Assessing Au-Al wire bond reliability using integrated stress sensors (original) (raw)
Related papers
Sensors and Actuators A: Physical, 2008
Two real-time, in situ methods to measure the breaking force of fine bonding wires while on the wire bonder are reported and compared. The first method uses a special test chip with a piezoresistive microsensor integrated next to the bonding pad. A 25 m diameter Au wire piece is attached with a ball bond to the test pad of the microsensor. The wire piece between the ball bond and the lower edge of the wire clamps is 15 mm in length. The clamps tear the wire at a speed of 2 mm/s. The wire breaks at the heat-affected zone (HAZ) next to the ball bond. The microsensor is calibrated using FE models. The numerical results show that the microsensor signal is highly sensitive to ball and pad geometry, values of the piezoresistive coefficients, and the z-location of the microsensor under the bonding pad. This results in a high estimated error of about 46% for the calibration factor of the microsensor.
IEEE Transactions on Components, Packaging and Manufacturing Technology, 2013
ABSTRACT A ball bonding process in wire bonding generally involves impact followed by ultrasonic (US) bonding prior to wedge bonding. During the ball bonding process, the impact force flattening the free-air ball introduces significant localized out-of-plane compressive stress on the pad and the low-k structure beneath. The subsequent process of US bonding induces in-plane and shear stresses to the structure. High induced stress during bonding is not desirable, as it may lead to pad damage or cratering of the silicon structure. In this paper, we report on studies conducted on using four piezoresistive sensors embedded underneath the center of the bond pad for the evaluation of in-plane and out-of-plane stresses, which covers both the impact and US stages during the ball bonding process. Different levels of impact force, bond force, bonding duration, and US power are investigated using gold wire bonding for feasibility and sensitivity studies of the stress sensors. Fast Fourier transform (FFT) and inverse FFT are used for noise filtering and to isolate the US signal yielding a continuous output signal from the in situ measurement of contact and US stages during the ball bonding process. It is found that the stress sensors are sensible to capture different impact force, bond force, bonding duration, and US power.
Measuring stress next to Au ball bond during high temperature aging
Microelectronics Reliability, 2009
A real-time signal of the stress caused by a ball bond is recorded during long-term high temperature storage (HTS) without destroying the ball bond, using a piezoresistive integrated CMOS microsensor located next to the Al bond pad (test pad) on a test chip. The sensor is sensitive to in-plane shear stress changes Ds xy that arise due to tensile or compressive stress at the test pad. While performing HTS at 200°C during 400 h, significantly different stress signals are observed with a ball bond (test structure) compared to those observed without a ball bond (reference structure). Simultaneous to Ds xy the contact resistance of the bond was directly measured with a four-wire method in which two connection paths lead to the test pad and a second wire bond is made on top of the test ball bond. The contact resistance values measured at room temperature (25°C) before and after HTS are 2.1 mX and 6.1 mX, respectively. Effects influencing the stress signal during HTS include volume changes by the growth of intermetallics. The stress increase initially observed during HTS shows bond shrinking corresponding to growth of Au-rich phases which was previously reported to result in volume shrinkage. A subsequent phase of signal drop is observed starting after 200 h, indicating the presence of a different mechanism partly reducing the stress built up previously, and attributed to lateral growth of Al-rich intermetallics, partially consuming the pad Al outside the bond region, and resulting in volume expansion. Finite element models are developed to support the interpretation of the stress signal features. One of the models simulates the shrinking of Au-Al material due to phase transformation. When calibrated to experimental data, the peak underpad Tresca stress level generated during such contraction is 53 MPa, located 2.4 lm inside of the 55 lm diameter bond zone.
IEEE Transactions on Components, Packaging and Manufacturing Technology, 2016
Gold wire bonding has been widely used as the first-level interconnect in semiconductor packaging. The increase in the gold price has motivated the industry search for an alternative to the gold wire used in wire bonding and the transition to a copper wire bonding technology. Potential advantages of transition to a Cu-Al wire bond system include low cost of copper wire, lower thermal resistivity, lower electrical resistivity, higher deformation strength, damage during ultrasonic squeeze, and stability compared with gold wire. However, the transition to the copper wire brings along some tradeoffs, including poor corrosion resistance, narrow process window, higher hardness, and potential for cratering. Formation of excessive Cu-Al intermetallics may increase the electrical resistance and reduce the mechanical bonding strength. Current state of the art for studying the Cu-Al system focuses on the accumulation of statistically significant number of failures under accelerated testing. In this paper, a new approach has been developed to identify the occurrence of impending apparently random defect fall-outs and premature failures observed in the Cu-Al wire bond system. The use of intermetallic thickness, composition, and corrosion as a leading indicator of failure for the assessment of the remaining useful life for Cu-Al wire bond interconnects has been studied under exposure to high temperature. Damage in the wire bonds has been studied using an X-ray micro-Computed Tomography (CT). Microstructure evolution was studied under the isothermal aging conditions of 150°C, 175°C, and 200°C until failure. Activation energy was calculated using the growth rate of intermetallic at different temperatures. An effect of temperature and humidity on a Cu-Al wire bond system was studied using the Parr bomb technique at different elevated temperature and humidity conditions (110°C/100%RH, 120°C/100%RH, and 130°C/100%RH), and a failure mechanism was developed. The present methodology uses the evolution of the intermetallic compound thickness and composition in conjunction with the Levenberg-Marquardt algorithm to identify accrued damage in wire bond subjected to thermal aging. The proposed method Manuscript
TEM microstructural analysis of As-Bonded Al–Au wire-bonds
Journal of Materials Science, 2007
In this study the interface morphology of a model 99.999% (5N) Au wire bonded to Al pads in the as-bonded state was examined by scanning/transmission electron microscopy with energy dispersive spectroscopy. Specimens for transmission electron microscopy were prepared using the lift-out method in a dual-beam focused ion beam system. Analysis of the bond microstructure was conducted as a function of the Al pad content and as a function of the bonding temperature. Additions of Si and Cu to the Al pad affect the morphology and the uniformity of the interface. A characteristic-void line is formed between two intermetallic regions with different morphologies in the as-bonded samples. According to the morphological analysis it was concluded that a liquid phase forms during the bonding stage, and the void-line formed in the intermetallic region is the result of shrinkage upon solidification and not the Kirkendall effect.
