3. The instrument

3.1. The telescope

The THEMIS telescope is an alto-azimuthal mounting where all the post-focus instrumentation is contained in a cylindrical tank rotating with the telescope around a vertical axis. The primary mirror is 0.9 m in diameter with an equivalent focal length of 54.861 m (f/61). A transfer optics forms an image, in diameter, on a secondary focus (F2), where the scale is 0.266 mm arcsec^-1 and the FWHM of the point spread function, for diffraction limited optics, is 0.15'' at 5500 Å.

When it was decided to install the IPM on THEMIS, the telescope was already completely designed and partially built in Orsay (Paris). Many mechanical constraints, therefore, have been imposed to the instrumental project, as: i) a 90 folding (M2 in Fig. 4 (click here)) at 610 mm from F2, breaking the instrument in a horizontal and a vertical path, ii) a maximum volume for the vertical path of 500 mm 500 mm 3100 mm (WPH), iii) a 300 mm avoidance zone for access to the spectrograph at 1650 mm from M2, iv) a reduction of the available volume on the last 400 mm, for a second access to the spectrograph.

The IPM was therefore mounted, in part to a horizontal rail jutting out from the tank, in part to vertical rails fixed on the external wall. A moveable mirror, just before F2, horizontally folds the sunlight coming from the telescope towards the IPM. This mirror (M1 in Fig. 4 (click here)) can be replaced by a dichroic, allowing to carry out, in different wavelength ranges, simultaneous observations with the IPM and the French spectrographs.

  
Figure 4: Schematic drawing of the instrumental layout. The solid line represents the principal optical path, while the secondary ones are showed by dotted lines. Dashed boxes represent moveable mirrors, which can be pulled in or out the optical path. The meaning of the labels is as follows. C: polarizing beamsplitter cube, CCD: CCD camera, D: diaphragm, FPI: Fabry-Perot interferometer, IF: interference filter, L: lens, LP: linear polarizers, M: mirror, ND: neutral density, P: pinhole, PMT: photomultiplier, S: electronic shutter, SL: spectral lamp, TV: TV camera, UBF: Universal Birefringent Filter, W: filter wheel, WP: wave plate

3.2. The principal optical path

The principal optical path is shown by the solid line in Fig. 4 (click here), schematically representing the instrumental layout.

The image is formed by the telescope on the field diaphragm D1, 13.5 mm in diameter, corresponding to 51''. An optical target can be mounted on this diaphragm, to allow the positioning of the two CCD cameras on the focus and on the same field of view. Moreover, two perpendicular tongues on D1 can be used as geometrical references, to superimpose a posteriori the monochromatic and the reference images taken by CCD1 and CCD2. The whole acquisition system will be described in a separate paper; a preliminary description is already available (Cantarano et al. 1993).

The image on D1 is then collimated by a first lens L1, followed by the shutter S1, allowing to optically isolate the IPM from the telescope, and by the achromatic polarizing beamsplitter cube C1, sending one half of the incoming radiation to TV1 and CCD2. The cube C1 is then followed by the achromatic half wave plate WP1, which rotates the transmitted p polarized radiation to align its direction with that of the linear polarizer at the entrance of the UBF. Because the UBF has a fixed linear polarizer at the exit, to better exploit the incoming radiation, one half of it, that would remain unused, is folded at the entrance by C1 and then the UBF, which cannot work in linearly polarized light, is used back to front.

Going on along the principal optical path, we find the pinhole P1, which, by isolating the image of the entrance pupil formed by L1, reduces the instrumental stray light, then the shutter S2, used to separate the principal optical path from that of the spectral lamp SL, and then the mirror M2, folding the radiation in the vertical part of the instrument.

A second lens L2 then collimates the pupil, forming on the diaphragm D2 an enlarged image of the solar region selected by D1. Two perpendicular tongues on D2, opposite to those on D1, can be used as geometrical references to superimpose the monochromatic images to the spectral dishomogeneities maps, obtained by means of the Cd spectral lamp (see Sect. 4).

After D2, in an image space where the pupil is collimated, we find the interferometer, used in axial mode and in telecentric mounting.

A third lens L3, just after the interferometer, then collimates the image, forming beyond the UBF a new image of the pupil on the pinhole P2, used to remove the ghost images produced by reflections on the rear surfaces of the interferometer plates. To this purpose, the plates are wedged by , and as the minimum deflection is 2, for a relative aperture smaller than f/115, all the spurious images are separated and can be easily removed by a suitable pinhole, as P2, used to isolate the on-axis pupil image.

Finally, a forth lens L4 and a third folding mirror M3 form a monochromatic image on the CCD1 camera. This camera is preceded by the shutter S4, controlling the exposure time, and by a filter wheel W1, carrying the interference filters for the wavelength ranges where the UBF is calibrated (see Table 2 (click here)).

 

Calibrated ranges 5184 Å (Mg b1), 5380 Å (C I), 5576 Å (Fe I), 5890 Å (Na D2), Spectral resolving power at 5500 Å Wavelength drift Spectral homogeneity (after correction)

Field of view 51''

Detectors Thomson CCDs A/D converter: 12 bit Full-frame read time: 2.6 s Read noise: 30 e^-

Image scale 0.065'' pixel^-1 Wavelength setting time Exposure time (1% photometric precision on the continuum) Acquisition rate including: wavelength setting, exposure, frame reading, display and storing

 

The CCD is a Thompson 9.7 mm9.7 mm, with 512512 square pixels, 19 in size. Obviously, the image scale on the CCD must be such as not to reduce the telescope spatial resolution. As the overall field is 51'' and the FWHM of the point spread function is 0.15'', the maximum number of resolved elements is 340. Assuming therefore two pixels per resolved element, and an useful square field inscribed in the total admitted circular field, an at least 481481 pixels detector and an image scale of 0.077'' pixel^-1 or smaller are required. In our case, we have an useful field of with an image scale of 0.065'' pixel^-1, allowing therefore to exploit all the telescope spatial resolution.

