Research On High Performance Metamaterial Sensor

As the gap of the electromagnetic spectrum and with the frequency ranging from0.1THz ～ 10 THz,terahertz waves own unique features of instantaneity,high transmission and low photon energy.Terahertz frequency band is precisely within the vibration and rotation frequencies of numerous organic macromolecules.Moreover,terahertz waves cause low ionization damage to biological tissues,mainly due to low energies.Therefore,terahertz sensors have significant applications in biomedical and chemical fields.Metamaterials are artificial materials with a several electromagnetic properties that cannot be obtained in nature,such as negative conductivity,and negative relative permittivity,negative permeability.By artificially adjusting the structure,size and material of metamaterials,different electromagnetic properties can be acquired.Based on the structural characteristics of terahertz wave and metamaterial,this thesis studies the performance of terahertz metamaterial refractive index sensors.The main contents are as follows:1.A metamaterial refractive index sensor with simple semiconductor InN material as the unit resonant structure is designed.When there is no sample to be measured,the central resonant frequency of the sensor is 4.32 THz,with the absorption rate of 99%achieved at this point,and the Q value of the absorption curve is 96.1.The resonance mechanism of the sensor at the resonance point is analyzed,and the geometric parameters of the sensor are optimized.The variation of the electromagnetic spectrum when the object to be measured with a different refractive index is loaded on the sensor’s surface is studied.The simulation results show that the sensitivity of the sensor is 897 GHz / RIU and the FOM value is 19.9,which can realize high-performance sensing.2.The Q value,sensitivity and FOM value of the metamaterial refractive index sensor obtained by changing the above InN unit resonance into microfluidic cavity structure are improved.99% and 308.3 are obtained at the frequency point of 4.317 THz.By changing the refractive index of the medium in the cavity and simulating the refractive index changes of different objects to be measured,it is measured that the structural sensitivity of the microfluidic cavity can reach 950 GHz / RIU and the FOM values is 40.6.The microfluidic cavity can realize high-performance liquid sensing.3.As the design and optimization of a metamaterial sensor require a lot of computing power and time,a method to learn deeply is introduced for the simulation prediction and inverse parameter design of semiconductor structure metamaterial sensors.Taking the InN terahertz metamaterial biosensor studied in this thesis as an example,the forward prediction network model and inverse design network model based on a fully connected neural network are established.The model can be pre-screened by inputting the parameter characteristics of a similar structure and immediately outputting the approximate electromagnetic spectrum and sensor performance parameters through the neural network.Meanwhile,it can input the target electromagnetic spectrum and obtain the relatively optimal parameters and structure through the neural network.