I ordered a folded cavity spacer and related parts from NAKAO SEIKI.
They will be made of invar (IC-DX) which can be used under cryogenic temperature.
The drawings are uploaded on Wiki page (explanation are written in Japanese though) (https://gwpo.nao.ac.jp/wiki/CryogenicThermal/ActivityNAOJ/Spacer).

[Aritomi, Yaochin]

We measured squeezing and anti squeezing with new faraday to estimate loss and phase noise.

Attached picture shows the result. Loss is 25.4% and phase noise is 21.4 mrad. Compared with previous measurement, we have 4.5% more loss. We should have 3% more loss from faraday and HWP, but we should have ~3% less loss from dichroic mirror (entry 1613). So this 25.4% loss is higher than we expected. Maybe we have worse visibility after installation of faraday. We'll check visibility tomorrow.

If I remember correctly, after improving reflectivity of dichroic mirror, squeezing level didn't change. It is also better to check the reflectivity of dichroic mirror again.

Good news is that even though we turned on lasers today, we have only 21.4 mrad of phase noise and squeezing spectrum is very flat with new faraday.

I tuned the lens position and alignment, then the diffracted beam power became 5.0mW which seemed to be enough (input power was 5.7mW).
Then I installed a PBS in front of the AOM, and adjusted the alignment.
At this moment, still the diffracted beam has 5.0mW.

Tomorrow, I gonna install convex lens and re-align the double-pass AOM.

Why the correction saturates when it becomes negative and not when it becomes positive? 

I changed the lenses' position to modify the beam profile for HOMs, especially to make the beam width larger.
The target beam width was 500um though the initial beam width was ~200um.
To ahieve this beam size, I changed the lenses and their position as attached by monitoring beam profile.
I put f=-75mm and f=200mm lenses instead of f=-50 and f=150mm ones.
Then the beam width around AOMs became about 500um and it was beam waist.

After that, I played with STMs and maximize the diffraction efficiency (but not adjusted the alignment of AOM itself).
Eventually, the input power to AOM was 5.7mW and the 1st order diffraction power was 4.5mW which corresponded to ~80% diffraction efficiency.
Actually, the lens position was slightly disturbed when I clamped them.
So I will tune their position and adjust the alignment to maximize diffracted beam power as a next step.
Then I will put a PBS, QWP, and mirror to compose double-pass AOM configuration.

I installed the another AOM (S/N:149256) as attached to check its diffraction efficiency with larger beam size compared to previously installed one.
The input beam power was 5.5mW and diffracted beam power was 3.5mW though the RF level was not tuned.
So the diffraction efficiency can be enhanced by larger beam size.

Then I put a concave lens in front of the AOM previously installed whose efficiency was about 50%.
Actually the efficiency was increased to 60%.

I gonna increase the beam size at AOMs to modify the position of lenses to enhance the diffraction efficiency (according to spec sheet, about 85% can be achieved).

[Aritomi, Yaochin]

We removed old faraday and aligned IR to filter cavity. Current mode matching is 86% as follows.

We found that IR transmission with parametric amplification is more stable than before as attached movie. We also know that IR transmission is stable without parametric amplification. So IR fluctuation is related to parametric amplification and replacement of faraday. We suspect that IR fluctuation we had before came from back reflection although we have two faradays. Isolation ratio of new/old faraday is almost same (38dB), but new faraday is closer to OPO and this may reduce back reflection from optics between old and new faraday.

We'll make GR and IR overlap next week.

[Aritomi, Yaochin]

We measured loss of new faraday. We made BAB top on resonance of OPO by hand. Loss of new faraday and HWP and f = 100 mm lens is 3% as follows.

We also measured loss between PBS and PR chamber. Loss between PBS and PR chamber is 1.3%.

I attach a video of CC2 loop correction signal when squeezing is reflected by Input mirror.

It seems that the high frequency oscillation (from suspension pitch, due to beam miscentering on suspendend optic?) are not large enough to saturate the actuator, but probably 1 Hz pendulum motion does.

Would an increase of a factor 3 in piezo correction that we could gain by exploiting the whole piezo dynamic be enough to avoid CC2 unlock?

Aritomi, Yaochin and Yuhang

As suggested by Matteo, we checked the correction of the CC2 loop. The check is done before we replace the new Faraday isolator.

Actually, this correction signal surprised me because it evolves in very low frequency and the correction is quite large. I took two segments of time, in the second segment, we see the correction signal even saturates.

Simon, Pengbo

Today we covered both side of the sample OSTM (Sigma koki) with the first contact, and then put it on the sample holder.

After that, we started the polarization measurement of the KAGRA #7 sample. First,  we found an arrow on the barrel and some dirty marks on one side, so we used optical tissues with alcohol to clean it. Then, we put the sample on the holder with the arrow at the top, and start the p-polarization map.

Aritomi, Yaochin and Yuhang

To see the effect of better isolation on phase noise, we measured the CC loop phase noise just before and after the replacement of the Faraday isolator.

The result is shown in the attached figure 1. From this measurement, we could see the CC2 loop phase noise is reduced by more than a factor of 2.

In entry 1432, we measured once the phase noise and PD dark noise. From that entry, we could that dark noise is very close to the phase noise level in the kHz region. We want to confirm the situation, so we did the measurement of dark noise and shot noise again. The dark noise is measured when there is no light arriving at PD, all the others are the same(including demodulator, RF amplifier). The shot noise is measured when there is light arriving at PD, but we made sure OPO is locked at the same time. Because if we don't lock OPO, the light arriving CC1 PD has some frequency component. Also, all the others are the same setting for the measurement of shot noise(including demodulator and RF amplifier).

