From microscopes to data transfer via optical fibres all the way to modern quantum technologies, light plays an important role in science and industry. Particularly methods for changing the colour - and hence the frequency and wavelength - of light are of great importance in modern applications. Those methods require the use of nonlinear crystals. In such crystals, two photons of a particular frequency can, for instance, be turned into one photon having twice that frequency - say, two red ones into a single blue one.
For that to work, however, the light typically has to hit the crystal in a specific direction and with a specific polarization. This so-called phase matching often severely limits practical applications. A team of researchers led by ETH professor Rachel Grange at the Institute for Quantum Electronics, together with the group of Lucio Isa at the Department of Materials, have now developed a method by which efficient frequency doubling can be obtained without such fine tuning, and which also has other advantages compared to conventional methods.
Seemingly irreconcilable approaches
The researchers' recipe can be roughly summarized like this: rather small than large, and a mess is better than order. That sounds mysterious, but the task that Grange's team had set themselves was an equally big riddle: "For a better and more widely applicable frequency doubling, we wanted to combine two approaches that don't really fit together", says Romolo Savo, who led the project as a postdoc in the framework of a Marie- Skłodowska-Curie fellowship.
In the first approach, instead of a single large crystal one uses an assembly of lots of mini-crystals whose individual crystal axes point in random directions. In that way, it is no longer necessary to strictly control the directions of the incoming light beams. Among the many mini-crystals there will be some that are oriented favourably and some that are oriented unfavourably, but overall they will always produce a significant amount of frequency-doubled light. "It sounds counterintuitive", Savo admits, "and some of our colleagues were truly puzzled by the idea of using disorder in that way - but it works!"