1、JEDEC STANDARD Failure-Mechanism-Driven Reliability Monitoring JESD659B (Revision of JESD659A, September 1999) FEBRUARY 2007 (Reaffirmed: JUNE 2011) JEDEC SOLID STATE TECHNOLOGY ASSOCIATION NOTICE JEDEC standards and publications contain material that has been prepared, reviewed, and approved throug
2、h the JEDEC Board of Directors level and subsequently reviewed and approved by the JEDEC legal counsel. JEDEC standards and publications are designed to serve the public interest through eliminating misunderstandings between manufacturers and purchasers, facilitating interchangeability and improveme
3、nt of products, and assisting the purchaser in selecting and obtaining with minimum delay the proper product for use by those other than JEDEC members, whether the standard is to be used either domestically or internationally. JEDEC standards and publications are adopted without regard to whether or
4、 not their adoption may involve patents or articles, materials, or processes. By such action JEDEC does not assume any liability to any patent owner, nor does it assume any obligation whatever to parties adopting the JEDEC standards or publications. The information included in JEDEC standards and pu
5、blications represents a sound approach to product specification and application, principally from the solid state device manufacturer viewpoint. Within the JEDEC organization there are procedures whereby a JEDEC standard or publication may be further processed and ultimately become an ANSI standard.
6、 No claims to be in conformance with this standard may be made unless all requirements stated in the standard are met. Inquiries, comments, and suggestions relative to the content of this JEDEC standard or publication should be addressed to JEDEC at the address below, or refer to www.jedec.org under
7、 Standards and Documents for alternative contact information. Published by JEDEC Solid State Technology Association 2011 3103 North 10th Street Suite 240 South Arlington, VA 22201-2107 This document may be downloaded free of charge; however JEDEC retains the copyright on this material. By downloadin
8、g this file the individual agrees not to charge for or resell the resulting material. PRICE: Contact JEDEC Printed in the U.S.A. All rights reserved PLEASE! DONT VIOLATE THE LAW! This document is copyrighted by JEDEC and may not be reproduced without permission. For information, contact: JEDEC Solid
9、 State Technology Association 3103 North 10th Street Suite 240 South Arlington, VA 22201-2107 or refer to www.jedec.org under Standards-Documents/Copyright Information. JEDEC Standard 659B -i- FAILURE-MECHANISM-DRIVEN RELIABILITY MONITORING CONTENTS Page Introduction ii 1 Scope 1 2 Terms and definit
10、ions 1 3 Process controls 3 4 Identifying failure mechanisms 3 5 When to establish a monitor 4 6 What a monitor includes 4 7 Control action system 5 8 Monitor optimization 6 Annex A (informative) Differences between JESD659B and JESD659A 7 JEDEC Standard 659B -ii- FAILURE-MECHANISM-DRIVEN RELIABILIT
11、Y MONITORING Introduction This standard presents a methodology for monitoring component and subassembly reliability. It can be of use to suppliers and users interested in known reliability as an attribute of the component through the production life. Under this standard, the suite of metrics is tail
12、ored to monitor the failure mechanisms which limit the reliability. This is distinguished from a stress-driven monitor approach in which a fixed suite of acceptance stresses or other testing are prescribed and applied without customization according the failure mechanisms for the component or subass
13、embly. JEDEC Standard No. 659B Page 1 FAILURE-MECHANISM-DRIVEN RELIABILITY MONITORING (From JEDEC Board ballot JCB-07-19, formulated under the cognizance of JC-14.3 Committee on Silicon Devices Reliability Qualification and Monitoring.) 1 Scope This standard describes essential requirements for a re
14、liability monitor for components and subassemblies based on the measurement of failure mechanisms which limit reliability. It applies through the post-qualification production period. Both intrinsic (wearout and systematic) and extrinsic (defect-based) sources of failure are addressed. 2 Terms and d
15、efinitions For the purposes of this standard, the following definitions apply. Definitions marked by a asterisk (*) are taken from EIA-557-A, Statistical Process Control Systems. They are replicated here for completeness. characteristic*: A distinguishing feature of a process or its output on which
16、variables or attributes data can be collected. common cause*: A source of natural variation that affects all the individual values of the process output being studied. In control chart analysis it appears as part of the random process variation. control limits*: The maximum allowable variation of a
17、process characteristic due to common causes alone. Variation beyond a control limit may be evidence that special causes affecting the process. Control limits are calculated from process data and are usually represented as a line (or lines) on a control chart. They are not to be confused with enginee
