A Novel Wafer-Yield PDF Model and Verification With 90–180-nm SOC Chips

In this paper, we describe a new wafer-yield distribution model, which agrees well with experiment using fabricated products with various process technologies. To investigate physical reasoning of the proposed model, we firstly measure effective defect density of chips regarding to spatial dependency in a wafer. It is clarified that the defect density near wafer edge is a couple of times larger than that at the rest of wafer area. Note that the increase of defect at wafer edge causes a significant yield loss in production process lines.      

A 56–65 GHz Injection-Locked Frequency Tripler With Quadrature Outputs in 90-nm CMOS

A sub-harmonic injection-locked tripler multiplies a 20-GHz differential input to 60-GHz quadrature (I/Q) output signals. The tripler consists of a two-stage ring oscillator driven by a single-stage polyphase input filter and 50-$Omega$ I and Q-signal output buffers. Each gain stage incorporates a hard limiter to triple the input frequency for injection locking and a negative resistance cell with two positive feedback loops to increase gain. Regenerative peaking is also used to optimize the gain/bandwidth performance of the 50-$Omega$ output buffers. Fabricated in 90-nm CMOS, the tripler has a free-running frequency of 60.6 GHz. From a 0-dBm RF source, the measured output lock range is 56.5–64.5 GHz, and the measured phase noise penalty is 9.2 $ pm 1~$dB with respect to a 20.2-GHz input. The $0.3times 0.3~ hbox{mm}^{2}$ tripler (including passives) consumes 9.6 mW, while the output buffers consume 14.2 mW, all from a 1-V supply.      

A 3.1–10.6 GHz Ultra-Wideband Low Noise Amplifier With 13-dB Gain, 3.4-dB Noise Figure, and Consumes Only 12.9 mW of DC Power

A 3.1-10.6 GHz ultra-wideband two-stage pseudomorphit high electron mobility transistor low noise amplifier is presented. The first stage of the amplifier employs a resistive shunt feedback topology and two T-network sections to provide wideband input matching to a 50-Omega antenna. The current-sharing dc bias topology is used to ensure the low power consumption under fixed 3-V battery operation. The amplifier exhibits state of the art performance consuming only 12.9mW of dc power with a power gain of 12.5dB, plusmn0.5dB gain flatness, and 3.4-4.0dB noise figure. Input match is better than -12.0dB, output match is better than -15dB, and group delay is 184pSplusmn28pS     

Low voltage stress induced leakage current and time to breakdown in ultra-thin (1.2–2.3 nm) oxides

Ultra-thin gate-oxide reliability is an essential factor in CMOS technologies. The low voltage gate current in ultra-thin oxide of metal–oxide–semiconductor devices is very sensitive to electrical stresses. It can be used as a reliability monitor when the oxide thickness becomes too small for traditional electrical measurements. In this paper, the low voltage stress induced leakage current (LVSILC) for various oxide thicknesses ranging from 1.2 to 2.3 nm is investigated during constant voltage stress (CVS). From the LVSILC measurements, we shown that time to breakdown can be deduced as a function of the stress voltage. We also study the effect of elevated stress temperature on the time to breakdown. We show that temperature dependence of the time to breakdown is non-Arrhenius and decreases in a drastic way with a slope of 0.036 decade/°C.     

Electrical characterization and quantum modeling of MOS capacitors with ultra-thin oxides (1.4–3 nm)

Electrical characterization of MOS capacitors with ultra-thin oxides (1.4–3 nm) has been carried out. The validity of the correction to C–V data, needed to take into account the series resistance and leakage current, is discussed. The gate current in accumulation and in depletion regions has been investigated and properly modeled based on detailed analysis of tunneling from the polygate.     

