Electromigration in integrated circuits

We review the mechanism of integrated circuit failure known as “electromigration.” Electromigration is a physical wearout process that operates primarily in the on-chip interconnections of an integrated circuit. It is caused by the currents that run through those interconnections. The first part of the review introduces the founding principles of electromigration and the research history upon which those principles are based. The second part introduces and reviews the increasingly relevant issue of pulsed current electromigration, and the final part contains a relatively detailed review of a recent study of pulsed current electromigration.     

Ferroelectric materials for dynamic-memory integrated circuits

The possibilities of using ferroelectric materials for new generations of integrated circuits for high-density dynamic memory (up to 1 Gbit per crystal) are discussed. The correspondence of the specific capacitance and leakage currents of thin film ferroelectric capacitors to the requirements for integrated circuits with various information capacities is examined. It is shown that the capacitance-voltage characteristic of the ferroelectric strongly influences the specific capacitance and the rate of decrease of the voltage across the capacitors when they are discharged in the process of storing information. The prospects for increasing the specific capacitance of memory capacitors using relaxor ferroelectrics are examined. Zh. Tekh. Fiz. 69, 89–92 (May 1999)     

Polymer-based photonic integrated circuits

A myriad of passive and active guided-wave devices has been successfully demonstrated using the photolime gel polymer. These include high density linear and curved channel waveguide arrays, electro-optic modulator and modulator arrays, highly multiplexed waveguide holograms for wavelength division demultiplexing and optical interconnects, waveguide lens, and rare-earth ion-doped polymer waveguide amplifiers. A single-mode linear channel waveguide array with device packaging density of 1250 channels cm^-1 has been achieved. The first 12-channel wavelength division demultiplexer working at 830, 840, 850, 860, 870, 880, 890, 900, 910, 920, 930 and 940 nm on a GaAs substrate is also described in this paper. A polymer-based electro-optic travelling wave modulator with 40 GHz electrical bandwidth is further delineated. A rare-earth ion-doped polymer waveguide amplifier working at 1.06 μm with 8.5 dB optical gain is also achieved using this polymer matrix. The tunability of the waveguide refractive index of photolime gel polymer allows the formation of a graded index (GRIN) layer. As a result, these active and passive guided wave devices can be realized on any substrate of interest. High quality waveguides (loss<0.1 dB cm^-1) have been made on glass, LiNbO3, fused silica, quartz, PC board, GaAs, Si, Al, Cu, Cr, Au, Kovar, BeO, Al₂O₃ and AIN.     

Polarization management for silicon photonic integrated circuits

Polarization management is very important for photonic integrated circuits (PICs) and their applications. Due to geometrical anisotropy and fabrication inaccuracies, the characteristics of the guided transverse‐electrical (TE) and transverse‐magnetic (TM) modes are generally different. Polarization‐dependent dispersion and polarization‐dependent loss are such manifestations in PICs. These issues become more severe in high index contrast structures such as nanophotonic waveguides made of silicon‐on‐insulator (SOI), which has been regarded as a good platform for optical interconnects because of the compatibility with CMOS processing. Recently, polarization division multiplexing (PDM) with coherent detection using silicon photonics has also attracted much attention. This trend further highlights the importance of polarization management in silicon PICs. The authors review their work on polarization management for silicon PICs using the polarization independence and polarization diversity methods. Polarization issues and solutions in PICs made of SOI nanowires and ridge waveguides are discussed.     

耦合电路中的复杂振荡行为分析

讨论了两个非线性电路适当连接后的耦合系统随耦合强度变化的演化过程.给出了两子系统各自的分岔行为及通向混沌的过程,指出原子系统均为周期运动时,耦合系统依然会由倍周期分岔进入混沌,同时在混沌区域中存在有周期急剧增加及周期增加分岔等现象.而当周期运动和混沌振荡相互作用时,在弱耦合条件下,受混沌子系统的影响,原周期子系统会在其原先的轨道邻域内作微幅振荡,其振荡幅值随耦合强度的增加而增大,混沌的特征越加明显,相反,周期子系统不仅可以导致混沌子系统的失稳,也会引起混沌吸引子结构的变化. 关键词： 非线性电路 耦合强度 分岔 混沌     

