Development of a thermosonic wire-bonding process for gold wire bonding to copper pads using argon shielding

To improve the bondability and ensure the reliability of Au/Cu ball bonds of the thermosonic (TS) wire-bonding process, an argon-shielding atmosphere was applied to prevent the copper pad from oxidizing. With argon shielding in the TS wire-bonding process, 100% gold wire attached on a copper pad can be achieved at the bonding temperature of 180°C and above. The ball-shear and wire-pull forces far exceed the minimum requirements specified in the related industrial codes. In a suitable range of bonding parameters, increasing bonding parameters resulted in greater bonding strength. However, if bonding parameters exceed the suitable range, the bonding strength is deteriorated. The reliability of the high-temperature storage (HTS) test for Au/Cu ball bonds was verified in this study. The bonding strength of Au/Cu ball bonds increases slightly with prolonged storage duration because of diffusion between the gold ball and copper pad during the HTS test. As a whole, argon shielding is a successful way to ensure the Au/Cu ball bond in the TS wire-bonding process applied for packaging of chips with copper interconnects.     

Mechanical reliability of Au and Cu wire bonds to Al, Ni/Au and Ni/Pd/Au capped Cu bond pads

This work is an assessment of the mechanical reliability of Au and Cu ball bonds to Al, Ni/Au and Ni/Pd/Au surfaces in terms of high temperature storage. All systems show very good shear strength after thermal storage for up to 120 days at 150 °C. The Au ball bonds on Al surface show Kirkendall voiding starting from 60 days. This did not decrease their mechanical strength but it is expected to become a reliability issue in the long run. The Cu wire bonds on Al caps show a higher initial strength, much slower intermetallics formation and no Kirkendall voiding. This makes them a potentially better industrial solution. Excellent bond strength was found for Cu- and Au-bonds on Ni/Au and Ni/Pd/Au caps. No intermetallics formation or other microstructural change have been found on these interfaces up to 120 days at 150 °C, which was related to the full solubility of the materials along these interfaces. This result suggests that they can be a successful industrial solution for the next generation of packages.     

In situ ultrasonic force signals during low-temperature thermosonic copper wire bonding

Ultrasonic in situ force signals from integrated piezo-resistive microsensors were used previously to describe the interfacial stick-slip motion as the most important mechanism in thermosonic Au wire ball bonding to Al pads. The same experimental method is applied here with a hard and a soft Cu wire type. The signals are compared with those obtained from ball bonds with standard Au wire. Prior to carrying out the microsensor measurements, the bonding processes are optimized to obtain consistent bonded ball diameters of 60 μm yielding average shear strengths of at least 110 MPa at a process temperature of 110 °C. The results of the process optimization show that the shear strength c_pk values of Cu ball bonds are almost twice as large as that of the Au ball bonds. The in situ ultrasonic force during Cu ball bonding process is found to be about 30% higher than that measured during the Au ball bonding process. The analysis of the microsensor signal harmonics leads to the conclusion that the stick-slip frictional behavior is significantly less pronounced in the Cu ball bonding process. The bond growth with Cu is approximately 2.5 times faster than with Au. Ball bonds made with the softer Cu wire show higher shear strengths while experiencing about 5% lower ultrasonic force than those made with the harder Cu wire.     

Silver pick up during tail formation in thermosonic wire bonding process

Cu ball bonding processes are significantly less robust than Au ball bonding processes. One reason for this are higher variations observed, e.g. in the free-air ball (FAB) formation process. There is a strong influence the tail bond process has on subsequent FAB formation. Tail tips and bond fractographs made with Cu and Au wires are investigated using scanning electron microscopy. Au and Cu wires pick up Ag material (Ag pick up) from the Ag metallization of Cu leadframe diepads during wire tail breaking. Ag pick up by Cu wire is more dominant than that by Au wire. Ultrasonic friction energy is necessary in order for Ag material pick up to occur. The impact force plays an important role for Ag pick up; less pick up is observed with a higher impact force. The hardness of the free-air ball (FAB) with Ag pick up is reduced by up to 4% compared to that of a FAB made from a tail broken at the neck of a ball bond and therefore having no Ag pick up. No significant change of FAB diameter is observed in these two cases.     

