【作者】 窦娜；

【导师】 李淳飞；

【作者基本信息】 哈尔滨工业大学 ， 光学， 2008， 博士

【摘要】 为了解决目前通讯网络中电子开关对网络容量升级的瓶颈限制,人们开始广泛研究光开关。较早研发的电控光开关目前已有商用产品,但是这类开关不能满足未来全光网络的要求,因此光控光开关,即全光开关,始终是人们研究的重点。全光开关一般基于介质的非线性光学特性,通过控制泵浦光或者信号光本身的功率来改变信号光的输出状态。目前已研发的全光开关,一般都需要很大的开关功率,远远超过信号功率。特别是用纯石英材料制备的全光开关,由于其非线性系数极小,需要的开关功率更大。因此我们的研究围绕着如何降低全光开关的开关功率使之达到与信号功率同数量级而展开。本文提出用硅基的双耦合器波导(包括光纤)环腔作为1×2的全光开关。利用耦合模方程,建立了考虑环腔损耗的各个端口光电场之间关系的方程,分析发现双耦合器环共振器(DCRR)的损耗严重影响了开关的工作,甚至无法实现开关。因此我们提出将环腔的一臂用掺铒光纤(EDF)代替,在980nm泵浦光作用下形成掺铒光纤放大器(EDFA),用以补偿环腔内的损耗。并针对双耦合器环腔内的克尔(Kerr)非线性,计算得出环腔的精细度、反射率和透射率及开关功率。在补偿损耗后,环腔的精细度被大大提高,同时环材料的光学非线性也被很大地提高(约5个数量级)。因为开关功率与环腔的精细度和非线性折射率系数成反比。因此开光功率被大大降低,可低达毫瓦量级。研究含有掺稀土光放大器的低开关功率全光开关是我们的创新研究的成果之一。双耦合器环共振腔由于具备了非线性和反馈两个条件,所以具有光学双稳特性。我们系统研究了含EDFA的双耦合器环共振腔的反射双稳和透射双稳特性。普通光纤环腔,实现光双稳性需要的开关功率为1000瓦量级。用掺铒光纤取代普通光纤并补偿腔内损耗,用数值模拟的方法,得出EDFA-RR双稳开关的开关功率大约是60毫瓦,比普通光纤环形谐振腔(即不含EDFA的光纤环腔)双稳开关功率低5个数量级。并在放大器增益,耦合器反射率,环长及环腔初相位等多个不同参数情况下,给出了双耦合器环形共振腔的光双稳特性曲线。我们的理论分析对未来的光纤环腔双稳器件有着实际的指导意义。此器件的开关时间由环腔的寿命决定,我们计算得到的结果是低于20ns。可以通过调整980nm泵浦光功率来控制环腔的精细度,使得开关时间和开关功率得到优化。研究含光放大器的双耦合器环腔的光双稳开关特性是我们的创新成果之二。掺饵光纤放大器的引入固然可以降低开关功率,但是给器件结构带来了复杂化,而且掺饵光纤的吸收比普通光纤的吸收大得多。我们的研究表明,光开关的开关功率还与光波导的横截面积成正比,因此压缩光波导的横向尺寸,从微米量级减至纳米量级,也可大大降低光开关的开关功率。在这种情况下不必进行材料的掺杂,就可以使开关功率降低到瓦量级。采用100微米环直径,就可实现14瓦的开关功率。而且环形光开光器件环长的减小对于提高光开关的开关速度(降低开关时间)有利。研究具有低开关功率和高开关速度的纳米尺寸的微型环腔是我们的有创新性的研究成果之三。硅材料在超短脉冲(飞秒)激光的作用下,会产生双光子吸收效应,此效应可强到把产生光克尔效应的单光子吸收效应加以忽略。我们提出,将泵浦光与信号光同时进入波导双耦合器环共振腔。由于双光子吸收引起自由电子-空穴对的产生。自由载流子浓度的变化致使硅的折射率发生变化,从而环内的相移也随之变化。非线性折射率与泵浦平均功率的平方成正比。当输入泵浦平均功率低于10毫瓦,就可以使环内相移变化达到π,实现双耦合器环共振腔两输出端口间的转换。所以研究纳米波导微环光开关在飞秒脉冲作用下的双光子吸收光开关机制是我们创新性研究成果之四。更多还原

