Intermetallic Coverage Research Articles

High Robustness of Coated-Ag Wire Bonding

Formation of free air ball (FAB) within the range of certain processing parameters leads to excellent bondability (distributed intermetallic coverage (IMC) at the center and periphery of the bonded ball interface, bonded ball concentricity, etc.) on bonding to aluminum (Al) bond pad and Au/Ag plated surfaces. The wire revealed zero stoppage and 0.8M bond touchdowns when tested under laboratory conditions. Thermal aging of bonded interface showed absence of Kirkendall voids until 192h at 250°C where Al diffuses into Ag bonded ball up to a depth of 5.3µm, consuming the complete Al bond pad and thus transforming into silver-aluminide containing 5 to 12wt% of Al and rest silver (Ag). Traces of coated gold (Au) and alloyed palladium (Pd) are noticed at the bond interface.

Resolution of Low IMC on Multi-Stacked Dice Device with Different Bond Pads

The continuing advances of die technology of integrated circuits (IC) miniaturization bringing more complexity in the product. At unusual conditions one may encounter new challenges intrinsic to the structure of the package. The study aims to qualify a product such as multi-stacked dice configurations with baseline reference using same die technology. The only difference is the substrate layout in which defined for the electrical purpose. The challenge is to understand and resolve low intermetallic coverage (IMC) on each die which may lead to manufacturability and reliability problems over time.

Elimination of Intermetallic Coverage Over-etching on Aluminum Pad

Intermetallic coverage (IMC) is one of the critical wirebond output responses that is usually checked to ensure the ball to pad integrity. The success of wirebonding relies on the formation of an interfacial intermetallic growth of ball bond to ensure it can withstand reliability stresses. The challenging approach in IMC analysis detect as over-etching around IMC area that leads to inaccurate IMC data collection. To address the over-etching, we generate a new method which is backside polishing that results to a reliable IMC data collection and help reduced the cycle time of IMC data gathering.

Bond Pad Probe Marks Effect on Intermetallic Coverage

The paper presents a study on the effect of semiconductor die’s bond pad probe marks on the interface between the wire and the bond pad. The probe marks are quantified in terms of percentage area in the bond pad, while the interface strength between the wire of Gold material to the Aluminum bond pad is measured through the intermetallic coverage (IMC). Actual evaluation showed that the size of the probe marks has significant impact on the bond pad area, especially on IMC. Validations were made comparing the IMC and shear strength performance of the wire ball bonded on the worst-case probe marks and on the standard probe marks. The learnings on this study could be used on future works with similar requirement.

Influence of Initial Shear Strength on Time-to-Failure of Copper (Cu) Wire Bonds in Thermal Aging Condition

Abstract Copper (Cu) wire bond is used in a majority of microelectronic devices but has not been fully accepted by all industries due to potential reliability issues. Good quality bond is believed to provide high reliability. Shear strength and intermetallic (IMC) coverage are being used as indicators, however there is no developed model. High shear strength is a result of large IMC coverage laterally under the ball, but the reliability of the bond is also related to IMC longitudinal growth (thickness). This work involves studying effect of shear strength on reliability of a variety of experimental Cu ball bonds by performing temperature aging experiments on test devices, in QFN packages. The ball bonds are made by altering ultrasonic power and time to obtain different IMC coverage. All test packages are monitored for resistance change at specific intervals by performing four point resistance measurement. Resistance increase is analyzed with initial shear strength to determine whether “good quality” bonds always lead to high reliability.