The result of CC1 phase noise and dark/shot noise comparison is attached in the second figure. We could see that this time, phase noise is at least 5 times higher than dark/shot noise above 300Hz. I checked the code I wrote for entry 1432's measurement, and I found the measurement of dark noise is consistent with the nowadays measurement. This proves that the CC1 phase noise is larger than before. Why the phase noise is higher now is still unknown.

Pengbo, Simon

Yesterday, we started mapping the birefingence distribution of the (yet) unknown Shinkosha substrate which turned out to have coated surfaces (apparently both HR and AR).

As reported already, due to this issue, we cannot do a quantized analysis regarding the polarization angle.
Therefore, I present here just the ratio of the S and P polarized fields which should resemble the basic distribution in any case.

As can be seen, we have a very structured map which looks like the spider-web structure we know already from Shinkosha#7 sample. In addition, we got a lot of spots with some exaggerated measurement values. Those are most likely due to the coatings and represent defects.

According to the spec sheet, the diffraction efficiency of AOM is 87.7% at 1060nm and 500um beam diameter with 1.50W RF power.
This morning, I tweaked the STMs to increase the diffracted beam power.
However the diffraction efficiency was about 50%.
I'm suspecting the beam diameter is so small that the efficiency is low.

Since I have another double-pass AOM, I will input larger diameter beam to another AOM and see the effect of beam diameter on efficiency.
Also I will check the RF power.

This entry is just a log of Today's work.
I am revising the design of folded cavity spacer.
The company will come to NAOJ on next Tuesday and I will order the spacer with final design.

After that I will upload the drawings on GWSPwiki.

[Aritomi, Yuhang and Yao-Chin]

According to the previous R&D result (entry 1616), we found some issues for old isolator (IO-3-1064-VHP) including its aperture size too small and position too far relative to OPO cavity. Today, we installed isolator (FI-1060-5SC-HP) and put closer to OPO cavity to reduce back reflection. I also checked that the input/output aperture (Ø-5mm) of isolator was larger than old isolator. Fig 1 shows new optical layout. We main installed optical elements in blue frame zone.

Because SQZ and p-pol light were different polarization after OPO cavity, we could use PBS to separate them. Thus, the p-pol light is reflected from input PBS of isolator to right angle prism mirror (MRA10-K13) and its height was 54mm as shown in Fig 2. However, the polarization of SQZ light was rotated 45 degree after isolator. We used half wave plate to rotate its polarization keeping S polarization. In addition, we also re-installed the TAMA PD's position in order to add more space putting isolator.

We also check SQZ and LO overlap by using alignment mode cleaner, moving the lens (f:100mm) before isolator to improve mode matching.

[Aritomi, Yuhang, Yaochin]

Today we installed new faraday (FI-1060-5SC-HP) between OPO and PBS to reduce back reflection from homodyne. Yaochin will report the detail. We measured frequency independent squeezing before/after installation of faraday. CC2 demodulation phase is as follows.

Attached picture shows squeezing and anti squeezing spectrum with/without faraday. After installation of faraday, squeezing spectrum got amazingly better. Yuhang will report the phase noise measurement.

Today I worked on the alignment for AOM.
The following is the procedure what I did.

1. Adjusted the lens position before AOM in order to locate the beam waist around the AOM.
2. Played with STMs and AOM stage and maximized diffracted beam power.
3. Then adjusted RF level to maximize diffracted beam power.
   At that moment, the beam power input to AOM was 8.6mW and that of diffracted beam was 4.6mW.
   So the efficiency was about 55%.
4. Then put a lens after the AOM to make collimated beam.
5. Put an iris to eliminate 0 order diffracted beam and pick -1st order one.
6. Put a mirror to make the reflection beam enter the AOM again.
7. Played with the end mirror to see the double-passed 1st order diffracted beam.
8. Eventually I found 2 beams are picked off by PBS.
   One is 0 order double-passed beam and the another is -1st order double-passed one.
   I mean the former one is frequency shifted only once, and the latter one is shifted twice.
9. Measured the double-passed beam power and it was 2mW.

Double-passed beam has the same beam path as input beam.
I think this may be reasonable for double-pass AOM configuration.

I think the next step is to optimize the alignment to maximize double-passed beam and minimize the beam jitter.
Also I should check the spec sheet to confirm the diffraction efficiency and estimate the requirement for beam jitter.

In addition, I connected the tube between scroll vacuum pump and cryostat chamber (attached).

Simon, Pengbo

Today we started the polarization map on Shinkosha KAGRA-size sample.

First, we took the sample out of the transportation box and inspected the mirror visually, we found coating on both sides and an arrow on the barrel, which we think might indicates the HR coating. (please check the attached photo). So we place the sample with the HR side facing the laser.

Then, We measured the AC and DC value under s- and p- polarization without the sample and move the sample back to the center. Due to the HR coating, we increase the laser power to recieve a measurable signal (--> 2A input current + removing ND filters.) However, After this adjustment, we cannot confirm the polarization status becasue the laser power to too high to measure. For now, we assume it was still the p polarization and ran the map.

Apart from that, we did a test about the output beam. We put another HWP in the output beam and adjusting so that it is maximized in p polarization.

As can be seen, the results did't change so much.