Microfabricated rubidium vapour cell with a thick glass core for small-scale atomic clock applications

TLDR: In this paper, anodic bonding of silicon and relatively thick glass wafers was used for the fabrication of atomic reference cells with dimensions larger than standard micromachined cells for use in compact atomic devices such as vapour-cell atomic clocks or magnetometers.

Abstract: This paper presents a new fabrication method to manufacture alkali reference cells having dimensions larger than standard micromachined cells and smaller than glass-blown ones, for use in compact atomic devices such as vapour-cell atomic clocks or magnetometers. The technology is based on anodic bonding of silicon and relatively thick glass wafers and fills a gap in cell sizes and technologies available up to now: on one side, microfabrication technologies with typical dimensions <= 2 mm and on the other side, classical glass-blowing technologies for typical dimensions of about 6-10 mm or larger. The fabrication process is described for cells containing atomic Rb and spectroscopic measurements (optical absorption spectrum and double resonance) are reported. The analysis of the bonding strength of our cells was performed and shows that the first anodic bonding steps exhibit higher bonding strengths than the later ones. The spectroscopic results show a good quality of the cells. From the double-resonance signals, we predict a clock stability of approximate to 3 x 10(-11) at 1 s of integration time, which compares well to the performance of compact commercial Rb atomic clocks.

Figures

Figure 8. Number of chips broken on the interface of each anodic bonding step.

Figure 10. Absorption spectra of the 4 mm thick rubidium cell (grey) and the reference cell (black).

Figure 7. Pictures of the obtained cell. (a) Top view. A small spot of alkali metal can be seen in the cavity. (b) Isometric view. (c) Glass-blown vapour cell of the same dimensions; the deviations from the ideal cylindrical shape are clearly visible here.

Figure 9. Picture of the absorption spectrum analysis setup. The light path (for Rb) is shown in yellow and the recorded spectrum can be seen on the oscilloscope (top left).

Figure 1. CAD image of the designed 4 mm thick vapour cell.

Figure 11. Experimental setup for DR spectroscopy.

Frequently asked questions

Q1. What have the authors contributed in "Microfabricated rubidium vapour cell with a thick glass core for small-scale atomic clock applications" ?

This paper presents a new fabrication method to manufacture alkali reference cells having dimensions larger than standard micromachined cells and smaller than glass-blown ones, for use in compact atomic devices such as vapour-cell atomic clocks or magnetometers. The fabrication process is described for cells containing atomic Rb and spectroscopic measurements ( optical absorption spectrum and double resonance ) are reported. From the double-resonance signals, the authors predict a clock stability of ≈3 × 10−11 at 1 s of integration time, which compares well to the performance of compact commercial Rb atomic clocks. 

Q2. What is the way to achieve a stable and hermetic bonding?

For high enough voltages, the electric field created through the entire stack of additional substrate layers is sufficient to obtain a stable and hermetic anodic bonding. 

Q3. Why is the cell designed to be bigger than the previous two mm thick microfabricated cells?

Their cell design is driven by the goal to sample more Rb atoms, and in a bigger cell than their previous 2 mm thick microfabricated cells developed for a small-scale atomic clock [16], in order to achieve a better signal for improved clock stability. 

Q4. What is the reason why the VCSEL was not observed in the cavity?

Alkali metals being very reactive with oxygen, in case of a leak, the entering air would create Rb oxide, which was notobserved in the cavity. 

Q5. How is the VCSEL swept through the Rb D2 line?

The VCSEL’s light output wavelength is swept through the Rb D2 line (around 780 nm) using a sawtooth signal from an external function generator. 

Q6. What are the compact realizations of atomic clocks?

Vapour-cell atomic clocks (‘Rb clocks’) [1] are the most compact realizations of atomic clocks [2] and serve as precise frequency and time references in numerous applications such as telecommunication, network synchronization or satellite navigation, with several thousands of units sold every year. 

Q7. What is the disadvantage of the etching of glass?

the melting of the glass prevents obtaining a flat window surface quality at these small dimensions, which results in undesirable intensity losses for the light sent through the cell for clock operation. 

Q8. What are the advantages of using these cells in atomic devices?

These cells are of interest for use in novel compact or miniaturized atomic devices using small vapour cells, such as Rb atomic clocks, atomic magnetometers and gyroscopes. 

Q9. How can the authors achieve a flat window surface?

By bonding several of these wafers together to form a hollow cell cavity, the obtained sidewalls have very little roughness and it is possible to obtain almost perfectly flat and polished window surfaces. 

Q10. What is the disadvantage of the miniaturization of atomic clocks?

This miniaturization has many advantages (e.g. reducing the power consumption, the mass and the production cost of the clock’s physics package) but a drawback is that it degrades the short-term stability (expressed in terms of the Allan deviation) [4] when the size of the alkali vapour cell used in the clock is reduced [5]. 

Q11. What is the advantage of a photolithographic mask?

One of the major advantages of this technology for producing alkali cells is the possibility of processing many samples in parallel at wafer level and the freedom in the shapes that can be given to even very small structures as they aredefined by a photolithographic mask. 

Q12. What was the first step in the development of MEMS?

MEMS technology development started a couple of decades ago with the anisotropic wet etching of silicon, at first mainly based on the microelectronics industry processes. 

Q13. What is the evidence of the absence of voids in the stack?

In spite of the presence of voids and the weaker bonding interfaces evidenced, the whole assembly is hermetic, proven by the absence of oxidation of Rb in the cavity.