A Critical Review of Wirebond Visual Inspection Criteria

Abstract Wirebond engineers and technicians primarily rely on three means by which to evaluate the integrity and reliability of a wirebond interconnect. Destruct/non-destruct pull test, ball shear, cross-sections, intermetallic coverage under the ball are tests used to evaluate the mechanical strength as part of process development and in line process controls. Reliability testing and evaluation involves these same tests after prolonged environmental exposure such as, temp cycling, burn-in, humidity testing etc. according to standards such as AEC-Q100. Much has been written about the above. A third and very important consideration in wirebonding is visual inspection of the deformed wire after wirebond formation. The amount of wire squash-out, placement of the bond on the pad relative to nearby circuitry, looping profiles and heel integrity are all very important aspects of wirebond interconnect inspection. Poor workmanship can lead to noncompliant bonds per visual inspection criteria and can directly impact the reliability of the finished product. Bond pad inspection prior to bonding is also a critical aspect that is often overlooked. Historically, MIL-STD-883 has served as a baseline set of visual specs related to wirebond, and package assembly in general. Many company quality documents reference MIL-STD-883 TM 2017 and TM 2010 and/or cut and paste the bond inspection criteria into internal documents. This paper critically reviews the wirebond visual inspection criteria contained in MIL-STD-883 relative to the requirements of a fully optimized modern day wirebond process.

A Study on the Possibilities of Obtaining Intermetallic Coating of Al-Cu and Cu-Zn Systems Deposited on Metal Carrier

The research aimed at investigating the possibility of obtaining thin intermetallic films, which can be used as thermal control coatings for spacecraft. The coating films build the passive part of the overall thermal control system of a spacecraft, thus they must provide strong adhesion with the carrier material and have high functional characteristics, optical ones in particular. The study concerned issues of synthesis of stable intermetallic phase of Al4Cu9, Al2Cu, Cu5Zn8 on aluminum and copper carriers by magnetron layer-wise sputtering of reagents. They have looked into patterns of emergence of intermetallic coating applying modes of "rapid" and "slow" sputtering of reagents, applying various thicknesses of sputtered layers of reagents, temperature of the carrier, heat treatment of sprayed coatings. Incomplete and complete modes of synthesis of intermetallic coverage have been discovered. Obtained cross-sectional images of coatings, outcomes of microanalyzer scanning of reagent distribution through the thickness of sputtered coatings, microhardness values, optical absorption and emission ratios, unit of electric resistances, adhesion to the carrier. Obtained prototypes of intermetallic thermostatic coatings could be used in "solar reflectors" and "solar absorbers" classes. The results of measurements of optical and strength characteristics revealed that the intermetallic films can be used not only as thermostatic coatings for space technology, but also in general mechanical engineering, due to high-end mechanical properties.

Effects of Probe Marks on Shear Test of Copper Ball Bonds in Two Pad Aluminum Thicknesses

Abstract This paper continues the work of reference [1], evaluating shear test results of Cu ball bonds over a variety of probe marks in two different pad aluminum (Al) thicknesses (0.8μm and 3μm). The presence of invasive probe marks on thick Al bond pads lowers certain shear force results.. Lower values of shear force imply reduced Cu bond reliability. Physical factors are investigated relating to poor intermetallic (IMC) formation in the Cu wirebond and bond shear force. Optical microscopic image analysis of Cu bonds, bond contact areas and Al “splash” are studied for correlation with the shear test results. Percent IMC coverage of bond contact areas decrease when invasive probe marks are present beneath the bond, which in turn may reduce the shear force. Probe mark features are studied to discover the characteristics of greatest influence on % IMC coverage and shear test values in each of the pad metal thicknesses.

Fine Pitch Cu Wire Bonding Capability – Process Optimization and Reliability Study

Wire bonding has been around for over 70 years and it is still growing. The most recent advances in wire bonding are the wide proliferation of fine pitch Cu wire bonded applications and the development of fine pitch Cu capability extending to the newest technology nodes. Around 2008, fine pitch Cu wire bonding started to take off driven by the skyrocketing Au price. Since then, Cu wire bonding capability has improved dramatically so that today's advanced technology devices such as 28nm and 20nm nodes are being bonded with Cu wire including Pd coated and AuPd coated Cu wire. It turns out that not only is Cu wire cheaper, but it is more suitable for high I/O counts, and fine pitch advanced node applications due to its better electrical and mechanical properties. This paper examines the most advanced Cu wire bonding capability using a Cu optimized process called ProCuTM process. 40um pitch and 35um pitch capability are demonstrated using this new ProCu process. Process responses such as process bonding windows, intermetallic coverage (IMC), and free air ball size control are studied in detail. The ProCu process achieved fine pitch capability for 40um as a robust production process and it also shows that 35um pitch with 13um (0.5mil) wire is a possibility using the latest technology. The JEDEC Reliability test has been a challenge for Cu wire bonding especially for fine pitch applications. As part of this paper, we also examined reliability aspects of fine pitch Cu wire bonding. TEM analysis was used to understand the major factors that affect Cu wire bonding reliability.

