Two New Interesting Carbon-Based Micro-Electronics Developments: Memristors from Coal Feedstock and the New Graphene Chip

 

Amorphous Carbon Nanomembranes from Coal-Based Carbon Dots

     Coal is the feedstock for new atom-thick high-purity materials that can significantly improve the performance of micro-electronics. More specifically, they can improve the performance of tiny transistors known as memristors. In the process, coal char is converted into nanoscale disks known as “carbon dots,” which are used to create membranes used in two-dimensional transistors and memristors. Research by the NREL has shown these memristors to deliver superior performance. “These atomically thin carbon dots form the foundation for creating membranes essential in advanced electronic technologies, notably two-dimensional transistors and memristors, crucial components for the next generation of microelectronics.” These ultra-thin carbon layers act as insulators for ultra-thin semi-conductors. Experiments with the new carbon dots have revealed more than doubled processing speed, energy savings/increased efficiency, and low leakage. The Abstract of the paper in Nature Communications Engineering summarized the new research: “These atomically thin amorphous carbon films are mechanically strong with modulus of 400 ± 100 GPa and demonstrate robust dielectric properties with high dielectric strength above 20 MV cm−1 and low leakage current density below 10−4 A cm−2 through a scaled thickness of three-atomic layers. They can be implemented as solution-deposited ultrathin gate dielectrics in transistors or ion-transport media in memristors, enabling exceptional device performance and spatiotemporal uniformity.”

     Another potential benefit is better data storage reliability for AI processes. The NREL researchers are working with the company Taiwan Semiconductor to develop these two-dimensional devices. NREL described the implications of the research as follows: “Memristor computer memory will enable machine learning and artificial intelligence by making data storage and processing devices smaller, faster, and more energy efficient. The memory devices made by the team reduce energy consumption by 5-20-fold over conventional memristors and overcome device variation issues that have plagued the field.” The unique atomic arrangement of the carbon atoms in the materials is what makes the devices perform so well.

     I won’t pretend to understand the chemistry, or the electronics involved in this research but below is an image from the paper that shows “coal-derived carbon dot precursors,” and another that shows “application of quasi-2D amorphous carbon bilayers as the switching media in memristors to enable small cycle-to-cycle variability.”

 

 

 

 

Memristors

 

     According to Wikipedia: “A memristor (a portmanteau of memory resistor) is a non-linear two-terminal electrical component relating electric charge and magnetic flux linkage. It was described and named in 1971 by Leon Chua, completing a theoretical quartet of fundamental electrical components which also comprises the resistor, capacitor and inductor.” The concept was later expanded to the idea of memristive systems, or circuits. While the idea is still considered experimental and scale-up into real-world applications remains a hurdle, there is much interest and effort: “Memristors remain a laboratory curiosity, as yet made in insufficient numbers to gain any commercial applications. Despite this lack of mass availability, according to Allied Market Research the memristor market was worth $3.2 million in 2015 and was at the time projected to be worth $79.0 million by 2022. In fact, it was worth $190.0 million in 2022.”

 

“A potential application of memristors is in analog memories for superconducting quantum computers.”

 

     In the September 2023 issue of Science magazine, Chinese researchers “described the development and testing of a memristor-based integrated circuit, designed to dramatically increase the speed and efficiency of Machine Learning and Artificial Intelligence tasks, optimized for Edge Computing applications.” They noted that memristor-based technology has the potential to overcome the so-called “von Neumann bottleneck” which impedes conventional computing architecture. This is described as “on-chip”" learning for certain applications. This technology has the potential to streamline the training of neural networks.

 

 

 

Georgia Tech’s Graphene Chip May Enable Graphene Electronics

 

    Researchers at Georgia Tech have created the world’s first functioning graphene-based semiconductor chip. This may be the first breakthrough that paves the way for a post-silicon era in semiconductors. Specifically, the researchers were able to create a bandgap in graphene to make it function as a semiconductor. The bandgap allows the graphene chip to be turned on and off like silicon chips. The researchers were able to grow graphene on silicon carbide wafers, creating what is known as epitaxial graphene. Graphene has 10 times the electron mobility, or conductivity of silicon and is very durable. The discovery should enable much faster computing speeds. Growing graphene under high temperatures enabled the graphene to develop semiconductor properties. Like the atom-thick carbon dots mentioned in the above section, epitaxial graphene is considered to be a two-dimensional semiconductor.  Lead researcher Walt de Heer noted: “One main aspect of graphene electronics is that we can utilize the quantum-mechanical wave properties of the electrons and [electron] holes which are not accessible in silicon electronics.” The new research opens the door to graphene electronics, which may one day replace silicon electronics as it replaced vacuum tubes. It also has the potential to enable advanced quantum computing devices.

