首页> 外文OA文献 >Electroabsorption modulators used for all-optical signal processing and labelling
【2h】

Electroabsorption modulators used for all-optical signal processing and labelling

机译:电吸收调制器用于全光信号处理和标记

代理获取
本网站仅为用户提供外文OA文献查询和代理获取服务,本网站没有原文。下单后我们将采用程序或人工为您竭诚获取高质量的原文,但由于OA文献来源多样且变更频繁,仍可能出现获取不到、文献不完整或与标题不符等情况,如果获取不到我们将提供退款服务。请知悉。

摘要

This thesis concerns the applications of semiconductor components, primarily electroabsorption modulators (EAMs), in optical signal processing and labelling for future all optical communication networks. An introduction to electroabsorption modulators is given and several mechanisms that form the basis of electroabsorption are briefly discussed including Franz Keldysh effect, Quantum-Confined Stark Effect (QCSE) and Quantum-Confined Franz-Keldysh effect. QCSE is found to be more effective for absorption modulation than FKE at room temperature due to the quantum confinement of electrons and holes. Experimental investigations on electrical-to optical (e/o) modulation of the EAM are presented. From the measured power transfer curves, static extinction ratios larger than 20 dB were obtained for wavelengths in the C-band. It is also shown that the insertion loss and static extinction ratio decrease with the signal wavelength, indicating that an optimum wavelength can be found as a trade-off between the on-off ratio and the signal-to-noise ratio. The chirp property and the small signal bandwidth for electrical-to-optical modulation of the EAM are investigated. It is found that the measured chirp α–parameter ranges from –0.4 to 0.8 depending on the reverse bias; the higher the bias, the smaller the chirp becomes. Negative chirp may be achieved by sacrificing the extinction ratio and the output power. The small signal bandwidth was measured to be as large as 24 GHz. Cross absorption modulation (XAM) in an EAM is discussed including an introduction to the carrier effects and a simple model that simulates the carrier dynamics. Based on this model the static characteristics of an EAM under optical excitation are investigated theoretically; the results demonstrate the capability of an EAM used for wavelength conversion and 2R regeneration. The optical-to optical (o/o) modulation bandwidth and frequency chirp are experimentally investigated. It is found that the o/o modulation bandwidth drastically depends on the quantum well depth while the e/o modulation bandwidth is mainly influenced by the electrical bonding pad size. A device having a small pad and shallow wells shows 24 GHz bandwidth for both e/o and o/o modulation. In the o/o chirp measurements very small chirp α–parameters are obtained. Depending on the operating wavelength and the bias the chirp α–parameter ranges from –0.6 to 0.2. It is also found that higher bias voltages and shorter wavelengths are preferred to obtain a small or negative chirp α–parameter. The principle of EAM-XAM based wavelength conversion is discussed and a wavelength conversion experiment at 40 Gb/s is presented. The influence of some operation parameters, including pump light power, reverse bias of the converter and probe light wavelength, is experimentally investigated for the wavelength-converted light, including its chirp performance. As a result of this investigation, a higher pump power (up to 20 dBm) and a relatively larger reverse bias (-2.5 V) are preferred for obtaining both larger extinction ratio and lower chirp of the converted signal. A multi-wavelength conversion scheme (8 × 40Gb/s) is demonstrated, where the receiver sensitivity for the back-to-back case is –33 dBm and the average power penalty for the eight converted channels is 9.2 dB. The best channel at 1555.7 nm has a power penalty of 8 dB. The wavelength dependence of the power penalty is explained by studying the impact of the extinction ratio and the average power of the converted signal on the Q parameter. Physical explanations for the optimum pump power and device length is given