Woon-Il Choi

Decrease of Dark Current by Reducing Transfer Transistor Induced Partition Noise With Localized Channel Implantation

The local increase of the threshold voltage of the transfer transistor is proposed to reduce the dark current in a CMOS image sensor. It is suggested that the local increase of the threshold voltage controls the partition noise which contributes to the dark current. The dark current is shown to be reduced considerably by the proposed structure. The proposed method induces little change in the hot carrier reliability as well as in the device performance.

Effects of High-Pressure Annealing on Random Telegraph Signal Noise Characteristic of Source Follower Block in CMOS Image Sensor

The effect of high-pressure deuterium (D) and hydrogen (H) annealing on random telegraph signal (RTS) noise characteristics of source follower (SF) block, SF, and row selector (SEL) transistors in CMOS image sensor (CIS) active pixel sensor (APS) was comparatively analyzed in depth. RTS noise characteristics of SF transistor (M1) and SEL transistor (M2) with forming gas (FG) annealing showed the smallest Δ<i>I</i>_D/<i>I</i>_D, whereas FG annealing was not efficient to reduce the RTS noise of SF block (M1 + M2). Although Δ<i>I</i>_D/<i>I</i>_D of SF block was reduced by high-pressure H₂ annealing, high-pressure D₂ annealing showed the greatest reduction in Δ<i>I</i>_D/<i>I</i>_D of SF block (M1 + M2), which was believed to attribute to the effective passivation of interface traps by the isotope effect of D. Therefore, high-pressure D₂ annealing is potentially significant for reducing RTS noise characteristics and thermal budget as well as improving device performance in CIS APS.

Effect of Nitrogen Concentration on Low-Frequency Noise and Negative Bias Temperature Instability of p-Channel Metal--Oxide--Semiconductor Field-Effect Transistors with Nitrided Gate Oxide

In this paper, the dependence of negative bias temperature instability (NBTI) and low-frequency noise characteristics on the various nitrided gate oxides is reported. The threshold voltage shift (ΔVT) under NBTI stress for thermally nitrided oxide (TNO) was greater than that of plasma nitrided oxide (PNO), whereas the slopes of ΔVT versus stress time for PNO were similar to those for TNO. The flicker noise (1/f noise) characteristic of PNO was better than that of TNO by about 1 order of magnitude, although the 1/f noise of PNO showed almost the same dependence on the frequency as that of TNO. The carrier number fluctuation model due to the trapping and detrapping of electrons in oxide traps was found to be a dominant mechanism of flicker noise. The probability of the generation of drain current random telegraph signal (ID–RTS) noise shows similar values (70–78%) for all nitrided oxides, which shows that the generation of RTS noise is not greatly affected by the nitridation method or nitrogen concentration.

Decoupling of RTS noise in high density CMOS image sensor using new test structures

In this work, new test structures are proposed to characterize the RTS (Random Telegraph Signal) noise of the CMOS image sensor. The RTS noise of the driver transistor and the source follower transistor, as well as the source follower block itself, are measured using the proposed test structures. The probability of monitoring the RTS noise of the driver transistor and the source follower transistor is 76 % and 52 %, respectively. However, the probability of the generation of the RTS noise for the source follower block is about 74 %. Therefore, it can be said that the driver transistor dominates the RTS noise of the source follower block.

A Study of Dielectric Relaxation and Capacitance Matching of

Key analog characteristics such as dielectric relaxation and capacitance matching for Al2O3/HfO2/Al2O3 (AHA) metal-insulator-metal (MIM) capacitors were analyzed for high-performance analog circuit applications. Although the dc characteristics of AHA MIM capacitor were acceptable for analog operation, the variation of the quadratic voltage coefficient (α) under constant voltage stress (CVS) and the dielectric relaxation remained problematic. The dependence of α for AHA MIM capacitor decreased gradually under CVS and the dielectric relaxation of AHA MIM capacitor was greater than that of conventional MIM capacitor, which was due to the effect of preexisted traps in high- k dielectric. The extracted matching coefficient of AHA MIM capacitor was, however, 0.698% μm, which was low enough to be used for analog/mixed signal/radio frequency application.

{\rm Al}_{2}{\rm O}_{3}/{\rm HfO}_{2}/{\rm Al}_{2}{\rm O}_{3}

This paper investigated pixel characteristics with a split of transfer gate structure in CMOS image sensor. It is shown that the transfer gate (TG) should be as wide as possible for the given pixel structure for lag–free pixel operation. This investigation uncovered relationships between transfer gate geometry and pixel performance that could provide guidance when designing pixel structure. In particular, it shows how sensitive pixel fixed pattern noise is to the transfer gate width. Experimental results indicate that pixel fixed pattern noise can be reduced to be less than 1%.

MIM Capacitors

FPN (fixed-pattern-noise) mainly comes from the device or pattern mismatches in pixel and color filter, pixel photodiode leakage in CMOS image sensor. In this paper, optical stack module related pixel FPN was investigated and the classification of pixel FPN contribution with the individual optical module process was presented. The methodology and procedure would be helpful in reducing the greater pixel FPN and distinguishing the complex FPN sources with respect to various noise factors.

Trends in pixel fixed pattern noise characteristics with transfer gate geometry

Metal-insulator-metal (MIM) Capacitors are widely used for radio frequency (RF) and analog mixed signal (AMS) circuits, and dynamic random access memory (DRAM) applications [1]. Although high capacitance density of MIM capacitor for RF and AMS circuits is required to reduce the chip size and system cost, the conventional dielectric materials, such as SiO 2 and Si 3 N 4 , as the thickness of dielectric scales down, emerge as key limitations due to high leakage current and reliability issues. High-k (HK) dielectric, such as HfO 2 , Al 2 O 3 , ZrO 2 , Ta 2 O 5 , and La 2 O 3 , is necessary to achieve the high capacitance density and to reduce the leakage current [2]. Among the various HK dielectric candidates, hafnium-based MIM capacitors are widely used due to their high capacitance density, good thermal stability, and high band gap [3]. However, HK dielectric for MIM capacitors induces high voltage linearity and poor reliability characteristics after electrical stress, due to defects in the metal-insulator interface. The reliability characteristic of MIM capacitor under electrical stress is related to the injected charge in the interface of HK dielectrics [4]. These degradations cause the distortion of MIM capacitors, which limits the performance of the RF and AMS circuits. Hence, an accurate analysis of the degradation of MIM capacitor under electrical stress is required in order to stabilize these characteristics. Moreover, various kinds of voltage shape can be applied to the MIM capacitors. However, there was little study on the effect of stress voltage type on the reliability of advanced MIM capacitors. In this work, reliability of MIM capacitor with high capacitance density was analyzed using three kinds of voltage stress; constant voltage stress (CVS), unipolar voltage and bipolar voltage stresses.