This series has examined the semiconductor back-end process in detail, covering everything from the different types of semiconductor packages to the packaging process and materials. This final installment of the series will introduce the tests and standards that determine the reliability of semiconductor packages. In addition to detailing how these standards are evaluated and set, this article will also cover the assessments that test semiconductor packages’ expected lifetime, reliability in various surrounding conditions, and mechanical reliability.
What Is Product Reliability?
The quality of a semiconductor product is determined by how successfully it meets the required standards and properties. Meanwhile, the reliability of a semiconductor device is defined as the probability that it can maintain its quality without failures over a set period of time, thereby increasing customer satisfaction and the likelihood of repeat purchases. In this case, failures refer to errors that occur during the use of the product, while defects are errors that occur during the product’s manufacturing or inspection. Therefore, the product is said to be of poor quality if it has many defects, while it is considered to be unreliable if failures occur frequently or while it is under warranty.
▲ Figure 1. A table showing the differences between quality and reliability (Source: Hanol Publishing)
Figure 1 shows a more comprehensive breakdown of the differences in the meanings and characteristics of quality and reliability. To give a more detailed definition, reliability is the ability of a system, part, or material to maintain its initial quality and performance over a specific amount of time, distance, or amount of use. This must be achieved without experiencing failures under given conditions such as manner of usage or environmental factors. As a result, it is crucial for semiconductor companies to assess whether their quality and reliability levels have reached industry standards before commencing mass production. These levels should also be regularly examined while products are being mass produced.
As an important step to evaluating product reliability, the standards of reliability must be clearly defined in advance. In the case of a company that ships 100 products, some of the following questions would need to be considered: How many of the products should be operational after three years? What are the observable patterns in the product’s operation time? Can it be guaranteed that 90 out of 100 products will be operational after five years? How long will 95 out of 100 products still function properly?
Tests are required to verify these standards. Although it would be ideal to run tests for three years, five years and even longer to determine a product’s reliability at these stages, concentrating so much time on the product’s evaluation would delay mass production for too long. Therefore, companies use accelerated testing and statistical techniques to assess reliability. Calculations such as reliability functions, lifetime distributions, and the average lifetime are also used to complete a reliability validation in a relatively short amount of time.
JEDEC Standards
Companies that design and manufacture semiconductors evaluate the reliability of their own products and provide customers with the results. Customers can use the results to determine whether the product meets their needs or decide to perform their own reliability assessments. However, if a semiconductor company and its customers have different evaluation criteria, they must go through the trouble of aligning their different standards. As a solution, semiconductor companies usually adhere to standards from the JEDEC Solid State Technology Association1, commonly referred to as JEDEC, to satisfy the needs of the company and its customers.
1JEDEC Solid State Technology Association: The leading organization in developing open standards and publications for the microelectronics industry.
JEDEC’s primary role is to enable manufacturers and organizations to jointly review and establish uniform standards for electronic devices such as integrated circuits (ICs). As the specifications set by JEDEC have become widely accepted as the international standard, JEDEC is in effect the standards-setting organization for the semiconductor industry worldwide.
The organization’s board of directors (BoD) determines the policies and procedures and has the final approval of JEDEC standards. Additionally, there are also numerous JEDEC committees (JC) which set standards for the respective fields that they are specialized in. For the semiconductor industry, the following are some of the important committees and their roles: JC-14 Quality and Reliability of Solid State Products is responsible for reliability standards; JC-11 Mechanical Standardization sets standards for outlines of modules and semiconductor packages; JC-42 Solid State Memories sets standards for DRAMs; and JC-63 Multiple Chip Packages decides on standards for mobile multi-chip packages.
If a company needs to have a standard created for one of its products, the company submits a standard proposal to be voted on by the respective committee members. Each company has the right to one vote, regardless of its size. Proposals that are approved by a committee are then voted on again by the BoD, and proposals that are passed by the board are finally made into JEDEC standards and publicized to the respective industries.
Reliability Tests for Assessing Product Lifetimes
In addition to international evaluation standards, there are also numerous metrics used to assess product reliability. These include the various assessments which evaluate the lifetime of semiconductor products.
Early Failure Rate
The early failure rate (EFR) estimates the number of device failures which occur within a year in the user environment. For some product lines, however, this period can be reduced to six months or increased to over a year depending on the system’s lifetime or products that specifically require higher reliability. Burn-in2 screens products that are likely to fail in a short amount of time, while the EFR is used to verify whether the potential failure rate of these screened products is maintained at an acceptable level (see Figure 2). The test conditions are set and evaluated using the acceleration factor for the temperature and voltage of the relevant semiconductor product.