by considering impacts on the extinction ratio, average power and pulse width of the wavelength-converted signal. Other wavelength conversion schemes such as fibre-based cross phase modulation (XPM) and optical filtering, fibre-based Kerr switch, fibre based four-wave-mixing (FWM) and semiconductor optical amplifier (SOA)-based cross gain modulation (XGM), are briefly discussed. As a result of the comparison, it is suggested that fibre-based solutions have relatively lower power penalties and have great potential for ultra-high speed operation while single semiconductor devices can so far operate at 40 Gb/s and are more attractive in terms of compactness, stability and integration. An important advantage of the EAM-based wavelength conversion scheme is that the frequency chirp of the converted signal is very small, which is desirable for long distance transmission and optical labelling systems. An all-optical demultiplexing experiment from 160 Gb/s to 10 Gb/s using a single EAM with a very simple waveguide structure is presented. All 16 demultiplexed tributary channels are error free with an average receiver sensitivity of –25.3 dBm. An improvement of up to 6 dB in the receiver sensitivity by regeneration of the demultiplexed channel by an additional EAM acting as a saturable absorber is also demonstrated. A discussion of 2R regeneration based on a non-linear intensity transfer function is given. It is reiterated that a 2R regenerator can not reduce the BER but can inhibit its accumulation. The non-linear transfer function of an EAM is frequency dependent and the main improvement from an EAM-based regenerator is the enhancement of the ER and the suppression of the noise in a space bit. Applications of EAMs in optical label processing using various orthogonal labelling schemes are discussed. Through EAM-based wavelength conversion label encoding and recognition are realised for two-level labelled signals consisting of a 10 Gb/s Amplitude Shift Keyed (ASK) payload and a 2.5 Gb/s Differential Phase Shift Keyed (DPSK) label. The receiver sensitivities for the payload/label in back-to-back case and after label encoding are –25.6/-28.1 dBm and –23.7/-21 dBm, respectively. Using an EAM for optical label insertion and a MZ-SOA for optical label erasure and payload regeneration in the ASK(10 Gb/s)/ Frequency Shift Keying (312 Mb/s) orthogonal modulation format, the complete functionality of a network node including two-hop transmission and all-optical label swapping is experimentally demonstrated. The cascaded transmission and label swapping result in 1.9 dB power penalty for the payload and 1.8 dB penalty for the label. Operating as external modulators, two EAMs are used to encode and erase the optical label in the return-to-zero (RZ)-DPSK/ASK and non-return-to-zero (NRZ)-DPSK/ASK format. We experimentally demonstrated label encoding, transmission over a 50 km SMF link, and label erasure of a 40 Gb/s RZDPSK modulated payload with an orthogonal 2.5 Gb/s ASK label. The penalties for the payload and label due to labelling and transmission are 4.5 dB and 2 dB, respectively. A similar experiment was carried out for a NRZDPSK/ASK labelled signal. Compared to the RZDPSK/ASK scheme, NRZDPSK/ASK has a smaller spectrum width and the labelling penalty for the payload is 2.7dB larger while the label performance is almost the same. The modulation cross talk between the ASK payload and the DPSK label is theoretically analysed. As a result it is found that for a noiseless ASK/DPSK system with an infinite ASK extinction ratio, error free detection of the label can be obtained when the payload bit rate is at least 73 times larger than that of the label in case of balanced detection. For the single-ended DPSK receiver an even larger bit rate ratio (~130) is needed. Since real DPSK systems work at a relatively high bit rate this condition is normally not met. To solve this problem, instead of using a moderate ASK extinction