Design and Analysis of a 55–71-GHz Compact and Broadband Distributed Active Transformer Power Amplifier in 90-nm CMOS Process

A 55–71-GHz fully integrated power amplifier (PA) using a distributed active transformer (DAT) is implemented in 90-nm RF/MS CMOS technology. The DAT combiner, featuring efficient power combination and direct impedance transformation, is suitable for millimeter-wave (MMW) PA design. Systematic design procedures including an impedance allocation plan, a compensation line, and a gain boosting technique are presented for the MMW DAT PA. The monolithic microwave integrated circuit (MMIC) performs a high and flat small-signal gain of ${hbox{26}} pm {hbox{1.5}}~{hbox{dB}}$ from 55 to 71 GHz, which covers a full band for 60-GHz wireless personal area network applications. Using cascode devices and a DAT four-way power combination, the CMOS PA delivers 14.5- and 18-dBm saturated output power with 10.2% and 12.2% power-added efficiency under 1.8- and 3-V supply voltage, respectively, at 60 GHz. The maximum linear output power ( $ P _{1~{rm dB}} $) is 14.5 dBm. To the best of our knowledge, the MMIC is the first demonstration of a $V$-band CMOS PA using a DAT combining scheme with highest linear output power among the reported 60-GHz CMOS PAs to date.      

Dielectric resolution enhancement coating technology (DiRECT) - a sub-90 nm space and hole patterning technology using 248-nm lithography and plasma-enhanced polymerization

A plasma polymerization coating process named Dielectric Resolution Enhancement Coating Technology (DiRECT) is proposed to shrink critical dimensions (CDs) of space and hole patterns. Fluorocarbon plasmas are used as the precursors to coat a polymer layer on the patterned photo-resist. By adding only one processing step, we are able to shrink poly space and contact hole to sub-90 nm-level using 248-nm lithography. The results of our extensive tests have demonstrated the production-worthiness of this technique for its consistent lot-to-lot repeatability, tight within-wafer CD uniformity, and low defect level.     

Hot-Carrier and Fowler–Nordheim (FN) Tunneling Stresses on RF Reliability of 40-nm PMOSFETs With and Without SiGe Source/Drain

In this brief, RF reliabilities of hot-carrier and Fowler–Nordheim (FN) tunneling stresses on 40-nm PMOSFETs with and without SiGe source/drain (S/D) are studied and compared in detail. The results show that even the strained device with SiGe S/D has a better RF performance; however, its RF reliabilities after both hot-carrier and FN tunneling stresses are deteriorated remarkably by strain-induced defects. In addition, because the SiGe S/D strain-induced defects are mostly located at the surface of the extension of S/D, but not at the channel, the degradation difference between the strained and nonstrained PMOSFETs becomes less as the stress changes from the hot carrier (gate stress voltage $V_{rm gstr} = hbox{drain stress voltage} V_{rm dstr}$) to the vertical-only, i.e., the FN tunneling stress $(V_{ rm dstr} = hbox{0} hbox{but with some} V_{rm gstr})$.      

Device decapsulated (and/or depassivated) – Retest ok – What happened?

One of the very first steps to enter into physical device failure analysis is the device decapsulation. In some cases, an additional depassivation follows to give access to contact needles for internal probing. However, it happens from time-to-time that the device has been cured from its failure behaviour. Such cases often end as non-conclusive analysis result. Our principle investigations and results, however, will help to understand the mechanisms and allow in many “hopeless” cases to draw useful conclusions on the root causes. These are linked in most cases to metal related failures. Besides the “classical case” of touching bond wires, typical root causes are metal filament shorts in the nanometer-order-of magnitude, which can be removed easily by mechanical and/or chemical effects of any delayering procedure. Surface metal shorts may also be generated by bump metal or pad-interface-metallisation-redeposition onto the passivation surface and by metal residue-related recombinations of trimming fuses. In both latter cases, nanometer metal films short-circuit neighbouring pads or fuses in trimming-fuse-arrays. This paper describes these and some second-order mechanisms in detail, which sometimes let the chip recover after decapsulation and/or depassivation.     