2Burn-in: A test that stresses a product with voltage and temperature in order to eliminate potential defects in the product at an early stage. Burn-in performed after packaging is called test during burn-in (TBDI).
▲ Figure 2. The early failure rate (EFR) zone within a bathtub curve, which represents three phrases of a product’s failure rate over time (Source: Hanol Publishing)
High & Low Temperature Operating Life Tests
The high temperature operating life (HTOL) test is one of the most common types of product lifetime evaluations. It assesses issues that arise from temperature and voltage stress during a product’s operation. The HTOL test is also considered to be comprehensive as it not only assesses premature failures but also identifies failures caused by accidents or wear and tear. Meanwhile, the low temperature operating life (LTOL) test assesses the likelihood of failures due to the impact of hot carriers3. However, due to the application of voltage and temperature, there is also the possibility of other failures occurring.
3Hot carrier: Traveling electrons that become excessively mobile due to the high electric field deriving from transistors shrinking in size and causing the channel length to become shorter, which increases the electric field. This short channel effect occurs in semiconductor transistors.
High Temperature Storage Life
The high temperature storage life (HTSL) test evaluates the reliability of a product under high-temperature storage conditions. Such conditions can affect the lifetime of a product due to diffusion, oxidation, intermetallic growth, and chemical degradation of package materials.
Endurance & Data Retention
The endurance test evaluates how many program/erase (PE) cycles products such as NAND flash memory can withstand. Regarding NAND, a key reliability factor of these products is data retention as it measures how long data will be stored in a cell for a given period of time even in the absence of power.
Reliability Tests for Various Surrounding Conditions
There are numerous surrounding conditions that may contribute to the failure of a semiconductor product. This is why tests are required to assess the ability of products to withstand certain surrounding conditions before they are shipped to their respective destinations.
Preconditioning
After a product has been shipped and stored, preconditioning assesses potential problems that may occur during a customer’s manufacturing process as hygroscopic4 and thermal stress can result in reliability issues. Preconditioning evaluates packaging reliability in humid conditions by replicating conditions where the product is sold, shipped to the customer, unpacked from its vacuum packaging, and mounted in the system. This treatment is used as a precursor for reliability tests of surrounding conditions, including temperature humidity bias (THB), the highly accelerated stress test (HAST), and thermal cycle (TC).
4Hygroscopic: The phenomenon of absorbing moisture from the air. In semiconductors, this absorption of moisture can lead to failures.
The evaluation follows the order of TC, bake, soak, and reflow. Figure 3 shows how preconditioning applies to the processes of packaging, transportation, and mounting to the system.
▲ Figure 3. Relationship of production, transportation, and usage to preconditioning test conditions (Source: Hanol Publishing)
Thermal Cycle
A thermal cycle (TC) assesses a product’s tolerance to instantaneous temperature changes that may occur in different user environments. Semiconductor packages and modules are composed of various materials with different coefficients of thermal expansion5 (CTE) which can lead to stress-induced fatigue failures. This stress is caused by shrinkage and expansion which occurs following thermal changes.
5Coefficient of thermal expansion (CTE): A material property that indicates the extent to which a material expands upon heating.
The primary purpose of a TC is to measure the stress tolerance of a semiconductor package in relation to temperature changes, but high and low temperature stresses can cause many other failures. Prolonged thermal shocks verify the potential for interfacial delamination6, internal and external package cracks, and chip cracks caused by factors such as stress from each package material or thermal expansion. In addition, as the importance of solder joints is increasing due to restrictions on the use of hazardous materials including lead and the expansion of applications such as mobile devices, a TC can effectively evaluate the reliability of solder joints.
6Interfacial delamination: The separation of interfaces in a semiconductor package.
Temperature Humidity Storage & Temperature Humidity Bias Tests
The temperature humidity storage (THS) test assesses the tolerance of semiconductor products in high temperatures and humidity. To determine the appropriate exposure time, it is recommended to simulate the actual user environment by measuring the amount of moisture absorption after opening the moisture-proof package. Meanwhile, temperature humidity bias (THB) evaluates moisture resistance by subjecting the product to an electrical bias7. Although most failures are caused by aluminum corrosion, other potential problems can arise from temperature stress. This test also identifies package reliability issues such as pad metal corrosion due to moisture penetrating through micro gaps between leads and mold pores, and failures caused by moisture infiltrating through pores or holes in the protective film.