【Abstract】 In order to break through the bottle limitation of the electronic switches to net capacity-upgrading in modern communication systems, optical switches are widely investigated. The electronic-controlling optical switches (EOS) initially studied are commercially obtained now. The switch properties do not meet the requirement of future all optical network. So the optic-controlling optical switches, ie. all optical switches (AOS), are the focus of the researchers. AOS are based on the nonlinearity of the medium. The output states of the signal are controlled by the pump power or the signal power itself. However, because of the small nonlinear coefficient of the common silica fiber, all the suggested AOS need high switch power. This dissertation is centered on how to decrease the switch power to the same level of signal power.Firstly, we suggest double-coupler waveguide (including fiber) as 1×2all-optics switch. According to coupler mode equation, we get the relationship of all ports electrical field with losses. From these relationships, we find the losses of DCRR seriously affect the switch work. And it can’t realize switch. So we suggest to use EDF instead of common fiber. It will form EDFA to compensate the losses when 980nm pump light is input. We calculate the finesse, reflectivity, transmission and switch power based on Kerr effect. After compensating losses, finesse and switch efficiency of RR are increased and switch power are decreased. Switch power is proportional to cross section of fiber and inverse proportional to finesse, nonlinear refraction index and length of RR. Nonlinear refraction index of EDF is 5 orders larger than the common silica fiber. So the switch power is decreased 5 orders. Switch time depends on material. It is on nanosecond order. It is the second research achievement that study all-optics switching of low switching power.Double-coupler ring resonator (DCRR) has optics bistability because it has the characteristics of nonlinear and feedback. We study the reflection bistability and transmission bistability of DCRR with losses. Common fiber RR needs 1000W power to gain bistability. But EDF can instead of common fiber and compensate the losses, the switch power of EDFA-RR bistability switch is decreased to 60mW. It decreased 5 orders than common fiber ring resonator. And we analyze the bistability curses of DCRR optics switch for different amplifier gain, coupler reflectivity, RR length and initial phase in the RR. Theoretical analysis has guide purpose for the experiment in the future. The switching time of this device depends on lifetime of ring cavity. The result that we calculates is lower 20ns. We can control finesse of ring resonator according to adjusting 980nm pump power and can gain optimal switching time and switching power. It is the second research achievement that studies the DCRR bistability with EDFA.EDFA can decrease switching power, but the device could be complicated. And EDF absorption is larger than ordinary fiber. We demonstrate that switching power of optics switches is proportional to cross section of waveguide. So it can decrease dramatically switching power of optics switches that cross section of waveguide decrease from millimeter to nanometer. In this situation, switching power of optics switches can decrease to watt when the material is not doped. It can realize switching power of 14W adopted100μm diameter ring. Decreasing length of device can improve switching speed (decreasing switching time). It is the third research achievement that study nano-size ring resonator of low switching power and high speed.Silicon can induce two-photon absorption (TPA) under fetosecond laser. This effect is so powerful that one-photon absorption of Kerr effect can be neglected. Pump light and signal light are input waveguide DCRR at the same time. Electron-hole pair are generated due to TPA effect. Free-carrier concentration change cause silicon refraction index change and phase shift change in the RR. Nonlinear refraction index change is proportional to square of mean pump power. When phase shift change in the RR achievedπ, the switching can be realized between two output ports. Refraction index change caused by TPA is 10-3mW. However, Refraction index change caused by Kerr effect is smaller when other parameters are the same. It is the fouth research achievement that study TPA optics switching when nano waveguide microring pump by fetosecond pulse更多还原