Federici agrees that the lack of a compact, solid-state terahertz source is "probably one of the major issues that's preventing terahertz from progressing, in many different areas." But to compete with other terahertz sources, such as quantum-cascade lasers and mixers, he says, the power will have to be even larger--closer to milliwatts. The low temperature needed by the superconductors is also a problem, Federici says, especially for portable equipment: "As soon as you say, I've got to cool it down to 50 degrees kelvin, a lot of people that are interested in the application of terahertz will say, 'Well, forget it, then.' "

Such cooling is also needed for quantum-cascade lasers, which have recently been used to produce frequencies in the range above one terahertz. (See "T-Rays Advance Toward Airport Screening.") Although these devices are not likely to operate at much lower frequencies, they are well established and continually improving.

The superconducting emitters are "beautiful science, absolutely," says Michael Pepper, a professor of physics at the University of Cambridge, in England, and scientific director for Cambridge-based TeraView, which employs different sources in terahertz equipment that it sells for medical and security uses. But Pepper sees little need for a constant-output, fixed-frequency terahertz source. Three-dimensional imaging needs pulses, whose delayed reflection indicates distance, he says. Spectroscopic identification of chemicals--the other major technique--requires a tunable source, whereas in the current work, the cavity that enhances the superconductor's emission also locks its frequency.

Welp acknowledges that "to make a useful device, we need maybe a milliwatt." But he says that the team has ideas for modifying the device's geometry to change how well individual junctions contribute to the radiation. "I think we can gain a lot in power, just by improving the efficiency of coupling to the cavity mode," he says.