Silicon nanowire devices: from a biosensor to an artificial neuron


NISE Seminar | event contribution
April 18, 2018 | MPI of Microstructure Physics

Synergy between, physics, material sciences and biotechnology during last decade has led to a tremendous scientific progress in the fields of biodetection and nanomedicine. This tight interaction led to the emergence of a new class of bioinspired systems that enables to bring the area of biosensorics e.g. for cell or molecular diagnostics and analytics to the new level. The advances are expected in terms of (i) possibility of early diagnostics of diseases due to the increased sensitivity of the detectors, (ii) real time and high throughput analysis offered by combination of integrated electronics and microfluidic approach, and (iii) establishing the new functional formats for the bioassays. Most promising candidates for the future diagnostics are the electronic nanobiosensors that have attracted great attention in the last decades since they provide rich quantitative information for medical and biotechnological assays without pre-treatment and specific optical labelling of the detected species. One dimensional nanostructures, in particular semiconductor and metallic nanowires, have attracted attention as highly efficient sensor elements due to their high surface-to-volume ratio, which simplifies the detection of biochemical species down to single molecules.
Apart from biodetection, the semiconductor nanowire devices have a great potential to make revolution in the digital electronics, offering the possibility of new architecture for the system. It is certain that accomplishing the complex human-like tasks like recognition, judgement or inference cannot be reached only by means of aggressive scaling of the current CMOS technology, but rather via breaking the paradigm in the device architectures and their functionalities. Brain-inspired neuromorphic architectures do open new hardware opportunities that are envisioned to merge learning and memory functions within one unit cell, similar to a neuron. In this respect, an artificial neuron emerges as a key element of a physical neural network and an element which is able to emulate neuroplasticity, critical to assure the simultaneous processing and memorizing of information. Surprisingly, in contrast to artificial synapses that can be considered as two-terminal devices, the structure of artificial neuron has striking similarity to a transistor and is out of reach by now.


Authors

Silicon nanowire devices: from a biosensor to an artificial neuron


NISE Seminar | event contribution
April 18, 2018 | MPI of Microstructure Physics

Synergy between, physics, material sciences and biotechnology during last decade has led to a tremendous scientific progress in the fields of biodetection and nanomedicine. This tight interaction led to the emergence of a new class of bioinspired systems that enables to bring the area of biosensorics e.g. for cell or molecular diagnostics and analytics to the new level. The advances are expected in terms of (i) possibility of early diagnostics of diseases due to the increased sensitivity of the detectors, (ii) real time and high throughput analysis offered by combination of integrated electronics and microfluidic approach, and (iii) establishing the new functional formats for the bioassays. Most promising candidates for the future diagnostics are the electronic nanobiosensors that have attracted great attention in the last decades since they provide rich quantitative information for medical and biotechnological assays without pre-treatment and specific optical labelling of the detected species. One dimensional nanostructures, in particular semiconductor and metallic nanowires, have attracted attention as highly efficient sensor elements due to their high surface-to-volume ratio, which simplifies the detection of biochemical species down to single molecules.
Apart from biodetection, the semiconductor nanowire devices have a great potential to make revolution in the digital electronics, offering the possibility of new architecture for the system. It is certain that accomplishing the complex human-like tasks like recognition, judgement or inference cannot be reached only by means of aggressive scaling of the current CMOS technology, but rather via breaking the paradigm in the device architectures and their functionalities. Brain-inspired neuromorphic architectures do open new hardware opportunities that are envisioned to merge learning and memory functions within one unit cell, similar to a neuron. In this respect, an artificial neuron emerges as a key element of a physical neural network and an element which is able to emulate neuroplasticity, critical to assure the simultaneous processing and memorizing of information. Surprisingly, in contrast to artificial synapses that can be considered as two-terminal devices, the structure of artificial neuron has striking similarity to a transistor and is out of reach by now.


Authors