Today Daniele, Ronic and I cleaned the mirrors and locked the cavity. However, the finesse was only 13,000 because of the not clean enough environment and not pure enough alcohol and water.

We will carefully clean the environment, clean the mirrors again with pure alcohol and water and measure the finesse when I return. If it doesn't work, we will use plasma to clean the mirror. We have gone to the lab to confirm the plasma device and then we will study the best parameter settings: polarity, time, and current.

Have a nice weekend!

- Today we cleaned the environment and put the spin coater and microscope inside the airflow.
- Tomorrow, Daniele and I will clean the mirrors one by one using pure alcohol and water, and measure the finesse each time. If it does not improve, we will clean them with plasma.

Today, Daniele and I cleaned mirrors one by one using pure water, alcohol, and the spin coater. Here are the measurements of finesse each time:

1. Initial value: 14,076

2. Clean Mirror 2: 20,606

3. Clean Mirror 4: 18,750

4. Clean Mirror 3: 18,762

5. Clean Mirror 1: 18,563

6. Reclean Mirror 4: 15,226 (unstable lock)

7. Reclean Mirror 4 again: 16,563 (unstable lock)

The finesse reached a maximum of 20,606 but finally was down. For the last two measurements, the locking state was unstable and noisy. Tomorrow we will optimize the locking status and re-measure.

- Today, Daniele and I cleaned the cavity inside, recleaned the M2 and M4 and their mounts, optimized the locking, and the finesse is now about 25,000.

- Although it's lower than the expected 30,000-40,000, we decided to move on to the next step. In addition, the mount of M4 is near the end of the tuning range and may cause instability at high power.

- We adjusted the cavity length to match the repetition rate of the pulsed laser, and the FSR in air is 160.265 MHz.

- Tomorrow, we'll turn on the vacuum and use the pulsed laser to get resonance.

These days, Ronic, Daniele and I achieved stable cavity locking with the menhir pulsed laser.

- After vacuuming, the current cavity finesse is now about 23,000. The diameter of the cavity mode is w_x=2.2mm, w_y=2.7mm.

- We had to compensate for frequency drift by manually adjusting the cavity length to keep locking.

Now the problem is that CEP's compensation range is not enough. The laser CEP is drifting from day to day. We adjusted the CEP by tuning the pump current of the menhir laser, but the adjustment range was not enough.

- The laser pump current is varied in the locking state and the variation of repetition rate is recorded. The current range is 850mA to 950mA and the repetition rate changes by 24 Hz. The calculation process is shown in Figure 3.

- By calculation, the variation of CEP caused by the variation of laser current is only π/2, which we hope is 2π.

- For Gamma Factory, the target FSR is 40 MHz, so the 4-pulse stack provides 4 times CEP tuning range to meet the requirements. But for our experiment, it seems not enough now.

The next step is to evaluate the gap to the maximum gain and draw the curve of CEP. Then we will discuss solutions.

Here is a simulation of the relative FP-cavity gain vs the CEP for a Finesse of 23000 and taking into account the Menhir laser optical spectrum and several CVBG parameters.

I added the commented Matlab code to produce this plot.

Last week, Ronic and I focused on CEP measurements of the menhir laser.

1. Measurements without Cavity Locking:
   • Direct measurement of repetition rate (Frep) with a spectrum analyzer. Altering the laser pump current from 950mA to 850mA, Frep changed by +28Hz.
   • Measurement of the variation of carrier-envelope frequency (Fceo) by beating with CW laser. Altering the laser pump current from 950mA to 850mA resulted in a beating frequency of n0*dFrep + dFceo = +/-2.4MHz, so dFceo ~ 50MHz.
2. Measurement with Cavity Locking:
   • Maintaining cavity locking, we changed the laser pump current and AOM frequency to record the transmitted power of 5 consecutive fundamental mode (TEM00) resonances.
   • The pump currents were set to 850 mA, 900 mA and 950 mA, and the AOM frequency were set to 210 MHz and 250 MHz. We then plotted the measured transmission amplitude values against the theoretical gain curve (see Figure 1).
   • By adjusting the CEP, we reach the top point on the curve, which is the maximum gain. At this point, the coupling frequency increases from 10% to 50% (see Figure 2).
   • We observe that a 100mA change in pump current adjusts the CEP for pi/2, while changing the AOM frequency by +/-40MHz adjusts the CEP for pi. In summary, our CEP tuning range is about 3pi/2 (130 MHz) - not the full 2pi, but still probably giving us maximum gain.
3. Next Steps:
   • Investigate factors associated with changes in CEP, such as laser temperature or pressure.
   • Discuss with Menhir the feasibility of expanding the laser pump current adjustment range (now limited to 100mA).
   • Optimize AOM frequency and locking status, connect the amplifier.

here is the code to get this last curve

In the last two days, Ronic and I connected the amplifier and locked the cavity.

- We installed an iris on the output to remove a part of the pump.

- We turned on the second stage of the amplifier. When locking, the injected power is 220 mW and the transmitted power after M2 is 26 uW.

- Low gain and coupling efficiency due to bad mode matching and CEP.