Effect of plasma treatment of Au-Ni-Cu bond pads on process windows of Au wire bonding

The wire bondability of Au-Ni-Cu bond pads with different Au plating schemes, including electrolytic and immersion plates, are evaluated after plasma treatment. The plasma cleaning conditions, such as cleaning power and time, are optimized based on the process window and wire pull strength measurements for different bond pad temperatures. Difference in the efficiency of plasma treatment in improving the wire bondability for different Au plates is identified. The plasma-cleaned bond pads are exposed to air to evaluate the recontamination process and the corresponding degradation of wire pull strength. The changes in bond pad surface characteristics, such as surface free energy and polar functionality, with exposure time are correlated to the wire pull strength, which in turn provides practical information about the shelf life of wire bonding after plasma cleaning.     

Microstructural study of copper free air balls in thermosonic wire bonding

Copper wires are increasingly used in place of gold wires for making bonded interconnections in microelectronics. In this paper, a microstructural study is reported of cross-sectioned free air balls (FABs) made with 23 μm diameter copper bonding wire. It was found that the FAB is comprised of a few columnar grains and a large number of fine subgrains formed within the columnar grains around the periphery of the FAB. It was determined that conduction through the wire was the dominant heat loss mechanism during cooling, and the solidification process started from the wire-ball interface and proceeded across the diameter then outward towards the ball periphery.The microstructure of the Cu ball bond after thermosonic bonding was investigated. The result showed that the subgrain orientations were changed in the bonding process. It is evident that metal flow along the bonding interface was from the central area to the bond periphery during thermosonic bonding.     

Thermosonic bonding of gold wire onto silver bonding layer on the bond pads of chips with copper interconnects

A copper pad oxidizes easily at elevated temperatures during thermosonic wire bonding for chips with copper interconnects. The bondability and bonding strength of a gold wire onto a bare copper pad are seriously degraded by the formation of a copper oxide film. A new bonding approach is proposed to overcome this intrinsic drawback of the copper pad. A silver layer is deposited as a bonding layer on the surface of copper pads. Both the ball-shear force and the wire-pull force of a gold wire bonded onto copper pads with silver bonding layers far exceed the minimum values stated in the JEDEC standard and MIL specifications. The silver bonding layer improves bonding between the gold ball and copper pads. The reliability of gold ball bonds on a bond pad is verified in a high-temperature storage (HTS) test. The bonding strength increases with the storage time and far exceeds that required by the relevant industrial codes. The superior bondability and high strength after the HTS test were interpreted with reference to the results of electron probe x-ray microanalyzer (EPMA) analysis. This use of a silver bonding layer may make the fabrication of copper chips simpler than by other protective schemes.     

Effect of gas type and flow rate on Cu free air ball formation in thermosonic wire bonding

The development of novel Cu wires for thermosonic wire bonding is time consuming and the effects of shielding gas on the electrical flame off (EFO) process is not fully understood. An online method is used in this study for characterizing Cu free air balls (FABs) formed with different shielding gas types and flow rates. The ball heights before (H_FAB) and after deformation (H_def) are responses of the online method and measured as functions of gas flow rate. Sudden changes in the slopes of these functions, a non-parallelity of the two functions, and a large standard deviation of the H_FAB measurements all identify FAB defects. Using scanning electron microscope (SEM) images in parallel with the online measurements golf-club shaped and pointed shaped FABs are found and the conditions at which they occur are identified. In general FAB defects are thought to be caused by changes in surface tension of the molten metal during EFO due to inhomogeneous cooling or oxidation. It is found that the convective cooling effect of the shielding gas increases with flow rate up to 0.65 l/min where the bulk temperature of a thermocouple at the EFO site decreases by 19 °C. Flow rates above 0.7 l/min yield an undesirable EFO process due to an increase in oxidation which can be explained by a change in flow from laminar to turbulent. The addition of H₂ to the shielding gas reduces the oxidation of the FAB as well as providing additional thermal energy during EFO. Different Cu wire materials yield different results where some perform better than others.     

