David Savastano, Editor04.13.22
Flexible electronics are a reality today, but the ability to drive these electronics with flexible microprocessors is a key opportunity for countless new applications, from sensors to the Internet of Things. The microprocessors wouldn’t need to match the complexity of silicon CMOS-based microprocessors, but added functionality would be a huge benefit.
There is research being done in the area, and during the 2022 International Solid-State Circuits Conference (2022 ISSCC), imec, KU Leuven and PragmatIC Semiconductor showcased their new 8-bit flexible microprocessor for low-power applications.
Imec designed the flexible 8-bit microprocessor in 0.8µm indium-gallium-zinc-oxide (IGZO)-transistor technology, and is capable of running real-time complex assembly code. The partners noted that PragmatIC Semiconductor was able to integrate the approximately ~16,000 metal-oxide thin-film transistors on a 24.9mm2flexible chip at its FlexIC foundry.
This work has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme under grant agreement No 716426 (FLICs project).
Imec principal scientist Kris Myny noted that imec and KU Leuven have been collaborating on flexible electronics design activities for more than a decade. The specific research towards flexible microprocessors started with Prof. Myny’s ERC starting grant FLICs (2017-2021) awarded by the European Research Council on the development of flexible integrated circuits and systems.
“Within KU Leuven, it is mainly the MICAS department, led by Prof. Wim Dehaene who was already present at the start of the collaboration,” said Myny. “The main goal of the ERC project is to enable sub-micrometer transistors on flexible substrates, while keeping an eye on robustness and low power consumption of envisioned complex flexible chips such as the processor.
Myny noted that the relationship between PragmatIC and imec can be traced back for more than a decade as well, and was as an example visible in an European funded project PING (2015-2017) for the development of metal-oxide based NFC tags.
“Recently, during the ERC project, imec and KU Leuven have been designing in PragmatIC’s downscaled thin-film transistor technologies, which led to this flexible microprocessor. In general, the effective work specifically towards the microprocessor may be counted less than one year, however, this is excluding the longer pre-work and pre-studies on this topic,” added Myny.
Ph.D. researcher Hikmet Çeliker noted that the teams of imec and KU Leuven collaborating on this topic are design experts for Si CMOS and thin-film transistor technologies.
“Therefore, the design studies towards the flexible microprocessor, the library design and the full implementation of the digital flow have been completed by those teams, bringing the technology-aware design know-how to this project,” observed Çeliker. “In addition to this, the imec and KU Leuven teams have focused on the characterization of the designed chips after production by PragmatIC.
“The main role of PragmatIC in this work is to offer a foundry model similar to the existing Si CMOS models,” Çeliker added. “In such a model, the foundry PragmatIC would provide a process development kit (PDK) for their specific technology to enable third-parties to design in their technologies, after which PragmatIC would fabricate the designed circuits. The quality of their PDK and of their fabrication have enabled imec and KU Leuven to realize successfully the microprocessor.”
The 8-bit flexible microprocessor offers significant benefits, beginning with low processing temperature.
“The key advantage of thin-film transistors is their low process temperature, which enables the manufacturing on flexible substrates,” Myny reported. “This is therefore one of the key differentiators to Si CMOS technologies, as flexible chips can be physically flexible, conformable and bendable thanks to the flexible substrate.
“Thin-film transistor technologies result in ultrathin chips that can be seamlessly embedded in everyday objects,” added Myny. “Last, but not least, the technology has the potential to result in very low cost per unit area for the final chip, paving the way to enable low-cost internet-of-things applications.”
Çeliker said that while flexible microprocessors are not meant to replace the current Si CMOS processors due to the performance differences, flexible technologies exhibit intrinsic lower performance (about 100x) compared to Si CMOS, have larger device sizes compared to Si CMOS and lack the availability of complementary devices.
“This leads inevitably to a strong performance gap in terms of power, speed and area,” Çeliker oserved. “Nonetheless, the advantages, such as flexibility and low cost per area, pave the way for low-cost and low-power Internet of Things (IoT) applications which do not require very high performance. The main purpose of this flexible microprocessor chip could be to perform basic digital calculations at the sensor node. Another potential market would be wearable healthcare patches to monitor vital signs of patients, which again requires basic signal processing at the patch level.”
Myny believes the performance of the demonstrated microprocessor is on par with the envisioned applications in terms of clock speed requirements.
“A crucial improvement to be proposed would be to reduce the static power consumption in order to enable RF-powered and battery-powered systems,” added Myny. “The power consumption is today still large due to the availability of only n-type transistors. Improvements in power consumption would require either to disrupt in p-type devices or to elaborate deeper on the digital cell library reducing the static leakage currents.
Myny said that another key element for a real application would be the memory.
“At the moment of writing, a good non-volatile memory is not available to be integrated in this transistor technology, which needs several developments,” Myny pointed out. “As an alternative concept, the implementation of a ROM (read-only memory) or WORM (write-once read many) memory can be targeted. These memory concepts are available and can be programmed by means of laser ablation.”
Çeliker reported that the partners are continuing their work on the flexible microprocessor, with an eye on commercializing this technology in a few years.
“The key benefit of foundry model collaboration will allow a closer go-to-market strategy, as designs are being validated immediately in the appropriate technology,” Çeliker noted. “After validation of the designs during the characterization, the chips can be commercialized for the specific markets of interest.
