A Study of the Dichroic Atomic Vapor Lock Signal

By Jiehang Zhang


Laser frequency locking is an essential topic in atomic physics.Dichroic Atomic Vapor Lock (DAVL) signal is a convenient choice. A 780nm laser is used to drive the Rb D2 transition. By applying a magnet, the difference of the two Zeeman shifted absorption profiles can be used for locking. In this experiment, several factors that may affect the stability of the are studied: birefringence effect for the PBS is tested, and temperature effect on the PBS, the Rb cell and the magnet are measured.


In Fig[1] below, the Gaussian distribution represents Doppler-broadened absorption profile. With the existance of magnetic field, the left-handed polarized and right-handed polarized light will have different Zeeman shifts. Thus the two absorption profiles are seperated. If we take the difference of the two profiles, we will get a signal like the one shown in Fig[2].



In Fig[2], the center of the slope just correspond to the absorption peak of the original profile. We can use this slope to lock the laser frequency.

This Dichroic Atomic Vapor Lock signal has several advantages:

1.The experimental setup is simple(as can be seen in the next section), there's no need for external modulation(spared lock-in amplifiers & other stuff). Actually the magnetic field is already providing the modulation.

2.It is less sensitive to laser intensity fluctuation. Take the saturation absorption for comparision, intensity fluctuation will have a much larger effect on the signal.

3.The height of the signal is large. Still comparing to saturation absorption, it's height is maximally 1/3 of the overall height of the absorption peak.

Experimental Setup:

Fig[3] shown below is a schematic of the experimental setup:

Fig[3] experimental setup(schematic)

A 780nm laser is fiber-coupled. It goes through a half wave plate and a PBS, therefor we can controll the intensity through the half wave plate. The laser drives the D2 transition of Rb. A magnet is put right besides the Rb vapor cell, providing Zeeman shift. The laser then goes through a quater wave plate, which turns the left-handed & right-handed polarized light into respectively vertical and horizontal linearly polarized light. Then the PBS can seperate them, for the subtraction of the two detectors.

Fig[4] below is a real picture of the experimental setup.

Fig[4] experimental setup(real picture)

Experimental Results:

Here is the actual signal we observe

Fig[5] Actual signal

The laser frequency is scanned back and forth, thus the signal is symmetric along the axis where the "Begin Scan" arrow is. We are observing at the 85Rb F=2 ground state resonance signal, which is indicated in the graph. This signal is relatively seperated from the others.

First, I studied qualitatively the birefringence effect on the front window of the Rb vapor cell. I used a translational stage mounted upon the cell, to pressure the front window. Since the cell is put on a platform, the displacement of the translational stage is just the change in shape of the cell front window.

The points in the upper two graphs correspond to at each measurement, the higher peak and lower peak of the signal. The points in the left-down graph correspond to the magnitude of the signal(i.e.Higher peak minus lower peak).

Fig[6] Birefringence Effect of the front window of Rb cell

It is found that both the magnitude of the signal, and the overall middle level is decreased when pressure is applied to the front window. While releasing the pressure, the signal rises back by a even lager amount, i.e. it bounces back like a spring. It has memory of the past.

However, it is very hard to convert the displacement of the translational stage to any quantity like pressure, or tilt angle of the window. Therefore later on, to be more careful, a quantitative study of the temperature effect on different parts is carried out.

In the following graphs, besides the peak values, the middle level(zero offset) is indicated. And also it is used for caliberating laser frequency. The reason for that is: First, the slope of the lock signal is found out to be not changing. Second, if the zero offset is changed, it can correspond to a laser frequency shift.


Temperature effect on the PBS is found out to be the most significant. And also, due to the way I heated the PBS up, initially there was a very obvious heat imbalance, which resulted(as can be seen in Fig[7]) the fast rising and there after falling of the signal. The sensitivity is caliberated to be 30.1MHz/C. With a uncertainty of 50%.


As for the Rb cell, the heat induces a very large fluctuation of the overall level. The sensitivity is 13.2MHz/C, with uncertainty 300%. The big amount of error also reveals the fluctuation.


Finally the temperature effect on the magnet is tested. Comparing to the former two tests, this sensitivity is much smaller: 2.1MHz/C. Basically the magnetic field does not contribute the instability of the signal.


Qualitative: Squeezing the front window of the vapor cell will decrease the signal magnitude, and shift the middle level. And also it has memory when pressure is released

Quantitative: Temperature drift speed for PBS£¬Rb cell, and magnet are respectively: 30.1MHz/C, 13.2MHz/C, 2.1MHz/C, with big systematic errors. Therefore we have to be very careful with the temperature.


I would like to thank Prof. Luis Orozco for providing the project, helping me very patiently, and the opportunity for taking this course. David Norris for his great amount of help, guidance and advice in the lab.


[1]K.L.Corwin, Z.T.Lu, C.F.Hand, R.J.Epstein, &C.E.Wieman, ”°Frequency-Stabilized Diode Laser with the Zeeman Shift in an Atomic Vapor”± Applied Optics, Vol. 37, Issue 15, pp. 3295-3298 (1998)

[2] F. E. Becerra,?R. T. Willis,?S. L. Rolston,?L. A. Orozco, Two-Photon Dichroic Atomic Vapor Laser Lock Using Electromagnetically Induced Transparency and Absorption, J. Opt. Sco. Am. B Vol. 26, No. 7(2009)