Projecting gate oxide reliability and optimizing reliabilityscreens

The effect of time-dependent stress voltage and temperature on the reliability of thin SiO₂ films is incorporated in a quantitative defect-induced breakdown model. Based on this model, design curves which can be used along with a breakdown voltage distribution for an oxide technology to determine optimal burn-in conditions are presented. The tradeoff between improved reliability and lower burn-in yield for different gate oxide technologies can also be examined quantitatively using the model     

Burn-in effect on yield

By removing infant mortalities, burn-in of semiconductor devices improves reliability. However, burn-in may affect the yield of semiconductor devices since defects grow during burn-in and some of them end up with yield loss. The amount of yield loss depends upon burn-in environments. Another burn-in effect is the yield gain. Since yield is a function of defect density, if some defects are detected and removed during burn-in, the yield of the post-burn-in process can be expected to increase. The amount of yield gain depends upon the number of defects removed during burn-in. In this paper we present yield loss and gain expressions and relate them with the reliability projection of semiconductor devices in order to determine burn-in time     

缺陷引起的可靠性和成品率关系研究

为了在满足最低可靠性要求的同时尽量提升Ic的成品率,基于缺陷的泊松分布模型及负二项分布模型研究了由缺陷引起的Ic可靠性和成品率这两者之间的关系,并分别建立了相应的成品率一早期可靠性关系模型。基于成品率一可靠性模型,针对氧化层缺陷模型,采用模拟运算的方法,得到了随时间变化的成品率一可靠性关系模型。模型表明,在满足最低可靠性要求的同时,合理设计老化实验参数,可以最大限度地提高成品率,降低Ic制造成本。最后根据这一模型对Ic老化筛选实验的参数选择提出了优化的建议。     

An overview of manufacturing yield and reliability modeling forsemiconductor products

This paper presents an overview of yield, reliability, burn-in, cost factors, and fault coverage as practiced in the semiconductor manufacturing industry. Reliability and yield modeling can be used as a foundation for developing effective stress burn-in, which in turn can warranty high-quality semiconductor products. Yield models are described and their advantages and disadvantages are discussed. Both yield reliability relationships and relation models between yield and reliability are thoroughly analyzed in regard to their importance to semiconductor products     

Relating integrated circuit yield and time-dependent reliability for various defect density distributions

The issue of developing a model to estimate field reliability from process yield has received a growing interest in recent years. Thus far, only Poisson and negative binomial relationships have been obtained, assuming that the number of yield defects is independent of the number of reliability defects in a device. In this paper, we derived explicit yield-reliability relationships for various defect density, such as Erlang, uniform, and triangle distributions, using a multinomial distribution to consider a correlation between the number of yield defects, and the number of reliability defects. The proposed model has advantages over previous models for any defect density distribution to determine the optimal burn-in time.     

Analysis of burn-in time using the general law of reliability

Empirical lifetime distributions sometimes have a bathtub-shaped failure rate. This paper deals with some models having a bathtub-shaped failure rate. The root-mean-square criterion is proposed for selection of the best model. Besides two criteria of optimum burn-in time are proposed. The comparison of some models with the general law of reliability is given to determine a burn-in time in a number of examples.     

Modeling and maximizing burn-in effectiveness

System burn-in can get rid of many residual defects left from component and subsystem burn-in since incompatibility exists not only among components but also among different subsystems and at the system level. Even if system, subsystem, and component burn-in are performed, the system reliability often does not achieve the requirement. In this case, redundancy is a good way to increase system reliability when improving component reliability is expensive. This paper proposes a nonlinear model to: estimate the optimal burn-in times for all levels, and determine the optimal amount of redundancy for each subsystem. For illustration, a bridge system configuration is considered; however, the model can be easily applied to other system configurations. Since there are few studies on system, subsystem, and component incompatibility, reasonable values are assigned for the compatibility factors at each level     

Reliability Enhancement Through Optimal Burn-In

Burn-in is an important screening method used in predicting, achieving, and enhancing field reliability. Although electronics burn-in has been studied qualitatively, no comprehensive quantitative approach exists for determining optimal burn-in periods. This paper presents a cost-optimization model from a system viewpoint, with burn-in periods for the components as the decision variables. This model is applied to an electronic product recently developed which uses many ICs. State-of-the-art ICs have high early-failure rates and long infant mortality periods. Proper use of burn-in reduces early failure rates and reduces system deployment costs. The total cost to be minimized is formulated as a function of the mean costs of the component, device burn-in, shop repair, and field repair, which in turn are functions of the mean number of failures during and after burn-in. Component and system reliability are constraints that have to be satisfied. The early device failures are assumed to have a Weibull distribution. The formulated problem, with failure rates and cost factors, is optimized. Some basic properties of reliability and cost functions are discussed.     