7Electrical bias: The deliberate application of direct current (DC) between two points for the purpose of controlling a circuit.
Pressure Cooker Test
The pressure cooker test (PCT), which is more rigorous than THS and THB, is an ideal early assessment of moisture resistance. Also known as an autoclave8, the test evaluates the moisture resistance of a plastic mold compound and assesses the reliability of the mold structure by infiltrating moisture using 100% relative humidity and high pressure. Additionally, it identifies problems arising from moisture penetration between the leads and through pores in the mold.
8Autoclave: A high-pressure cooker. When water is added while the cooker is sealed at high temperature, the water evaporates and increases the pressure and humidity to create the conditions necessary for the specimen inside the autoclave.
Similar to THS, the PCT used to be an essential reliability test for thick semiconductor packages. However, recent international trends, including JEDEC’s assessment, suggest that the stress magnitude is too high for current packages. Thus, the test is used selectively depending on the type of package. The PCT is used for leadframe products, while the unbiased highly accelerated stress test (UHAST) is used for substrate products.
Unbiased Highly Accelerated Stress Test, Highly Accelerated Stress Test & High Accelerated Life Test
The UHAST evaluates reliability by applying stresses similar to those applied by a PCT to thin packages of substrate-type products, such as fine-pitch ball grid array (FBGA) packages. There are also similarities in their ability to identify and discover types of failures. While the PCT applies stress by using saturated humidity, or 100% relative humidity, the UHAST uses unsaturated humidified conditions at 85% relative humidity that is similar to the customer user environment. Galvanic corrosion9 or direct chemical corrosion are mainly employed for this evaluation.
9Galvanic corrosion: An electrochemical process whereby a more active metal (anode) corrodes in preference to a more resistant metal (cathode) that it is in contact with through an electrolyte.
An additional evaluation is the highly accelerated stress test (HAST) which is used to assess the reliability of non-hermetic packages operating in humid environments. It uses the same method as THB, as pins with static bias applied undergo temperature, humidity, and pressure stresses. Lastly, the highly accelerated life test (HALT) is a quick stress test that helps identify and correct defects during the product design phase.
Reliability Tests for Mechanical Factors
Semiconductor products are subjected to environmental stresses caused by mechanical, climatic, and electrical factors during their handling, storage, transportation, and operation. These loads significantly affect the design reliability of the equipment. For this reason, it is necessary to evaluate products under development or in mass production to identify abnormalities. Manufacturers can subject products to physical stresses such as vibration, shocks, and drops during the assessment process.
Shock
Shock testing evaluates resistance to simulated impacts that may occur during handling and transportation. Typical shock tests include the hammer shock test, which involves fixing a test sample in place and striking it with a hammer, and the drop test, in which a product is subjected to a free-fall drop. The hammer shock test assesses how well a product withstands the force and pulse of the hammer, as well as the number of impacts it can endure. In the case of the drop test, the test subject is dropped in free fall from a height of 1 to 1.2 meters to reflect the actual working environment of the user.
Vibration, Bending & Torsion
Vibration is an assessment of the product’s resistance to vibrations that may occur when the product is being transported. It usually employs sine vibration10 testing in compliance with JEDEC standards.
10Sine vibration: A vibration that varies in frequency over time.
Other tests include the bending test, which assesses solder joint defects caused by warping or bending on the printed circuit board (PCB), as well as the torsion test. Also referred to as the twist or torque test, the torsion test evaluates the resistance to solder joint problems and product warpage that may occur on the PCB when it is subjected to torsional stress.
Ensuring Reliable Semiconductor Products
The reliability tests and standards discussed in this episode serve as the foundation for ensuring that these vital components meet the stringent demands of today’s technology-driven world. From tests for surrounding conditions and mechanical factors to product lifetime evaluations, the various assessments show that the semiconductor industry is committed to producing reliable and durable products. In particular, SK hynix is doing its utmost to ensure its products meet the highest reliability standards and exceed customer expectations. Going forward, the company will continue refining and reviewing its reliability testing to keep pace with the ever-evolving technological landscape.
Read articles from the Front-End Process series
Read articles from the Back-End Process series
