We are aiming to stabilize intensity of laser to a fluctuation range of plus-minus 1% using Acousto-Optic Modulator (AOM) for a long timescale (in the order of days). This stabilization will improve the precision of experiments that use laser beams.
Laser intensity drifts with time, and the amplitude drift could result in un-precise energy level transition, which affects energy shift induced by ac-stark shift of detuned laser. This unstable laser intensity affects the precision of Ra-226 excitation.
Our goal for this research is to find a laser stabilization method that works for any lasers with any desired intensity. Users are able to manually set the desired intensity. The proportional and integral constants in the PID controller will be able to change on its own corresponding to the input signals.
- Photo Diode: PDA10A2
- Acousto Optic Modulator: ATM-802DA1
- Laser: SYNCROLASE-100
- Voltage Controlled Oscillator: ZX95-100-S+
- Voltage Variable Attenuator: ZX73-2500-S+
- RF Amplifier: ZHL-1-2W+
First, the laser beam [1] is transmitted by the fiber and goes through the two fiber ports [2]. It then is received by the Acousto-Optic Modulator (AOM) [3], which diffracts the light to different intensity. We use the first order beam in our project. Next, the beam coming out from the AOM goes through the PBS (Polarizing Beam Splitter) [4] and the NPBS (Non-Polarizing Beam Splitter) [5]. These two beam splitters split the light and control the amount of light that goes into the ion trap [6] and the photodiode [7]. We choose an NPBS that splits 90 percent of the light and lets 10 percent of the power go into the photodiode. The photodiode is the detector of laser intensity in our project. It converts the light signal into voltage, which can then be received and analyzed by the PID program in Red Pitaya. The photodiode generates a voltage signal in the range of ± 4 volts and outputs it to the Red Pitaya. A ¼ divider [8] is put between the photodiode and the Red Pitaya to compensate the limitation that Red Pitaya can only take a voltage from -1 to 1 volt. A PID program is built in Red Pitaya [9]. It collects signals from the photodiode and compares it to a setpoint voltage to create an error signal. The Red Pitaya will then return this DC gain error signal and send it to a voltage variable attenuator (VVA) [10]. VVA receives this DC gain signal as well as an RF reference signal sent from the VCO (voltage control oscillator) [11], and it will output an attenuated RF gain with attenuation determined by the input control voltage (see the attenuation graph in the VVA section). The RF gain is then amplified by an RF amplifier [12] and sent back into the AOM to alter the intensity that goes through it.
An AOM uses the acousto-optic effect, which is similar to the Bragg diffraction, to diffract and shift the frequency of the incident light. The acousto-optic effect describes the diffraction of light by acoustic waves. The basis of this effect is a general effect of photoelasticity consisting in the change of the medium permittivity under the action of a mechanical strain. With the acousto-optic effect, the amount of light diffracted by the sound wave depends on the intensity of the sound. Thus, we use the intensity of the sound to modulate the intensity of the light.
For our project, we choose the first order beam from the AOM. The difference between each order beam is that they have different dynamic ranges. The dynamic range is the ratio between the largest and smallest values of the intensity that can be outputted by the AOM. For the zeroth order beam, the intensity can be attenuated from 100 percent to 20 percent, which provides a dynamic range from 100 to 300; while for the first order beam, the intensity can be attenuated from 85 percent to 0 percent, which provides a dynamic range from 100 all the way to infinity. Thus, we choose the first order beam for our project to have a larger dynamic range.
The intensity of the first order beam is controlled by the control signal that goes into the AOM. When the control signal increases, the intensity increases correspondingly. We adjust the intensity of the laser using this property of the AOM.
Red Pitaya is an affordable, PC based, multi instrument. It acts as the brain of our project, which takes up the role of PID controller for the feedback loop (Discussed in Feedback Loop Section). Signal coming from the photodiode is received by the DAC portal of the Red Pitaya, and after the PID program an output signal will be sent out from the ADC portal to the voltage variable attenuator (VVA).
Attenuators are transmission line components used to reduce the input power in a system by a predetermined amount. It is operated throughout its entire dynamic range. Different from fixed attenuators, variable attenuators can be controlled to vary the attenuation level of the device with a number of different methods. A voltage variable attenuator, the VVA, has its attenuation controlled by the voltage. The typical attenuation of the VVA we use, ZX73-2500+ attenuator, is shown in the figure below.
From the figure above, we can see that the attenuation of our ZX73-2500+ attenuator is non-linear, with the curve relatively linear and flat at control voltage range 0 to 0.5 volts and at high voltage. These linear regions imply that at these two control voltage ranges, the attenuation of the VVA does not change dramatically under large control voltage change. This property of the VVA helps to diminish the overcompensation of the PID program and prevent it from destroying the whole feedback loop. An overcompensation of the PID program means that the PID overcompensates for every positive error and creates an even larger error in the negative direction (and vice-versa). With a VVA in the circuit, the overcompensation from the PID feedback can be diminished and it will not cause serious failure in the stabilization. To achieve high precision, we require the input control signal at higher voltage. We choose 4-12 volts as our range of control voltage in our project.
