Doppler-free spectroscopy of rubidium atoms driven by a control laser

Abstract

A scheme of Doppler-free spectroscopy is experimentally demonstrated with a co-propagating control laser locking to an atomic hyperfine transition, and the differential transmission of the probe and the reference laser is detected. Crossover resonances are eliminated by selecting the class of atoms with zero velocity in the direction of beam propagation. In addition, the sub-Doppler spectrum experiences optical gain compared to the conventional saturated-absorption spectrum as a result of optical pumping.

Graphical abstract

Keywords

Doppler-free spectroscopy / crossover resonances / optical pumping

1 Introduction

The normal atomic absorption spectra are Doppler broadened due to random movements of atoms. Saturation absorption spectroscopy (SAS) [1–4] and polarization spectroscopy (PS) [5–8] are commonly used Doppler-free spectroscopic techniques to resolve the hyperfine structure of atomic spectra, and have proved useful tools in areas of laser frequency stabilization [8, 9] and the frequency reference application [10]. The SAS contains two counter-propagating beams splitted from one laser: One strong pump beam saturates atoms and the other much weaker probe beam is used to obtain the transmitted spectrum. The PS uses the similar setup but with a circularly polarized pump beam to induce the birefringence in the medium and a different detection scheme. In both of the two Doppler-free techniques, when the pump and probe beams are absorbed by atoms in the same velocity group simultaneously, several sub-Doppler absorption and dispersion features with the width approaching the atomic nature linewidth appear in the Doppler absorption profile. These sub-Doppler features can be classified into two types: The first one happens when the pump and probe beams interact with atoms of zero velocity or velocity perpendicular to the beam propagation direction, corresponding to the real transitions of atoms; the second type occurs when the laser is tuned exactly halfway between two hyperfine levels and interacts with the atoms moving with a certain longitudinal velocity, which is called crossover resonances. Crossover resonances appear in the multiple unresolved hyperfine transitions that have a common level (Fig. 1), and provide a way of laser frequency stabilization with a frequency detuning from the exact transition, which is favorable in atomic laser cooling and trapping experiments. However, the presence of the crossover resonances makes the real resonances of atomic spectra less resolvable, especially in the case that adjacent transitions are very close to each other. It is noted that the amplitude of crossover resonances are usually larger than that of real transitions in SAS and PS, and these spurious resonances render the identification of closely spaced atomic structures difficult.

There are several methods to eliminate crossover resonances in multilevel structures of atoms. It is suggested to use co-propagating probe and pump laser beams to avoid the appearance of crossover resonances [11]. The pump laser is locked to a hyperfine transition, and its transmission is monitored while the frequency of the probe laser is scanning over the range of the Doppler line width. When the probe comes into resonance with the same or other hyperfine transitions for atoms with zero longitudinal velocity, the pump laser transmission spectrum shows a Doppler-free dip due to the absorption reduction caused by the probe laser. In addition, elimination of crossover resonances can also be achieved with a nanocell [12, 13].

In this work we demonstrate a new Doppler-free spectroscopy scheme in the control of laser driven multilevel atoms. Three co-propagating laser beams are used. The frequency of the control laser is stabilized to an atomic hyperfine transition, while the probe and reference lasers are scanning frequency over a range of several hundreds MHz. Unlike the scheme mentioned above [11], the differential transmission of the probe and reference laser beams instead of the transmission of the control laser, is monitored. The key idea is based on the velocity-selective optical pumping. Since the control laser is fixed on a hyperfine transition, only the atoms with zero velocity or velocity perpendicular to the laser propagation direction could absorb the control laser, and this contributes to the variation of the differential transmission of the probe and reference beams, resulting in the formation of sub-Doppler peaks on the flat background of transmission spectra. The presented scheme has the main characteristics as follows: i) the Doppler background is eliminated; ii) crossover resonances disappear; iii) real atomic resonances are optically amplified through optical pumping.