3.3. The reference path

This secondary path originates from the achromatic polarizing beamsplitter cube C1. The reflected s polarized light passes through WP2, an achromatic half wave plate, which can be rotated to differently distribute the light between CCD2 and TV1 by means of the polarizing cube C2. The lens L6 forms on CCD2 an image of the same solar region seen by CCD1. LP1 are two crossed linear polarizers, used to adjust the light level, W2 is a filter wheel, identical to W1, and finally S5 is a shutter, controlling the exposure time. The CCD2 camera allows to obtain, simultaneously with CCD1, images of the same solar area on the same spectral region, but with a larger passband ( 50 Å). These white light images can be then used as reference and/or for destretching procedures, to correct a posteriori distortion and image motion produced by seeing.

The lens L7 then forms an image of all the admitted circular field selected by D1 on the TV camera TV1, preceded by the mirror M4, the interference filter IF1 ( = 5300 Å, FWHM = 50 Å) to increase the contrast, and a macrophoto lens. TV1, continuously showing the selected solar region, can be used to monitor the solar and atmospheric conditions.

3.4. The path of the spectral lamp

The spectral lamp SL is the same already described in Sect. 2.3. A macrophoto lens forms an 1:1 image of the lamp discharge tube on the pinhole P3, with the same diameter of the image of the entrance pupil, formed by L1. The two lenses L8 and the mirror M11 form then an 1:1 image of the pinhole, the position of which coincides with that of the entrance pupil, when the moveable mirror M10 is inserted on the principal optical path. In this case the FPI and the UBF see the lamp in the same way as the Sun. A second moveable mirror M5, which can be inserted on the optical path after L3, and the mirror M6 form on the pinhole P4 an enlarged image of P3. P4, as P2, removes the ghost images produced by reflections on the rear surfaces of the interferometer. The two lenses L5 form then: i) an 1:1 image of the pupil on the cathode of the photomultiplier PMT1, by M9 and the moveable mirror M7, ii) the same image in front of the color TV camera TV3, by the two moveable mirrors M7 and M8, iii) an image of the interferometer plates in front of the color TV camera TV2.

PMT1 is preceded by the two crossed linear polarizers LP2, to adjust the light level, by the interference filter IF2 ( = 6438 Å, FWHM = 50 Å) to isolate the Cd red line, and by the shutter S6. PMT1, which permits the measurement of the wavelength drift of the interferometer, is used together with the reference photomultiplier PMT2, which, via a fiber optic, sees the Cd line through the interference filter IF3, equal to IF2.

The TV3 camera can be used to verify and, if necessary, to adjust the orthogonality between the optical path and the interferometer plates. To this purpose, TV3 can show the entrance or the spectral lamp pupil, by pulling the moveable mirror M10 in or out the optical path. To compensate the much different luminosity of the two images, the neutral density ND is automatically inserted in front of the macrophoto lens, mounted on TV3, when the entrance pupil is shown. A double cross hair generator allows to verify on a TV monitor the relative position of the two pupil images, and by acting on the mirror M11, it is possible to move the lamp pupil until the two images coincide. When this condition is verified, if the interferometer plates and the optical path are not orthogonal, the image of the lamp pupil, showing the inner part of the ring pattern produced by the FPI, will appear asymmetrical. In this case, by changing the orientation of the interferometer, a symmetrical image can be obtained, securing therefore the correct orthogonality.

The TV2 camera can be used to verify and, if necessary, to adjust the parallelism of the interferometer plates. To this purpose, this camera can show the FPI plates lighted by the spectral lamp radiation. For a perfect interferometer, used in telecentric mounting, the passband wavelength position should be identical on each point of the plates, the image of which, therefore, should appear homogeneous. In practice, however, in different points of the plates the interference orders will be randomly and sistematically shifted, due to flatness and parallelism errors. For this reason, the plates will appear dishomogeneous in intensity and sometimes in color, if in different points different lines emitted by the lamp coincide in wavelength with some interferometer order. By observing the plate image shown by TV2 and changing the parallelism conditions up to obtain an image as homogeneous as possible, it is easy to secure the plate parallelism with a precision better than /1000.

3.5. Remote controls

All the procedures which need to adjust the instrument and to verify its correct behaviour must be performed from a remote control room. To this purpose, many instrumental components are provided with actuators for a remote positioning, while three TV cameras allow to verify the correctness of the operations. In the following, these remote controls are described.

Mirror M6: it allows to center on the monitor connected to TV3 the image of the entrance pupil with respect to an electronic cross hair.

Mirror M11: it allows to obtain the coincidence between the entrance and the Cd lamp pupil, verified by the TV3 camera (see Sect. 3.4).

Mirror M3 and cube C2: they allow to center the solar image on the two CCD cameras.

Mirror M4: it allows to center on the monitor connected to TV1 the image of the observed solar region.

FPI: the interferometer is provided with two actuators which allow to rotate it around one axis, normal to the plane of Fig. 4 (click here), and around a second axis, contained on this plane and normal to the optical path. In this way it is possible to secure the correct orthogonality between the interferometer plates and the optical axis, verified by the TV2 camera (see Sect. 3.4).

CCD1 and CCD2: each one of the two CCD cameras is mounted on a motorized translation stage, allowing to obtain the correct focusing.