A hybrid inorganic–organic neuromorphic sensor utilizing a thin film zinc oxide (ZnO) detector with organic neuromorphic pre-processing is developed to quantify ammonia in aqueous environments, including biological analytes. Impedimetric ZnO sensor, connected to an organic somatic circuit, reliably and accurately detects changes in electrical impedance to measure and quantify variations in the concentration of ammonia. The sensing mechanism of the ZnO thin film sensor is hypothesized to be the cause of the decrease in resistance of a solution with an increase in ammonia concentration. It is found that the surface oxide of the ZnO layer reacts with even very low concentrations of ammonia ($N{H}_3$), leading to changes in resistivity. This makes the sensor capable of detecting ammonia in a range of concentrations between 0.0001 and 0.1 M. A neuromorphic circuit converts the analog change of ammonia concentration expressed as a change of sensor impedance to the digitized frequency of spikes. Detecting such a low ammonia concentration is critical for environmental monitoring and medical diagnosis. The digitized nature of neuromorphic signal pre-processing makes it more resilient for signal transmission in the presence of noise and serves as a demonstration of “smart sensing.”

中文翻译：

---

具有神经形态检测功能的氨水溶液传感器


开发了一种混合无机-有机神经形态传感器，利用薄膜氧化锌 （ZnO） 检测器和有机神经形态预处理来量化水环境中的氨，包括生物分析物。阻抗式 ZnO 传感器连接到有机体细胞电路，可靠、准确地检测电阻抗的变化，以测量和量化氨浓度的变化。据推测，ZnO 薄膜传感器的传感机制是导致溶液电阻随着氨浓度的增加而降低的原因。研究发现，ZnO 层的表面氧化物甚至会与非常低浓度的氨 （$N{H}_3$） 发生反应，从而导致电阻率发生变化。这使得传感器能够检测浓度在 0.0001 至 0.1 M 之间的氨。神经形态电路将表示为传感器阻抗变化的氨浓度的模拟变化转换为尖峰的数字化频率。检测如此低的氨浓度对于环境监测和医学诊断至关重要。神经形态信号预处理的数字化特性使其在存在噪声的情况下对信号传输更具弹性，并作为“智能传感”的演示。