“Of course, as the chip still need some developments and improvements, we consider a time-to-market of three to five years, which includes a new generation microprocessor that consumes less power,” Çeliker added. “Besides this, the other elements in a system, such as an analog-to-digital convertor or memory block, need to be developed as well in parallel.”
There is research being done in the area, and during the 2022 International Solid-State Circuits Conference (2022 ISSCC), imec, KU Leuven and PragmatIC Semiconductor showcased their new 8-bit flexible microprocessor for low-power applications.
Imec designed the flexible 8-bit microprocessor in 0.8µm indium-gallium-zinc-oxide (IGZO)-transistor technology, and is capable of running real-time complex assembly code. The partners noted that PragmatIC Semiconductor was able to integrate the approximately ~16,000 metal-oxide thin-film transistors on a 24.9mm2flexible chip at its FlexIC foundry.
This work has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme under grant agreement No 716426 (FLICs project).
Imec principal scientist Kris Myny noted that imec and KU Leuven have been collaborating on flexible electronics design activities for more than a decade. The specific research towards flexible microprocessors started with Prof. Myny’s ERC starting grant FLICs (2017-2021) awarded by the European Research Council on the development of flexible integrated circuits and systems.
“Within KU Leuven, it is mainly the MICAS department, led by Prof. Wim Dehaene who was already present at the start of the collaboration,” said Myny. “The main goal of the ERC project is to enable sub-micrometer transistors on flexible substrates, while keeping an eye on robustness and low power consumption of envisioned complex flexible chips such as the processor.
Myny noted that the relationship between PragmatIC and imec can be traced back for more than a decade as well, and was as an example visible in an European funded project PING (2015-2017) for the development of metal-oxide based NFC tags.
“Recently, during the ERC project, imec and KU Leuven have been designing in PragmatIC’s downscaled thin-film transistor technologies, which led to this flexible microprocessor. In general, the effective work specifically towards the microprocessor may be counted less than one year, however, this is excluding the longer pre-work and pre-studies on this topic,” added Myny.
Ph.D. researcher Hikmet Çeliker noted that the teams of imec and KU Leuven collaborating on this topic are design experts for Si CMOS and thin-film transistor technologies.
“Therefore, the design studies towards the flexible microprocessor, the library design and the full implementation of the digital flow have been completed by those teams, bringing the technology-aware design know-how to this project,” observed Çeliker. “In addition to this, the imec and KU Leuven teams have focused on the characterization of the designed chips after production by PragmatIC.
“The main role of PragmatIC in this work is to offer a foundry model similar to the existing Si CMOS models,” Çeliker added. “In such a model, the foundry PragmatIC would provide a process development kit (PDK) for their specific technology to enable third-parties to design in their technologies, after which PragmatIC would fabricate the designed circuits. The quality of their PDK and of their fabrication have enabled imec and KU Leuven to realize successfully the microprocessor.”
The 8-bit flexible microprocessor offers significant benefits, beginning with low processing temperature.
“The key advantage of thin-film transistors is their low process temperature, which enables the manufacturing on flexible substrates,” Myny reported. “This is therefore one of the key differentiators to Si CMOS technologies, as flexible chips can be physically flexible, conformable and bendable thanks to the flexible substrate.
“Thin-film transistor technologies result in ultrathin chips that can be seamlessly embedded in everyday objects,” added Myny. “Last, but not least, the technology has the potential to result in very low cost per unit area for the final chip, paving the way to enable low-cost internet-of-things applications.”
Çeliker said that while flexible microprocessors are not meant to replace the current Si CMOS processors due to the performance differences, flexible technologies exhibit intrinsic lower performance (about 100x) compared to Si CMOS, have larger device sizes compared to Si CMOS and lack the availability of complementary devices.
“This leads inevitably to a strong performance gap in terms of power, speed and area,” Çeliker oserved. “Nonetheless, the advantages, such as flexibility and low cost per area, pave the way for low-cost and low-power Internet of Things (IoT) applications which do not require very high performance. The main purpose of this flexible microprocessor chip could be to perform basic digital calculations at the sensor node. Another potential market would be wearable healthcare patches to monitor vital signs of patients, which again requires basic signal processing at the patch level.”
Myny believes the performance of the demonstrated microprocessor is on par with the envisioned applications in terms of clock speed requirements.
“A crucial improvement to be proposed would be to reduce the static power consumption in order to enable RF-powered and battery-powered systems,” added Myny. “The power consumption is today still large due to the availability of only n-type transistors. Improvements in power consumption would require either to disrupt in p-type devices or to elaborate deeper on the digital cell library reducing the static leakage currents.
Myny said that another key element for a real application would be the memory.
“At the moment of writing, a good non-volatile memory is not available to be integrated in this transistor technology, which needs several developments,” Myny pointed out. “As an alternative concept, the implementation of a ROM (read-only memory) or WORM (write-once read many) memory can be targeted. These memory concepts are available and can be programmed by means of laser ablation.”
Çeliker reported that the partners are continuing their work on the flexible microprocessor, with an eye on commercializing this technology in a few years.
“The key benefit of foundry model collaboration will allow a closer go-to-market strategy, as designs are being validated immediately in the appropriate technology,” Çeliker noted. “After validation of the designs during the characterization, the chips can be commercialized for the specific markets of interest.
“Of course, as the chip still need some developments and improvements, we consider a time-to-market of three to five years, which includes a new generation microprocessor that consumes less power,” Çeliker added. “Besides this, the other elements in a system, such as an analog-to-digital convertor or memory block, need to be developed as well in parallel.”