Modeling of copper wire bonding process on high power LEDs

In this paper, a couple thermal mechanical transient dynamic finite element framework of copper wire bonding process on high power lighting emitting diodes (LEDs) is developed, which considers the thermal heating effects of friction and plastic deformation. The whole wire bonding process is simplified to consist of impact and ultrasonic vibration stages. Parametric studies are also carried out to examine the effects of ultrasonic vibration amplitude and bonding force on stress/strain distribution and friction thermal heating effect during wire bonding process. Different friction coefficients of interface between the free air ball (FAB) and the bond pad are taken in the simulation to examine the effects of friction on the stress and strain level of electrode structure. Modeling results show that the stress/strain distribution and temperature evolution of wire bonding system are significant influenced by the ultrasonic vibration amplitudes, bonding forces and friction coefficients. Discussion and comparison are conducted between the copper and the gold wire bonding processes on the high power LEDs by numerical simulation. The results have disclosed that higher stress/strain in the bond pad and the ohmic contact layer is induced during the copper wire bonding process. Therefore, the process parameters of copper wire bonding should be controlled carefully. This numerical simulation work may provide guidelines for the copper wire bonding process virtual window development of high power LEDs packaging.     

The evaluation of copper migration during the die attach curing and second wire bonding process

Copper migration on the silver plated surface of the lead-frames with various heat treatments was evaluated by X-ray photoelectron spectroscopy (XPS), transmission electron microscopy (TEM) and atomic force microscopy (AFM) methodologies. Copper migration may introduce copper oxidation and result in the wedge bonding failures due to the nonstick on lead (NSOL) phenomenon. Experiments were performed on the two kinds of TQFP leadframes with the stamped and etched manufacturing processes subjected to various heat treatments and bonding conditions to understand the underlying physics. TEM and AFM provided the additional insight of the grain structure and surface roughness of silver. XPS results showed that the etched leadframe was relatively better one that less copper oxide was detected on the silver surface after annealing process. However, more copper was observed to diffuse onto the silver surface after annealing in the stamped leadframe. In comparison between the stamped and etched lead-frames, the silver plated layer in later one is more efficient in blocking copper diffusion-either surface or bulk. Finally a full factorial design of experiment (DOE) with wedge bond pull strength as response was performed to verify the results of XPS, TEM, and AFM. The evaluations based on XPS, TEM, and AFM analyzes can really help to improve the yield of the wedge bonding process and optimize the IC manufacturing process window.     

Study of factors affecting the hardness of ball bonds in copper wire bonding

Copper wire bonding has gained popularity due to its economic advantage and superior electrical performance. However, copper is harder than gold, and replacing gold wire with copper wire introduces hardness related issues. This article reports investigations of the properties including microhardness of the copper balls bonded using ?25.4-μm copper wire and different combinations of electronic-flame-off (EFO) current and firing time settings with forming gas (5%H₂ and 95%N₂) as the inert cover gas. FABs with an identical diameter, obtained under different EFO firing conditions, were ball bonded with the same wire bonding parameters established using design of experiments. Microhardness tests were then performed on the cross-section of the bonded balls. The study revealed that ultrasonic generator current is the most significant factor to increase the bonded mashed ball diameter, ball shear and shear per unit area and to decrease the ball height. The microhardness of bonded copper balls is related to the EFO parameters, with FABs obtained by higher EFO current being softer. The lower hardness is attributed to the higher maximum temperature during the FAB melting state.     

