in figure 5-6) may be systematically adjusted to maximise the signal. Maximum signal can only be obtained when the system is correctly aligned internally and to the spectrometer.

The importance of proper alignment of the input beam into the the scatter plate/hemisphere combination is illustrated in figure 5-7. In the case where both surfaces have perfect reflectivity, it follows that the orientation of the input beam would have no effect since all radiation entering the hemisphere must also leave. This situation is not achievable in practice, and so it is important to keep the number of reflections within the hemisphere to a minimum and therefore to ensure the early of return scattered radiation to the centre of the scatter plate whence it may be reflected out of the hemisphere towards the collimating paraboloid.

Figure 5-7. Schematic illustrating the importance of alignment of the incident beam to throughput of the hemisphere. Left: the specular component of the misaligned beam undergoes multiple reflections without hitting that part of the scatter plate visible to the collimating paraboloid, while the correctly aligned beam (right) is incident on that point directly and on subsequent reflections (section 5.5.2).

Any radiation that reaches the collimating paraboloid must have have undergone its last reflection in the hemisphere/scatter plate combination at a small area of the scatter plate near the focal point of the collimator and therefore any reflection that does not occur in this area cannot contribute directly to the output of the system. Considering firstly the specular component of the misaligned beam: this may undergo multiple reflections, each of which will result in some loss of radiation, without hitting that part of the scatter plate which is visible to the collimator (figure 5-7, left), while the correctly aligned beam (figure 5-7, right) is incident on that point directly and after reflection from the hemisphere and so both the first and third reflections will result in diffuse radiation reaching the collimator. Similarly in the case of diffuse reflections, any ray that originates at the centre of the scatter plate and does not escape through either the input or the output holes will be returned to the centre of the scatter plate following its first reflection in the hemisphere, and so every reflection at the scatter plate may contribute to the output of the system. It should be emphasised that while figure 5-7 might suggest that a large part of the reflection is specular and hence will be lost through the input hole after its third reflection, this is not the case because of the scattering efficiency of the diffuser.

5.6 Validation of Infrared Laser Beam Profile

It is important to the validity of the infrared measurement that the optics of the Fourier spectrometer be illuminated in the same way by the laser beam as by the solar beam so that the effects of any defects or misalignments are properly recorded. This requires the radial intensity profile of the laser beam to be at least as flat as that of the solar beam and to contain no ‘hot spots’.

In the absence of any beam profile monitor capable of measuring the expanded infrared beam, the following method was adopted. Measurements of laser power were made by using the Bruker to record a series of some 120 very low resolution spectra of the laser, using each of the Bruker’s 12 standard input apertures in turn. All solar measurements were made with the same optical filter in filter wheel in the detector optics (figure 3-2) to prevent the detector and electronics being overloaded at the larger apertures.

The operating software for the Bruker IFS 120M has a facility for displaying these spectra as they are retrieved, at about one every second and a half. There is no facility for writing these to disk but successive scans may be kept on screen while subsequent ones are plotted on the same axes² and in this way an overlay plot could be assembled of the 120 low resolution spectra measured. The plots so produced would typically consist of a group of emission lines at the frequency of the laser exhibiting some slight variation in intensity, with some five or six lines exhibiting significantly greater or lesser intensity; these were deemed to be outliers.

The half-way point between the peaks of these emission lines exhibiting the highest and lowest intensities (after outliers had been discarded) was taken as the total signal strength for that aperture. This procedure was repeated for each of the twelve apertures in the input aperture wheel. From this set of readings an ‘onion-peeling’ method was used to derive an estimate of the beam profile. It is possible to define a set of annuli centred on the principal axis of the instrument using successive apertures; the proportion of the total signal that is due to radiation passing through these annuli may then be found by taking the difference of the signal strengths obtained with successive pairs of apertures. Since the diameters of these apertures are known it is possible calculate the area of each of the annuli and hence normalise the readings by dividing the signal strength for each annulus by its area; in this way it is possible to convert these annular signal strengths into an estimate of the radial beam profile.

The profiles of the solar and laser beams thus obtained are shown in figure 5-8; since this method imposes axial symmetry on the profile, only half the beam is shown, with the y-axis at the centre. The profile obtained for the solar beam has a central dip which is clearly not physical (as this would imply a lower radiation output from the centre of the solar disc compared to the edges), and intensities in the second and third annuli that are suspiciously high compared to the fourth to sixth annuli (given that the Sun’s intensity profile is effectively flat across the central 85% (Kift [private communication]), however qualitatively similar results are seen for the three inner annuli in the case of the laser beam suggesting that this is an artefact resulting from the sizes of the apertures deviating from their nominal values. With the exception of the inner three annuli, the solar beam profile is in good qualitative agreement with the appearance of the visible radiation at the aperture wheel.