Using Reed–Muller RM (1, m) Codes Over Channels With Synchronization and Substitution Errors

We analyze the performance of a Reed-Muller RM(1,m) code over a channel that, in addition to substitution errors, permits either the repetition of a single bit or the deletion of a single bit; the latter feature is used to model synchronization errors. We first analyze the run-length structure of this code. We enumerate all pairs of codewords that can result in the same sequence after the deletion of a single bit, and propose a simple way to prune the code by dropping one information bit such that the resulting linear subcode has good post-deletion and post-repetition minimum distance. A bounded distance decoding algorithm is provided for the use of this pruned code over the channel. This algorithm has the same order of complexity as the usual fast Hadamard transform based decoder for the RM(1,m) code     

Water‐Dispersible,Ligand‐Free,and Extra‐Small (<10 nm) Titania Nanoparticles: Control Over Primary,Secondary, and Tertiary Agglomeration Through a Modified “Non‐Aqueous” Route

Non‐aqueous routes to inorganic nanoparticles are supposedly based on the absence of water; here, this view is partially challenged, showing that the presence of water (or moisture) is probably necessary, and is surely useful to achieve a precise control over the growth/aggregation phenomena leading to titanium dioxide nanoparticles. This study is focused on the preparation of size‐controlled and ligand‐free titania (anatase) nanoparticles in water dispersion. This is achieved through a three‐step process: 1) production of primary (3–4 nm) nanoparticles from titanium alkoxides (Ti(OnPr)₄, Ti(OnBu)₄ or Ti(OⁱPr)₄) in benzyl alcohol through the controlled addition of water; 2) thermal growth phase, where the aggregation of primary nanoparticles at 80 °C leads to secondary nanoparticles with a typical fractal dimension of 2.2–2.4; the primary particles are still identifiable as the individual crystallites composing the secondary nanoparticles; 3) precipitation/re‐dispersion in water, where secondary nanoparticles further agglomerate to yield tertiary nanoparticles. The size of the latter and their photocatalytic efficiency is primarily controlled by the nature of residual alkoxide chains; in particular, isopropoxide groups allow to produce anatase nanoparticles with an average size of 7–8 nm in water dispersion and in the absence of any stabilizing ligand, which is an unprecedented result.     

Multi-Channel Field-Effect Transistor (MCFET)—Part I: Electrical Performance and Current Gain Analysis

Multi-Channel Field-Effect Transistor (MCFET) structures with ultralow $I_{ rm OFF}$ (16 $hbox{pA}/muhbox{m}$) and high $I_{rm ON}$ (N: 2.27 $ hbox{mA}/muhbox{m}$ and P: 1.32 $hbox{mA}/muhbox{m}$ ) currents are obtained on silicon on insulator (SOI) with a high-$ kappa$/metal gate stack, satisfying both low-standby-power and high-performance requirements. The experimental current gain of the MCFET structure is compared with that of an optimized planar FD-SOI reference with the same high-$kappa$/metal gate stack and is quantitatively explained by an analytical model. Transport properties are investigated, and the specific MCFET electrostatic properties are evidenced, in particular a higher $V_{rm Dsat}$ for MCFETs compared with the planar reference. Finally, through 3-D numerical simulations correlated with specific characterizations, the influence of the channel width on the electrical performance is analyzed. For narrow devices, the parasitic bottom channel increases the total drain current of the MCFET structure without degrading the electrostatic integrity.      

Synthesis of Microscale Raft‐shaped Zinc(II)–Phenylalanine Complexes and Zinc(II)–Phenylalanine/dye Hybrid Bundles with New Optical Properties

Novel raft‐like zinc(II)–phenylalanine complexes and zinc(II)–phenylalanine/acid green 27 (AG27) hybrid radial bundles have been successfully synthesized by a simple refluxing reaction. The formation processes of the morphologies and the superstructures of the hybrid bundles were proposed based on the time‐dependent evolution process. The AG27 molecules act as both the inclusion compound and the controller of the morphologies and the superstructures of the final hybrid. The combination of the zinc(II)–phenylalanine complex and AG27 leads to distinct optical properties compared with the individual component materials. This approach opens a new and effective way for the fabrication of amino acid/dye hybrid materials with unique optical properties and is expected to allow access to other organic/organic hybrid materials with structural specificity and functional novelty.     

