Summary: Just after we opened up the chamber at EY, I got the chance to take photos of the P-Cal periscope from the point of view of the ETMY beam spot. The figures show that, in addition to glints associated with the camera mirrors, there were also glints from the periscope structure. A similar glint might explain why removing the mirrors at LLO did not completely get rid of the peaks in DARM produced by the periscope, and the glints from the structure motivate the work we are undertaking to baffle the structure. In addition to baffling the periscope structure, we may also need to baffle or angle the P-cal beam port.
The sources of the glints are finishing groves, which retro-reflect light where parts of their walls are normal to the direction to the beam spot. For linearly symmetric structures, this can happen only where the structures are tangential to circles around the specular reflection point, the place where a plane containing the line segment is normal to the path to the beam spot. I demonstrate glint control for such grooves by sanding samples in the radial direction, and suggest similar ways to control glints from other linear structures like corners, edges and folds. Such glint control might improve baffles or reduce the need for baffles in future upgrades.
Glints that retro-reflect scattered light are important because they can greatly increase scattering noise. Figure 1 shows the P-Cal periscope and indicates the various structures on the periscope. The expected glints from the two camera-periscope paths are visible and, in addition, there are a couple of glints from the support structure that are indicated. The two smaller images on the right side of the second page show that the major glint from the structure disappears when the camera and flash are moved only a few cm off-center of the test mass.
Figure 2 is a close up of one of the mirror regions on the periscope. An image of the second mirror of the ETM camera periscope is reflected in the closest mirror, and the reflection of the flash off of the camera port window and off of the flange around the window are also apparent. Periscopes are scattering dangers because they can image the test mass beam spot directly in front of and normal to the port, while the beam spot is at an angle to ports without periscopes.
In contrast with the camera periscopes, no similar glint is visible in the P-cal beam periscope mirrors. Unfortunately, the lack of glints does not completely alleviate the scattering concern for the P-cal beam periscope. The P-cal beam mirrors are narrow-band reflectors, unlike the broad-band camera mirrors, and there is a chain of 3 such mirrors before the camera flash reaches the port (note that the periscope structure can be seen through the mirrors). There has been a suggestion that the P-cal port flanges are producing scattering noise ( https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=39121 ) and I think we should baffle the flange and possibly angle the window unless it can be shown that the acceptance angle is small enough that it excludes the beam spot region of the test mass (the P-cal beams are, by design, offset above and below the beam spot on the test mass). It might be possible to check this by shining a laser in through the port and seeing if it can reach the location of the beam spot.
Figure 3 is a close up of one of the major glints seen on the first page. The concentric circles are drawn around the point on the periscope where the surface is normal to the direction to the test mass and so would reflect the flash right back to the camera located at the test mass. The glint can be seen to be in regions where the scratches or grooves are tangential to these circles. This is the case even in the upper right hand corner where irregular scratches are tangential and reflective. In the lower left hand corner of the image there is an abrupt change in groove direction and the glint ends abruptly.
The reason the glints only appear where the grooves are tangential to the circles around the point where the plane is normal, is that the walls of linearly symmetric grooves (linear symmetry in the sense that short translations along the line do not change the structure) can only tilt in directions perpendicular to the linear axis of the groove. So, surfaces of the groove can only retro-reflect at places where the linear axis of the groove is normal to the direction to the light source.
Figures 4 and 5 are images of the periscope taken when the camera was not at the ETM beam spot and show that the glints are again where the groves are tangential to the circles centered around the specular reflection point, where the plane of the surface is normal to the direction to the camera. The glints can change dramatically with only slight movements of the camera and flash as illustrated in the comparison of Figures 4 and 5. The reflection point has moved only a few cm from one side of a screw to the other.
Figure 6 shows the generality of this glint prediction, with the expected glints indicated by yellow arrows. Glints appear where the linear axis of linearly symmetric structures, such as folds, milling grooves, corners, bellows and edges, are tangential to circles around the point where a plane containing part of the linear axis would produce a specular reflection, where it is normal to the direction to the beam spot. Since many structures in the photo are in planes perpendicular to the main beam axis, this point is the same for all of the glints indicated by yellow arrows.
The corner reflection from the ACB, at the top right of Figure 1, is another example of a glint that occurs at the point where the linear axis (of the corner) is perpendicular to the direction to the beam spot.
The mechanism producing these glints suggests ways of controlling glints from baffles and in un-baffled structures. Figure 7 is a demonstration of glint control, comparing pieces that are milled or lathed to the same pieces after sanding so that the grooves could be oriented radially. This demonstration shows, for example, that if the down tube of the elliptical baffle were sanded along its long axis, it wouldn’t glint nearly as much.
Suggestions to help minimize glints for future upgrades:
The basic idea is to minimize surface regions which retro-reflect light to the beam spot because they are normal to the direction to the beam spot.
For linearly symmetric structures like edges, corners, or grooves, normal surfaces can be minimized by ensuring that the linear axis is never normal to the direction to the scattered light source, the optic.
An equivalent rule is to minimize regions that are tangential to circles around the specular reflection point of planes containing line segments of the linearly symmetric structure. This can be done by making the linear structures tend towards radial to the reflection point.
For the many structures that are in planes normal to the beam axis, the reflection point is on the beam axis, so this reduces to minimizing regions that are tangential to circles around the beam axis. Linear structures may be hierarchical, and the axes of linear substructures, like milling groves, also need to be considered. Where the structures must be tangential, glints can be minimized by overlaying radial grooves, as in Figure 7. Radial grooves (made by milling or sanding) in the glinting tangential regions of the frame of the elliptical baffle in Figure 6 would reduce or eliminate the glints.
Photographs like those in Figure 7, taken with the camera in the relative position of the beam spot, might be useful to fore-warn of glints from linear structures.
Robert S., Rick S.