Insulated Cu wire free air ball characterization

Insulated Cu wire technology has immense potential for fine pitch wire bonding interconnection. Understanding the behavior of the insulated Cu free air ball (FAB) formation is crucial for wire bonding process. The FAB formation, size, shape and cleanliness under different conditions for 20μm insulated Cu wire were investigated using SEM, FESEM and FTIR surface analysis. The results were compared with that of bare Cu wire. Consistently spherical residue free FAB of insulated Cu wire were formed using forming gas. The samples with insulated Cu wire consistently produced larger FAB than that of bare Cu wire, indicating that the energy required for free air ball formation is lower. Basic bonding performances in terms of ball bond strength, intermetallic (IMC) coverage growth and stitch bond strength of insulated Cu wire at time zero are also discussed in the paper.

Copper Wire Bonding: R&D to High Volume Manufacturing

During the past two years copper wire bonding has entered high volume manufacturing at a number of leading edge OSATs and IDMs. Usage of copper wire has achieved 20% market share and is expected to exceed 50% within three years. Products spanning the range from low pin count devices with relatively large wire diameter to FPGA's with nearly one thousand wires at 20 μm or even 18 μm wire are now using copper wire. This paper will discuss the requirements for developing a robust copper wire bonding process and then moving it to high volume manufacturing. Process optimization begins with the selection of the appropriate wire diameter, ball diameter, bonding tool and bonding process type. These are functions not only of the bond pad opening, but also of the pad aluminum thickness and relative sensitivity of the pad to damage. Proper optimization depends on the availability of new and modified bond quality metrologies, such as extensive reliance on cross-sectioning and intermetallic coverage measurements. The bonding window of a copper wire bonding process is defined in substantially new terms compared to optimization in gold wire bonding. Once an optimized process has been developed in the lab on a single bonder, it needs to be verified. Copper wire bond processes are much less forgiving with respect to the acceptable variability on the manufacturing floor. To ensure that the process is stable, a low volume pre-manufacturing test is highly recommended. This not only makes sure that the process is stable across multiple bonders, but also ensures the adequacies of manufacturing controls, incoming materials quality and sufficient equipment calibration and maintenance procedures.

Bond pad surface quality for reliable wire bonding

The impact of fluoride and AlO x contaminants on the quality of wire bonding (Au to Al welding) has been measured. No or minor impact of these contaminants has been found with a standard ball shear test. Despite an accelerated lifetime test, the shear strength remains above the required specification limits. Intermetallic coverage, the percentage of the interface between the bond ball and the bond pad where welding resulted in intermetallic diffusion, gives a better picture of the quality decay, but this method is too elaborate for production control and specification limits are unclear.

Measurement of gold ballbond intermetallic coverage

An important characteristic of gold ballbonds is intermetallic coverage (IMC), defined as the percentage of the total area on which gold aluminide compounds are formed by reaction between the gold ball and aluminium alloy metallization during thermosonic bonding. IMC measurement complements mechanical measurements such as ball pull and shear strength and in general, IMC should be maximised to aid in achieving a high degree of initial mechanical robustness so that devices can withstand subsequent processing steps and maintain high reliability. Too little IMC can result in weak bonds and erratic intermetallic growth patterns during subsequent thermal processing of bonded devices and may result in mechanical and electrical malfunction and failure. Optical microscopy of the underside of ballbonds is often used to determine IMC, but usually without consideration of certain key physical effects such as lighting, resolution and surface roughness. This paper compares the relative merits of IMC measurement by scanning electron microscopy (SEM) and optical microscopy and includes practical guidelines for the use of each technique.