Figure 5-8. Profiles of solar (left) and laser (right) beams obtained by ‘onion peeling’ method described in section 5.6.

5.6.1 Error Estimates for Beam Profiles

This method has three main sources of error associated with it. Firstly the use of this ‘onion-peeling’ method will introduce errors regardless of the precision of the signal strength measurement or the tolerance of the apertures themselves, simply because each measurement will be an average of the radiation over the area of that aperture. This will result in a ‘quantised’ profile with inherent axial symmetry that may be artificial. We rely on careful alignment of the system to ensure that the infrared beam is symmetrical and that both beams are coaxial with the interferometer so that no artificial asymmetry is introduced by the measurement technique.

Secondly the readings of signal introduce errors which depend mainly on the signal-to-noise ratio. If we do not make any assumptions about the distribution of the peak heights, then the error might be taken as half the difference between the highest and lowest peak heights as tabulated below.

If we assume that the distribution of peak heights stored on screen is Gaussian, and that we have rejected five percent of these as outliers, then the spread between the highest and lowest peaks becomes the two-sigma limits of the sample, and the error estimates will reduce by half. The estimate can be further reduced by a factor of root-n (here n = 120) since the individual measurements are independent. This results in an error of ± 2.3% for the 0.3 mm aperture and less than ± 1% for the larger ones. If the assumptions made are valid then these errors may be discounted as they are obviously insignificant compared to the error arising from the limited number of apertures available This is illustrated by figure 5-9 which shows a plot of the signal strength (in terms of counts per minute) versus the aperture diameter at which this strength the signal strength was recorded for the infrared laser. It can be seen that the beam diameter is approximately 2.4 mm and that all the increase in signal strength therefore occurs over the central eight of the twelve apertures only with no further increase occurring over the largest four; very similar results were obtained for the solar beam.

Figure 5-9. Plot of the signal strength for IR laser versus aperture diameter showing the beam diameter as 2.4 mm and that the change in signal strength occurs over only eight of the twelve apertures (section 5.6.1).

The third source of error arises from the fact that the apertures are manufactured to a tolerance of ±50 mm (Jewkes (private communication)) which could introduce a quite considerable error in the smaller apertures. The fact that the profiles of both the Sun and the laser show a dip in the centre of the beam suggests that this is indeed the case. If we make the assumption that the profile of the solar beam is flat across the central fifty percent then we may, in theory, use the measurement of the solar beam to retrieve an estimate of the true diameters of the smaller apertures³. This may be achieved by permitting the diameters of the apertures to vary during the ‘onion peeling’ beam profile calculations in such a way that the profile of the solar beam appears flat across the inner apertures, and then using the aperture diameters so derived in the calculation of the laser beam profile. In practice this is problematic since there is no unique solution, due to the measurement procedure; one possible solution is illustrated in figure 5-10. The plots show the profiles of solar beam (left) and laser (right) beam obtained by this method, where the solar beam profile was constrained to be flat across the 1.1 mm diameter aperture.

From these results it was thought that the laser beam profile was sufficiently flat beyond that which is required for the use of the 0.65 mm aperture, although the lack of a unique solution leaves the validity of results thus obtained open to question.

Figure 5-10. Profiles of solar (left) and laser (right) beams obtained by ‘onion peeling’ method but with laser beam modified by changing the assumed diameters of the apertures (within manufacturing tolerances) to achieve a solar beam profile flat across the 1.1 mm diameter aperture (section 5.6.1).

5.7 Conclusions

A passively-stabilised infrared laser system developed for direct measurement of the instrument line shape of high-resolution Fourier spectrometers has been described. This consisted of a red, helium:neon alignment laser and a 3.39 mm, helium:neon laser with optics to expand the beam, flatten its radial intensity profile, and monitor its power. The system was enclosed so as to minimise the effects of draughts in the laboratory. The infrared laser is constrained to single longitudinal mode operation by the short cavity length, while off-axis modes are prevented by careful alignment. Measurements have been carried out to confirm single longitudinal mode operation and that the profile of the output beam was adequately flat.

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