     The crystal structure of graphene is chemically bonded to silicon carbide (SiC) and has been dubbed semiconducting epitaxial graphene (SEC), or epigraphene. The SiC is heated to over 1000 degrees Celsius, which causes the silicon to evaporate and the carbon to form graphene. The heating is done in an argon quartz tube with a copper coil through induction. De Heer also noted: “The chips we use cost about [US] $10, the crucible about $1, and the quartz tube about $10.” Thus far, this seems to be the most significant method to introduce the needed bandgap into graphene electronics. This bandgap has long been the main difficult-to-overcome hurdle to enable graphene electronics. Other approaches have been unsatisfactory. According to an article in IEEE Spectrum: “Our research is distinct from these other approaches because we have produced large areas of semiconducting SEC on defect-free, atomically flat SiC terraces,” says de Heer. “SiC is a highly developed, readily available electronic material that is fully compatible with conventional microelectronics processing methods.” The researchers do note, however, that it will take time to develop these techniques further, to incorporate graphene-based electronics with silicon-based electronics, and to eventually develop graphene electronics that can replace silicon electronics. One way being researched to integrate SEC with conventional electronics is coating SEC with boron nitride. The figures below are from the recent paper in Nature. 

 

 

 

References:

New study discovers coal's unexpected role in next-gen microelectronics. Abdul-Rahman Oladimeji Bello, Interesting Engineering. January 5, 2024. New study discovers coal's unexpected role in next-gen microelectronics (msn.com)

Ultrathin quasi-2D amorphous carbon dielectric prepared from solution precursor for nanoelectronics. Fufei An, Congjun Wang, Viet Hung Pham, Albina Borisevich, Jiangchao Qian, Kaijun Yin, Saran Pidaparthy, Brian Robinson, Ang-Sheng Chou, Junseok Lee, Jennifer Weidman, Sittichai Natesakhawat, Han Wang, André Schleife, Jian-Min Zuo, Christopher Matranga & Qing Cao. Communications Engineering volume 2, Article number: 93. December 20, 2023. Ultrathin quasi-2D amorphous carbon dielectric prepared from solution precursor for nanoelectronics | Communications Engineering (nature.com)

Coal-derived Carbon Material Improves Performance and Efficiency of Computer Microelectronic Devices. National Energy Technology Laboratory. December 20, 2023. Coal-derived Carbon Material Improves Performance and Efficiency of Computer Microelectronic Devices | netl.doe.gov

Georgia Tech's Groundbreaking Graphene Chip Could Herald a New Era in Electronics. Ethan Brown. Trendy Digests. January 22, 2024. Georgia Tech's Groundbreaking Graphene Chip Could Herald a New Era in Electronics (msn.com)

Memristor. Wikipedia. Memristor - Wikipedia

Edge learning using a fully integrated neuro-inspired memristor chip. Wenbin Zhang, Peng Yao, Bin Gao, Qi Liu, Dong Wu, Qintian Zhang, Yuankun Li, Qi Qin, Jiaming Li, and Huaqiang Wu. Science. September 14, 2023. Vol 381, Issue 6663. pp. 1205-1211. Edge learning using a fully integrated neuro-inspired memristor chip | Science

Ultrahigh-mobility semiconducting epitaxial graphene on silicon carbide. Jian Zhao, Peixuan Ji, Yaqi Li, Rui Li, Kaimin Zhang, Hao Tian, Kaicheng Yu, Boyue Bian, Luzhen Hao, Xue Xiao, Will Griffin, Noel Dudeck, Ramiro Moro, Lei Ma & Walt A. de Heer. Nature. volume 625, pages60–65 (2024). Ultrahigh-mobility semiconducting epitaxial graphene on silicon carbide | Nature

Researchers Claim First Functioning Graphene-Based Chip The semiconductor bests silicon alternatives for electron mobility. Dexter Johnson. IEEE Spectrum. January 18, 2024. Researchers Claim First Functioning Graphene-Based Chip - IEEE Spectrum