Coulomb blockade-like transport and enhanced memory in organic transistors embedded with sub-nm Pt nanoparticles for neuromorphic computing

The human brain does not store information like a computer hard drive. Instead, it relies on a vast network of connections between nerve cells, where the strength of each link can change based on experience. This ability to strengthen or weaken connections over time is called plasticity, and it is the physical basis of learning and memory. Some connections change instantly but fade quickly, allowing us to hold a phone number in our mind for a few seconds. Others change slowly and last for years, forming our long-term knowledge. To build computers that think more like humans, scientists are trying to create electronic devices that can mimic this flexible behavior. The challenge lies in making these devices simple, cheap, and capable of holding many different states of memory, rather than just the simple on or off switches used in current technology.

A team of researchers at the University of Missouri has taken a significant step toward this goal by creating a new type of electronic switch that behaves remarkably like a biological synapse. They built a transistor, a fundamental component of modern electronics, using a special organic plastic material. To give this device its memory, they inserted a layer of tiny platinum particles, each smaller than a single nanometer, sandwiched between two ultra-thin sheets of aluminum oxide. This arrangement sits right at the boundary where the plastic meets the insulating layer. The result is a device that can remember information for a long time, but also display short-term memory effects that are crucial for complex thinking. What makes this discovery particularly striking is that the tiny platinum particles cause the flow of electricity to behave in a way that resembles quantum physics, a phenomenon usually only seen at temperatures near absolute zero, yet here it happens at room temperature.

The researchers constructed these devices using a bottom-gate design, where a metal gate sits at the base, covered by a ferroelectric plastic layer that can hold an electric charge. On top of this, they grew a precise layer of aluminum oxide, deposited the platinum particles, and capped it with another thin layer of aluminum oxide. Finally, they added a layer of a semiconducting plastic called DPP-DTT and metal contacts to complete the circuit. When they tested these transistors, they found something unexpected. As they swept the voltage up and down, the current flowing through the device did not change smoothly. Instead, it moved in distinct steps, with flat regions followed by sharp spikes. These plateaus and spikes are signatures of a quantum effect known as Coulomb blockade, where the tiny particles resist letting electrons pass until a specific amount of energy is reached. Because the platinum particles are so small, they act like individual islands that can trap single electrons, creating discrete levels of memory that are very well separated from one another.

This quantum behavior translates into a very large and stable memory window. The device can be switched between different conductance states, representing different memory values, with a voltage range of over 20 volts. This is much larger than what is typically seen in similar organic devices. The researchers observed that the platinum particles could hold a negative charge, which effectively shifts the voltage needed to turn the transistor on. When the voltage is changed, the particles release or capture electrons in a controlled manner, creating the distinct steps in the current. The team also noticed that the spacing between the current spikes was regular, suggesting that the particles were behaving in a unified way, almost as if they were a single sheet of charge that screened the electric field in a predictable pattern.

Beyond its electrical memory, the device showed a remarkable ability to be programmed by light. When the researchers shined a green or blue laser on the transistor, the flow of current increased, effectively writing new information into the device. This is because the light creates extra charge carriers in the plastic layer, some of which get trapped by the platinum particles. To erase this information, they applied a specific electrical pulse to the gate, which released the trapped charges. This dual capability allows the same device to be written with light and erased with electricity, mimicking how biological synapses can be influenced by different types of signals. The researchers tested this by using the device to recognize handwritten numbers from a standard database. When trained with the light-induced changes, the system achieved an accuracy of about 83 percent, demonstrating that the memory states were stable and useful for complex tasks.

Perhaps the most exciting finding for future brain-like computers was the device's ability to exhibit short-term plasticity. In biology, if two signals arrive at a synapse in quick succession, the second signal often produces a stronger response than the first. This is called paired-pulse facilitation. The researchers tested this by flashing two light pulses at the device with a very short gap between them. They found that the second response was more than twice as strong as the first, a level of facilitation that matches biological synapses. This short-term effect faded away over a few seconds, just like in the brain, suggesting that the trapped charges were slowly relaxing back to their original state. The device could therefore handle both long-term memory, which persists for a long time, and short-term memory, which is essential for processing immediate information.

The study rules out the idea that these effects are caused by simple defects or random noise. The control devices, which lacked the platinum particles, showed none of these behaviors. They had no memory window, no current spikes, and no short-term plasticity. This confirms that the unique properties come directly from the engineered layer of platinum nanoparticles. The researchers also found that the size of the particles mattered; slightly larger particles produced different spacing in the current steps, aligning with theoretical predictions about how the energy levels of these tiny clusters change with size. The consistency of the results across different devices suggests that the method of creating these layers is robust and repeatable.

By combining a large memory window, quantum-like transport, and dual-mode programming, this work opens a new path for neuromorphic computing. The device proves that organic materials, often considered too disordered for precise quantum effects, can be engineered to support highly controlled charge trapping. The ability to use both light and electricity to control the memory states offers a versatile platform for future systems. While the current devices are still in the research phase, the demonstration of both long-term and short-term plasticity in a single, simple component suggests that we are getting closer to building electronic systems that can learn and adapt in ways that mirror the natural world. The findings suggest that with further refinement, these materials could become the building blocks for computers that are not only powerful but also efficient and capable of handling the complex, fluid nature of human thought.