2 Experimental setup

The experimental setup is shown in Fig. 2. Two external cavity diodes lasers (Toptica DL100) with the linewidth of \sim 1 MHz are used. One laser beam from DL 2 is split into three parts. Two parallel weak beams act as the probe and the reference laser respectively, and the other much stronger counter-propagating beam is used as the pump laser. The output of DL 1 serves as a control laser. All the three beams co-propagate through a 10-cm-long room-temperature natural rubidium vapor cell with magnetic shielding. The probe and control lasers have orthogonal polarizations. They are combined with a polarizing beamsplitter cube (PBS 1), and then the control laser is separated by another polarizing beamsplitter cube (PBS 2) when it has passed through the Rb vapor cell. The profiles of the probe and control beams are reshaped by anamorphic prism pairs to obtain beam diameters of 1 mm and 3 mm respectively. The two beams are fully overlapped in the Rb vapor cell in order to achieve better resolution. Switching of the pump beam is controlled mechanically to provide reference spectra. Neutral density filters are used to vary powers of the two parallel beams independently. Transmissions of the probe and reference beams are detected by two photodiodes and the differential signal output is recorded by an oscilloscope.

As shown in Fig. 3, in ⁸⁷Rb D₂ line transition, the control laser drives the 5²S{}_{1/2}{}^{}F=1\to 5^2P{}_{3/2}{}^{}F_{}^{\prime }=2 transition and is actively locked to it using the SAS frequency modulation method. Most of the atoms are optically pumped to the 5²S{}_{1/2}{}^{}F=2 state in the presence of the control laser. The probe laser has a frequency scanning range of several hundreds MHz around the 5²S{}_{1/2}{}^{}F=2\to 5^2P_3/2 transition. The powers of the probe and reference beams are adjusted to be approximately equal, so that the differential transmission between the two beams should be zero in the absence of the control laser, through which the Doppler broadened background is eliminated.

3 Results and analyses

In Fig. 4 (a), the conventional saturated-absorption spectrum is presented when the control laser is blocked. The individual hyperfine transitions are well-resolved including real atomic transitions and crossover resonances, and the amplitude of crossover resonances are much stronger than that of real transitions. As the control laser is resonant with the 5²S{}_{1/2}{}^{}F=1\to 5^2P{}_{3/2}{}^{}F_{}^{\prime }=2 transition, only the atoms with zero velocity in the beam propagation direction can absorb the control laser and are optically pumped to the 5²P{}_{3/2}{}^{}F_{}^{\prime }=2 state. Then, the spontaneous emission from the 5²P{}_{3/2}{}^{}F_{}^{\prime }=2 state to the 5²S{}_{1/2}{}^{}F=2 state causes the increase of the population of the zero-velocity atoms in the 5²S{}_{1/2}{}^{}F=2 state, resulting in the enhancement of the probe absorption and further the change of differential transmission of the probe and reference beams. While the probe laser is scanning around the 5²S{}_{1/2}{}^{}F=2\to 5^2P{}_{3/2}{}^{}F_{}^{\prime }=1,2,3 transition, a re-pumping process from the state 5²S{}_{1/2}{}^{}F=2 to 5²S{}_{1/2}{}^{}F=1 also occurs. In the case of the probe frequency is near to the 5²S{}_{1/2}{}^{}F=2\to 5^2P{}_{3/2}{}^{}F_{}^{\prime }=2 transition, and the control and the probe laser transitions form a three-level \Lambda system. If the control laser intensity is high enough, the atomic medium would become transparent for the probe in the vicinity of two-photon resonance as a result of the coherence induced by the control laser, which is known as electromagnetic induced transparency (EIT) [14, 15]. In our scheme, the control laser power is 200 μW while the probe laser power is 20 μW, therefore the optical pumping effect rather than the quantum coherence dominates. As shown in Fig. 4 (a), the 5²S{}_{1/2}{}^{}F=2\to 5^2P{}_{3/2}{}^{}F_{}^{\prime }=3 transition has the most prominent increments because it is a closed transition and re-pumping from state 5²S{}_{1/2}{}^{}F=2 to state 5²S{}_{1/2}{}^{}F=1 is forbidden. The line width of sub-Doppler features are slightly broadened due to the power broadening, the frequency stability of the control laser and the imperfect overlap of the co-propagating probe and control beams.

Figure 5 shows the Doppler-free spectrum in ⁸⁵Rb D₂ line transition with the same experimental setup. For ⁸⁵Rb, the control laser is locked to the 5²S{}_{1/2}{}^{}F=2\to 5^2P{}_{3/2}{}^{}F_{}^{\prime }=3 transition, and the probe scans around the 5²S{}_{1/2}{}^{}F=3\to 5^2P{}_{3/2}{}^{}F_{}^{\prime }=2,3,4 transitions. As shown in Fig. 5 (b), three sub-Doppler peaks with optical gain are clearly identified. It can also be seen that the closed transition of 5²S{}_{1/2}{}^{}F=3\to 5^2P{}_{3/2}{}^{}F_{}^{\prime }=4 has stronger amplitude than other open transitions.

