GRATING GHOSTS

Section 10.06 reported on the effects of reflections from the unwanted orders with the 1200 g/mm gratings, especially in the "two grating concave arrangement" we want to use. In m=0 the one grating illuminates the other and the subsequent m=0 reflection reaches the camera. This, and other possible problems, prompted us to switch to 600 g/mm gratings.

The grating-grating reflection does not occur for this case. The worst unwanted reflection is 3% at m=0, then 2% at m=2 - these may require some baffling (email N.D. May 31, 1999).

GHOSTS IN THE OPTICS

As there are at least 22 lens elements and other optical components in the beam train, it is most important that the design targets of narrow band, high efficiecy AR coatings be applied to all optical element surfaces. A design goal of transmission of T= 0.998 is not out of the question. This also is approx the value for the contacted (oiled or cemented) surfaces as well, due to the index step across the interface.

The other inefficient surfaces in the system are the grating, the narrow band filters, the CCD dector itself, and the dewar window. The window (silica) is not dedicated to the PNS optical system, and as such wont be coated specifically for minimum losses at the given wavelength. The grating has an efficiency of 75 to 80% at best, and the stray light effects has been reported in another section. Some of the losses from the filter will be absorptive and some reflective. As the filters are in collimated space, these reflections should be returned to the telescope, and collimator body. The H-alpha beamsplitter, when inserted, may be a source of a double reflection, which as it is in collimated space will be returned to the camera, and be in focus at the detector. Thus the specification that the rear surface of this plate be narrow band AR coated for the 501 nm wavelength band is critical.

The major problem is thus seen to be the CCD detector area. An assumption of the efficiency of the detector of 75% to 90 %, will determine the intensity of the source for the ghosts that may be in the camera. The window ( in close contact to the CCD) will also be another source for return reflections. A series of Zemax ray traces were done to find potential surfaces in the camera, that would return ghosts to the detector. Note that if the surfaces are AR coated as above then maximum intensity of this ghost will be 0.002 x the source intensity, integrated over the geometrical area of the ghost spot.

I (GB) traced from the CCD surface as the source, to elements in the camera, and up to collimated space, ie this is light reflected off the detector array, from the original image points, back into the optical system, and then returned to the detector. The surfaces that produce " pupil ghosts" or "halo ghosts" on the detector, are noted below, and attachments are provided of the zemax layout and footprint on the detector. Surfaces that give rise to return footprints that are larger than the detector area, are assumed to be not significant, and are not listed.

Raytrace out put for the following is held on file (by project office).

Case A: Front surface of SK16 element, radius 138.39 mm convex. Pages 2, 3, 4. This is the most serious ghost returned. Case A-1 show the source beams, the sources are at heights of 0, 10, and 17mm . Case A-2 shows the return ghost. CaseA-3 shows the footprint on the detector, the square approximates the area of the detector. The axial return ghost is approx 4.2mm diameter. Return beams outside this are not shown. This surface is the most critical for the best possible AR coating, ie not only does the specific AR coat design contribute to the high thru-put of the camera (efficiency), but it is critical in controlling this ghost.

Case B: Front surface of dewar window shown, the back side will also produce similar "halo ghosts". The intensity of these returns will be larger than for the lens surfaces, given that the window will not be narrow band AR coated. Pages 5, 6. Case B-1 show the source and return beams. Case B-2 show the ghost footprints. The axial return ghost is approx 18.0 mm diameter. There is no option available to reduce these ghosts, unless dedicated dewars with narrow band coated windows are used.

Case C: Rear surface of SF1 lens, concave radius 283.01mm. Pages 7, 8. Again these near halo type ghosts will have their intensities reduce by the AR coat. Case B-2 shows the axial return ghost is approx 23 mm.

Case D: Front surface of SF1 lens, concave radius 119.92mm. Pages 9, 10. Only the near axial sources are returned, and these footprints are of area comparable to the detector.

All other surfaces in the camera, produce return ghosts that are very much larger than the detector area. The other area of note is the return beams from the first spilt doublet in the camera, ie the BK7/SF2 narrow airgap assembly. The returns from the first three surfaces are "focused" at or near the contacted (oiled or cemented) interface of the second doublet.

Case E: Shows layout plots of these focused returns. Pages 11, 12, 13. This shows the return beams from front surface of BK7 lens, radius convex 252.2mm. It is focused at or near the contacted interface of the BK7/SF2 doublet. Case E-3 shows the footprint on this surface, where all returns are concentrated near the middle of the surface. Case E-1 shows the source beam, and case E-2 shows the ghost beam. Although the intensity of the return reflection from the AR coated radius 252.2mm surface, will be low, this trace illustrates the care needed to have this contacted interface, bubble and defect free. Otherwise this point will be a new source for scatter for stray light.

In collimated space, return beams are possible from any plane surface, at or near orthogonal to the optical axis, as this space is conjugate with the detector plane.

Mechanical mounts for the gratings will need to be non reflective, or angled such that returns do not enter the camera. The O-III filter will need to be placed, so that the rear surface is AR coated, and the "reflective side" is facing the collimator. This way the source beam should pass straight through. (note filter is approx 70% effiecent). The beamsplitter, when inserted, will reflect some fraction to the upper mechanical body space, and care is required that this is absorbed, and not returned to the cameras.

Similar work on lens element ghosts (stray light) has not been done for the H-Alpha cameras as of this date.

The above text was "frozen" on Nov 29, 1999

Text added: June 5, 2001

Date: Tue, 05 Jun 2001 17:13:03 +1000
From: Gabe Bloxham and Damien Jones ...There is one thing though. We have found a detector ghost, hitherto unsuspected. It arises by reflection from the left detector, say, then a zeroth order reflection from the right grating followed by one from the left grating and thence to the right detector. Murphy's law means that all the angles are within a degree or two of the theoretically perfect geometry to do this.

(The "right" grating belongs to the "right" camera but is actually on the left.)

The good news is that we have made measurements and modelled the situation and are satisfied that baffles can be made to prevent this ghost from causing any harm. The baffles will be made and installed before the instrument ships.