18、ring specification limits. critical failure mechanism: In semiconductor devices, any potential physical failure mechanism that exhibits one or more of the following: intermittency (e.g., bond lifts), increasing failure rate (e.g., electromigration), and inconsistent or unpredictable failure kinetics
19、 (e.g., stress-induced metal voiding) extrinsic failure mechanism: (1) A failure mechanism caused by an error occurring during the design, layout, fabrication, or assembly process or by a defect in the fabrication or assembly materials. (2) A failure mechanism that is directly attributable to a defe
20、ct created during manufacturing. failure: (1) The loss of the ability of a component to meet the electrical or physical performance specifications that (by design or testing) it was intended to meet. (2) A component that has failed. failure mechanism from fabrication processes: A physical failure me
21、chanism in which all products with the same wafer fabrication process, design rules, and processing line are treated as a homogeneous population for the purpose of statistical reliability monitoring independent of package technology, material, construction, and type. JEDEC Standard No. 659B Page 2 2
22、 Terms and definitions (contd) failure mechanism from assembly: A physical failure mechanism in which all products with the same assembly technology, including assembly material, assembly construction, and package type and built on the same assembly line are treated as a homogeneous population for t
23、he purpose of statistical reliability monitoring independent of the fabrication process and line. intrinsic failure mechanism: (1) A failure mechanism caused by a natural deterioration in the materials or the manner in which the materials are combined during fabrication or assembly processes that ar
24、e within specification limits. (2) A failure mechanism attributable to natural deterioration of materials processed per specification. node*: A definable point in the process at which form, fit, or function of the product or service is altered. nonconformity*: A specific occurrence of a condition th
25、at does not conform to specification. Such an occurrence is sometimes called a discrepancy. parameter: A measurable characteristic. physical failure mechanism: A physical or chemical process that ultimately results in failure. process*: (1) A combination of people, procedures, methods, machines, mat
26、erials, measurement equipment, and/or environment for specific work activities to produce a given product or service. (2) A repeatable sequence of activities with measurable inputs and outputs. sample:* A set of individuals taken from a population. special cause; assignable cause:* A source of varia
27、tion that is intermittent, unpredictable, or unstable, and affects only some of the individual values of the process output being studied. statistical reliability monitoring (SRM): A statistically based methodology for monitoring and improving reliability involving identification and classification
28、of failure mechanisms, development and use of monitors, and investigation of failure kinetics, allowing prediction of failure rates at use conditions. statistical reliability monitor family (SRMF): A product or group of products whose process similarities make them a homogeneous population for the p
29、urpose of statistical reliability monitoring. A homogeneous population of product from one SRMF shall have similar propensity towards the physical failure mechanisms being monitored when that product is stressed by accelerated tests or operated in its intended system application. Each product in an
30、SRMF will have the same failure rate for each mechanism only when the factors affecting a failure mechanism are identical for each product type. variables data*: A measure of a characteristic where every value within a given interval is possible. JEDEC Standard No. 659B Page 3 3 Process controls The
31、 supplier shall define a Statistical Process Control (SPC) system for all critical process nodes in accordance with EIA-557-A, Statistical Process Control Systems. This shall include establishing critical equipment capabilities, preventive maintenance, and calibration procedures. SPC metrics can int
32、ersect with those used for monitoring reliability, but are not required to do so. An SPC system alone might not address all requirements for a failure-mechanism driven reliability monitor program. 4 Identifying failure mechanisms The supplier shall assess the failure mechanisms which contribute to t
33、he failure rate and institute respective monitors for those mechanisms. The identification of failure mechanisms may be based on mechanisms observed during development characterizing product sensitivities or process capabilities, mechanisms observed during qualification, mechanisms identified as par
