Quantum sensing

Quantum sensing with neutral atoms

Sensors are a big part of our world today, with improvements in cost and manufacturability enabling their widespread deployment.  Atomic clocks and atom interferometer gyroscopes are precise instruments with the potential to revolutionize everyday timekeeping and navigation, particularly for applications to GPS denied environments.  However, these are large instruments, typically occupying an entire lab bench.  They are also expensive to manufacture since most components must be assembled by hand.  In our laboratory, we explore the scientific and engineering issues surrounding the miniaturization of atomic technology for future applications to the production of handheld atomic instruments.  Specifically, we explore micro-electromechanical system (MEMS) technology to develop compact and robust sensors using miniaturized atomic beams.  These include chip-scale clocks with enhanced long-term stability and portable gyroscopes without deadtime.  Using this silicon platform, we are now examining ways to build precision sensors that can one day be mass-produced on-chip.  An example of a miniature atomic beam collimator fabricated and tested at Georgia Tech is shown below on top of a penny for comparison:

[Fig. 1: A cascaded silicon collimator sitting on a penny. H x L x W 1 mm x 3 mm x 5 mm.]


An atomic beam collimator is typically an array of well-aligned microcapillaries each of which has a large aspect ratio L/D.  L is the length, and D is the diameter of an individual channel.  One end of the collimator can be attached to an atomic reservoir maintained at elevated temperatures.  Molecular flow predicts that thermal atoms should exit from the other end as a directed beam with a small divergence angle of  0.84D/L.  For the silicon collimator shown above, we first laid out the 2D pattern of the microchannel array via photolithography on a standard base wafer.  Using deep reactive ion etching (DRIE), we then etched twenty grooves with a 100 um x 100 um cross-section and a length of 3 mm for each collimator.  Then, another capping wafer was eutectically bonded from the top (see Fig. 2a).  Finally, we diced the bonded wafer into miniature collimator units.  The use of photolithography to define the atomic paths allows for flexible customization and control with a high spatial resolution of less than 1 um–for example, in our work we made focusing collimators that could deliver a continuous stream of atoms to a targeted region with pinpoint accuracy, as well as generate multiple atomic beams.  In addition, the MEMS process is compatible with CMOS technology.  Therefore, temperature sensors, heaters, and other functional optoelectronics can in future be integrated on-chip.  This paves the way towards shrinking size and reducing power consumption for compact atomic instruments.

[Fig. 2: (a) SEM end view of an actual wafer used to collimate atoms, with 20 microchannels of 100 um width and 3 mm length.  (b) Artist’s rendition of a device in which rubidium atoms emerge as a highly collimated output from etched microchannels and are probed by laser light.  (c) Measured atomic Doppler distribution.  Microfabrication allowed us to create a “cascaded” collimator (red curve) that outperforms conventional collimation (blue curve) by dramatically reducing the undesired wings of the emission (spectral wings in the data) by a factor of 40.  This reduction can significantly increase the signal-to-noise ratio (SNR) of chip-scale devices.  U.S. patent application No. 62/672,709 is pending.]

Using the above lithographic approach, we achieved the “cascaded collimator” that is shown in Fig. 1, and which dramatically exemplifies the customization that is possible.  This collimator addresses a major limitation of capillary beams, which is a long “tail” in the angular distribution function extending to large angles  (seen in Fig. 2c above).  The collimator is a cascaded series of shorter tubes in series with one another, and the on-axis beam flux will be unaffected by the cascade.  Such a device would be difficult to form using machined collimators due to the need to carefully align each stage to the next with micrometer precision.  However, for planar devices it is straightforward to fabricate lithographically by either etching relief regions within the base wafer, or by cutting such reliefs into the capping wafer.  The key innovation of this device is that an atom, which enters a gap region is more likely to leave immediately rather than passing all the way down the subsequent tube and can be removed within the source itself, which greatly reduces the off-axis emission that can contaminate downstream chip components.

We used Doppler-sensitive laser spectroscopy on the Rb D2 line to experimentally demonstrate this effect (see Fig. 2c above).  Our experimental data show that the wings of the distribution at 200 MHz detuning, corresponding to atoms moving at an angle of 30 degrees from the main axis, are reduced by a factor more than 70.  For future applications employing miniature atomic devices, we have also studied the long-term behavior of these sources to prove their robustness and suitability for applications outside of a laboratory.  We did so by taking long-term data on the output flux and spectral width using a sample that was made to operate under a 24-hour-per-day, 7-day-per-week regimen for over 4200 hours, a total period of nearly 6 months, over which time the collimator was extremely stable.  This test demonstrates the usefulness and reliability of these microfabricated atomic sources for applications in quantum device fabrication.


Li, C., Chai, X., Wei, B., Yang, J., Daruwalla, A., Ayazi, F., & Raman, C. (2019). Cascaded collimator for atomic beams traveling in planar silicon devices. Nature communications, 10(1), 1-8.

Li, C., Wei, B., Chai, X., Yang, J., Daruwalla, A., Ayazi, F., & Raman, C. (2020). Robust characterization of microfabricated atomic beams on a six-month time scale. Physical Review Research, 2(2), 023239.