ratio, we introduced a base band coding scheme named mark-insertion coding for the ASK payload and using this coding scheme we realised label swapping based on a RZASK payload at 40 Gb/s and a DPSK label at 622 Mb/s using an EAM as the label swapper. A new polarization modulation scheme is proposed and various signal-processing functions based on Polarization Shift Keying (PolSK) modulation format are demonstrated. Polarization modulation is implemented by a normal Mach Zednder Modulator operating in a special but simple way. Detection and erasure of polarization information are realised by a device that is comprised of a polarization controller and a polarization beam splitter. A new orthogonal labelling scheme based on a 40 Gb/s DPSK payload and a 2.5 Gb/s PolSK label is proposed and experimentally demonstrated. The most striking feature of this new labelling format is that there is no modulation crosstalk between the payload and the label, in contrast with all previous orthogonal modulation formats. Swapping penalties are 0.15 dB and 0.6 dB for the payload and the label respectively. Penalties due to swapping and 40 km SMF transmission are 2.2 dB and 2.9 dB for the payload and the label respectively.
机译:本文涉及半导体组件,主要是电吸收调制器(EAM)在未来所有光通信网络的光信号处理和标记中的应用。介绍了电吸收调节剂,简要讨论了构成电吸收基础的几种机理,包括弗朗兹·凯尔迪什效应,量子限制的斯塔克效应(QCSE)和量子限制的弗朗兹·凯尔迪什效应。由于电子和空穴的量子限制,发现QCSE在室温下比FKE对吸收调制更有效。提出了对EAM的电光(e / o)调制的实验研究。根据测得的功率传输曲线,对于C波段的波长可获得大于20 dB的静态消光比。还表明插入损耗和静态消光比随信号波长而减小,表明可以找到最佳波长作为开-关比与信噪比之间的折衷。研究了线性调频脉冲的线性调频特性和小信号带宽。发现,根据反向偏置,测得的线性调频脉冲的–参数范围为–0.4至0.8。偏差越大,线性调频脉冲越小。可以通过牺牲消光比和输出功率来实现负chi声。测得的小信号带宽高达24 GHz。讨论了EAM中的交叉吸收调制(XAM),包括对载流子效应的介绍和模拟载流子动力学的简单模型。基于该模型,从理论上研究了光激发下EAM的静态特性。结果证明了EAM用于波长转换和2R再生的能力。实验研究了光到光(o / o)调制带宽和频率chi。已经发现,o / o调制带宽极大地取决于量子阱深度,而e / o调制带宽主要受电键合焊盘尺寸的影响。具有小焊盘和浅阱的设备对e / o和o / o调制均显示24 GHz带宽。在o / o线性调频测量中,可以获得非常小的线性调频α参数。取决于工作波长和偏置,线性调频脉冲的–参数范围为–0.6至0.2。还发现,较高的偏置电压和较短的波长对于获得较小或负的线性调频α参数是优选的。讨论了基于EAM-XAM的波长转换的原理,并提出了40 Gb / s的波长转换实验。对于波长转换的光,包括线性调频性能,通过实验研究了一些工作参数的影响,包括泵浦光功率,转换器的反向偏置和探测光波长。研究的结果是,较高的泵浦功率(最高20 dBm)和相对较大的反向偏置(-2.5 V)对于获得较大的消光比和较低的转换信号chi是优选的。演示了一种多波长转换方案(8×40Gb / s),其中背对背情况的接收器灵敏度为–33 dBm,八个转换通道的平均功率损失为9.2 dB。 1555.7 nm处的最佳通道的功率损失为8 dB。通过研究消光比和转换信号的平均功率对Q参数的影响,可以解释功率损失的波长依赖性。通过考虑对波长转换信号的消光比,平均功率和脉冲宽度的影响,给出了最佳泵浦功率和设备长度的物理解释。其他波长转换方案,例如基于光纤的交叉相位调制(XPM)和光学滤波,基于光纤的Kerr开关,基于光纤的四波混合(FWM)和基于半导体光放大器(SOA)的交叉增益调制(XGM) ,简要讨论。比较的结果表明,基于光纤的解决方案具有相对较低的功率损失,并且具有超高速运行的巨大潜力,而单个半导体器件迄今为止可以以40 Gb / s的速度运行,并且在以下方面更具吸引力:紧凑,稳定和集成。基于EAM的波长转换方案的一个重要优点是,转换后信号的频率chi非常小,这对于长距离传输和光学标记系统是理想的。提出了使用具有非常简单的波导结构的单个EAM从160 Gb / s到10 Gb / s的全光解复用实验。所有16个解复用支路信道均无错误,平均接收机灵敏度为–25.3 dBm。还证明了通过充当可饱和吸收体的附加EAM通过对多路分解后的通道进行再生,可将接收器灵敏度提高多达6 dB。对基于非线性强度传递函数的2R再生进行了讨论。重申了2R再生器不能降低BER,但是可以抑制其积累。 EAM的非线性传递函数与频率有关,基于EAM的再生器的主要改进是ER的增强和空位噪声的抑制。讨论了EAM在使用各种正交标记方案的光学标记处理中的应用。通过基于EAM的波长转换,可以对由10 Gb / s幅移键控(ASK)有效载荷和2.5 Gb / s差分相移键控(DPSK)标签组成的两级标记信号实现编码和识别。在背对背情况下以及标签编码后,有效负载/标签的接收器灵敏度分别为–25.6 / -28.1 dBm和–23.7 / -21 dBm。通过以ASK(10 Gb / s)/频移键控(312 Mb / s)正交调制格式使用EAM进行光学标签插入和MZ-SOA进行光学标签擦除和有效载荷再生,网络节点的完整功能包括实验证明了两跳传输和全光标签交换。级联的传输和标签交换导致有效负载的功率损耗为1.9 dB,标签的功率损耗为1.8 dB。作为外部调制器,两个EAM用于以归零(RZ)-DPSK / ASK和不归零(NRZ)-DPSK / ASK格式编码和擦除光学标签。我们通过实验证明了标签编码,在50 km SMF链路上的传输以及使用正交2.5 Gb / s ASK标签擦除40 Gb / s RZDPSK调制有效载荷的标签。由于标记和传输,有效载荷和标记的惩罚分别为4.5 dB和2 dB。对NRZDPSK / ASK标记的信号进行了类似的实验。与RZDPSK / ASK方案相比,NRZDPSK / ASK具有较小的频谱宽度,有效载荷的标记损失大2.7dB,而标记性能几乎相同。理论上分析了ASK有效载荷和DPSK标签之间的调制串扰。结果发现,对于具有无穷大ASK消光比的无噪声ASK / DPSK系统,当有效负载比特率至少为标签的73倍时,可以实现标签的无差错检测检测。对于单端DPSK接收器,需要更大的比特率比(〜130)。由于实际的DPSK系统以相对较高的比特率工作,因此通常不满足此条件。为了解决这个问题,我们没有采用适度的ASK消光比,而是为ASK有效载荷引入了一种称为标记插入编码的基带编码方案,并使用该编码方案,以40 Gb / s的RZASK有效载荷实现了标签交换,并且使用EAM作为标签交换器以622 Mb / s的速度发送DPSK标签。提出了一种新的极化调制方案,并说明了基于极化移位键控(PolSK)调制格式的各种信号处理功能。偏振调制是通过以特殊但简单的方式运行的普通Mach Zednder调制器实现的。偏振信息的检测和消除是通过由偏振控制器和偏振分束器组成的设备实现的。提出并基于40 Gb / s DPSK有效载荷和2.5 Gb / s PolSK标签的新的正交标记方案。与以前的所有正交调制格式相比,这种新标记格式最显着的特征是有效负载和标签之间没有调制串扰。有效负载和标签的交换惩罚分别为0.15 dB和0.6 dB。交换和40 km SMF传输造成的罚款分别为有效载荷和标签2.2 dB和2.9 dB。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号