Recent advances have expanded an old material beyond niche telecommunications applications and put it at the forefront of a potentially thriving new Australian industry.
When lithium niobate was first synthesised in 1949 at Bell Laboratories in the United States, the extent of the hopes of the engineers working on it were that it might prove useful in telecommunications.
The material appeared to have ferroelectric properties, allowing its polarity to be manipulated, and is piezoelectric, meaning a charge can induce movement in it — or vice versa. As researchers explored these properties, they began to introduce the material into a variety of technologies, and it became a well-established part of the telecommunications industry.
“There was one in every television by the 1970s,” RMIT Distinguished Professor Arnan Mitchell told create.
“But they were niche applications. You have a lithium niobate device, and it does one thing; it’s not an integrated circuit.”
For integrated circuits, the basis of the sophisticated electronics that permit modern computing, silicon became the norm. But researchers like Mitchell, the Director of RMIT’s Integrated Photonics and Applications Centre, thought lithium niobate still held potential.
For a start, it has valuable uses in photonics, due to its ability to produce and manipulate the full spectrum of electromagnetic waves, and not just visible light. And recent advances in manufacturing have meant it is much more feasible to produce it in a form suitable for semiconductor wafers.
“In the past, lithium niobate was available as a bulk material, so this meant rather thick substrates, and what became recently possible is to have these as very thin films,” Dr Andy Boes, a University of Adelaide Senior Lecturer who has been collaborating with Mitchell, told create.
“This enables its integration with other platforms or confining the light and enabling these tight circuits to integrate more components.”
It also means lithium niobate could address some of the shortcomings of silicon. In large data centres, for instance, the amount of information that needs to be transmitted demands the use of optical fibres.
“The industry is focused on adding photonics to silicon electronic chips, and this has become an industrial reality,” Mitchell said.
“You essentially need a really high bandwidth communication system from every computer to every other computer, and so that is what’s bringing photonics into the electronics world — the need to have all of these computers able to talk to each other in data centres and do this at a reasonable price.”
That drove Mitchell and other photonics researchers to re-examine whether lithium niobate could have uses beyond its niche telecommunications applications.
“It’s only really in the last five to 10 years that actual integrated circuits made out of lithium niobate and marrying them with silicon photonics have become possible,” Mitchell said.
“Let’s go back to this tried-and-true material that’s been used in bulk form and try and think about how it could be used the same way that people are now looking at integrated circuits and silicon. It seems that you can have the best of both worlds by bringing those two things together.”
One company attracted by these possibilities is Advanced Navigation, an Australian robotics and artificial intelligence firm. It is using the lithium niobate chips developed at RMIT in its line of Boreas digital gyroscopes, which will help NASA vehicles navigate on the moon, where it is not possible to use GPS.
“Boreas is the first ever digital fibre optic gyroscope, and it has some pretty drastic advantages over traditional fibre optic gyroscopes,” Xavier Orr, Advanced Navigation’s CEO, told create. “That technology — part of which is born out of RMIT — will be the first Australian tech landing on the moon.”
US aerospace company Intuitive Machines is building the lunar lander for NASA, and Advanced Navigation is contracted to provide it with the gyroscopes it will use in its navigation. Orr said that the lithium niobate chips allow them to be cheaper, use less power, and take up 40 per cent less space.
“That reduction in size, weight and power — the increased performance — represents about $85 million in cost savings when they launch,” he said.
“To get to the moon, every gram has quite a high cost, so when you’re able to cut size and weight down, it results in pretty substantial cost savings. And it also allows in new capabilities.”
These include an additional Intuitive Machines vehicle named the Hopper, which also uses Advanced Navigation’s technology.
“It’s utilising this D-FOG [digital fibre optic gyroscope] technology, as well as some of our other photonics technology,” Orr said.
“It hops around the moon; it’s doing little take-offs and landings again and again. So that kind of vehicle is really enabled by our technology.”
To get the necessary precision for technology like this would require extremely high-precision fibres if analogue gyroscopes were used, Mitchell said. The digitisation offered by lithium niobate allows for more leeway.
“If you have digital signal processing then you can be a little bit less careful about it; you can use ordinary fibre, and you have better yield so you can wind them more easily and then just use digital signal processing to understand how each coil works,” he explained.
“It becomes cheaper, so you can be a bit more aggressive in making the coils more compact; you can use more standard materials and use the information processing to overcome any imperfections in the materials.
“This is really what a complex integrated circuit allows you to do, and ultimately that makes it a bit more compact. So the structure can be quite a lot smaller and do all three axes with one chip rather than having a separate box for each, X, Y and Z axis.”
While that reduced cost matters a lot when launching equipment into space, it also gives the technology a broader range of applications here on Earth. After all, most companies in the market for navigation equipment do not have a NASA-sized budget.
“Traditional fibre-optic gyroscopes, if you go and buy them, cost tens of thousands of dollars, and the idea is to reduce that by a factor of 10 so that you can start thinking about using these in many more applications,” said Mitchell. “One of the companies we are working with does surveying with drones to look at the integrity of railway lines … You can have a cheaper drone with a cheaper navigation system on it and still get millimetre-scale accuracy information about where the rails are.”
Terrestrial use is on Orr’s mind too, and Advanced Navigation’ Boreas gyroscope is finding use in a broad range of industries.
“We sell that into marine, automotive, aerospace applications, a lot of surveying applications and a lot of autonomous systems,” he said.
A feature that could have positive long-term implications for local industry, and one that is important to Advanced Navigation, is that the lithium niobate chips are produced in Australia — right there at RMIT.
“The D-FOG technology is seen as a sovereign capability by the Australian government,” Orr said.
“So there’s a real desire to keep that all onshore and have them fully built here.”
When Advanced Navigation took this requirement to Mitchell and Boes’s team, the academics used the opportunity to let them know how much more they could do for the company.
“We said, ‘Well yes, we could do that, but you might be interested to know that it’s now possible to make much more complicated integrated circuits and so we could think about more sophisticated systems than what people have been doing for 30 years,’” Mitchell said.
“This is what we’ve been doing with Advanced Navigation — and also researchers at ANU [the Australian National University], who have more sophisticated sensing mechanisms using digital information processing.”
Orr also appreciates how closely his company can work with the university researchers.
“They produced all the prototypes at that facility for us, which allowed us to do the testing really quickly,” he said.
The “Valley of Death”
Mitchell sees the collaboration as one of a type that should play a bigger role in Australia’s industrial landscape. He believes that while the country has a solid research base, it’s lacking the funding and institutional support that can take innovations to fully developed technology ready for exploitation.
“To give Advanced Navigation their due, they’re one of the few companies that has really said, okay, we’ll take a punt, and we’ll actually step into that valley of death with you if you’ll meet us halfway,” he said. “This is what Europe and America do differently to Australia: they focus a lot of their resources — their government resources — on filling that gap.”
Orr sees these collaborations as working well for everyone.
“A university might have been working on something for 10, 15, 20 years and it’s a really deep technology that takes a very long time to develop, and then they have incredible results that can really change a whole industry,” he said.
That’s why Mitchell and Boes hope more businesses think about what they could do with lithium niobate chips. Mitchell estimates that around 100 companies in Australia could use the technology, enough to build a local industry.
“We are probably still talking about, ambitiously, tens to hundreds of thousands of [chips] a year, … we could probably push out that many chips with a facility about the size of the one we’ve got at RMIT University,” he said.
Boes also sees opportunity for ambitious companies.
“[Get] familiar with what is currently possible, but also what will be possible in the near future,” he urged. “There are lots of opportunities in this space.