Optical vortex beams can be used for particle manipulation, quantum optics, communications and other applications. In communications, throughput can be increased by spatially overlapping modes with different orbital angular momentum. The modes can be separated, or demultiplexed, by various bulky approaches including by pairs of spatial light modulators, or centimetre-scale diffractive elements made by a computer numerical control (CNC) router or related milling approaches. However, such components are relatively large for integrated applications, for example, for use on fibre tips.

Now, Shlomi Lightman and colleagues in Israel, have employed direct laser printing to make micrometre-scale structures for multiplexing and demultiplexing orbital angular momentum states of light (Optica 4, 605–610; 2017). A first demonstration uses two diffractive phase elements, whereas a second demonstration shows an integrated device (pictured) combining the required elements.

Credit: OSA

The team used a commercial 3D direct laser writing system to print structures from a material with a refractive index of about 1.51 at 780-nm wavelength. After printing, the structures were rinsed in developer and treated with isopropanol for clean-up. The volumes of the two components (transformer and corrector) are 60 × 60 × 32 μm3 and 60 × 30 × 51.5 μm3, respectively; the two components were separated by a distance of 60 μm.

Vortex beams were created using a phase-only reflective spatial light modulator illuminated by a Ti:sapphire laser in continuous-wave mode; beams with orbital angular momentum from −3ℏ to 3ℏ were generated. In the demonstration, beams were spatially separated 9.4 μm per angular momentum unit. In the system, the topological charge must be less than 5, but this can be increased by adjusting the dimensions between the two elements; the authors claim that vortex beams with a charge of 10 can be processed. The team adjusted the printing process to analyse the effect of surface roughness, which affects the performance of the devices. As expected, smoother surfaces result in less side-lobes and crosstalk.

The experiments were repeated for the wavelengths 690, 790, 890 and 990 nm and demonstrated a 300-nm bandwidth, with the main limitation being dispersion of the photoresist. The team believes that design optimization can enable operation over the wavelength range ∼300–2,000 nm. However, beyond this range the polymer will become opaque and other materials should be employed.