Interfacial evolution and bond reliability in thermosonic Pd coated Cu wire bonding on aluminum metallization: Effect of palladium distribution

In this paper, the growth kinetics of Cu–Al intermetallic compounds formed during isothermal annealing of Pd–Cu wire bonds with different palladium distribution at 175 °C are investigated by electron microscopy and compared to bare Cu wire bonds. Transmission electron microscopy (TEM) was used to provide high resolution imaging of the Cu–Al IMCs in the as-bonded state and TEM-EDX used to analyze the concentrations of Pd at the bond interface in the as-bonded state. Cu–Al IMCs were found to grow thicker with increasing annealing duration. The growth kinetics of the Cu–Al IMCs were correlated with the diffusion process during thermal annealing. The IMC thickness for Pd–Cu wire bonds with Pd at the bond interface was found to be thinner as compared to that for Pd–Cu wire bonds with no Pd at the bond interface. Thus, the presence of palladium at the bond interface has slowed down the IMC growth. Nano-voids were found in the Pd–Cu wire bonds with Pd at the bond interface, but not in the Pd–Cu wire bonds with no Pd at the bond interface. The IMC growth rate for the Pd–Cu bonds with no Pd was found to be close to that for bare Cu for the initial annealing durations. Corresponding bond pull testing showed that Pd–Cu wire bonds containing Pd have best preserved the bond strength after 168 h aging at 175 °C due to the beneficial presence of Pd.     

Effects of Cu and Pd addition on Au bonding wire/Al pad interfacial reactions and bond reliability

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.     

Evaluation of wire bonding performance, process conditions, and metallurgical integrity of chip on board wire bonds

Chip on board wire bonding presents challenges to modern wire bonding technology which include smaller, closely spaced wire bond pads; bonding to soft substrates without special processing and pad construction; and diverse first bond and second bond metallurgies. These challenges are addressed by extensive bonding accuracy tests, a design of experiments approach for optimizing wire bond process parameters, reliability testing, and detailed materials characterization of the metallurgical integrity of the wire bonds. The thermo-mechanical integrity of the wire bond interconnects was evaluated by wire pull and hot storage tests. Hot storage testing allowed for detection of samples with an electrolytic gold surface finish that was too thin, and exhibited a contamination-corrosion condition of the nickel under-plating. Other samples with an excessively thick, rough textured nickel under-plating layer exhibited poor wire bond-ability. The methodology of materials analyses of the metallurgy of the wire bond interconnects is described. The paper illustrates a wire bond lift technique that is used to inspect for cratering damage and the “area-uniformity” of gold aluminum intermetallics. An improved understanding of the wire bonding process was achieved by showing the dependence of the visual appearance of the wire bonds on wire bond process parameters.     

Optimizing the fine-pitch copper wire bonding process with multiple quality characteristics using a grey-fuzzy Taguchi method

This paper presents an application of the grey-fuzzy Taguchi method to derive the optimum parameters for the fine-pitch copper (Cu) wire bonding process with multiple quality characteristics. Cu wire has become an important replacement material without the high manufacturing cost of gold for wire bonding processes in integrated circuit (IC) packaging. However, it sometimes fails to take advantage of the Cu material properties due to an insufficient understanding of the relationship between process inputs and outputs of the fine-pitch Cu wire bonding. To assure the accomplishment of cost savings without losing yield, the proposed grey-fuzzy Taguchi method is utilized with an L18 (2¹ × 3⁷) experimental design and grey relational analysis (GRA) to evaluate the degree of relationship between process inputs and responses, followed by transforming the multiple quality characteristics into a single performance index using a fuzzy inference system. Finally, the process parameters are optimized by the Taguchi method. The proposed optimization methodology provides superior optimization performance compared to the grey-Taguchi method and the parameter setting used in mass production in terms of the conformation experiments.     