Next steps:

- Turn on the third stage of the amplifier, measure the beam parameters, and adjust the telescope.

- Check the adjustment range of AOM frequency that enables the amplifier to operate safely.

- Measure consecutive fundamental mode resonances to determine the direction of AOM frequency tuning.

1st version of the document.
if some information is not correct or missing, give any comment by replying to this post.

- These days, Ronic, Fatematuj and I measured the beam parameters of the output of the third-stage amplifier.

- We used 2 wedges and reflection filters to reduce the intensity on the CCD.

- We measured multiple points at pump current of 2 A (output power ~10 W). The waist diameter of the output is w_x = 792.26 um, w_y=873.90 um.

- The next step is to design the telescope and improve the coupling efficiency.

Today, Ronic and I installed the new telescope and locked the cavity.

- We locked at the amplifier current of 1 A and obtained 32% of coupling. (see Figure 1)

- The telescope was designed for a current of 2 A (output power ~10 W). To inject this power, we need to add some filters to devices.

- For CEP tuning, when we changed the AOM frequency while cavity locking, sometimes it caused unlock and power drops. It will be dangerous in high-power cases. So it's better to optimize the AOM frequency in low power and just tune the laser current in high power. Now the current variation range of the menhir laser is 750mA to 950mA.

- Today, Ronic and I locked at the amplifier current of 2 A and obtained ~60% coupling after optimizing the CEP (see Figure 1).

- The injected power is 10 W at 2 A. We measured only 14 kW inside the cavity, which corresponds to an effective gain of 1,400 and a full gain of 2,300. The cavity finesse is 23,000 and the normal gain should be around 6,200.

- We found fluctuations in transmission, possibly because of mode degradation. Tomorrow we will use D-shape mirrors to suppress high-order modes and optimize alignment and locking.

Today, Ronic and I optimized the locking at the amplifier current of 2 A and obtained ~ 21 kW inside the cavity.

- When all the iris open, the injected power is 10 W and the coupling is ~ 40%, corresponding to an effective gain of 2,100 and a full gain of 5,250. But the coupling may not be the true value because there is a large spot around the output beam.

- We have optimized the CEP, alignment, D-shaped mirrors and locking state. We optimized alignment after leaving the iris open and the inside power went from 14kW to 21kW.

- The transmission and reflection signals both have some same fluctuations, and they seem to come from the cavity. It's possible that the over-angled mirror mount could be the cause, but not sure. We will check in different power and see the stability of the signal.

- In addition, we found that the design values of the mirror incidence angles for the SBOX (3.359°, 5.900°) are different from the mirror ratings (1.146°). This may result in parameters such as reflection and transmission being different from the datasheet. It will also change the estimated maximum finesse, gain, and power inside the cavity. It might be better if the mirror parameters could be recalculated based on the actual angle of incidence.

These days, Ronic, Aurélien, Fatematuj and I have been doing some tests on polarization issue, trying to see if it is possible to obtain higher gain under other polarization conditions.

• We installed an additional half-wave plate + PBS + PD at the transmission. By rotating the waveplate of the injection laser, we can compare the resonance signals of single and full polarization. Figures 1 and 2 demonstrate this comparison. The yellow curve is full polarization and the green one is single polarization. The intensity ratio of the different polarizations is unstable in the open-loop state.
• Based on PZT scan frequency = 3.1 Hz, amplitude = 10 V, sensitivity = 3.7 Hz/V, time difference between two polarization peaks = 100 us, we can calculate △Frep = 10mHz and △v = 42kHz, which means the frequency variation between two polarizations. We see two polarizations only at the main resonance.
• By the way, we found two spots behind the M3 window (see Figure 3) and the power of both is related to the intra-cavity power. We moved the position of D-shaped mirror and the second spot became weaker and larger like mirror's edge. Maybe the D-shaped mirror is causing a part of laser to be reflected through the window, but it's unclear exactly how the optical path works.

Next steps:

• We will lock the cavity in different polarization and see if there is higher gain.
• We will move D-shaped to the maximum and see if the second spot disappears.
• We will check the coupling value and try to optimize the telescope using adjustable stages.

just to add some details :

about the S and P polarization frequency shift:
the PZT scan is 10Vpp at 3.1Hz => the slope is 62 V/s because of the triangle shape of the PZT scan.
so 100µs of separation of the 2 polarization is equivalent to 6.2mV on the PZT.
as the PZT sensitivity is 3.7Hz/V on Frep, the separation of the 2 polarization is equivalent to 23mHz on Frep.
△Frep/Frep = △v/v =>  △v = 41.8 kHz

about the possibility to separate S and P polarization states on secondary resonances:
the total spectral width is ~2nm which is equivalent to 565GHz and contain about 3500 laser harmonics at 160MHz.
the central spectral harmonic is roughly the number n0=1.82M, so with the first secondary resonance condition, the Frep/FSR detuning corresponds to n0*Frep = (n0+1) FSR
so (Frep - FSR) = FSR/n0 ~ 88Hz.
then if the central frequency, related to n0, is on an S-polarization resonance, the harmonics at (n0+475) will be on the P-polarization resonance and so on...
the conclusion is the power detected by a photodiode on secondary resonances are a mix of S and P polarizations (if the laser input beam is also a combination of S and P polarizations)
and we cannot make an observation of different peaks with different polarizations in transmission of the FP-cavity.
for that, it is mandatory to be on the main resonance with Frep = FSR (CEP~0) as condition of resonance.