Burn-in optimization under reliability and capacity restrictions

Burn-in is a method to screen out early failures of electronic components. The burn-in problems that minimize the system life-cycle cost have been investigated reasonably well in many applications, but physical constraints during the decision process have not been considered. The authors search for optimal burn-in time and develop a cost-optimization model. Two types of constraint are to be satisfied during decision making: (1) the minimum system reliability requirement, and (2) the maximum capacity available for burn-in. Guidelines are suggested for making burn-in decisions. An example is given to illustrate a practical application. The model generalizes the burn-in problems that were oversimplified in a previous study     

The failure rate of software: does it exist?

This paper first overviews some foundation issues in reliability and argues that generally, it does not make mathematical sense to talk about the failure rate of a piece of software. Furthermore, even when it does make sense, the failure rate is personal, and does not exist outside of the software engineer's mind. Also discussed are matters pertaining to what constitutes a software reliability model, the Markov assumption underlying many of the existing models, the need for models with point masses, and the notion that under subjectivist thinking, reliability models are indexed by two time-scales: (1) mission time; and (2) time at which the software reliability is assessed.<>     

On the Relationship of Semiconductor Yield and Reliability

Traditionally, semiconductor reliability has been estimated from the life tests or accelerated stress tests at the completion of manufacturing processes. Recent research, however, has been directed to reliability estimation during the early production stage through a relation model of yield and reliability. Because the relation model depends on the assumed density distribution of manufacturing defects, we investigate the effect of the defect density distributions on the predicted reliability, for a single-area device without repair and for a two-area device with repair, respectively. We show that for any device, reliability functions preserve an ordering of yield functions. It is also pointed out that the repair capability improves only yield but not reliability, resulting in a large value of the factor that scales from yield to reliability. In order to achieve a reliable device, therefore, we suggest to improve yield and to perform the device test such as burn-in if the scaling factor is large.     

Determination of Optimum Burn-In Time: A Composite Criterion

The object of this paper is to present a mathematical model capable of determining the optimum amount of time that semiconductor devices, which have specified life characteristics, must be placed on burn-in to obtain maximum performance versus total cost. To make the model operational and realistic, the traditional assumption of an exponential (more recently, Weibull) distribution of life is omitted in favor of the generalized gamma distribution (GGD). This is done because the GGD includes, as special cases, such distributions as the normal, Rayleigh, Maxwell, chi, chi2, Weibull, exponential, ordinary gamma, etc. The use of the greater representational capability of the GGD is justified in the results of the studies showing that (other things being equal) small changes in parametric values of life characteristics can cause vast differences in the optimum burn-in time and maximum system effectiveness. The physical performance sector of the model incorporates system effectiveness that includes such factors as availability, expected time to repair, mission reliability, system use coefficient, storage survival probability, and operational readiness. The costs considered are those due to burn-in operation, production, and sales. The model has been studied by use of computer runs from the standpoint of critical analysis and parametric sensitivity analysis.     

Dielectric breakdown distributions for void containing silicon substrates

Distributions of gate oxide failure in various types of silicon substrate materials have been investigated for a wide range of oxide thicknesses. Silicon substrates containing various well-characterized void distributions along with defect-free materials were tested using special low-series resistance capacitor structures. Results of both ramped field tests of variable ramp rate and constant field tests were performed and analyzed within the framework of Weibull statistics. Ramped field tests are not “time zero dielectric breakdown” tests as is commonly asserted. They can in fact be very useful in extrapolating time dependent failure. The same set of Weibull parameters can be used to describe both ramped field and constant field wearout tests if an appropriate model for the time dependent damage accumulation during the field ramp is used. There are implications for reliability predication and the burn-in screening of device populations containing such defects.     

COTS器件空间应用的可靠性保证技术研究

 为保证COTS(Commercial off-the-shelf)器件成功应用于我国高可靠的航天产品中,本文总结分析了国内外空间应用COTS器件的发展情况,提出了空间应用COTS器件的可靠性保证总体思路.该思路特点是首先进行与任务剖面相结合的可靠性评估试验,再对COTS器件进行可靠性筛选.对于热环境的可靠性保证,通过研究NASA的筛选规定,提出了不同任务要求下的老练温度-老练时间的对应关系式;实例表明温度可靠性评估试验方法能够指导工程.对于电离总剂量的可靠性保证,提出了不同任务条件下的回归筛选模型,并进行了验证,相对误差为1%.结果表明本文能够系统性地指导空间工程应用COTS器件.     

