Image Slicer

Astronomical spectrographs are mainly housed in thermal and mechanical stabilized enclosures and linked via one or a set of optical fibres to the telescope. The fibre core diameter determines the width of the entrance slit and respective the resolution of the instrument. Because of focal ratio degradation (short FRD), which describes the change in focal ratio due imperfections and microbends in an optical fibre, we won’t get an equal input/output light cone. Therefore it is useful to adjust the output cone using a microlens at the end of the fibre. Unfortunately this leads to a change of the entrance slit width due to magnification. This can be easily calculated using the simple lens equation. Since the smallest step-index fibre core diameter is 50µm, one might use an image slicer in the focal plane of the microlens to reduce fibre disc significantly. Such image slicers are basically used to cut the projected fibre disc into two equal slices which reduces the slit width by a factor of 2 and increases the resolution of a spectrograph. Not only the fibre diameter can be significantly reduced, also the focal ratio is enlarged by the lens magnification factor which allows using collimators with a long focal length.

Figure (a) shows an usual fibre output whereas in figure (b) the disc is already sliced into two pieces. The black bars indicate the corresponding slit size.

Gerado Avila et. al. has developed an image slicer using two front-surface mirrors instead of an expensive Bowen-Walraven slicer made out of two prisms. The basic functionality is to slice the circular fibre output into two separate pieces and to align the sliced pieces on top of each other. This is done by creating a mirror stack  so that the entrance beam is cutted into two pieces where one part of the beam is reflected twice.  A general CAD model of such a two-mirror image slicer is shown in the following figure.

The fibre-link between telescope and spectrograph is realized by a 50µm fibre (SMA connectors) that transfers the injected starlight into an output beam of  f/4.5. The fibre output is then reimaged onto the image slicer using a small achromatic lens from Marcony Designs with a focal length of 8mm. With a distance of  13mm from the fibre end to the center of the doublet one will get a spot size of roughly 80µm in the image plane of the lens and a focal ratio of approximately f/7.2. As G. Avila already specified, both mirrors should be aligned laterally under an angle of 67.8° with respect to each other.  A mirror separation of nearly 60µm is necessary to slice the magnified fibre output, impinging the image slicer under 45°, into two equal slices.

A complete assembly of the image slicer is shown in the figures below.

Exploded view of image slicer and fibre injection unit.

Front view including slicer mount.

Detailed view of the mirror separation.

Projection optics:

I have tried to model an acromatic doublett with 8mm focal length and 4mm diameter to verify the realistic distances between fibre end and  image slicer.  Since Marcony Designs doesn’t puplish all lens properties I have therfore made the lens mount slightly adjustable along the optical axis in a range of +/- 1mm.

Polishing and grinding the mirrors:

Instead of using aluminized microscope cover glass I ordered two front-surface mirrors from Astromedia with a thickness of 1.2-1.3mm. Cover glass edges are well defined which is obviously not the case for these mirrors from Astromedia. Therefore I decided to sharpen the slicing edges using a self made polishing tool as it is shown in the left figure below. The mirrors where dry-grinded using sandpaper with grit sizes starting from 600 to 1200 grits per square centimetre. Afterwards I used special abrasive sheets (fibre polishing paper) to polish the edges so that  it looked sharp and well polished over the whole length. Unfortunately after removing the protective foil the edges looked a bit tattered. But some more polishing steps should help to overcome this problem.

Front-surface mirror mounted in homemade polishing tool.


The mirror assembling was done while observing the image plane of the slicer using a WATEC 120N camera. As the perfect mirror orientation was found a jig with rubber end was used to hold the mirror in position. Both mirrors were glued successively with UHU Plus (schnellfest) onto the Aluminium body.

Assembled image slicer.

First light:

Image of the slit.

A first light spectrum including the image slicer as the new entrance slit is shown below. The spectrograph was aligned roughly by holders made from aluminium blocks and cardboard which were glued together with double-sided tape and hot glue. In the blue spectral region (top) one can see clearly the overlap between the single orders. This comes because the mirror image slicer prototype was mainly build just to show that it works.  Some improvements are required to adjust the spacing between the single orders.

First light spectra after assembling the mirrors.

State of the art:

Different tests have shown that the upper mirror must have a thin and well defined edge where the beam passes by to provoke a propper cut. Furthermore there are some more modifications necessary to improve the mechanical stability and the quality of the slit:

– modify the design of lens mount and fibre injection
– improve the mirror separation
– use FC connectors instead of SMA
– anodize all parts made from aluminium
– design a pillar to protect the fibre against bendings


Since the very first prototype showed a good slicing performance the next iteration is to narrow the slit and to adjust the focal ratio by changing the lens position. To fit the new requirements most of the mechanical parts has been revised or replaced by better matching materials. Especially the fibre holder was the most critical part because it was made out of POM, a material with a thermal expansion coefficient five times higher than Aluminium. Therfore the refurbished part will be made out of brass (Messing). At the end all parts must match with a good centering w. r. t. the optical axis to keep the optical aberrations as small as possible.

Pillar design to stabilize the fibre input.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s




%d bloggers like this: