Through the detection of Coherent Population Trapping (CPT) resonances, we demonstrate the temperature-dependence cancellation of the Cs clock frequency in microfabricated vapor cells filled with a mixture of Ne and Ar. The inversion temperature at which the Cs clock frequency temperature sensitivity is greatly reduced only depends on the partial pressure of buffer gases and is measured to be lower than 80°C as expected with simple theoretical calculations. These results are important for the development of state-of-the-art Cs vapor cell clocks with improved long-term frequency stability.
© 2011 OSA
A key point towards the development of vapor cell atomic frequency standards with improved long-term frequency stability is to reduce or even cancel the sensitivity of the clock frequency to temperature fluctuations . In most of such frequency references, the alkali vapor cell is filled with a buffer gas or mixture of buffer gases. The presence of buffer gas causes three main phenomena: a reduction of the resonance linewidth through the Lamb-Dicke effect by reducing the free mean-path of alkali atoms , a homogeneous broadening and red shift of optical transitions [3,4] and a temperature-dependent shift of the clock frequency due to slightly dephasing interactions between alkali and buffer gas atoms or molecules . The latter point explains the necessity to fill the cell with an optimized content to improve the clock performances.
The pressure and temperature shift coefficients of several pure buffer gases have been published in the literature. Nevertheless, reported values by different authors often suffer from discrepancy and lack of accuracy. This key problem makes difficult the realization of vapor cells with predetermined characteristics. Using available data, it has been demonstrated that a strong quadratic dependence of the Rb clock frequency on temperature can be obtained by using two buffer gases shifting each the clock transition in opposite directions. This technique is well-known towards the development of optically pumped passive rubidium frequency standards .
In the case of the Cs atom, most of the literature only reports measurements of linear coefficients due to a lack of resolution of frequency measurements in the experimental setup or to a limited temperature range studied [7, 8]. Recently, O. Kozlova et al. established first measurements of the quadratic coefficient for various buffer gases (Ne and N2)  thanks to a high-resolution CPT Cs clock combining an original excitation scheme, a temporal Ramsey interrogation [10, 11] and allowing a great reduction of the light shift sensitivity . At the same time and particularly, it has been measured in microcells  and cm-scale cells  that the Cs clock transition frequency shows a quadratic dependence on temperature around an inversion temperature of about 80°C using only Ne as a unique buffer gas. In this case, it has been shown experimentally that the inversion temperature does not depend on the buffer gas pressure  as expected theoretically. This result could be of great interest for the development of simple configuration Cs CPT chip scale atomic clocks (CSAC) [15, 16] operating typically in this temperature range.
In the frame of the challenging european MAC-TFC project aiming to develop the first european CSAC prototype, Cs vapor microcells filled with a mixture of buffer gases (Ne and Ar) were realized in Wroclaw University as described in . Mixtures with N2 were not allowed due to the absorption of this gas by the Cs dispenser used for Cs activation and filling process . Several microcells have been characterized in FEMTO-ST and UniNe laboratories using similar laboratory-prototype Cs CPT clocks.
In this article, we mainly investigate the temperature-dependence of the Cs clock frequency in vapor microcells filled with a mixture of Ne and Ar. A basic theoretical study based on temperature coefficients reported in  is given to describe the expected change of Cs frequency with temperature in such cells. Experimental tests based on a Cs CPT clock demonstrate the possibility to obtain a quadratic dependence on temperature of the Cs clock frequency in cells filled with a mixture of Ne and Ar. While quadratic dependence on temperature of the Cs frequency with N2 and Ar buffer gas mixture has already been reported , this is to our knowledge the first reported observation of the temperature-dependence cancellation of the Cs clock frequency in cells containing a mixture of Ne and Ar.
2. Basic theoretical estimations
In vapor cell frequency standards, main phenomena that can shift the clock frequency are the Zeeman shift, the light shift and the buffer gas pressure shift. The frequency ν of the Cs hyperfine transition |F = 3, mF = 0〉 → |F = 4, mF = 0〉 can be described by:5]:
When a mixture of two buffer gases 1 and 2 is used, the Eq. (2) becomes:
Another way to study cells filled with a mixture of two buffers gases is to define the ratio a between the pressure of buffer gas 2 and buffer gas 1 such as:
Nulling the expression of Eq. (8), we find that the pressure ratio a required to obtain an inversion temperature at T = Ti is given by:
Using the same principle, the expression of the inversion temperature Ti for a given pressure ratio a is obtained by:
Figure 1 shows the estimated frequency shift of the Cs frequency versus the cell temperature for two different Ne-Ar pressure ratio and a total pressure of 6.6 kPa. Coefficients published in  are used.
Figure 2 shows the frequency shift of a Cs-Ne-Ar cell as well as the expected inversion temperature versus the Ne and Ar partial pressures in the cell. The total pressure is imposed to be 6.6 kPa.
It is clearly shown that there is no way to obtain any inversion temperature above about 80°C for Cs cells filled with a Ne-Ar mixture. For a = 0 (pure Ne cell), the temperature-dependence of the Cs frequency is expected to be cancelled at about 80°C as recently observed experimentally in chip-scale cells  and cm-scale glass blown cells . Compared to pure Ne cells, the addition of Ar always shifts the inversion temperature to lower temperatures. The pressure ratio between Ar and Ne determines the inversion temperature. For example, the frequency of a Cs clock based on a vapor cell filled with a 85%Ne-15%Ar mixture is expected to exhibit a quadratic dependence on temperature around an inversion temperature of 20°C. This explains why the control of the filling pressure ratio and filling temperature is crucial when vapor cells with a mixture of buffer gases are realized. This issue is particularly critical in microfabricated cells because high temperatures are required to operate the cell sealing with anodic bonding. Such values can be of great interest for the development of vapor cell Cs frequency standards using cm-scale cells. The total frequency shift depends on both the pressure ratio and the total pressure and is increased with the quantity of Ne.
3. Experimental set-up
Two Cs CPT clocks have been implemented in FEMTO-ST and University of Neuchâtel (UniNe) laboratories respectively. Both clocks are quasi-identical and based on similar components. Figure 3 shows a typical schematic of the experimental Cs CPT clock.
The optical source is a 1.5 MHz-linewidth Distributed Feedback (DFB) laser resonant with the Cs D1 line at 894 nm. Two phase-coherent optical sidebands needed to perform the CPT spectroscopy are generated by modulating an external temperature-controlled pigtailed electro-optical phase modulator (EOM) at 9.192 GHz. At the output of the EOM, the laser intensity is stabilized using an acousto-optical modulator (AOM)-based power servo loop. For this purpose, a fraction of the laser power is extracted using a beam splitter cube and detected by the low noise photodiode PD2. The detector output voltage is compared to a high-stability reference voltage. An error signal is generated, processed into a simple analog PI controller and sent into the amplitude modulation (AM) input port of the synthesizer driving the AOM. This servo loop allows to reduce the relative laser intensity fluctuations at the output of the EOM by a factor 25 at 1000 s integration time. At the output of the AOM, the collimated laser beam is sent through a neutral density filter to attenuate the optical power and a quarter-wave plate to circularly polarize the beam.
CPT interaction occurs in a 2-mm diameter and 1.4-mm long microfabricated cell realized according to the process flow described in [17–19]. The cell temperature is stabilized at the mK level using a commercial temperature controller. The cell is surrounded by a solenoid to apply a static magnetic field of 10 μT and raise the Zeeman degeneracy. The ensemble is isolated from spurious magnetic fields using two concentric cylindrical mu-metal magnetic shields. The optical power transmitted through the cell is detected by the low noise Si photodiode PD1. Two servos loops are then implemented from the output voltage of the photodiode PD1. The first one is used to lock the laser frequency near the center of the homogeneously broadened absorption line by modulating the DFB laser current at 80 kHz and demodulating it with the lock-in amplifier 1. The second servo loop aims to lock the frequency of the 9.192 GHz signal driving the EOM on the Cs 0-0 hyperfine transition by modulating at 1.2 kHz the microwave synthesizer frequency. The lock-in amplifier 2 allows to increase significantly the signal to noise ratio of the detected CPT resonance. Once the 9.192 GHz signal locked on the atomic resonance, the clock frequency is measured by comparison with the signal from a microwave synthesizer referenced by a reference hydrogen maser.
The main difference between FEMTO-ST and UniNe clocks are the following. The microwave modulation frequency is 9.192 GHz and 4.596 GHz in FEMTO-ST and UniNe respectively. However, this does not change the physics of the buffer gas related temperature coefficient studied here. Also, in UniNe experiment, the DFB laser exhibits a linewidth of 15 MHz and no laser power servo loop is implemented.
4. Experimental results
Several microcells have been realized. Experimental results on four of them are reported here. All the cells are expected to be filled with 85 % of Ne and 15 % of Ar. Various total pressures are expected. Cells 1, 2 and 4 were tested in FEMTO-ST laboratory while the cell 3 was tested in UniNe.
Figure 4 shows a typical CPT resonance observed in a microcell filled with a buffer gas mixture of Ne and Ar.
The linewidth of the dark resonance is measured to be 3.9 kHz. The contrast of the CPT resonance, defined as the ratio between the height S of the dark resonance and the DC background voltage B, is measured to be about 0.9 %. According to theoretical expectations, we measured that the CPT resonance linewidth is found to be broader in a Ne-Ar microcell than in a pure Ne microcell for identical experimental configurations.
The buffer gas pressure shift measurement procedure is similar to the one described in . The microcell temperature is changed. For each temperature, the clock frequency is measured for various laser intensities to determine the linear light shift function. The frequency shift is then estimated by extrapolation at zero intensity and substraction of the Zeeman shift.
Figure 5 shows the frequency shift of the Cs 0-0 hyperfine resonance as a function of the temperature for different microcells.
The circles, stars, squares and triangles are the experimental data while solid lines are a computer fit of the Eq. (3) to the data. For the four cells, it is clearly shown that the temperature-dependence of the Cs clock frequency is canceled for inversion temperatures lower than 80°C as expected in Fig. 2. Data fits are respectively 24915 + 11.16T – 0.10T2, 36472 + 36.49T – 0.39T2, 36170 + 15.39T – 0.14T2 and 51823 + 29.10T – 0.47T2 for the cells 1, 2, 3 and 4 respectively. The inversion temperature is measured to be 51.6, 46.4, 53.0 and 31.7°C for the cell 1, 2, 3 and 4 respectively.
Table 1 resumes measured characteristics of the tested Cs-Ne-Ar microcells. All the preseented cells have been realized separately in different runs.
For the different cells, the linewidth of the CPT resonance tends to be lower for increased buffer gas total pressure as expected for this buffer gas pressure range. In general, the measured total pressure is found to be about 2–3 times smaller than the expected one. We note also that the measured pressure ratio is not in correct agreement with the expected one. Further tests are in progress to control accurately the actual filling total and partial pressures. Nevertheless, experimental data (measured on different experimental Cs clocks) are in correct agreement for cells 2 and 3 expected to be filled with the same pressure of 24 kPa. This indicates a good repeatability of the cell fabrication process and that only a re-scaling of the apparatus is needed to obtain cells with desired characteristics.
We studied the frequency-temperature dependence of the Cs clock frequency in microcells filled with a mixture of Ne and Ar. Simple theoretical estimations have been done and can be extended to others buffer gas mixtures. It has been measured that a quadratic temperature dependence of the Cs frequency can be obtained with such microcells for inversion temperatures lower than 80°C. The inversion temperature only depends on the Ne-Ar pressure ratio and is shifted to higher temperatures for cells with increasing partial pressure of Ne. The total frequency shift depends on both the total and partial pressures. These properties highlight the necessity to control accurately the filling pressure and temperature during the cell filling process. The presented results are of high relevance for the development of any type of Cs vapor-cell frequency standards.
This work is supported by a collaborative project MAC-TFC of the European Commission (FP 7, ICT program, grant n° 224132). The authors thank C. Rocher for technical assistance.
References and links
1. S. Knappe, V. Gerginov, P. D. D. Schwindt, V. Shah, H. G. Robinson, L. Hollberg, and J. Kitching, “Atomic vapor cells for chip-scale atomic clocks with improved long-term frequency stability,” Opt. Lett. 30(18), 2351 (2005). [CrossRef] [PubMed]
2. R. H. Dicke, “The Effect of Collisions upon the Doppler Width of Spectral Lines,” Phys. Rev. 89, 472–473 (1953). [CrossRef]
3. E. Bernabeu and J. M. Alvarez, “Shift and broadening of hyperfine components of the first doublet of cesium perturbed by foreign gases,” Phys. Rev. A 22(6), 2690 (1980). [CrossRef]
4. G. A. Pitz, D. E. Wertepny, and G. P. Perram, “Pressure broadening and shift of the Cs D1 transition by the noble gases and N2, H2, HD, D2, CH4, C2H6, CF4 and 3He,” Phys. Rev. A 80, 062718 (2009). [CrossRef]
5. J. Vanier and C. Audoin, The Quantum Physics of Atomic Frequency Standards, (Adam Hilger, Bristol, 1989). [CrossRef]
6. J. Vanier, R. Kunski, N. Cyr, J. Y. Savard, and M. Tetu, “On hyperfine frequency shifts caused by buffer gases: Application to the optically pumped passive rubidium frequency standard,” J. Appl. Phys. 53(8), 5387 (1982). [CrossRef]
7. F. Strumia, N. Beverini, A. Moretti, and G. Rovera, “Optimization of the buffer gas mixture for optically pumped Cs frequency standard,” Proc. of the 1976 Freq. Contr. Symp. , 468–472 (1976).
