Zeeman Shifted Modulation Transfer Spectroscopy in Atomic Cesium Modulation transfer spectroscopy (MTS) is a useful technique for locking a laser on one of the closed cesium D transitions. We have focused on the F = F = 5 transition as it is used for laser cooling and creating a magneto optical trap (MOT) when it comes to cesium atoms. Commonly, the frequency is shifted with the help of an acousto-optic modulator (AOM). However, the results of our experiment show that this could also be done with the help of the Zeeman effect. signal gen. EOM amplifier Cs cell PD f +6 mixer Figure 1: The basic optical scheme for MTS. The probe beam is drawn in blue. It s polarisation is vertical. The pump beam (after the first PBS drawn in red) has a horizontal polarisation (be aware that the EOM changes it) [1]. We started with a basic MTS setup as shown in Figure 1. First, the beam is split into a probe and a pump beam with the help of a λ/ plate and a polarised beam splitter (PBS). The pump beam is horizontally polarised and therefore passes straight through the PBS. Next, it proceeds through the electro-optic modulator (EOM) and another λ/ plate, which is used to symmetrise the final MTS signal. After the beam hits the mirror, it is expanded by a factor of 3 with the help of a pair of lenses. Afterwards, the beam travels through another PBS, hits two mirrors and passes the cesium cell. Meanwhile, the probe beam is vertically polarised and gets reflected on the first PBS. It first gets expanded and than passes through the cesium cell, where it is aligned with the pump beam. Afterwards, it hits the two mirrors and is again reflected by the second PBS. In the end, a lens is used in order to collect the expanded beam on a fast photodiode (PD). A signal generator is used to drive the EOM with the resonant frequency at 5.6 MHz. The generator s output is first sent to a power splitter. A part of it is led to the EOM, while the other part is mixed with the photodiode s output with the help of a frequency mixer. The output from the mixer is amplified by a factor of, while the used amplifier (seen in figure 3) also acts as a low pass filter. The signal is than observed on an oscilloscope. As the laser s frequency is swept, the sub-doppler transitions can be observed. The resulting signal can be seen in Figure where the saturated absorption signal from Toptica s compact Doppler free absorption unit (CoSy) is used for reference. From our experience the MTS spectrum is more suitable for locking the laser 1
Modulation Transfer and Saturated Absorption Spectroscopy F'=3 F'= F'= F'=5 F=3 F= - - Frequency HMHzL Figure : The MTS spectra (red and blue lines) compared with saturated absorption spectroscopy (orange lines). +V kω + VIN - 1Ω 7 OP7 1Ω 3 6 OUT kω -V Figure 3: The scheme of the amplifier used in the MTS setup. The amplification factor of approximately is achieved. OP7 s low frequency bandwidth also conveniently serves as a low pass filter. on the F = F = 5 transition, because the saturation absorbtion signal for this transition is sometimes barely visible, as seen in Figure. As a part of a Friday afternoon experiment, we wanted to observe the Zeeman shift of the F = F = 5 line. It proved to be a handy method for shifting the locking frequency of the laser. The adapted MTS scheme is shown on Figure. The only modifications were the addition of a coil surrounding the cesium cell and two λ/ plates. The improvised coil used in our setup measured 11 cm in diameter and comprised of two 5 cm long segments of 75 windings. The segments were separated by a.5 cm gap necessary for fitting the coil around the post carrying the cesium cell. The first λ/ plate between the first PBS and the cesium cell can be used to change the probe beam s polarisation. The second λ/ plate allows the adjustment of the polarisation before the second PBS in order to get more light reflected onto the photodiode. Another observation is that it can affect the asymmetry of the MTS signal when using linearly polarised light.
signal gen. EOM Coil f +6 amplifier mixer Cs cell PD Figure : The modified MTS scheme used for observing the Zeeman effect. A coil surrounding the cesium cell and two λ/ plates were added. Figure 5: A photo of our experimental setup. The spectrum seen on the oscilloscope was acquired with linearly polarised light and at 6A. With the coil we achieved a relatively homogeneous magnetic field parallel to the probe beam passing through the cesium cell. When it was turned on and the current was gradually increased, the single cesium F = F = 5 line was split in two as expected according to the Zeeman effect. The frequency shift of the linearly polarised light can be seen in Figure 6. The response is linear and the slopes of the fitted lines are approximately 16 MHz/A. This shows that the lines can be easily shifted by up to 1 MHz and that the MTS spectra are relatively resilient when it comes 3
to the use of a non-ideal coil. The Zeeman Shift for F= F'=5 Transition Frequency Shift HMHzL 1 5-5 -1 Current HAL 6 Figure 6: The Zeeman shift of the F = F = 5 line. The black dots represent the frequencies of the split lines acquired from the MTS spectra at different currents. The gray lines are linear fits of the experimental data. Next, we observed the effect of the magnetic field on differently polarised probe beams []. The results can be seen in Figure 7. Linear polarisation can be achieved without the λ/ plates or by setting their orientation angle to degrees. The circular polarisations σ + and σ are achieved by rotating the λ/ plate next to the cesium cell by 5 and -5 degrees. The motivation behind using circular polarisation is the stronger signal of the single line and the fact that it is more symmetric. Therefore, it has a steeper slope which is preferred for precise locking of the laser frequency. From our experience with laser locking with the help of Toptica s DigiLock software it is also helpful to have only one strong distinct line. Otherwise, problems with the lock hopping onto a neighbouring line can sometimes occur. In conclusion, MTS is a suitable method for observation of the Zeeman effect. Furthermore, the Zeeman effect on the MTS signal could also serve as a method for shifting the laser locking frequency. This can be seen as a low cost alternative to the use of an AOM. In adittion, some ideas for further research come to mind. It might be interesting to test the limits of this method, when it comes to the range of the frequency shift. The improvised coil used in our experiment could be replaced with a better one that could handle higher electrical currents and produce stronger magnetic fields. In addition, it might also be interesting to test if the laser frequency can be rapidly modulated by modulating the current running through the coil. References [1] D. J. McCarron, S. A. King and S. L. Cornish, Modulation transfer spectroscopy in atomic rubidium, Meas. Sci. Technol. 19, 1561 (8). [] K. L. Corwin, Z.-T. Lu, C. F. Hand, R. J. Epstein and C. E. Wieman, Frequency-stabilized diode laser with the Zeeman shift in an atomic vapor, Appl. Opt. 37, 395 (1998).
MTS, Linear Polarisation HaL..1. -.1 -. -15-1 -5 5 1 15. MTS, Circular s + HbL.1. -.1 -. -15-1 -5 5 1 15 MTS, Circular s -. HcL.1. -.1 -. -15-1 -5 5 1 15 Figure 7: The MTS spectra around the F = F = 5 transition when introducing an external magnetic field parallel to the laser beams. The response varies when different probe beam polarisations are used. 5