34、t of Failure Mode and Effects Analysis (FMEA), or mechanisms observed in previous test or field experience with products using like processes, materials set, or tooling and production facilities. JEP131 provides guidance on FMEA. Stress-driven qualifications (e.g., JESD47) and evaluations in which s
35、ampling is structured to show compliance to an acceptance criterion without generating failures will not provide an adequate basis for identifying the set of reliability-limiting failure mechanisms for a component. See section 6 regarding continuing acceptability to the original qualification criter
36、ia. The supplier shall be able to identify the rationale for the adequacy of the set of failure mechanisms identified for monitoring. The failure mechanisms requiring monitoring shall be reassessed for completeness and appropriateness when the process is changed or the product undergoes modification
37、s which may alter its sensitivity to the existing process. JEDEC Standard No. 659B Page 4 5 When to establish a monitor A monitor for each failure mechanism which contributes non-trivially to the failure rate for the component shall be established. Non-trivial contributions shall be ascribed to crit
38、ical failure mechanisms and systematic process risks as well as to extrinsic (defect-based) detractors that contribute significantly to the overall failure rate. Since extrinsic (defect-based) detractors may exist at trivial non-zero levels, a threshold of significant contribution at 10% of the over
39、all failure rate expected in the field is recommended. Monitors may take any of several forms. Common forms are as in-process metrics and as product-level life tests or accelerated stress tests. Some monitors may be used to address more than one failure mechanism, e.g., a functional stress on an int
40、egrated product designed to monitor multiple extrinsic failure mechanisms. The vehicle for the monitor measurement shall be the product or another vehicle which shares the risk for this failure mechanism by similarity in process and construction (i.e., from the same SRMF). 6 What a monitor includes
41、For mechanisms needing a monitor (See Section 5), this section identifies features that monitor shall include. Sampling for the respective monitors shall be sufficient to allow the supplier to identify an excursion to the total component failure rate (driven by the composite of failure mechanisms).
42、Sampling for any given monitor should be scaled to yield non-zero fallout (attributes monitors) or sufficient parametric shift for extrapolation to failure (variables monitors) as an opportunity for learning enablement and to validate the continuing presence, abatement, or elimination of particular
43、mechanisms. NOTE 1 To yield non-zero results at levels compatible with a component failure rate, monitors for intrinsic failure mechanisms typically will require smaller samples than those for extrinsic failure mechanisms do. As such, sampling for extrinsic failure mechanisms may require the aggrega
44、tion of monitor data across broader sampling (especially in time). The aggregation of extrinsic data should be confined to no more than an annual basis; shorter periods are encouraged. NOTE 2 When a monitor is designed to measure well beyond the required minimum and no discernable failures (for attr
45、ibutes monitors) or shifts (for parametric monitors) are produced in a substantive sample history, it may be impractical to force non-zero data and unnecessary to do so for the purpose of measuring variations over time. Monitors shall include targets which reflect acceptable levels. Ideally, these t
46、argets directly correlate to quantified failure rate levels. Regardless of whether they have an established correlation, the supplier shall provide for the quantification of the impact when the total component failure rate is in jeopardy for an excursion as signaled by the combined monitors. Regardl
47、ess of the qualification approach used, targets for a qualified device shall be in compliance with the qualification acceptance criteria used or equivalent. JEDEC Standard No. 659B Page 5 6 What a monitor includes (contd) For failure mechanisms from fabrication processes that require a monitor, each
48、 wafer fabricator shall have its own monitor or SRMF. For failure mechanisms from assembly that require a monitor, each assembly site monitor or SRMF shall have its own monitor or SRMF. For mechanisms requiring a monitor and driven by the combined effects of wafer fabrication and assembly, each comb
49、ination of fabricator and assembly site shall have its own monitor or SRMF. 7 Control action system The suppliers control system shall address situations in which characteristics reach or exceed statistical warning limits or characteristics exceed the control limits. The suppliers control action system will: Document the definition of nonconformities as well as the procedure used to identify, verify, and determine the cause of nonconformities reported. Corrective actions shall remedy the situation causing the nonconformity without adversely affecting the reliability of the produc
