Ultrasonic Wire Bonding Research Articles

Machine learning-based shear force quality prediction of ultrasonic wire bonds: utilizing process data and machine data without additional sensors

Ultrasonic wire bonding is a highly automated production process that is used billions of times a year in the electronics and electromobility industries. Due to the complexity of the process and the large number of influencing parameters, there are currently no automated methods that can be used without additional sensors to evaluate the shear force bond quality quantitatively and non-destructively with sufficiently high precision. For this reason, this paper presents a new methodology that uses machine learning to enable quantitative, non-destructive prediction of bond quality without additional sensors. For this purpose, a machine learning algorithm was developed that uses various machine data and process data from existing sensors to quantitatively predict the shear force of the bonded joint. In addition, features are extracted from process time series, such as current, power, and frequency of the ultrasonic generator as well as deformation during bonding. Of the total of 2,090 features considered, the number of features could be reduced to 26 by recursive feature elimination, while maintaining the prediction accuracy. By using optimized deep neural networks, on average, a prediction precision of the regression on the shear force of the source bond of over 89.6% R2-score and a mean absolute error of 241 cN can be achieved.

A Miniature Ultrasound Source for Neural Modulation.

This work describes a unique ultrasound (US) exposure system designed to create very localized ( [Formula: see text]) sound fields at operating frequencies that are currently being used for preclinical US neuromodulation. This system can expose small clusters of neuronal tissue, such as cell cultures or intact brain structures in target animal models, opening up opportunities to examine possible mechanisms of action. We modified a dental descaler and drove it at a resonance frequency of 96 kHz, well above its nominal operating point of 28 kHz. A ceramic microtip from an ultrasonic wire bonder was attached to the end of the applicator, creating a 100- [Formula: see text] point source. The device was calibrated with a polyvinylidene difluoride (PVDF) membrane hydrophone, in a novel, air-backed, configuration. The experimental results were confirmed by simulation using a monopole model. The results show a consistent decaying sound field from the tip, well-suited to neural stimulation. The system was tested on an existing neurological model, Drosophila melanogaster, which has not previously been used for US neuromodulation experiments. The results show brain-directed US stimulation induces or suppresses motor actions, demonstrated through synchronized tracking of fly limb movements. These results provide the basis for ongoing and future studies of US interaction with neuronal tissue, both at the level of single neurons and intact organisms.

Toward Ultrasonic Wire Bonding for High Power Device: A Vector Based Resonant Frequency Tracking and Constant Amplitude Control

The electrical packaging for high power devices such as IGBT or MOSFET is achieved by heavy aluminum ultrasonic wire bonding. In this process, the amplitude stability of ultrasonic transducer determines the bonding quality, which demands more accuracy in resonant frequency tracking (RFT) and constant amplitude control (CAC). In this investigation, we proposed a vector based resonant frequency (VRFT) tracking and vector based constant amplitude control (VCAC) to control the ultrasonic transducer. The current of dynamic branch was conducted by calculating the voltage and current vector of ultrasonic transducer. Small signal model of ultrasonic transducer was established and applied into RFT and CAC. A control strategy of fuzzy self-tuning PID closed loop was utilized to RFT and CAC. The PID parameters was calculated by the linear control theories and modified by the fuzzy controller which improved the vibration stability of the transducer. A prototype of 100Watt/60kHz ultrasonic driver was self-developed to verify the VRFT and VCAC. It demonstrated that the locked frequency of VRFT can be controlled within less than 0.04 millisecond when staring from the swept frequency, and the vibration response is fast with the varying bonding pressure. The amplitude of fuzzy PID VCAC is more stable and robust compared to the classical PID control, which improved the bonded joint quality in heavy aluminum wire packaging. Note to Practitioners—This article was motivated by the reliability and efficiency of IGBT/ MOSFET device packaging through ultrasonic aluminum wire bonding. This bonding process requires more amplitude stability of the ultrasonic transducer to achieve acceptable bonded quality. Currently, the resonant frequency tracking (RFT) by classical PID control is in slow speed, and the amplitude of the transducer cannot be controlled constantly due to the inherent static capacitor of the transducer. These issues for the ultrasonic control are originated from the engineer’s experience or simple heuristic rule instead of advanced approaches. To improve the vibration performance, we proposed a vector based model consisting of the relationship the dynamic branch current and the total current of the transducer, to control the RFT and constant amplitude control (CAC). Moreover, the vector based is utilized into the fuzzy PID control, where the calculation is reduced by area coefficient defuzzy. The verification experiments prove that the response time of the proposed vector based RFT and CAC, and the bonded joint quality is optimized obviously. It is noted that in our control strategy, the RMS value, the phase of voltage and current of the transducer are acquired by DSP processor, and the calculation of fuzzy PID controller is not complicated, which is feasible to implement in application. This significantly benefits the demand of accurate control to the ultrasonic power in electronic packaging.

