NanoThermoMechanical Thermal Computing: NanoPhotonics Metamaterials for High Temperature Memories and Logic Devices
Sidy Ndao (Mechanical and Materials Engineering)
NASA has been interested in the question: how to record data on the surface of planet Venus when its average surface temperature is about 400oC? Clearly current state-of-the-art computing technologies can not function at such elevated temperatures. In the pursuit of alternative technologies, research has been focused on two main approaches, namely material research (i.e. alternative wide bandgap semiconductor materials such as silicon carbide and gallium nitride), and NanoElectroMechanical memory and switches, both of which still depend on Semiconductor properties and/or electricity. In his MRSEC project, Prof. Ndao’s group flipped the question and decided to use heat instead of electricity for computing: Thermal computing.
With this approach in mind, Prof. Ndao’s group developed the concept of NanoThermoMechanical memory and logic devices. Unlike electronic memory, a NanoThermoMechanical memory device uses heat instead of electricity to record, store, and recover data. Memory function is achieved through the coupling of Near-Field Thermal Radiation (NFTR) and thermal expansion resulting in negative differential thermal resistance and thermal latching. NFTR is a mode of transferring heat via thermal radiation between two surfaces, which occurs when the vacuum gap separating them becomes very small (i.e., comparable to the radiation wavelength). NFTR’s intensity has been shown to increase exponentially with decreasing separation gap. The increased NFTR intensity results from the tunneling of the evanescent surface waves between the two surfaces as they are brought close enough to each other. Thermal memory is a thermal system that has two equilibrium stable points for a given set of boundary conditions, which is not typical for thermal systems. The bistable thermal system requires nonlinear thermal resistance. Nonlinear thermal resistance, also called negative differential thermal resistance (NDTR), which can be thought of as increase in heat transfer from a hot to a cold body with decreasing temperature difference between the two (which seems to be counterintuitive). NanoThermoMechanical memory (shown in Figure A1) uses the coupling between thermal expansion and near-field radiative heat transfer to achieve NDTR. The memory device, as shown in Figure A1-a, consists of two terminals: a top terminal and a bottom terminal (base, stem, and head). The top terminal is kept at a high temperature (the hot terminal). For the bottom terminal, temperature is allowed to vary along its length with the base kept at a relatively low temperature (the cold terminal). The base is fixed while the head is free to move up and down as a result of thermal expansion. Data writing is done by slightly heating (to write ONE) or cooling (to write ZERO) the head about a threshold temperature, Tth, using a probe. Data reading is accomplished by measuring the temperature of the head. During memory function, heat is transferred from the hot terminal to the head by radiation; from the head to the cold base, heat is transferred by conduction through the stem. The NanoThermoMechanical Memory was shown to work at temperatures are high as 1,400 K. Results of this work have been published in Applied Physics Letters (Elzouka, M., Ndao, S., “Near-Field NanoThermoMechanical Memory,” Applied Physics Letters, 105, 243510, 2014).
Prof. Ndao also designed and prototyped a NanoThermoMechanical diode, an important building block in any thermal computing technology. His group showed experimentally—for the first time—the use of near-field thermal radiation (NFTR) to achieve thermal rectification at high temperatures, which can be used to build high temperature thermal diodes for performing logic operations in harsh and extreme environment. They fabricated and tested a proof-of-concept NanoThermoMechanical device that has shown a maximum rectification of 10.9% at terminals’ temperatures of 375 and 530 K.
The NanoThermoMechanical rectifier is conceptually composed of a fixed terminal (the upper part), a moving terminal (the lower part), and a thermally-expandable structure connected to the moving terminal, as shown in A2-a. They achieved rectification through the coupling between NFTR and the size of a micro/nano gap separating the two terminals, engineered to be a function of heat flow direction. The principle of operation is illustrated in A2-a. The fabrication of a NanoThermoMechanical diode was achieved using conventional microfabrication techniques. A2-c,d show SEM images of the NanoThermoMechanical rectifier. The maximum thermal rectification achieved is 10.9% for Tlow = 375K and Thigh = 530K. The microdevice has also shown thermal rectification at Tlow as high as , with a rectification of 5.3%. This suggests that their technology can operate at elevated ambient temperatures without cooling. Results of this work have been published in Nature Scientific Reports (Elzouka, M., Ndao, S., “High Temperature Near-Field NanoThermoMechanical Rectification,” Nature Scientific Reports, 7, 44901, 2017). Their most recent work on the MRSEC project, published in the Journal of Quantitative Spectroscopy & Radiative Transfer, with the use of meshed doped Silicon Photonic Crystals shows thermal rectification as high as 2500% (Elzouka, M., Ndao, S., “Meshed Doped Silicon Photonic Crystals for Manipulating Near-Field Thermal Radiation,” Journal of Quantitative Spectroscopy & Radiative Transfer, 204, 56–62, 2018). (Figure A3)
The thermal computer can be powered by radio isotopes or even waste heat. Nearly 60 percent of the energy produced for consumption in the US is wasted in heat, if you could harness this heat and use it for energy in these devices, you could obviously cut down on waste and the cost of energy.
Prof. Ndao’s group is basically creating the world's first thermal computer, and hopefully one day it will be used to unlock the mysteries of outer space, explore and harvest our own planet's deep-beneath-the-surface geology and harness waste heat for more efficient energy utilization.
Figure A1: NanoThermoMechanical Memory
Figure A2: NanoThermoMechanical rectifier
Figure A3: (TOP) Heat transfer enhancement from meshed Photonic Crystals; (BOTTOM) Photonic Crystal dispersion relation