【作者】 梁锐；

【导师】 周晓军；

【作者基本信息】 电子科技大学 ， 光学工程， 2011， 博士

【摘要】 模数转换器将连续的模拟信号转换为离散数字信号,是信息处理系统中的重要器件。随着数字化技术的高速发展,模数转换器在高速通信、实时监测、雷达信号处理、图像处理以及空间通信等领域的作用也愈加突出。同时,人们对信息处理系统中模数转换器性能的要求也日益提高。目前实际应用最广泛的是电子模数转换器,但是由于受到电子迁移率的限制,电子模数转换器在超高速模数转换中遭遇电子瓶颈,限制了其性能的进一步提高。超导材料模数转换器要求低温的工作环境,这大大限制了其应用范围。而全光模数转换技术可以克服电子模数转换和超导材料模数转换的限制,将是未来信息处理系统的关键技术。全光量化是全光模数转换器的重要组成部分,是全光模数转换器的重要研究课题。本论文针对基于拉曼自频移和谱压缩的全光量化进行了详细研究,其主要内容可概括为以下几点:(1)简要介绍了全光模数转换器和全光量化的研究背景与意义,并回顾了光量化的发展历程以及国内外研究现状。(2)构建了由高非线性光纤、色散渐增光纤和阵列波导光栅组成的全光量化结构。首先,窄脉宽、高功率的脉冲在高非线性光纤中产生拉曼自频移效应,实现功率—波长的转变,是此量化结构的主要部分;自频移后脉冲经过色散渐增光纤时,谱宽压缩,可有效提高量化精度;最后,阵列波导光栅用来空间分离谱压缩后的脉冲,为全光量化的下一步——编码做准备。描述了全光量化的数值分析基础——广义非线性薛定谔方程的推导过程,并简要介绍了求解广义非线性薛定谔方程的数值方法——分步傅里叶法。(3)采用矩量法求解脉冲参量如脉宽、啁啾、能量、延时和中心波长频移量沿传输方向的演化公式,来分析脉冲在光纤中的传输情况。纠正了J. Santhanam和G. P. Agrawal在推导过程中的错误,获得更准确的表达式;并针对在亚皮秒脉冲入射的情况,将修正后的结果和修正前的结果进行了对比。(4)超短脉冲在光纤中的拉曼自频移现象是由光纤的拉曼延迟响应特性引起的,其自频移幅度不仅与脉冲特性有关,还与传输光纤有很大的关系。本文中详细研究了亚皮秒脉冲在三种不同光纤(如光子晶体光纤、保偏光纤、普通高非线性光纤等)中产生的拉曼自频移现象,对比了不同形状、不同脉宽、不同峰值功率的脉冲在光纤中发生拉曼自频移效应后的输出脉冲,并对如何选择合适的工作光纤提出了合理化建议。采用两根不同的高非线性光纤进行了拉曼自频移实验,改变入射功率,获得140nm的自频移范围。(5)基于拉曼自频移后形成的拉曼孤子特性,首次详细描述了亚皮秒孤子在色散渐增光纤中的谱压缩过程,并对比了基孤子在五种不同类型(即直线型、指数型、高斯型、双曲型和对数型)色散渐增光纤中的演化过程和谱压缩效果。结果表明,直线型、指数型和对数型色散渐增光纤更适合压缩拉曼自频移后脉冲的谱宽。(6)模拟了亚皮秒脉冲入射时,由光子晶体光纤、色散渐增光纤和阵列波导光栅组成的全光量化系统的性能,论证了此全光量化结构的可行性。利用矩量法获得理想状况下孤子延时的表达式,并详细讨论了脉冲在拉曼自频移和谱压缩过程中伴随的脉冲延时对采样速率的影响。更多还原

【Abstract】 Analog-to-digital converters (ADCs) which convert analogue signals into digital ones, play an important role in the signal processing system. With recent tremendous growth of the high-speed digital technique, the effect of the ADCs on the high-speed communications, real-time measurements, radar systems, image processing, and space communications enhances. And the ultrawide-bandwidth applications also encourage the demands of high-speed and high-resolution ADCs. Recently, electrical ADCs are most widely used in practical applications. But to electrical ADC, the enhancement of performance is limited by its electron mobility. Superconducting material ADCs need to work under cold condition, which limits their applications. All-optical ADCs can overcome the disadvantage of the electrical and superconducting material ADCs. All-optical analog-to-digital conversion, which is characterized by high-speed and high-resolution will be extremely beneficial to signal processing systems in future. And all-optical quantization is a key technology to the realization of the all-optical ADCs. The researches on the all-optical quantization are highly significant. The works presented in the dissertation focus on the investigation of all-optical quantization based on the Raman self-frequency shift and spectral compression. The main contents of the dissertation are shown as follows:(1) Research backgrounds of the all-optical ADCs and all-optical quantization are simply introduced. The development history and status of the all-optical quantization at home and broad are reviewed.(2) An all-optical quantization configuration is constructed with highly nonlinear fiber (HNLF), dispersion-increasing fiber (DIF) and arrayed waveguide grating (AWG). Firstly, the Raman self-frequency shift (RSFS) occurs as pulses with narrow pulse widths and high powers propagation in the HNLF, which can realize the conversion from power to wavelength. Where, the RSFS is the mainly part of the all-optical quantization. Then, the spectra of the shifted pulses are compressed as propagating along the DIF, which can effectively improve the resolution of the all-optical quantization. Finally, the AWG separates the compressed pulses, which prepares for the following coding process.. Generalized nonlinear Schr?dinger equation (GNLS equation) is used to describe the propagation of the pulse in fiber. In this dissertation, the inference of the GNLS equation is shown. And the split-step Fourier method, the mostly used numerical method to analyze the GNLS equation, is depicted.(3) Pulse propagation in an optical fiber is analyzed through using the moment method to solve the evolution equation of the pulse parameters, such as pulse width, chirping, energy, time delay, and frequency shift. More accurate expressions are obtained by modified the mistakes in the reference by J. Santhanam and G. P. Agrawal. And the results before and after modification are contrasted for the case of subpicosecond pulses injected.(4) The physical origin of the RSFS in fibers is related to the delayed nature of the Raman response. So the RSFS depends on not only the pulse characteristics, but also the propagating fiber.The RSFS of subpicosecond pulses in three different fibers, viz. photonic crystal fiber (PCF), polarization maintaining fiber (PMF) and commonly HNLF, are detailedly investigated. When the shape, width and peak power of input pulses are different, the output pulses after the RSFS are compared. And several rational suggestions are given for choosing suitable fibers. Experimental results about the RSFS in two different HNLFs are presented. By changing the input peak power, the frequency shift of 140nm can be obtained.(5) Based on the characteristics of the Raman solitions origined from the Raman self-frequency shifted pulse, the spectral compression processes of subpicosecond solitons are firstly described. The fundamental solitons are injected into the DIF as input pulses. Five different DIFs, viz. linear-type, exponential-type, Gaussian-type, hyperbolic-type, and logarithm-type DIFs, are used to compress the spectra of fundamental solitons. The evolutions and spectral compression in five DIFs are investigated. The results show that the linear-type, exponential-type, and logarithm-type DIFs are more suitable for compressing the spectra of the shifted pulses.(6) The performance of the all-optical quantization system composed of PCF, DIF, and AWG is analyzed as subpicosecond pulses input. The feasibility of the suggested all-optical quantization scheme is testified. Under ideal condition, the time delay of soliton is derived first time by analyzing the GNLS equation using the moment method. And the influence on the sampling velocity from the time delay which is accompanied with the RSFS and spectral compression is discussed.更多还原