Finite element modeling and analysis of ultrasonic bonding process of thick aluminum wires for power electronic packaging

Ultrasonic wedge bonding of aluminum (Al) wires is a widely applied interconnect technology for power electronic packaging. The joint quality of the wedge bonding is mainly affected by the process parameters and material properties. Inappropriate process parameters will lead to failure modes such as chip surface pit, metal layer peeling off, wire cracking, non-sticking to the pad, etc., which limits the long-term stability of power devices. In order to reach the desired reliability, the design of experiment (DoE) is generally deployed which is costly in terms of time and related materials. Therefore, simulation-assisted analysis is in demand to rapidly narrow down the process windows. In this paper, an ultrasonic bonding model involving thick Al wires (300 μm) was established based on the Finite Element Method (FEM), to optimize process parameters effectively with reduced time and cost. The model was designed in ANSYS utilizing the transient structural mechanics module with various stresses and ultrasonic power, to simulate the relative deformation of the bonded wires and the displacement against the substrate. The result was then verified by ultrasonic wedge bonding samples with 9 sets of process parameters. The stress distributions were simulated and analyzed with the failure modes of tensile strength tests, while the deformation of wires under various process parameters was measured and compared with shear strength tests. Further, the relationship between the failure modes of the joint and the deformation was then analyzed by Response Surface Method (RSM), and the regression equation of the wire deformation and related process parameters was established and fitted with the actual sample's data. Such analysis not only found the optimum range of the deformation of thick Al ultrasonic wire bonds but also quickly provided a range of optimized processes for Al thick wires applying ultrasonic wedge bonding techniques.

Cell Interconnections in Battery Packs Using Laser-assisted Ultrasonic Wire Bonding

Abstract This paper presents the results of a series of bonding tests using a laser-assisted ultrasonic wire bonding process. Aluminium and copper wire, both 500 μm (20 mil) thick, were bonded to nickel-coated steel caps of type 21700 battery cells. Mechanical bond strength tests prove that laser-assisted wire bonding has significant advantages over room temperature wire bonding. For example, it can be used to reduce the process time with aluminium wire or to increase the bondability of copper wire on nickel-coated steel. The results show a direct relation between tool tip temperature and measured bond strength. The quality of the joints was effectively improved by heating the tool tip up to 430 °C. These advantages are the same as in classic thermosonic wire bonding, but without the major disadvantage of having to heat to whole package. The cell temperature was shown to remain safely below the critical 60 °C in any application.

A Novel Approach to Control of Piezo-Transducer in Microelectronics Packaging: PSO-PID and Editing Trajectory Optimization

This article presents a novel optimization algorithm to control the ultrasonic vibration of piezo-transducer in microelectronics packaging. The control of resonant frequency tracking to piezo-transducer is significant to the bonding quality. A conventional PID controller is applied to track the resonant frequency of piezo-transducer dynamically; however, there are shortcomings, such as tedious parameter tuning and uncertain optimal parameters. This article proposes the particle swarm optimization (PSO) to search the PID parameters intelligently. The MATLAB-Simulink model of PSO-PID is established to analyze the phase and frequency response of the piezo-transducer by a closed loop, where the optimal control parameters are determined by the PSO search. To solve the conventional abrupt ultrasonic power, an editable voltage trajectory with analytical geometry is proposed to provide various driving voltage trajectories. Finally, the control experiments with the proposed optimization are carried out in ultrasonic wire bonder. The results validate that the PSO-PID and editing trajectory optimizations can attain competitive bonding quality, stability, and robustness.

