NAOJ GW Elog Logbook 3.2
I finished the English version of the manual for the back-scatterometer.
It can be downloaded from the JGW document server (I and Torii-san are the authors).
Members: Tatsumi, Manuel
Measured the bulk reference sample: Pump_power=30mW, Chopper=430Hz; AC=74mV; DC=3.94V; Abs=1.16/cm; --> calibration factor R=0.55W-1. fig.
Measured the surface reference sample: Pump_power=30mW, Chopper=430Hz; AC=380mV; DC=4.75V; Abs=0.22; --> calibration factor R=12W-1. fig.
Both values of R are consistent with the calibrations we made last year (considering a repeatability of 5-10%).
=================================================================================================================
Measured the shinkousha little sample (1.5"diameter) with pump maximum power 9W.
Scan along z axis: fig. Today's measurement looks noisier because I didn't apply the filter
Resolved map: fig. It show the same unhomogeneus pattern al last time we measured it.
=================================================================================================================
Measured the 2 sapphire tama-sized samples Hirose-san sent us few months ago:
the translation stage is too weak to bear 100mmx60mm samples, so we tried to put a height tunable jack on the table and the new mirror mount over it. Unfortunately, because of the translation stage position, there was not room enough to to reach the center of the sample. So we changed the jack with some heavy blocks and a V-shape. Then we could reach the center of the sample. We measured the probe AC noise and we found that with the probe passing through the sample, the noise (~50microV) is double respect to the noise without the sample (25microV). This means that a large part of the noise is due to the sample vibrations. So we put a rubber sheet below the block. But the noise doesn't change considerably. fig.
So we have ~50microV of noise level, which correspond in ppm/cm for sapphire at 9W pump power to ~50e-6V*3.34/6.5/9W/0.55W-1 = ~6ppm/cm
The specs say 5ppm/cm for silica at 1W, which for sapphire (3.34 correction factor) and 9W, correspond to about 2ppm/cm. So we have a noise level higher than what the specs say.
To measure the signal I take the mean of the AC signal (lock-in time constant 1s) over 100s
| AC mean 0W | AC mean 9W | absorption | |||
| [microV] | [microV] | [ppm/cm] | |||
| sample#1 | point A | 35 | 43 | 1 ± 4 | |
| sample#1 | point B | 37 | 53 | 2 ± 4 | |
| sample#1 | point C | 36 | 78 | 5 ± 4 | |
| sample#1 | point D | 43 | 49 | 1 ± 5 | |
| sample#1 | point E | 42 | 79 | 4 ± 5 | |
| sample#2 | point A | 37 | 48 | 1 ± 4 | |
| sample#2 | point B | 44 | 51 | 1 ± 5 | |
| sample#2 | point C | 35 | 43 | 1 ± 4 | |
| sample#2 | point D | 46 | 16 | -3 ± 5 | |
| sample#2 | point E | 44 | 30 | -1 ± 5 |
The point positions are shown in the picture. As best as I could do with this setup, the positioning has a 3mm error.
The measured absorption value is always below the noise.
On friday afternoon Flaminio and me noticed that the chopper frequency is quite unstable, about 5Hz of standard deviation. The old one was not that unstable. We should check it.
Two days later, I found that the wheel was stopped.
Therefore, I and Manuel replace the new motor with chopper also.
I made a simulation of the temperature distribution on a GaAs sample of 0.4mm in thinckness and 5% of absorption.
First with the absorption distributed on all the substrate (bulk) and then only on the coating (surface).
Plot the map of the temperature, as a function of the depth and of the radius, for both cases (bulk and surface).
Used the temperature distribution in the optical simulation with a 1310nm wavelength probe.
Simulated a scan along the sample for both cases and found the same value of the AC/DC PD signal.
I and Akutsu-san went to Hashimoto today and insected the final results on the Doughnut Baffle production.
So far, it seems that the results are OK and within the constraints of bKAGRA. However, due to the baking procedure, scratches appeared on at least on side of each baffle. Yet, they will probably have no big impact on the scattering, according to the simulations.
The results plus some pictures taken at the facility, are summarized in a JGW document.
Because the mechanical chopper on absorption measurement bench was broken,
we bought new one (Stanford research instruments SR540).
Today I repalce the controller (motor driver with local oscillator) of the chopper.
Old chopping motor and wheel are unchanged, since we want to keep the setup
as much as possible.
After the replacement, I checked all of cable connections.
And then, I found that many cables are unplugged.
Tomorrow I will check the absorption measurement system by using reference samples.
On friday afternoon Flaminio and me noticed that the chopper frequency is quite unstable, about 5Hz of standard deviation. The old one was not that unstable. We should check it.
Two days later, I found that the wheel was stopped.
Therefore, I and Manuel replace the new motor with chopper also.
Finished mounting the part of SHG crystal. Next step is to build a Fabry-Perot Cavity outside the crystal.
All the details will be showed in the attachment below.
We tested the correspondence between the laser output power and the beam size.Then tested the beam profile to get the exact position of lens.Detail shows in the attachment.
But for now the beam size we got is not small enough for the SHG and the transmittance is a little bit low, so today we will reset the position of all components, add one more lens, mount the SHG and try to get the green light.
Today I installed an optical lever to sense the motion of the suspended input mirror of the filter cavity. Such lever uses a lens and a pair of PSD respectively placed in the focal plane and in the image plane. In these scheme the PSD in the focal plane is only sensitive to mirror tilts while the one on the image plane can only sense its shift. A picture of the optical lever is shown in the figure attached.