Organic-inorganic materials for fabrication of integrated optical circuits

In the last few years, ORganically MOdified SIlicates (ORMOSILS) [R_xSi(OR)_4−x] prepared by sol-gel process were particularly attractive for integrated optics fabrication. A composition based on 3-(trimethoxysilyl)propylmethacrylate (MAPTMS) has already allowed the industrial fabrication of optical integrated devices. For this kind of materials, the polymerization of the organic network is typical of free radical curing.In this work, we try to obtain waveguides with another hybrid precursor [2-(3,4-epoxycyclohexylethyltrimethoxysilane)] using cationic polymerization. The main advantage of cationic polymerization is its ability to allow spontaneous cure reaction in presence of oxygen, in contrast with radical polymerization. We choose cycloaliphatic compounds because of their well-known high polymerization rates. The polymerization of the organic network of this hybrid material requires a cationic photoinitiator.The purpose of this paper is dedicated to the inorganic part of the material. Hydrolysis and polycondensation are followed by ²⁹Si NMR. The main objective is to obtain the highest reactive multifunctional oligomer with the lowest OH groups content.Based on our results, we obtained 3D waveguides with a cross-section of 5 μm × 5 μm.     

集成电路微互连结构中的热迁移

高工作电流在集成电路微互连结构中产生大量焦耳热,引起局部区域的温升、形成高温度梯度,金属原子沿着温度梯度反向运动发生热迁移.热迁移是集成电路微互连失效的主要原因之一.阐述了热迁移原理、失效模式及原子迁移方程.综述和分析了在单纯温度场、电场和温度场耦合等不同载荷条件下金属引线和合金焊料的热迁移研究.归纳并提出了集成电路微互连结构热迁移研究亟待解决的问题. 关键词： 集成电路 微互连 热迁移     

Amorphous silicon logic integrated circuits

Amorphous silicon thin-film integrated circuits, with between 4 and 18 transistor functions per chip, have been fabricated on glass substrates. The amorphous silicon and the dielectric layers are deposited by rf glow discharge. The circuits have been designed to realize basic logic functions such as inverters, NAND and NOR gates, and addressable memory cells. For the first time, an amorphous silicon flip flop requiring a supply voltage of only 4.5 V has been manufactured. The logic voltage levels of the flip flop are compatible with standard bipolar TTL circuits. Measurements on an inverter chain show a typical propagation delay time of 70 s and a power-delay-time product of 65 pJ. All of the circuits use n-channel enhancement type load transistors instead of integrated ohmic load resistors. The channel length of the driver transistors is 15 m with a gate source/drain overlap of 7.5 m. Experimental geometry ratios range from =2.25 to =21. Generally, the driver transistors exhibit on/off ratios greater than 10⁶ for supply voltages smaller than 5 V. At these voltages the measured on-currents per unit channel width are in the order of 5...10nA/m.The influence of the geometry ratio on static inverter characteristic and switching speed is discussed by means of a simple model. Two different manufacturing schemes for the fabrication of the integrated circuits are outlined. Mask layouts and experimental transfer characteristics of several integrated circuits are presented.     

石墨烯-硅基混合光子集成电路

近年来硅基光子学已经慢慢走向成熟,它被认为是未来取代电子集成电路,实现下一代更高性能的光子集成电路的关键技术.这得益于硅基光子器件与现代的互补金属氧化物半导体工艺相兼容,能够实现廉价的大规模集成.然而,由于受硅材料本身的光电特性所限,在硅基平台上实现高性能的有源器件仍然存在着巨大挑战.石墨烯-硅基混合光子集成电路的发展为解决这一问题提供了可行的方案.这得益于石墨烯作为一种兼具高载流子迁移率、高电光系数和宽带吸收等优点的二维光电材料,能够方便地与现有硅基器件相集成,并充分发挥自身的光电性能优势.本文结合我们课题组在该领域研究的一些最新成果,介绍了国际上在石墨烯-硅基混合光子集成电路上的一些重要研究进展,涵盖了光源、光波导、光调制器和光探测器四个重要组成部分.     

