The new semiconductor will allow you to assemble electronics without external processors.
The usual diode can only one thing: it transmits the current in one direction and blocks it in another. This principle is based almost all modern electronics. But for signal processing, storing data and working with images, additional elements are usually needed: memory, processors, sensors. The new development of Chinese and Canadian researchers proposes to combine these functions in one component.
A group from the University of Science and Technology of China and McGill University has introduced a p-n diode that combines three functions at once: he perceives light, stores information and participates in signal processing.
The basis is a well-known p-n-transition design. Such a transition forms two types of semiconductor: a p-type area with an excess of holes, that is, carriers of a positive charge, and an area of n-type with an excess of electrons. At the border of these areas, a transitional region arises, which sets a one-way current movement. In the classic version, the diodes end there.
The new scheme uses the same basic principle, but adds to it the fine-tuning of the material’s zone structure. Researchers have grown vertical nanoniths on a silicon substrate. Each thread consists of three layers: the layer of gallia nitride p-type (GaN), then a layer of aluminum-galle n-type n-type (AlGaN) layer and another layer of the GAN n-type.
The key detail here is the middle layer of AlGaN. It has a wider prohibited area, meaning the electrons need more energy to go through this area. As a result, the layer works as a barrier and at the same time as a charge trap. Electrons accumulate in a certain area and form a kind of charge tank inside the p-n-transition.
This design allows you to control the capture and release of charge carriers much more accurately than in the usual diode. Due to this, the same element begins to behave differently depending on the conditions. When illuminating, it works as a photodetector and responds to light with a sensitivity of about 10.45 mA / W. In certain modes of excitation, an effect similar to the behavior of synapses in neuromorphic systems is manifested: the response is enhanced by repeated pulses, and the coefficient of paired relief reaches 122%. In other modes, the device fixes the state and behaves like a memory with eight stable levels.
In fact, one element combines the functions of the sensor, memory and computing node. In traditional schemes, it would be necessary to collect several different components and connect them with additional logic.
In experiments, the researchers collected an array of such diodes and used it as a compact image sensor. The device not only fixed the picture, but also immediately performed basic processing: it suppressed noise and recognized the image without external computing blocks. This approach reduces the number of elements in the scheme and reduces energy consumption.
The work is based on the so-called zone structure engineering, that is, the management of energy levels within the material. It is this approach that allows you to embed several functions in one p-n-transition without adding new contacts or complication of architecture.
If the technology can be scaled, such diodes can form the basis of compact electronics, where the sensor, memory and processing are combined in one place. This is important for devices with strict size and power consumption restrictions: wearable electronics, IoT sensors and built-in data processing systems.
The usual diode can only one thing: it transmits the current in one direction and blocks it in another. This principle is based almost all modern electronics. But for signal processing, storing data and working with images, additional elements are usually needed: memory, processors, sensors. The new development of Chinese and Canadian researchers proposes to combine these functions in one component.
A group from the University of Science and Technology of China and McGill University has introduced a p-n diode that combines three functions at once: he perceives light, stores information and participates in signal processing.
The basis is a well-known p-n-transition design. Such a transition forms two types of semiconductor: a p-type area with an excess of holes, that is, carriers of a positive charge, and an area of n-type with an excess of electrons. At the border of these areas, a transitional region arises, which sets a one-way current movement. In the classic version, the diodes end there.
The new scheme uses the same basic principle, but adds to it the fine-tuning of the material’s zone structure. Researchers have grown vertical nanoniths on a silicon substrate. Each thread consists of three layers: the layer of gallia nitride p-type (GaN), then a layer of aluminum-galle n-type n-type (AlGaN) layer and another layer of the GAN n-type.
The key detail here is the middle layer of AlGaN. It has a wider prohibited area, meaning the electrons need more energy to go through this area. As a result, the layer works as a barrier and at the same time as a charge trap. Electrons accumulate in a certain area and form a kind of charge tank inside the p-n-transition.
This design allows you to control the capture and release of charge carriers much more accurately than in the usual diode. Due to this, the same element begins to behave differently depending on the conditions. When illuminating, it works as a photodetector and responds to light with a sensitivity of about 10.45 mA / W. In certain modes of excitation, an effect similar to the behavior of synapses in neuromorphic systems is manifested: the response is enhanced by repeated pulses, and the coefficient of paired relief reaches 122%. In other modes, the device fixes the state and behaves like a memory with eight stable levels.
In fact, one element combines the functions of the sensor, memory and computing node. In traditional schemes, it would be necessary to collect several different components and connect them with additional logic.
In experiments, the researchers collected an array of such diodes and used it as a compact image sensor. The device not only fixed the picture, but also immediately performed basic processing: it suppressed noise and recognized the image without external computing blocks. This approach reduces the number of elements in the scheme and reduces energy consumption.
The work is based on the so-called zone structure engineering, that is, the management of energy levels within the material. It is this approach that allows you to embed several functions in one p-n-transition without adding new contacts or complication of architecture.
If the technology can be scaled, such diodes can form the basis of compact electronics, where the sensor, memory and processing are combined in one place. This is important for devices with strict size and power consumption restrictions: wearable electronics, IoT sensors and built-in data processing systems.