Today, Ronic and I lock the cavity on the other polarization and achieved 50kW with 22W injection.

- After we optimize and lock the cavity in vertical polarization, we rotate the waveplate at the transmission to minimize the signal, then rotate the waveplate at the injection laser to maximize the signal. We lock the cavity in this horizontal polarization and optimize CEP and alignment. The results are: when injected at 10W, the circulating power in vertical polarization is 21.4kW and in horizontal polarization is 23.3kW. The coupling are both ~30%.

- The cavity reflection signal obtained in horizontal polarization is weaker than that in vertical polarization (1/10), so we usually lock on vertical polarization first.

- We then increase the power to see the change in coupling. From current 2A to 3A, coupling change from 30% to 40%. Finally, we obtained a circulating power of 50kW with 22W injection (3A current). In the initial stage of locking, high-order modes appear, but in stable locking, there is only fundamental mode and no mode degeneration. Although there are many fluctuations in transmission and reflection.

- Tomorrow, we will optimize the coupling and add removable stages under the telescope.

These days, Ronic and I achieved 200kW inside the cavity and 70% coupling efficiency.

- By optimizing the telescope, the coupling reached 70% with iris fully open and maintained 60%-70% coupling at high power.

- The cavity mode went from 2.2mm,2.5mm (38kW) to finally 2.3mm,2.8mm (200kW) without changing a lot.

- Gradually raising the power while optimizing alignment, CEP, and locking, we got the following stable power:

- Next steps:

• Explain the strange drop phenomenon that occurs at high power, where both transmission and reflection drop, as in Fig. 2.
• Maintains a half-hour locking at 200kW. Now the temperature of the amplifier at 8A is over 40 degrees, which may be risky.

all the injection power in the chart have not been measured recently but during the Loic thesis period.
and these old measurements stopped at 5.5A of pump current.... so, the data at "8A" is a pure estimation.

about the last measurement :
it was made at 6A/8A/8A/8A for the 4 pump diodes of the amplifier (because 1st stage has a Peltier issue and we cannot check its temperature), so the average current is 7.5A instead of 8A.
and the linear scale between pump current and amplifier power is ~ 12W/A, then the estimated amplifier power for the last measurement is 76W instead of 87W
and the estimated gain is more 2658.
for this current, the amplifier works out of its nominal limits (temperature set at 25°C but measured at 30°C !!!)  and the fans of the crate are making noise like hell.
so the last gain estimation should be treated very cautiously.

about the transmission and reflection signals behavior, one can write :
R + T + L = 1 => energy conservation for the cavity.
dR + dT + dL = 0 => dL = - (dR + dT)

if dX = Xfinal - Xinitial, dR and dT are < 0 on the last picture, then dL > 0.
it means that this picture seems to show that some losses are increasing from the beginning of the locking process.

several possibilities :
- we saw a strange D-shape effect on the large port of the cavity.
it seems that one of the D-shape mount/mirror is touching the intra-cavity beam producing some ghost effect on this large cavity port.
some cavity axis changing during the beginning of the lock could introduce some additionnal losses.
it can be easily tested by puting the D-shapes far from the beam.

- because of cavity axis changing at the beginning of the lock, the mirror losses are different.
but it is surprising that it is still going in the same direction... more losses at the end.
could be tested by slightly changing the optical axis of the cavity.

- "prior damage" behavior with a bump in the middle of the mirror due to thermal effect which introduces some losses at the end.
=> if it's the case, it's not a good behavior !!! :-(((
can be tested by looking at the wavefront phase in transmission.

- Non linear effect is the coatings.
but the field density seems not so much to produce this kind of effect

- A thermally induced change in the refractive index of the mirrors.
Daniele mentionned a relation between real and imaginary (related to absorption) parts of this refractive index which could explain that a reflectivity change could induce an absorption change.

Today, Ronic and I achieved 272kW inside the cavity at 7.5A. The coupling maintained 60%-70%.

- Compared to yesterday's experiment, we moved the position of the D-shaped mirrors farther in two directions to make the higher-order modes just disappear.

- Possible reasons for higher gain: D-shaped mirrors position, high power and pump vacuum cleaned cavity mirrors so that improve the finesse.

- We didn't see the strange drops like yesterday (Figure 1). However, in the window behind the M3, we can see 3 spots correlating with the intracavity power, even though moving the D-shaped very far does not make them disappear, only weakens them. We don't know where they came from. When this round of experiments is over, we can open the cavity and observe the optical paths.

- Next steps:

• Repeat the experiment to ensure that the gain does not drop.
• Long-term measurement at maximum power when the amplifier temperature is safe.
• Measure the transmittance of the cavity mirrors and the amplifier power.
• Open the cavity and observe the optical paths and the mirror surface.