Organic materials in future integrated optoelectronic circuits

During the last two decades, lithium niobate has been extensively studied for applications in integrated optical circuits. However, it is difficult to integrate lithium niobate optical devices with semiconductor electronic devices because the materials are incompatible. In recent years, semiconductor materials have been emerging as the main contenders in applications; these materials have the advantage of allowing both optical and electronic devices to be integrated. Further, the semiconductor technology has advanced rapidly, allowing us to engineer device parameters very precisely. In semiconductor optoelectronic devices, that is, bulk and quantum well structures, electroabsorption has mainly been used for amplitude modulation of light. The electrorefraction effect is the most useful for devices employing phase-modulation techniques, but this effect cannot be effectively utilized in semiconductors since the strongest electrorefraction effect is near the absorption edge of the material. Recently, organic materials have been shown to have electro-optic coefficients equal to or larger than that of lithium niobate. There are major advantages of organic materials: (1) the organics can be deposited on semiconductor substrates, and therefore both electronic and optical circuits can be integrated; (2) in organic materials the electrorefraction can be effectively utilized to obtain both amplitude and phase modulation; (3) the organic material composition can be adjusted to satisfy some device requirements. In this paper, a comparison of these material systems are made in terms of device applications.     

Stability and entropy production in electrical circuits

A number of the theorems expounded by Prigogine, Glansdorff and their collaborators are translated into electrical circuit terminology and their validity and significance discussed. The simultaneous occurrence of inductors and capacitors represents a situation not envisioned in the chemically oriented discussions and imposes some limitations. The electrical terminology also leads to “dual” theorems, in which voltage sources are replaced by current sources. The validity of the theorems in situations in which fluctuations are critical to the relaxation behavior is analyzed. The “excess entropy production” theorem is only valid if the circuit relaxation can be described by single-valued macroscopic variables, but not if it must be described by distribution functions. We stress that no purely local characterization, which examines a multistable system only in the neighborhoods where it occurs with high probability, can predict or characterize the steady state.     

The use of III-V compounds for integrated optical circuits

Recent progress in the development of integrated optical circuits using the III-V compound semiconductors is reviewed. The application of double heterostructure configurations is emphasized, not only for optical sources, but also for detectors, with reference to both the A/GaAs/GaAs system and the InGaAsP quaternary. Devices utilizing periodic corrugations are described briefly, whereas alternate attempts to fabricate optical circuits by etching or sputtering techniques are discussed in more detail. Recent advances in processing techniques suitable for optical integration, such as reactive-ion etching, and the use of lasers or electron beams for device processing, are described.     

The use of III-V compounds for integrated optical circuits

Recent progress in the development of integrated optical circuits using the III-V compound semiconductors is reviewed. The application of double heterostructure configurations is emphasized, not only for optical sources, but also for detectors, with reference to both the A/GaAs/GaAs system and the InGaAsP quaternary. Devices utilizing periodic corrugations are described briefly, whereas alternate attempts to fabricate optical circuits by etching or sputtering techniques are discussed in more detail. Recent advances in processing techniques suitable for optical integration, such as reactive-ion etching, and the use of lasers or electron beams for device processing, are described.     

Ion implantation into fused quartz for integrated optical circuits

Results obtained in the fabrication of slab and strip waveguides by ion implantation into fused quartz are discussed. Using a step-index waveguide model the increase in refractive index is calculated. The optical loss is smaller than 1 dB/cm at λ = 568 nm without annealing. The properties of strip waveguides fabricated by ion implantation through photoresist masks of thicknesses from 0.4 μm to 0.8 μm are described. A bright fluorescence is observed with emission at 530 nm and 640 nm and its dependence on ion fluence and ion energy is measured.     

Intermixing of multiple quantum wells for all-optical integrated circuits

All-optical waveguide switches that can potentially be implemented in all-optical integrated circuits are described. These switches are presently made in either single or multiple semiconductor quantum well structures. Selective-area intermixing of the quantum wells is presented as a most viable means of achieving the passive waveguide sections for the interconnections between the switches.     

Discrete-charge quantum circuits and electrical resistance

From the theory of quantum LC circuits with discrete charge, and semiclassical considerations, we obtain approximate energy eigenvalues, depending on the parameter . Next, we include electrical resistance for the quantum RLC circuit, obtaining a relation that strongly reminds us of the Landauer formula.     

Design method for high performance grating couplers in photonic integrated circuits

The optimization of grating couplers is usually realized by multiple simulations using specific computational software for this task. Many grating parameters must be analyzed and designed to get the maximum coupling efficiency and the transmission spectrum centred at the wavelength of operation. However, these simulations may take a long time and consume high computational resources depending on the simulation resolution. This work is focused on finding a method to optimize the grating parameters with the lowest number of simulations. In this way, closed-form expressions are presented to get the optimal values for the period and fill-factor, which are the main parameters in the grating design. The usefulness of the proposed approach is shown for the design of silicon grating couplers operating at 1.31 µm and 1.55 µm and both TE and TM polarizations.