Theoretical And Experimental Study Of Resonant Hollow-core Terahertz Fiber

With the development of terahertz(THz)technology,the demand of the THz fiber as an important THz device is increasing.Compared with other types of fibers,the hollow-core THz fibers have many advantages such as easy inner modification with functional materials,low loss,flexible dispersion and anti-electromagnetic interference.The resonant hollow-core THz fibers have attracted great attentions in the hollow-core fiber,it’s necessary to design a new type of resonant hollow-core THz fiber with low loss and high transmission efficiency.In this paper,the Bragg resonant hollow-core terahertz fiber(BR-HCTF)and anti-resonant hollow-core terahertz fiber(AR-HCTF)are proposed.We focused on the theoretical and experimental research of the performance with BR-HCTF and AR-HCTF,besides,a machine learning-assisted resonant hollow core terahertz fiber simulation method is established.The work includes following parts.1.The theoretical research based on the physical models and simulation methods of BR-HCTF and AR-HCTF is proposed.The structural design and light guiding mechanism of the BR-HCTF and AR-HCTF are introduced.The main ideas of some commonly numerical simulation methods are analyzed.Combined with the features of these numerical simulation methods,the finite element method is used to investigate the performance of the proposed fibers.2.The performance of the BR-HCTF and AR-HCTF are simulated and analyzed.For BR-HCTF,the influences of the cladding thickness and core diameter on fibers performance are analyzed.For AR-HCTF,the influences of the wall thickness,core diameter and short axis of cladding tube on fibers performance are analyzed.Besides,some quantitative analyses with the peak frequency of the confinement loss are proposed,and the influences of these structural parameters on the frequency and period of the resonance peak are studied.3.The performance of the BR-HCTF and AR-HCTF are characterized experimentally.Some common 3D printing methods are introduced.Combined with the features of these 3D printing methods,the stereo lithography appearance is used to fabricate the proposed fibers.The THz transmission time-domain spectroscopy system is used to realize the experimental characterization.The frequency and period of the resonance peak of the theoretical calculation,numerical simulation and experimental measurement are compared.4.The machine learning-assisted simulation method for resonant hollow core terahertz fiber is established.Three different machine learning methods are used to predict the experimental spectrum,and the mean square error,relative error,and correlation coefficient which is represented by R~2 are used to evaluate the prediction performance of the simulation method.The results show that the ELMAN neural network simulation method was used to simulate and reconstruct the spectra of BR-HCTF and AR-HCTF,and the obtained R~2 was about 72.25%and 74.76%,respectively.Compared with the physical model simulation,the R~2 increased by 28.73%and 11.01%,respectively.Therefore,the simulation accuracy of the model can be improved by machine learning.