Influence of Intermetallic Phases on Reliability in Thermosonic Au-Al Wire Bonding
2006 1st Electronic Systemintegration Technology Conference, 2006
Since roughly 2002, reliability problems occur often at the ball bonds after wire bonding or reliability testing procedures. Thefailure mainly appears as ball lift-offor reducedpullforce values. Applying Focused Ion Beam (FIB) techniques as well as Scanning and Transmission Electron Microscopy (SEM, TEM) alternative failure modes in addition to Kirkendall voids werefound that could be divided into five morphological categories. Thesefailures include the missing ofany Au-Al intermetallic compound (IMC) formation at the pads after wire bonding, no uniform horizontallvertical IMCformation after wire bonding, micro cracks between the interface IMC and Au ball during reliability stressing, micro cracks between the interface ofthe two IMCs during temperature aging andfinally corrosion ofthe interface IMC/Au ball during temperature aging or humidity testing. Although there is strong indication that part ofthesefailure modes are related to front endprocess steps, the practical limitations resultingfrom separation between semiconductor manufacturing and assembly line require a back end solution. It is shown that the use ofan improved high security (HS) or high reliability wire (HR) contributes significantly to reliability improvement. The wires contain special dopands which reduce the IMC and defect growth. The influence of these wires on the reliability ofbonds will be explained and shown in selected examples.
2004
This paper reports the design, fabrication, and characterization of a CMOS based two-dimensional stress sensor array. It is based on a stress sensor element exploiting the transverse pseudo-Hall effect in metal oxide semiconductor (MOS) field effect transistor (FET). In this work p-doped MOS devices (PMOS) were integrated in a 4 x 4 stress sensor array with a total area of only 120 x 120 pm2. The array is connected onchip to an analog multiplexer. The new device was used for the in-situ monitoring of a hall wedge wire bonding process. It gives access to position and force information. The piezo-FET sensor m a y was used to spatially resolve the stress distribution underneath and close to a bondpad. A sensitivity of the m a y to position variations of at least 5 p m is demonstrated. Compared to previous stress sensor arrays the sensor density was increased by a factor of 22 and the number of required hondpads was reduced by a factor of 3.5.
The Fracture Resistance of Bonded Au Wires for Interconnection
IEEE Transactions on Components and Packaging Technologies, 2009
This paper presents the effects of the local strength degradation due to recrystallization and grain growth and the large deformation induced in wire bonding process on the fracture resistance of the bonded Au wires for interconnection. The experimental and numerical results show that, for a wire to deform to the required shape without reduction in the fracture resistance, sufficient ductility is more important than the strength of the wire.
In the copper ball bond process, aluminum squeezes out under the ball deformation process since it is softer than copper. The final thickness of the copper ball decides the overall strength of the bond and hence knowledge of its final thickness helps to optimize the ball bond force during the bonding process. An etching method is developed for the analysis of the squeezed aluminum beneath the copper wire ball bond. The aluminum remnant obtained from the etching method is compared with that from the conventional cross-section method. It was found that the results are comparable and the present developed etched method is simple and elegant.
Effects of Cu and Pd addition on Au bonding wire/Al pad interfacial reactions and bond reliability
Journal of Electronic Materials, 2006
Finer pitch wire bonding technology has been needed since chips have more and finer pitch I/Os. However, finer Au wires are more prone to Au-Al bond reliability and wire sweeping problems when molded with epoxy molding compound. One of the solutions for solving these problems is to add special alloying elements to Au bonding wires. In this study, Cu and Pd were added to Au bonding wire as alloying elements. These alloyed Au bonding wires-Au-1 wt.% Cu wire and Au-1 wt.% Pd wire-were bonded on Al pads and then subsequently aged at 175°C and 200°C. Cu and Pd additions to Au bonding wire slowed down interfacial reactions and crack formation due to the formation of a Cu-rich layer and a Pd-rich layer at the interface. Wire pull testing (WPT) after thermal aging showed that Cu and Pd addition enhanced bond reliability, and Cu was more effective for improving bond reliability than Pd. In addition, comparison between the results of observation of interfacial reactions and WPT proved that crack formation was an important factor to evaluate bond reliability.