Lighting in Design February-March 2016

and – as an imperfect dark body emitter – radiates energy across a continuous spectrum. Only a very small part of that emission (about 2 to 3 per cent) is as visible light. The rest we experience as heat. And it is that lack of efficient visible light produc- tion that makes incandescent lighting so wasteful of electricity. There are two low-energy alternatives, only one of which you probably see in people's homes. Fluorescent lighting is produced by a tube of mercury-vapour which releases ultraviolet light in response to an electrical current.That UV causes a phosphor coating on the inside of the glass to glow. They’re about 7 to 15 per cent efficient. LEDs are a solid-state p-n junction diode which emits light as a result of an electric field applied to a semi-conductor.They’re about 5 to 20 per cent ef- ficient and remain expensive.There’s still a great deal of lost energy in each of these alternative solutions. In the world of solar panels (really just the re- verse of a light bulb – using broad-spectrum light energy to create electricity), researchers have been developing mechanisms to improve the range of en- ergy wavelength that can be absorbed by solar cells. Perovskite-structured materials increase energy absorption by over 20 per cent.These aremethylam- monium lead halides dissolved in a solvent and then coated onto a substrate using vapour deposition. Then there’s lanthanide doping to improve conver- sion of infrared photons into higher energy photons. Or light-absorbing dyes such as ruthenium. Only one problem: all these specialist materials start to fail after 1000 K. And our incandescent is most productive much closer to 3000 K. Professor Marin Solja Č i Ć and his team began working on interference systems sandwiched around the incandescent emitter (for want of bet- ter terms, they refer to these as the cold and hot sides respectively). ‘In general,’ they declare, ‘the emissivity of a high-temperature emitter depends on temperature and wavelength.’ By surrounding the emitter with an interference structure designed to transmit vis- ible light and recycle infrared light across a wide range of emission angles they hoped to improve the efficiency of the system. Simply put, all that radiated non-visible energy can be transformed into visible light if we can only absorb it in some way and recycle it back into the emitter to reduce the initial energy required to heat the light source. As a very bad analogy, consider how much less energy it takes to heat water in a thermos versus when the water is flowing past in a pipe.

‘An ideal lighting source would then be a hypo- thetical high-temperature black body that emits only visible wavelengths. Such an emitter would have a luminous efficiency of approximately 40-45%,’ says Solja Č i Ć . The problem, as mentioned earlier, is that exist- ing incandescents are not ideal black body emitters and the current approaches to absorption fail at high temperatures. What Solja Č i Ć and his teamwere looking for was an interference stack, like a thin film, that could ab- sorb a wide range of wavelengths and angles (since incandescent light is not particularly linear), that is able to efficiently reabsorb non-visible energy radia- tion while being subjected to resistive heating, and still be compatible with thermal emitters. They selected four materials: SiO 2 , Al 2 O 3 , Ta 2 O 5 and TiO 2 . The next step was to create a film that would act appropriately.They chose what is called a rugate filter, which is based on a dielectric coating where the refractive index varies continuously as part of its structure. Dielectric coatings themselves are thin-film or interference coatings composed of sub-micron layers of transparent dielectric materi- als. These can be placed by vapour deposition on a substrate (like glass). They are used to modify the reflective proper- ties of a surface by exploiting the interference of reflections from an optical source. At its simplest, rugate filters have a sinusoidal oscillation of the re- fractive index.This creates a notch filter (or multiple notches), blocking a limited wavelength range and reflecting it back onto the emitter. There is a great deal of mathematics in the Solja Č i Ć paper since, as you can imagine, fine-tuning the properties of the materials and the rugate filter (and its numerous notches) is supremely complex. Solja Č i Ć and co then went ahead and built a proof-of-concept. It’s actually quite pretty. The rugate filter consists of ninety layers of SiO 2 andTa 2 O 5 and the usual tungsten filament was made to maximise reabsorption (making it into a highly-polished and flattened radiator-like structure). They then measured the luminous efficiency as being 6.6 per cent, already significantly better than existing incandescents and reaching the levels of low-end LEDs. Theoretically, they could achieve 40 per cent. The more efficient the reflection of non-visible light back to the tungsten filament, themore heat ismain- tained in the system so the more energy-efficient it becomes at converting heat into visible light.

19

LiD FEB/MAR 2016