The impact of copper contamination on the quality of the second wire bonding process using X-ray photoelectron spectroscopy method

A lot of wedge bonding failures were observed on the leadframe (LF) type A due to non-stick on lead (NSOL) during the second wire bonding process of TQFP package. The copper ion contamination from the plating process was identified as one of the key factors that attributed to the NSOL failures. Surface analyses were performed in terms of X-ray photoelectron spectroscopy on the surface of LF type A. It was found that as received silver-plated surface of the copper LF type A was contaminated by the copper. After the copper contamination was solved in the plating process, the design of experiment was implemented for the verification of the influence of copper contamination on the quality of the second bonding process. It was confirmed that copper contamination dramatically reduced the strength of wedge bonding. The wedge pull test showed that NSOL failures were not observed after the copper ion contamination in the plating process was well controlled.     

Surface oxide films on Aluminum bondpads: Influence on thermosonic wire bonding behavior and hardness

Using indentation testing, wire bond tests and electron microscopy, the influence of increased oxide films on Al metallization surfaces on the wire bonding behavior and hardness was investigated. Oxide film thickness values larger than about 20 nm obstruct the bond contact and resulted in a poor bonding quality. The presence of such films causes also a hardness increase which can be detected by current sensitive indentation test methods. Therefore, an improved indentation testing technique can be applied during the quality control of bondpad metallizations prior to wire bonding.     

Increasing the bonding strength of chips on flex substrates using thermosonic flip-chip bonding process with nonconductive paste

Thermosonic flip-chip bonding process with a nonconductive paste (NCP) was employed to improve the processability and bonding strength of the flip-chip onto flex substrates (FCOF). A non-conductive paste was deposited on the surface of the copper electrodes over the flex substrate, and a chip with eight gold bumps bonded onto the copper electrodes by the thermosonic flip-chip bonding process.For the chips and flex substrates assembly, ultrasonic power is important in the removal of some of the non-conductive paste on the surface of copper electrodes during thermosonic bonding. Accordingly, gold stud bumps in this study were directly bonded onto copper electrodes to form successful electrical paths between chips and the flex substrate. A particular ultrasonic power resulted in some metallurgical bonding between the gold bumps and the copper electrodes, increasing the bonding strength. The ultrasonic power was not only to remove the NCP from the copper electrodes, but also formed metallurgical bonds during the thermosonic flip-chip bonding process with NCP.In this study, the parameters of the bonding of chips onto flex substrates using thermosonic flip-chip bonding process with NCP were a bonding force of 4.9 N, a curing time of 40 s, a curing temperature of 140 °C and an ultrasonic power of 14.46 W. The processability and bonding strength of flip-chips on flex substrates using thermosonic bonding process with NCP was verified in this study. This process has great potential to be applied to the packaging of consumed electronic products.     

Iterative optimization of tail breaking force of 1 mil wire thermosonic ball bonding processes and the influence of plasma cleaning

An online tail breaking force measurement method is developed with a proximity sensor between wire clamp and horn. The wire under the tensile load measures about 1.5 cm extending from the bond location to the wire clamp. To increase the sensitivity, the bondhead speed is reduced to 2 mm/s during breaking the tail bond. It takes roughly 10 ms to break the tail bond. The force resolution of the method is estimated to be better than 5.2 mN. An automatic wire bonder used to continuously bond up to 80-wire loops while recording the on-line proximity signals. All wires are directed perpendicular to the ultrasound direction. The tail breaking force for each bond is evaluated from the signal and shown automatically on the bonder within 2 min after bonding.Results are obtained for a typical Au wire and a typical Cu wire bonding process. Both wires are 25 mm in diameter and bonded on Ag plated diepads of standard leadframes at 220 °C. An average Cu tail breaking force of higher than 50 mN is obtained if the leadframe is plasma cleaned before the bonding with 100% Ar for 5 min. This result is comparable to that obtained with Au wire. The standard deviation of the Cu tail breaking force is about twice that obtained with Au wire. The tail breaking force depends on the bonding parameters, metallization variation, and cleanliness of the bond pad. The cleanliness of the bonding pad is more important with Cu wire than with Au wire.