Percolation Mechanism and Characterization of (CdO)y(ZnO)1–y Thin Films

(CdO)_y(ZnO)_1–y thin films have been prepared by the sol–gel process, based on precursor solutions used separately for such oxides. The Cd/(Cd + Zn) atomic ratio in solution ranged from 0 to 0.32. These compositions were selected on the basis of an observed abrupt fall, of ca. four orders of magnitude, in the resistivity of the films within this range. Such a resistivity drop, with a threshold value of around y = 0.17, is consistent with a percolation mechanism in a three‐dimensional, random, two‐phase system composed of isotropic, sphere‐like, conducting CdO regions embedded in a highly resistive ZnO matrix. Optical measurements show that the films are highly transparent, above 90 % transmission, for wavelengths ≥600 nm. The optical absorption edge shifts to longer wavelengths as the Cd content in the film increases. On the basis of the percolation mechanism observed in the multicomponent system (CdO)_y(ZnO)_1–y, possible future pathways are proposed for the design and construction of highly efficient, transparent, conducting oxides.     

Degradation Mechanism and Stability Improvement Strategy for an Organic Laser Gain Material 4,4′‐Bis[(N‐carbazole)styryl]biphenyl (BSBCz)

The organic material 4,4′‐bis[(N‐carbazole)styryl]biphenyl (BSBCz) is an excellent gain medium for laser devices. However, BSBCz laser output quickly degrades during photoexcitation, which is an issue that must be overcome before it can be used for practical applications. In this study, the photodegradation mechanisms of BSBCz are investigated with the aim of enhancing its excited‐state stability. The photodegradation of BSBCz is attributed to instability of the triplet excited states that would occasionally decompose into other species. This decomposition reduces absorption and introduces exciton quenchers. Incorporating the triplet managing material 9,10‐di(naphtha‐2‐yl)anthracene (ADN) into BSBCz films greatly improves photoluminescence and amplified spontaneous emission stability because of the effective removal of the unstable triplets by ADN. This triplet managing method makes it possible to increase operational stability for BSBCz‐based organic light‐emitting diodes. Therefore, these results will contribute toward the fabrication of stable optically and electrically pumped organic laser diodes.     

Synthesis,Morphology, and Properties of Poly(3‐hexylthiophene)‐block‐Poly(vinylphenyl oxadiazole) Donor–Acceptor Rod–Coil Block Copolymers and Their Memory Device Applications

Novel donor–acceptor rod–coil diblock copolymers of regioregular poly(3‐hexylthiophene) ( P3HT )‐block‐poly(2‐phenyl‐5‐(4‐vinylphenyl)‐1,3,4‐oxadiaz‐ole) ( POXD ) are successfully synthesized by the combination of a modified Grignard metathesis reaction ( GRIM ) and atom transfer radical polymerization ( ATRP ). The effects of the block ratios of the P3HT donor and POXD pendant acceptor blocks on the morphology, field effect transistor mobility, and memory device characteristics are explored. The TEM, SAXS, WAXS, and AFM results suggest that the coil block fraction significantly affects the chain packing of the P3HT block and depresses its crystallinity. The optical absorption spectra indicate that the intramolecular charge transfer between the main chain P3HT donor and the side chain POXD acceptor is relatively weak and the level of order of P3HT chains is reduced by the incorporation of the POXD acceptor. The field effect transistor (FET) hole mobility of the system exhibits a similar trend on the optical properties, which are also decreased with the reduced ordered P3HT crystallinity. The low‐lying highest occupied molecular orbital (HOMO) energy level (–6.08 eV) of POXD is employed as charge trap for the electrical switching memory devices. P3HT‐ b ‐POXD exhibits a non‐volatile bistable memory or insulator behavior depending on the P3HT / POXD block ratio and the resulting morphology. The ITO/ P3HT₄₄ ‐b‐ POXD₁₈ /Al memory device shows a non‐volatile switching characteristic with negative differential resistance (NDR) effect due to the charge trapped POXD block. These experimental results provide the new strategies for the design of donor‐acceptor rod‐coil block copolymers for controlling morphology and physical properties as well as advanced memory device applications.     