In the following we obtain the first-derivative error signal corresponding to the Doppler-free profile presented in Fig. 4 (b) by using frequency modulation techniques. As shown in Fig. 6, three dispersion features are clearly resolved and can easily be matched to the corresponding hyperfine transitions. The intensity of the derivative signal is also enhanced due to optical pumping by the control laser. This provides an important advantage in high resolution spectroscopy and is also beneficial for the laser frequency locking for atomic multilevel structures. For other Doppler-free spectroscopy techniques such as PS, the peaks of closely spaced transitions usually merge, making it difficult to match the dispersion features and the actual transitions exactly [6].

4 Conclusions

In summary, we have demonstrated a Doppler-free spectroscopy scheme by using the co-propagating probe and control lasers. The Doppler broadened background is eliminated by subtracting the probe transmission with that of a reference beam propagating through the same atomic vapor. The control laser locked to a hyperfine transition causes velocity selective optical pumping, and only the atoms with zero longitudinal velocity contribute to the enhancement of the probe absorption. We detect the differential transmission of the probe and reference beams and obtain sub-Doppler peaks on a flat background. The presented Doppler-free scheme has features of eliminating crossover resonances and offering well-resolved atomic hyperfine structures. Meanwhile the signal experiences optical gain as a result of optical pumping. The scheme is meaningful for the high-resolution spectroscopy of closely spaced hyperfine levels (e.g. Zeeman spectra), for which the existence of crossover resonances complicates the identification of the relevant spectral features. Combined with the frequency modulation techniques [16], the presented scheme can be applied to laser frequency stabilization as well, making it much easier to match exactly the dispersion features to the atomic hyperfine transitions in comparison with the PS technique.

References

[1] W. Demtröder, Laser Spectroscopy, Berlin: Springer, 1998

[2] K. B. MacAdam, A. Steinbach, and C. E. Wieman, Am. J. Phys., 1992, 60(12): 1098

[3] D. A. Smith and I. G. Hughes, Am. J. Phys., 2004, 72(5): 631

[4] O. Schmidt, K. M. Knaak, R. Wynands, and D. Meschede, Appl. Phys. B, 1994, 59(2): 167

[5] C. Wieman and T. W. Hänsch, Phys. Rev. Lett., 1976, 36(20): 1170

[6] M. L. Harris, C. S. Adams, S. L. Cornish, I. C. McLeod, E. Tarleton, and I. G. Hughes, Phys. Rev. A, 2006, 73(6): 062509

[7] H. D. Do, G. Moon, and H. R. Noh, Phys. Rev. A, 2008, 77(3): 032513

[8] C. P. Pearman, C. S. Adams, S. G. Cox, P. F. Griffin, D. A. Smith, and I. G. Hughes, J. Phys. B, 2002, 35(24): 5141

[9] Y. Yoshikawa, T. Umeki, T. Mukae, Y. Torii, and T. Kuga, Appl. Opt., 2003, 42(33): 6645

[10] A. Onae, K. Okumura, J. Yoda, and K. Nakagawa, Conference on Precision Electromagnetic Measurements, in CPEM Digest, 1996, 17–20: 317

[11] A. Banerjee and V. Natarajan, Opt. Lett., 2003, 28(20): 1912

[12] S. Briaudeau, D. Bloch, and M. Ducloy, Phys. Rev. A, 1999, 59(5): 3723

[13] A. Sargsyan, D. Sarkisyan, A. Papoyan, Y. Pashayan-Leroy, P. Moroshkin, A. Weis, A. Khanbekyan, E. Mariotti, and L. Moi, Laser Phys., 2008, 18(6): 749

[14] Y. F. Zhu and T. N. Wasserlauf, Phys. Rev. A, 1999, 54(4): 3653

[15] U. D. Rapol, A. Wasan, and V. Natarajan, Phys. Rev. A, 2003, 67(5): 053802

[16] G. C. Bjorklund and M. Levenson, Phys. Rev. A, 1981, 24(1): 166

Rights & permissions

© Higher Education Press and Springer-Verlag Berlin Heidelberg 2012