On-chip sensor solution for hydrogen gas detection with the anodic niobium-oxide nanorod arrays

Two types of anodic niobium-oxide nanofilms were synthesized via anodization of an Al/Nb bilayer sputter-deposited onto a SiO2-coated Si wafer. Type I nanofilm was composed of a 200 nm thick NbO2 layer holding the upright-standing 650 nm long, 50 nm wide, and 70 nm spaced Nb2O5 nanorods, of 7·109 cm−2 density, whereas the Type II nanofilm had similarly long but bigger Nb2O5 nanorods, 100 nm wide, 220 nm spaced, and of 8·108 cm−2 density, aligned directly on a niobium metal without any buffering oxide layer, which was achieved for the first time. Each film was then incorporated in an advanced 3-D architecture and multilayer layout on a silicon chip comprising 33 microsensors, with variable sizes and tuned electrical characteristics, by combining the high-temperature vacuum or air annealing, sputter-deposition and lift-off photolithography to form Pt/NiCr top electrodes and a multifunctional SiO2 interlayer, chemical etching, laser dicing, and ultrasonic wire-bonding. The proposed on-chip sensor solution allowed for a sensitive, fast, and highly selective (toward NH3 and CH4) detection of hydrogen gas. Comprehensive gas sensing tests performed for Type II nanofilm ultimately confirmed the presence of a Schottky-type sensing mechanism, the contribution, however, being substantially weaker than that due to reactions over the surface of the oxide nanorods, especially when the rods show a transition from fully to partially depleted states when interacting with H2 gas. The film formation and chip fabrication technologies may be transferable to other PAA-assisted 1-dimensional metal-oxide nanomaterials suitable for on-chip gas sensing.

Microstructural investigations of aluminum and copper wire bonds

Ultrasonic wire bonding is a widely used technology to interconnect chips in the packaging industry. It is known that the bond quality depends on the microstructure prevalant in the bond. Similarly, the materials used as well as the bonding parameters depend also on the microstructure. However, the actual used materials such as aluminum and gold are either characterized by poor conductivity or high costs. Due to the outstanding properties of copper, it is a promising candidate for wire bonds. Nevertheless, a thorough investigation on the interrelationship between the material combinations, the processing parameters and the resulting microstructure for copper and aluminum wire bonds need to be conducted. Facing these unresolved issues, samples of aluminum and copper wires were extracted during various states of the bonding procedure. The samples were analyzed to gather information about microstructural influences on the bonding parameters. As a result a conclusion can be drawn regarding the ultrasonic wire bond strength/load subject to the bonding parameters. Furthermore, the different deformation behaviour of copper and aluminum will be depicted by employing the following analytical methods: hardness tests, light microscopy, and scanning electron microscopy.

High temperature, high power piezoelectric composite transducers.

Piezoelectric composites are a class of functional materials consisting of piezoelectric active materials and non-piezoelectric passive polymers, mechanically attached together to form different connectivities. These composites have several advantages compared to conventional piezoelectric ceramics and polymers, including improved electromechanical properties, mechanical flexibility and the ability to tailor properties by using several different connectivity patterns. These advantages have led to the improvement of overall transducer performance, such as transducer sensitivity and bandwidth, resulting in rapid implementation of piezoelectric composites in medical imaging ultrasounds and other acoustic transducers. Recently, new piezoelectric composite transducers have been developed with optimized composite components that have improved thermal stability and mechanical quality factors, making them promising candidates for high temperature, high power transducer applications, such as therapeutic ultrasound, high power ultrasonic wirebonding, high temperature non-destructive testing, and downhole energy harvesting. This paper will present recent developments of piezoelectric composite technology for high temperature and high power applications. The concerns and limitations of using piezoelectric composites will also be discussed, and the expected future research directions will be outlined.

Predicting bond failure after 1.5 ms of bonding, an initial study

We employ a real-time nondestructive contact resistance method to rapidly predict the end state of ultrasonic wire bonds. We classified 64 bonds as either successful (bond attached) or failed (bond unattached) with a support vector machine method. This method predicts with 92% accuracy, 92% sensitivity, and 92% specificity as early as 1.5ms into the ultrasonic processing whether the final bond will be attached. The importance of detecting a tentative failure (on the single bond level) at an early stage is that corrective action can be taken to maximize the yield.

Molecular Dynamic Simulation on Mechanism of Ultrasonic Wire Bonding in Electronic Package

The microscopic mechanism of ultrasonic wire bonding is investigated by molecular dynamics simulation on the interfacial contact and adhesion. Considering that the real bonding area is in the state of plane strain, a two-dimensional atomic model is presented. Sutton-Chen potential is adopted for the interaction between gold atoms. Computational results indicate that a strong adhesion generates at the interface after intimate contact between the wire and the bond pad, and the adhesive force should be the mechanism of ultrasonic wire bonding. Combining the real contact area from finite element analysis with the adhesive force from molecular dynamics simulation, the bonding strength of ultrasonic wire bonding is estimated.