On optimal burn-in procedures - a generalized model

Burn-in is a manufacturing technique that is intended to eliminate early failures. In this paper, burn-in procedures for a general failure model are considered. There are two types of failure in the general failure model. One is Type I failure (minor failure), which can be removed by a minimal repair or a complete repair; and the other is Type II failure (catastrophic failure), which can be removed only by a complete repair. During the burn-in process, two types of burn-in procedures are considered. In Burn-In Procedure I, the failed component is repaired completely regardless of the type of failure; whereas, in Burn-In Procedure II, only minimal repair is done for the Type I failure, and a complete repair is performed for the Type II failure. Under the model, various additive cost functions are considered. It is assumed that the component before undergoing the burn-in process has a bathtub-shaped failure rate function with the first change point t/sub 1/, and the second change point t/sub 2/. The two burn-in procedures are compared in cases when both the procedures are applicable. It is shown that the optimal burn-in time b/sup */ minimizing the cost function is always before t/sub 1/. It is also shown that a large initial failure rate justifies burn-in, i.e., b/sup */>0. The obtained results are applied to some examples.     

Some considerations on system burn-in

The questions of whether or not to perform system burn-in, and how long the burn-in period should be, can be answered by developing a probabilistic model of the system lifetime. Previously, such a model was obtained to relate component burn-in information & assembly quality to the system lifetime, assuming that the assembly defects introduced in various locations of a system are capable of connection failures represented by an exponential distribution. This paper extends the exponential-based results to a general distribution so as to study the dependence of system burn-in on the defect occurrence distribution. In particular, a method of determining an optimal burn-in period that maximizes system reliability is developed based on the system lifetime model, assuming that systems are repaired at burn-in failures.     

一种星载电子产品老练试验加速因子估计方法

提出一种基于可靠性预计数据的星载电子产品老练试验加速因子的估计方法。在该方法中,温度对产品失效过程的影响通过器件失效率预计模型中的温度应力参数予以刻画。通过比较产品在工作环境温度与老练试验温度下的预计失效率数据来估计老练试验加速因子。该方法简单、易行,含义明确,有望增强可靠性评估结果与可靠性预计结果的可比性。     

Cost Effective Burn-In and Replacement Times

Bum-in and replacement policies have been used to reduce the cost of maintaining a system function. Existing solutions to the burn-in and the combined burn-in, replacement problems are typically difficult to apply. A simpler solution to this problem is given. The only distributional property required is the reliability function. Thus either parametric or nonparametric estimates of reliability can be used in the solution. The nonparametric solution is appealing in the case of a combined burn-in and replacement policy because the U shaped failure rate makes parametric estimation of reliability difficult.     

Relation between yield and reliability of integrated circuits and application to failure rate assessment and reduction in the one digit FIT and PPM reliability ERA

Clear relations have been established between E-sort yield and burn-in, EFR and field failure rates for nearly 50 million high volume products in bipolar, CMOS and BICMOS technologies from different waferfabs. The correlations obey a simple model that assumes that the reliability defect density is a fraction of the waferfab defect density and that rootcauses of failures are the same. The model allows a die size independent prediction and assessment of FIT and PPM reliability levels of an IC just based on its yield, eliminating the need for excessive lifetesting. ‘Maverick’ batches are identified by more than 2 to 3 rejects per batch and can not be eliminated by scrap of low yielding wafers alone. For non-mature technologies only correlations with functional yield are found, the parametric yield loss can be disregarded. Using the results, it is shown how reliability can be improved in a fast and controlled way, even in the 1 digit FIT and PPM reliability era, by reducing waferfab defect density, elimination of special causes and implementation of screens at product test like voltage screen and Iddq testing. As the effect of yield on PPM reject level is not that strong, the latter approach can be very effective in improving reliability.     

Effect of artificial-intelligence planning-procedures on systemreliability

For an embedded real-time process-control system incorporating artificial-intelligence programs, the system reliability is determined by both the software-driven response computation time and the hardware-driven response execution time. A general model, based on the probability that the system can accomplish its mission under a time constraint without incurring failure, is proposed to estimate the software/hardware reliability of such a system. The factors which influence the proposed reliability measure are identified, and the effects of mission time, heuristics and real-time constraints on the system reliability with artificial-intelligence planning procedures are illustrated. An optimal search procedure might not always yield a higher reliability than that of a nonoptimal search procedure. Hence, design parameters and conditions under which one search procedure is preferred over another, in terms of improved software/hardware reliability, are identified