Red Pitaya is remotely controlled using a client on a computer via an internal network. After receiving the signal, the pitaya runs a python script (discussed in Code) section, compares the input voltage to the setpoint voltage, and applies a control signal, C, which depends on the proportional, integral and differential terms of the error signal. The relationship can be seen in the equation above. We neglected the differential term for our setup as we weren’t concerned with stabilizing at high frequency (laser fluctuations tended to be low frequency). The control signal from the pitaya goes through an operational amplifier which increases the control range from ± 1 volt to 4 and 12 volts. The Voltage Variable Attenuator (VVA) section explains why we want to be in the 4 -12 voltage range. The control signal was also sent to an oscilloscope at the same time. The majority of the data was taken on an oscilloscope. The data from the oscilloscope was read by a server on the computer and sent to the red pitaya. The computer (server and client) are constantly communicating with the red pitaya during the feedback loop.
The programming was mainly done on the red pitaya using python and the PyRPL module, which has python defined classes for controlling the pitaya outputs and inputs. The PyRPL code was written onto a server, and a client was then made using that server. There was also a simpler script made for an independent laser intensity stabilization setup using an Arduino due. The code is above under “PI_Stabilization.ino”. The Arduino presents versatility as it uses C++, a widely accessible programming language. On the other hand, the red pitaya can either be programmed using the limited PYRPL module or the native language of the pitaya, Verilog, which is less accessible. Either the Arduino or the pitaya is sufficient for stabilization. Still, when it comes to pulsed feedback (discussed shortly), the Arduino is a better choice as C++ is easier to work with than Verilog.
We need to identify a clear distinction between continuous and pulsed feedback. In this experiment, we are applying continuous waveform (CW) feedback. We are continuously reading data and stabilizing it over time. The picture above illustrates a summary of our CW feedback in a rectangle on the right. Generally, in ion trap experiments, laser beams are sent to the ion trap in pulse sequences (on the order of microseconds). We can create a stabilized pulsing laser by placing a second AOM after the laser is stabilized. The second AOM will receive a signal that turns it “on” or “off,” This way, we have a pseudo pulsed feedback setup.
*Note only CW feedback was used in this project
Laser intensity was stabilized to ± 0.2%, which surpasses our goal of ± 1%. The blue plot is the signal received by the photodiode, the orange curve is the control signal output by the red pitaya, and the green line is the stabilized signal. We took the data in 8 hours and found no drifts in the stabilization, which indicates that our setup is robust since it maintains the same intensity while environmental parameters such as temperature change. Using only the P and I terms, we were able to minimize the initial noise by a factor of 200.
Looking more closely at the stabilized signal, we can see no long-term drift; the signal only oscillates at high frequency. This high-frequency noise can be accounted for by the internal noise of the photodiode. The ± 0.2% fluctuation in our data was partly correlated to the photodiode uncertainty; the rest of the noise comes from the control signal output by the pitaya. The above data set was preliminary data and demonstrated the sufficiency of our stabilization for the P 3/2 life measurement.
Using 802 nm laser light (20% intensity fluctuation without stablilization), we stablilized the laser to average within ± 0.1% of our setpoint intensity. Extreme peaks during stablization still stayed within ± 0.3% of setpoint. Long term drift of intensity from 802 nm is no longer a problem with this intensity stabilizer. The graph above was taken on the laser that was used for the P 3/2 life time measurement.
Observations:
- It was found that omitting the proportional term for PID control was more stable than including proportional control.
- The lower the control voltage (shown in the bottom graph as it slopes downward), the less stablity we get due to the VVA response curve at different voltages.
There are some issues with our current pseudo pulsed feedback setup. 1) We can’t stabilize the laser intensity right before the beam reaches the ion trap; it has to go through different optics before reaching the trap, which means it will incur some extra noise that we can’t currently stabilize. 2) The second AOM (the one receiving pulsed signal) changes how fast it can turn on and off based on the environmental temperature. 3) Our current setup, as you can see from the picture above, uses too many separate power supplies, which makes it difficult to set up and apply to any measurement. Potential solution: utilize an Arduino DUE to make independent pulsed feedback. This would eliminate the second AOM and allow us to stabilize the laser light right before it reaches the ion trap.We would also be able to make an integrated power supply, making our stabilization more compact and robust.
This project is made for CS 15C class by Chaoshen Zhang, Sam Gebretsadkan, Vivian Liao and Rei Landsberger.