Facile Controlled Synthesis and Spectroscopy of CdS1−xSex Alloy and (CdS)1−x@(CdSe)x Core–Shell Nanotetrapods

Nanotetrapods of alloy CdS_1?xSe_x and core–shell (CdS)_1?x@(CdSe)_x are fabricated easily in water using ethylenediamine as a solvent‐coordinating molecular template, and then their optical properties are investigated using diverse static and time‐resolved spectroscopic methods. The arms of the alloy nanotetrapods have single‐crystalline structures of CdS_1?xSe_x without showing staking faults, while the arms of the core–shell nanotetrapods display polycrystalline shell structures of CdSe. The optical properties of CdS_1?xSe_x, where Se atoms are isolated in the CdS lattice, are very different from those of (CdS)_1?x@(CdSe)_x, where banded CdSe passivates the CdS core. Compared with pure CdS nanotetrapods, the photoluminescence of CdS_0.9Se_0.1 shifts to the red by 40 nm, whereas that of (CdS)_0.9@(CdSe)_0.1 does so only by 5 nm. Although the mean luminescence lifetime of alloy CdS_1?xSe_x is shorter than that of pure CdS, it is still much longer than that of core‐shell (CdS)_1?x@(CdSe)_x.     

Dual-Color Electroluminescence ($lambda approx hbox{450} hbox{nm}$ and 650–700 nm) From a Silicon-Based Light Source

In this letter, bright "purple" electroluminescence is observed from a multilayer stack with thin amorphous silicon (alpha-Si)/SiO₂, due to the dual-color light emissions in blue and red. Under photoexcitation, the blue emission is negligible compared with the long wavelength red emission and is only present weakly after anneal in N₂ exceeding 900degC. However, under electrical excitation, the blue becomes obviously visible and shows faster increasing rate with increased carrier injection than the red. From analysis of the samples with different alpha-Si thicknesses (2-7 nm), the red emission is deemed to have resulted from the quantum-confinement effect within the thin alpha-Si layers, whereas the blue emission is speculated to have originated from the Si/SiO₂ interfacial defects.     

One‐Pot Synthesis of Au25(SG)18 2‐ and 4‐nm Gold Nanoparticles and Comparison of Their Size‐Dependent Properties

A one‐pot synthesis of glutathione (denoted as ‐SG) capped gold nanoparticles, including Au₂₅(SG)₁₈ (ca. 1 nm in diameter) 2‐ and 4‐nm particles is reported. These nanoparticles are isolated by methanol‐induced precipitation with a controlled amount of added methanol. Except for their particle size, these nanoparticles have an identical chemical composition (i.e., gold and ‐SG content), synthetic history, and surface conditions, which allows for precise comparison of their size‐dependent properties, in particular the magnetic property as this could be attributed to contamination by trace iron impurities. Specifically, the structure, optical, and magnetic properties of these gold nanoparticles are compared. A trend from non‐fcc (fcc = face centered cubic) Au₂₅(SG)₁₈ nanoclusters (ca. 1 nm) to 2‐ and 4‐nm fcc‐crystalline Au nanocrystals is revealed. The Au₂₅(SG)₁₈ nanoparticles resemble molecules and exhibit multiple optical absorption peaks ascribed to one‐electron transitions, whereas the 4‐nm nanoparticles exhibit surface plasmon resonance at around 520 nm related to the collective excitation of conduction electrons upon optical excitation. The transition from the non‐fcc cluster state to the fcc crystalline state occurs at around 2 nm. Interestingly, both 2‐ and 4‐nm particles exhibit paramagnetism, whereas the Au₂₅(SG)₁₈ (anionic) clusters are diamagnetic. The information attained on the evolution of the properties of nanoparticles from nanoclusters to fcc‐structured nanocrystals is of major importance and provides insight into structure—property relationships.