Fundamental research of the optical logic gates in a sub-teraherz frequency range

该基础研究项目的目标是实现原理验证，并建立200至400 GHz频率范围内的光逻辑门的频率响应模型。研究结果概要从(1)到(3)总结如下：(1)分析诊断并实验验证超高频操作下可能出现的输出波形畸变。首先，我们分析发现，根据我们的原始门模型，门输出可以固有地包含弱子脉冲成分，这是研究者第一次了解到的。其次，我们通过实验验证了分析发现。由于在NEC（2000-2001）、朗讯科技（2001-2002）和埃因霍温技术大学（2005-2006）最近进行的著名演示实验中，产生子脉冲分量的固有因素被认为是存在的，因此它们的输出波形应该包含特定数量的……更多的失真（即子脉冲分量）。我们推测，这种失真是其相对较高的噪声水平的来源之一。现在我们正在尝试重新设计栅极结构，以显著降低子脉冲分量的强度比。[2]为了获得二次弛豫时间常数的潜在影响，与相应研究者群体的传统推测相反，我们发现光逻辑门的二次弛豫时间常数的影响可以抑制（而不是放大）输出畸变分量，这取决于输出光分量的光谱滤波条件。发现后，我们设计并实验验证了光谱滤波条件。[这项研究成果（发表于2006年1月）很快在一篇会议摘要论文（最近于2006年3月发表）中被提及，埃因霍温技术大学将光逻辑门的无错误操作比特率从160 Gb/s（2000年由包括本研究者在内的作者首次实现）刷新到320 Gb/s。(3)利用光逻辑门产生超高频锁模脉冲，而不是直接研究超高频范围内的光逻辑门，我们可以像本研究者在2000年理论上提出的那样，将光逻辑门的输出反馈到其输入端口时，研究其自发脉冲的产生。在这个研究项目中，我们也在这个研究方向上取得了重大进展；我们利用光逻辑门和反馈环路实现了2ps， 40ghz锁模脉冲的产生。实验产生的2 ps， 40 ghz脉冲的值几乎等同于光复用160 ghz时钟信号，因为脉冲宽度足够短，可以进行这种光复用。在成功产生锁模脉冲的基础上，我们在10 GHz到160 GHz的等效频率范围内，以及2 ps到大约10 ps的时间常数范围内，交替研究了光逻辑门的几种新的特性。我们在2006年5月的CLEO/QELS 2006和一份工业报纸上发表了这一漂亮的研究成果。少

The goal of this fundamental research project was to achieve a proof of principle and to establish a frequency-response model of the optical logic gate in a frequency range from 200 through 400 GHz. The outlines of the research results are summarized from (1)through (3), as follows :(1)to analytically diagnose and experimentally verify the possible output waveform distortion, under the ultrahigh-frequency operation.First, we analytically discovered, according to our original gate model, that the gate output can inherently contain weak sub-pulse components, for the first time to the investigator's best knowledge. Second, we experimentally verified the analytical discovery. Because the inherent factors which generate the sub-pulse components are supposed to exist in the recent, famous demonstration experiments done by NEC (2000-2001), Lucent Technology (2001-2002), and Technical Univ. of Eindhoven (2005-2006), respectively, their output waveforms are supposed to contain a particular amou … More nt of the distortion (i.e., the sub-pulse components). We speculate that this distortion was one of the sources of their relatively high noise level. [Now we are trying to re-design the gate structure, for significantly reducing the intensity ratio of the sub-pulse components.](2)to access the potential impacts of the secondary relaxation time constantIn contrast to the conventional speculation in the corresponding researcher's community, we discovered that the impacts of the secondary relaxation time constant of the optical logic gate can suppress (instead of enlarge) the output distortion components, depending upon the spectral filtering conditions for the output optical component. After the discovery, we designed and experimentally verified the spectral filtering conditions. [This research result (which was published in January 2006) was referred to very soon in a conference digest paper (which was published more recently in March 2006) in which the Technical University of Eindhoven renewed the record-high error-free-operation bitrate of the optical logic gate from 160 Gb/s (first achieved in 2000 by authors including the present investigator) to 320-Gb/s.](3)Ultrahigh-frequency mode-locked pulse generation with the optical logic gateInstead of directly studying the optical logic gate in a ultrahigh frequency range, we can alternatively study its spontaneous pulse generation when feedbacking the output of the optical logic gate to its input port, as the present investigator theoretically proposed in 2000. In this research project, we have made a significant progress in this research direction, as well ; we achieved 2-ps, 40-GHz mode-locked pulse generation with using the optical logic gate and the feedback loop. The value of the experimentally generated 2-ps, 40-GHz pulses is nearly equivalent to optically multiplexed 160-GHz clock signal, because the pulse width is short enough for such optical multiplexing. Based on the successful mode-locked pulse generation, we have alternatively studied several new, characteristic properties of the optical logic gate in an equivalent frequency range from 10 GHz to 160 GHz, and in a time-constant range from 2 ps to approximately 10 ps. We presented this beautiful research result to CLEO/QELS 2006 and to an industrial newspaper in May 2006. Less