A laser so stable that its frequency varies by no more than two parts in 10,000 trillion represents an approach for constructing high-quality optical cavities that improve prior designs by more than an order of magnitude. The laser was developed and tested by an international collaboration at NIST/JILA and a group at Physikalisch-Technische Bundesanstalt (PTB), NIST's counterpart in Germany. It will accelerate progress in developing optical clocks that operate at frequencies more than 10,000 times higher than the approximately 9.2-GHz microwaves used as the current worldwide time standard, the researchers say. Many of the most critical experiments in atomic physics require lasers with extremely narrow linewidth and extremely stable output to interrogate clusters of ultracold atoms or single trapped ions. "Stable lasers such as the one reported are already unlocking some of the mysteries of minute atomic interactions that are otherwise hidden," said study member Jun Ye of JILA. The novel laser design also is expected to add a new level of precision to research in gravitational wave detection on Earth and in space, and to precision tests of relativity as well as fundamental physics research in cavity quantum electrodynamics and quantum optomechanics. "And it may be of intense interest to the communications industry, since our system was tested at a familiar telecom wavelength of 1.5 micrometers," Ye said. That is the wavelength with the lowest loss in fiber optic networks. "The previous stability limit of about 2 × 10-16 was good," said Ye, who also works in the Quantum Physics Div. of JILA's Physical Measurement Laboratory, "but it prevented us from exploring the full potential of modern optical atomic clocks, where the atomic coherence time can be exceedingly long. The potential of pushing the laser stability better by an order of magnitude will allow us to realize atomic clocks that have unprecedented stability approaching 1 × 10-17 over 1 second." Diagram of the device interior. (Image: Thomas Kessler) Research on ultrastable lasers typically employs some kind of optical cavity interferometer, comprising a spacer with mirrors at each end. That design severely restricts the range of optical frequency that can resonate in the cavity. By superimposing the cavity output beam on another highly controlled reference beam, the interference effects (periodic reinforcement "beats") can reveal stability with exquisite sensitivity. Such systems, however, have historically been subject to thermal fluctuations that alter the cavity dimensions, reducing frequency stability. "We addressed that problem in several ways," Ye said. "By far the most significant factor was our decision to substitute single-crystal silicon for the ultralow expansion glass (ULE) or fused silica customarily employed in the cavity mirrors and spacers." Single-crystal silicon has a coefficient of thermal expansion that approaches zero at 124 K, "so an all-silicon interferometer can be made insensitive to temperature fluctuations" at that point, the authors write in their report, published online in Nature Photonics. In addition, at 124 K, the crystal has a much lower mechanical loss compared with conventional optical glass, and a much higher Young's modulus. Both mechanical properties combine to minimize the fundamental thermal fluctuations. Graph comparing average instability over time of the new design (green) to two ultrastable ULE units: Ref. 1 (blue), a PTB laser, and Ref. 2 (orange), a JILA laser. The group used extensive computer modeling to create a design that reduced the effects of environmental vibrations on the 21-cm-long resonator, which they mounted in a vertical configuration after testing its response to external forces in all three dimensions. The German researchers devised a novel and simple cryostat design using evaporated liquid nitrogen gas as the coolant. The gas flows through superinsulated vacuum tubes to an outer heat shield, and careful control of the flow limits temperature deviations in the system to about 1 mK from the 124-K target. The new all-silicon unit was tested for 24 hours against two of the best-performing conventional ULE-based optical-cavity-stabilized lasers — one from JILA and one from PTB — with thermal noise variation in the range of 6 × 10-16 and 2 × 10-16, respectively. The results, Ye said, "show that the all-silicon system surpasses the performance of any other optical cavities ever reported." As a long-term stable frequency reference, a preliminary test showed that it is equivalent to the stability of a hydrogen maser at time intervals up to 1000 sec. The scientists are now working on improvements. Among other modifications, they will attempt to reduce feedback errors resulting from spurious amplitude modulation, to suppress noise from the cryostat, and to experiment with various optical coatings on the silicon mirrors. The very thin optical coatings now remain the only significant contribution to the thermal noise of the cavity, and new approaches are being investigated jointly with Markus Aspelmeyer's group at the University of Vienna. Ye, who co-authored the Nature Photonics article, began to collaborate on the project in 2007 while visiting PTB on a grant from the Alexander von Humboldt Foundation. The German team was headed by Uwe Sterr and Fritz Riehle; Ye also recruited a student, Michael Martin, and a visitor in his lab, Lisheng Chen. For more information, visit: www.nist.gov