8. N. Beverini, F. Strumia, and G. Rovera, “Buffer gas pressure shift in the mF =0 → mF =0 ground state hyperfine line in Cs,” Opt. Commun. 37(6), 394 (1981). [CrossRef]
9. O. Kozlova, R. Boudot, S. Guérandel, and E. De Clercq, “Measurements of Cs buffer gas collisional frequency shift using pulsed coherent population trapping interrogation,” Proc. of the 2010 Conf. Precision Electromag. Meas., Daejon, Korea (2010).
10. T. Zanon, S. Guérandel, E. de Clercq, D. Holleville, N. Dimarcq, and A. Clairon, “High Contrast Ramsey Fringes with Coherent-Population-Trapping Pulses in a Double Lambda Atomic System,” Phys. Rev. Lett. 94, 193002 (2005). [CrossRef] [PubMed]
11. R. Boudot, S. Guérandel, E. De Clercq, N. Dimarcq, and A. Clairon, “Current status of a pulsed CPT Cs cell clock,” IEEE Trans. Instrum. Meas.58(4), 1217 (2009). [CrossRef]
12. N. Castagna, R. Boudot, S. Guérandel, E. De Clercq, N. Dimarcq, and A. Clairon, “Investigations on continuous and pulsed interrogation for a CPT atomic clock,” IEEE Trans. Ultrason. Ferroelectr. Freq. Control 56(2), 246 (2009). [CrossRef] [PubMed]
13. D. Miletic, P. Dziuban, R. Boudot, M. Hasegawa, R. K. Chutani, G. Mileti, V. Giordano, and C. Gorecki, “Quadratic dependence on temperature of the Cs 0-0 hyperfine resonance in a single Ne buffer gas microfabricated vapor cell,” Electron. Lett. 46(15), 1069 (2010). [CrossRef]
14. R. Boudot, P. Dziuban, M. Hasegawa, R. K. Chutani, S. Galliou, V. Giordano, and C. Gorecki “Coherent population trapping resonances in Cs-Ne microcells for miniature clocks applications,” J. Appl. Phys., to be published (2011). [CrossRef]
15. S. Knappe, V. Shah, P. D. D. Schwindt, L. Hollberg, J. Kitching, L- A. Liew, and J. Moreland, “A microfabricated atomic clock,” Appl. Phys. Lett. 85(9), 1460 (2004). [CrossRef]
16. R. Lutwak, A. Rashed, M. Varghese, G. Tepolt, J. Leblanc, M. Mescher, D. K. Serkland, K. M. Geib, and G. M. Peake, “CSAC- The Chip Scale Atomic Clock,” Proc. of the 7th Symp. Freq. Standards and Metrology, Pacific Grove (CA), edited by Lute Maleki (World scientific), 454–462 (2008).
17. P. Knapkiewicz, J. Dziuban, R. Walczak, L. Mauri, P. Dziuban, and C. Gorecki, “MEMS Cs vapour cell for European Micro-Atomic Clock,” Proc. Eurosensors, Sept. 5–8, Linz, Austria (2010).
18. A. Douahi, L. Nieradko, J. C. Beugnot, J. Dziuban, H. Maillote, S. Guérandel, M. Moraja, C. Gorecki, and V. Giordano, “Vapor Microcell for Chip Scale Atomic Frequency Standard,” Electron. Lett. 43(9), 279 (2007). [CrossRef]
19. L. Nieradko, C. Gorecki, A. Douahi, V. Giordano, J. C. Beugnot, J. Dziuban, and M. Moraja, “New approaching of fabrication and dispensing of micromachined Cs vapor cell,” J. Micro/Nanolith. MEMS MOEMS 7, 033013 (2008). [CrossRef]
20. R. Lutwak, D. Emmons, W. Riley, and R. M. Garvey, “The Chip-Scale atomic clock: Coherent population trapping vs conventional interrogation,” Proc. 34th Annual Precise Time and Time Interval Systems Applications Meeting, Reston (Virginia), 3–5 Dec. 2002 (2002).
22. S. Knappe, J. Kitching, L. Hollberg, and R. Wynands, “Temperature dependence of coherent population trapping resonances,” Appl. Phys. B 74, 217 (2002). [CrossRef]