Keita, Dripta, Georgia
Today we continued investigating the poor mode matching of the ALS-Y green beam, following on from work eariler in the year. We took some new photos and used the wincam camera profiler to look at the beam. We haven't solved the problem but we noted that:
- The beam is entering BSC-10 close to the edge of the viewport
- Keita will post a photo
- This could be affecting the polarization of the beam
- Next week I will go to End-X and compare the beam position on the viewport there
- The beam on the transmission monitor primary mirror (as viewed through the viewport for the IR periscope on ISCT EY) seems pretty well centered...
- Keita will post a video
- The primary is the smallest aperture-to-beam-radius ratio in this path -- being centered here suggests we aren't clipping on the other transmission monitor optics
- ...Though there is still a lot of scattered green light visible in the chamber
- Dripta will post some photos of the transmon as viewed from the camera viewport, in which the secondary mirror is visible.
- The beam reflected off the test mass which arrives back at ISCTEY has a bad mode shape (not just elliptical)
- We used the Hartmann path to view the beam reflected off the test mass by putting a half wave plate just after ALS-M11 (see this table layout) to divert extra ALS light towards the Hartmann sensor. We then put the beam profiler just downstream of HWS-L1, see profile in first attachment.
- We also noted there was a bright ALS beam going down the HWS path even without the temporary half wave plate, which suggests there is something on the path to the test mass converting linear polarization to elliptical. This beam had a very ugly mode shape.
- We also put the camera profiler on the reflected path after the Faraday isolator, between ALS-M17 and ALS-BS6, see second attachment.
- As noted in earlier investgations, the beam entering the chamber is elliptical but not severe enough to explain the mode mismatch
- See third attachment
- There are ghost beams of the beam reflected off the test mass on both the ALS and HWS paths
- ALS-HWP5 is tilted and there is a bight reflection off it that hits the IR periscope
- The ALS beam leaking on the HWS path is reflecting from a lens at what looks like the position of HWS-POL in the table layout diagram, leaving two ghost beams
- These ghost beams aren't great but are probably unrelated to our mode matching problem
Next week we might monitor the profile of the reflected beam while adding offsets to the QPD servo, which will change the beam position on the transmon optics, to see if there is some clipping.
1. Viewport beam spot position
1st and 2nd picture show the beam position on the viewport. In the first picture you can also see the two AR ghost beams hitting the inside of the duct material.
That the beam is close to the edge of the viewport will NOT affect the polarization but could affect the beam shape as the edge of the viewport is not really flat and could act as an irregular lens. We don't know how much the effect is at the current beam position, though.
(In the past, Pcal team found that excessive amount of aberration observed in ITM pictures taken with large aperture telescope (see e.g. the last page of https://alog.ligo-wa.caltech.edu/aLOG/uploads/35327_20170404172516_IrLowPower_reduced.pdf for example) is not observed in their tests outside of the vacuum chamber, and that it is mitigated/eliminated by simply limiting the input aperture of the telescope, which means that the viewport surface is not flat at the edge.)
2. In-chamber TMS optics spot position
See this one VID_20191018_170216.mp4 on G1902047 showing the primary (big mirror at the bottom) and a flat mirror above the primary. Centering on flat mirrors doesn't matter that much. Anyway, it may be a bit low on the primary but I don't call this bad.
3. Green Polarization
It's only a minor problem for capturing the beam shape of the return beam (and a minor problem for ALS in general), and the first picture of Georgia's entry is entirely valid, but it might be somewhat interesting to think about the cause of wrong polarization.
When we placed a HWP upstream of ALS-M11 in D1400241 to minimize the transmission of the incoming beam on ALS-M11 (i.e. wrong polarization), the brightness of the return beam (i.e. coming back from the chamber) transmitted through ALS-M11 was still much brighter than that of the wrong-polarization incoming beam (and it didn't change much when we rotated the HWP because ALS-M11 is a PBS and we're not changing the polarization of the beam going into the chamber, just the power).
That the return beam in wrong polarization is brighter than the input beam in the wrong pol cannot be explained by geometric rotation of the polarization axes alone. If it's just geometric rotation, what is sent in comes back with the exact same polarization because ETM is retroreflecting and the return beam exactly follows the incoming path, cancelling the rotation effect.
It could however be explained by a combination of geometric rotation and polarization-dependent optics in chamber. For example dichroic mirrors are spec'ed for green S but not P (see e.g. E1000669). Green BSs seem to be spec'ed for P and S (e.g. E1000870), but no test data is supplied for P. Even though the ALS periscope on the ISCTEY (and ISCTEX) is designed to elliminate the geometric rotation such that S pol on ICST stays S pol on TMS, it cannot be perfect.
Attached are the pictures Dripta took of the transmon viewed from the camera viewport. There is a lot of scattered green light.
- The large mirror with the well-centered spot is a flat mirror (TEL-FM1 in figure 2 of T0900385)
- The small mirror with the bright spot above it is the telescope secondary (TEL-M2, -200mm ROC). It's hard to tell but perhaps this is not well centered.
- The right viewport with the green spots around it is the viewport we inject the green ALS beam from
- The final picture also shows some of the steering optcs and lenses on the transmon which steer the beam onto the QPDs.
From alog 51458 we know that the 00 mode is only about 70% of the total power.
If we want to explain this using some kind of weird lensing on ISCTEY of on the viewport, the lensing needs to be large. Just to have some understanding of how large, here's a calculation of the effect of the elliptic lensing of the viewport.
I assumed that the lens is cylindrical so the mode shape is only altered in one direction, and plotted the mode overlap of the lensed beam and an ideal beam in power, not amplitude, as a function of the lens focal length.
To degrade the power coupling to 70% just by using a cylindrical lens, you need the focal length of about 14 meters. That's a huge effect, unlikely to come from high quality viewport D1101006/E1100267 as the transmission wavefront error is specified to be within lambda/10 for 633nm over the clear aperture of 5.2".