A first injection of white noise has been performend in order to excite mirror resonances. After identifying the degree of freedom (tilt or shift) at which each resonance belongs it could be possible to fine-tune the PSD positions to optimize the signal decoupling.
I measured the beam profile of probe and pump near the crossing point.
Set a vertical blade to the translation stage and move it from PCI software.
Read data on the power meter and fit with erf function.
This table is a summary of the results.
| Probe | Pump | |||||
| z (mm) | y_0 (mm) | halfwidth (um) | y_0 (mm) | halfwidth (um) | ||
| 0 | 37.5 | 178 | 39.75 | 172 | ||
| 20 | 39.7 | 145 | 39.85 | 48 | ||
| 30 | 41 | 126 | 39.92 | 70 | ||
| 50 | 43 | 94 | 40 | 172 | ||
The probe waist is about 94um (radius at 1/e^2) but it is not near the crossing point. At the crossing point the beam size is about 290um (diameter).
The pump waist is about 48um (radius at 1/e^2) and is near the crossing point. At the crossing point the beam size is about 100um (diameter).
The angle between pump and probe is 0.105rad, or 6°.
The PCI manual said beamsizes of 200um and 70um. Which were the values I used until now in my simulations.
I will put those new measured parameters in the simulations.
We tested two properties of the laser will be used in the squeezing light experiment:
1. The correspondence between the current of diode and the output power of Laser
2. The position of the laser beam waist
The reports will be shown in the attachment.
Members: Eleonora, Manuel, Tatsumi, Yuefan
Some preliminary work has been done to prepare the installation of the oprical lever for controlling filter cavity input mirror (already suspended in NM2)
Mirror check
The suspendend mirror has been checked after the earthquakes of the past weekend. No damages have been found.
Coil drivers connection
The four coil drivers to be used to move the mirror have been connected.
PSD electronics
The electronics necessary to power and to read signals from the two PSD need for the optical lever has been assembled and checked. (See picture attached)
PSD calibration
PSD have been calibrated by shining a laser to them and looking at the change in the output voltage when gradually varying their position in x and y direction.
In the plot attached the voltage value has been normalized using the total output value (SUM). The (normalized) calibration value found is about 6.6*10-3 m/V
Laser check
We checked the dimension and the beam power of the laser to be use for the optical lever. Dimensions of the beam (measured with a beam profile after 30 cm) are about 1.1 mm. The recommended dimensions for the PSD are beetween 0.2 and 7 mm.
We also checked the PSD output as a function of the laser power and we observed that the signal saturates for a power higher than 0.15 mW.
The three mirror mounts has been delivered, unboxed and checked.
One mount for Kagra-size mirrors (ø220mm x 150mm)
One mount for Tama-size mirrors (ø100mm x 60mm)
One mount for 2inch samples (and smaller with adaptors that we already have).
All parts go well on the large translation stage and the stage can move without any obstruction.
-----------------------------------------------------------------------------------------------------------------------
The large translation stage, has been disassebled and packed.
Today the company person pick it up to clean it and to grease it.
I simulated the signal of the absorption system, in order to be able to compare the measurement reported in this post .
The reference sample is made of silica and absorbs 116%/cm. Thickness is 3.6mm. Chopper frequency 400Hz.
I attach an animated gif to show three aspects:
1. The colored image represents the temperature distribution inside the bulk reference sample, heated by the horizontal pump beam. The shadow represents the probe beam at an angle of 7° to the pump.
Moving the sample along the z axis (along the pump beam) make the crossing point of pump and probe to scan the depth of the sample, and it is equivalent to move the probe vertically respect to the pump.
2. The interference pattern is the difference between the intensity of the probe passing through the heated sample and the intensity of the probe not passing through the sample. This is because the heating is modulated. The signal is given by the integral of the interference pattern in the area of the photodetector. The area of the photodetector is represented by the black square.
3. The plot is the calculated signal from the lock-in as a function of the sample depth. The 0 is when the pump-probe crossin point is at the surface of the sample.
Members: Takahashi, Tatsumi, Guo, Manuel.
We suspended a temporary mirror in the NM2 vacuum tank.
Set the earthquake stoppers at about 1mm from the mirror.
Set the 4 coil actuators coaxially aligned to the magnets.
Closed the vacuum tank.
I made glueing of four magnets on a LISM mirror.
See attched PDF file.
Page 1: Magnet polarity check
Page 2: Anti-Reflection coating check by green lase pointer
Page 3: Overall view of the glueing JIG
Page 4, 5: Closer view of Magnet clamping parts
See JPEG file:
This is TAMA traditional magnet configuration.
I also follow this.
Members: Tatsumi, Guo, Manuel.
We moved and clean an optical table in the central room of TAMA. The attached image shows from where to where.
I and Torii-san finished writing a manual for the JASMINE scatterometer in the ATC.
It is uploaded on the JGW website and available in Japanese and English.
I study on modulation frequency for SHG cavity.
Because the cavity is very short (30mm or less),
it has FSR of 4 GHz and has a finesse of 75.
These correspond to resonant width of 53 MHz.
By using FINESSE simulator,
I checked the PDH signals for each modulation frequency.
fmod merit demerit
600MHz non-resonant sideband modulation frequency is too high
15MHz TAMA used this freq. The PDH signal get smaller by a factor of 10.
I think that modulation frequency of 15 MHz is acceptable for us.
