A breakthrough in biomedical engineering could pave the way for tiny implantable devices capable of diagnosing, monitoring, and treating a wide range of health conditions.
An international team of researchers led by the University of Glasgow has created a new type of ultra-small antenna that can wirelessly transmit data through tissue to external devices.
The prototypes, dubbed µBots (pronounced ‘microbots’), are smaller, lighter, less power-hungry, and produce less heat than many current implantable devices, which typically rely on radio-frequency antennas to carry data. Those devices tend to be bulky and generate significant heat, making them difficult to use comfortably as long-term implants and increasing the risk of infection at the implant site.
The sub-millimetre-wide µBots instead combine acoustic and electromagnetic physics to create magnetoelectric antennas that are smaller, cooler, and capable of carrying a much richer stream of data across a wide bandwidth. They could help diagnose neurodegenerative diseases earlier, deliver drugs on demand, or even provide neuromodulation treatments for conditions such as epilepsy or Parkinson’s disease.
In a new paper published in Science Advances, the team describes how the µBots were developed at the University of Glasgow’s James Watt Nanofabrication Centre and tested in the presence of biological tissue to demonstrate telemetry performance.
Exploiting an overlooked substrate
The key to the device’s performance lies in a part of piezo-antenna design that’s often overlooked. Rather than treating the material the device is built on as an inert base, the team instead harnessed the substrate to exploit its unusual acoustic resonances – a property that allows sound waves to bounce back and forth within the device, generating additional frequencies, or overtones, that can be used for both power transfer and data transmission.
The result is an antenna with a remarkably broad -10dB bandwidth of up to 22.6GHz, enabling it to transmit significantly more data than conventional antennas. To demonstrate this, the team successfully transmitted sonogram video and audio signals wirelessly in real time between two magnetoelectric antennas – a far more demanding test than the binary data streams typically used to benchmark devices of this kind.
Getting a clean signal from a device with no precedent
Getting to that result wasn’t straightforward. Dr Mahdieh Shojaei Baghini, the paper’s first and corresponding author and the devices’ creator, said one of the toughest hurdles was simply pinning down how to measure the antennas accurately in the first place: “The main challenge associated with the antennas initially was precise characterisation, as we did not have any prior literature to benchmark against,” she told Electronic Specifier. “Due to their low power, high frequency nature, it is difficult to isolate the results that we were obtaining from noise and other parasitics.”
That difficulty is compounded by the fact that the pristine conditions of a lab won’t reflect where these devices will eventually be used. “Anechoic chambers do not always exist in a real-world setting where these devices will be deployed, and it is very important to record the correct output in the presence of ambient noise, especially for integration with other circuits and embedded systems as we move away from laboratory settings,” she explained.
To solve the problem, the team built their own reference points from scratch. “We fabricated control piezoelectric resonators and developed a one-of-a-kind extraction technique that unlocks these low power signals and even allows us to control them,” she said, adding that further detail on the method is included in the paper itself.
How the data rate stacks up
On the sonogram video and audio demonstration, Shojaei Baghini told Electronic Specifier the µBots are already competitive with existing implant technology – at a fraction of the size. “Sitting at approximately 1Mbit/s gross, the performance is comparable to current systems at a size that’s four orders of magnitude lower,” she said. She noted that real-world performance will also depend on the electronics built around the antenna: “True performance also relies on the associated electronics and the RFICs integrated into the complete system.”
That said, the team designed the technology with more than a single device in mind. “The basis of this technology is also developed keeping in mind swarms of µBots that are capable of communicating with each other,” Shojaei Baghini said. Factoring in packaging effects, nine µBots working together could match the angular stability of a conventional RF antenna while occupying a far smaller effective area – closely mirroring the phased-array results reported in the paper.
There’s also considerable headroom still to exploit. “We have a significant amount of headroom available and are using only 0.003% of the antenna’s -10dB bandwidth in our demonstration,” she said. The reason the full band hasn’t been pushed further isn’t a limitation of the antenna itself: “Compact implantable RFIC electronics have not yet been integrated, and a single implant rarely needs such enormous capacity” – a gap she says opens up “avenues for exploitation even outside the biomedical industry.”
What stands between prototype and patient
Asked what the biggest remaining barrier is before the technology could be used in a real patient, Shojaei Baghini pointed squarely at the electronics rather than the antenna itself. “As these devices exploit acoustic losses to in turn increase bandwidth, the main engineering barrier lies in the development of RF integrated circuits,” she told Electronic Specifier. “Low-power RFICs are maturing thanks to the rapid shrinking of CMOS technology nodes. We are keen to carry this out soon for complete integration with individuals.”
The team has already taken a step in that direction. In partnership with colleagues in Italy, they tested the µBots’ performance using real biological tissue – rat brain tissue, human cortical brain slices, and controlled cell cultures – with performance holding up reliably across each, alongside cyclic stability tests to confirm the antennas remain reliable under conditions closely matching the real world.
“With our partners in Italy, Dr Adam Armada-Moreira and Professor Michele Giugliano, we have obtained ethical approvals to test these devices in human, ex-vivo [tissue], which puts us another step closer to in vivo tests in individuals,” Shojaei Baghini said.
Dr Adam Armada-Moreira of the University of Modena and Reggio Emilia, one of the paper’s corresponding neuroscientists, said the technology could open new possibilities for studying the brain over long periods: “This technology could let researchers and clinicians map and modulate neural circuits with higher spatial selectivity while stable telemetry supports chronic electrophysiology and neuromodulation studies.”
Next steps
Professor Hadi Heidari, who leads the University of Glasgow‘s Microelectronics Lab and is one of the paper’s corresponding authors, said the results build on years of collaborative work. “These are significant results, which build on previous research here at Glasgow and from our partners across Europe to demonstrate the significant potential for µBots to deliver transformative results for clinicians and for patients alike,” he said. “We’ll continue to explore the potential of these devices as we look to test them in further trials and work towards commercialising the underlying technology in the years to come.”