Hardening and Softening in AlSi1 Bond Contacts During Ultrasonic Wire Bonding

Ultrasonic wedge-wedge bonding of 25- mum AlSi1 wires is characterized as a dynamic process of hardening and softening. In the first phase of wire bonding, the wedge is predeformed and cold worked by the bondforce. After ultrasonic energy has been switched on, recrystallization starts at the interface. During the bonding time hardening and softening processes alternate and a maximum in hardness is measured. Hardening and softening processes correlate well with the grain structure, the measured grain sizes and a typical plateau in the z-deformation curve of the contact. At the end of the wire bonding process the wedge is recrystallized and softer than the predeformated wedge, but harder than the as-received wire.

In Situ Studies of the Effect of Ultrasound During Deformation on Residual Hardness of a Metal

A method is developed using an ultrasonic wire bonder to systematically study in situ the effect of superimposed ultrasound during deformation on the residual hardness of a metal wire. The method is applied during ultrasonic ball bonding of Au wire. It is found that 69 mW applied ultrasonic power during deformation of the Au balls leads to lower residual hardness than similarly deformed Au balls without applied ultrasound; this may be beneficial in reducing defects in applications sensitive to bonding stresses. It is confirmed that, although samples deformed with applied ultrasound experience larger strain rates than samples deformed without, strain rate hardening (if it does exist) does not change the finding that ultrasound applied during deformation of Au leads to lower residual hardness.

Lead-free alkaline niobate-based transducer for ultrasonic wirebonding applications

Lead-free piezoelectric ceramic rings (Na 0.475K 0.475Li 0.05)(Nb 0.92Ta 0.05Sb 0.03)O 3 added with 0.4 wt% CeO 2 and 0.4 wt% MnO 2 have been successfully prepared and used as the driving elements of the ultrasonic wirebonding transducers with titanium alloy front and back plates. The transducer operates at similar frequency of the commercial lead zirconate titanate (PZT) transducers (∼65 kHz). Because of the better matching of the acoustic impedances between the ceramic and titanium alloy, an effective transfer of vibration energy is achieved in the transducer, leading to a large axial vibration (1.73 μm at 0.1 W). Moreover, the transducer exhibits a small lateral vibration (0.05 μm at 0.1 W), which is essential for producing a small or narrow bond. The transducer has successfully bonded the aluminum wire on the standard die and gold-plated PCB. The bonds are of good quality, having a smaller deformation ratio (as compared to the commercial PZT transducer) and high bond strength (exceeding the industrial requirement). These clearly show that the lead-free piezoelectric ceramic is a promising candidate for replacing the lead-based ceramics as a driving element in the future generation of wire bonders.

Bismuth sodium titanate based lead-free ultrasonic transducer for microelectronics wirebonding applications

Bismuth sodium titanate based lead-free piezoelectric ceramic 0.885(Bi 1/2Na 1/2)TiO 3–0.05(Bi 1/2K 1/2)TiO 3–0.015(Bi 1/2Li 1/2)TiO 3–0.05BaTiO 3 (BNKLBT-1.5) rings (OD = 12.7mm, ID = 5.1 mm and 2.3 mm thick) were fabricated and characterized. Four ceramic rings were sandwiched between front and back metal plates to form a pre-stressed piezoelectric sandwich transducer and used as the driving element of an ultrasonic wirebonding transducer and the performance of the transducer was evaluated. The BNKLBT-1.5 transducer, when fitted with titanium front and back plates, was found to have axial vibration comparable to that of PZT transducer thus showing that BNKLBT-1.5 has the potential to be used in fabricating lead-free ultrasonic wirebonding transducers.

Vibration characteristics of the capillary in ultrasonic wire bonders

Currently ultrasonic wire bonders with precision capillary tips are widely used in the packaging industry for bonding wires to ICs and circuits. However, the quality and strength of a bond are affected significantly by the vibration of the bonding capillary. It is, therefore, necessary to study the vibration characteristics of the capillary for better understanding and for developing a non-perturbative and low-cost monitoring device for improving bonding quality control. In this paper, the vibration characteristics of the capillary are presented in terms of natural frequency and vibration modes using finite-element models. The predicted displacement profile along the capillary is compared with the measurement using a non-contact laser interferometer. In general, a good agreement between the prediction and measurement is obtained, and it shows that the interferometer may be a good monitoring tool for characterization of capillary tip vibration, as well as providing a reference for other monitoring devices.

0.90(Bi1/2Na1/2)TiO3–0.05(Bi1/2K1/2)TiO3– 0.05BaTiO3 transducer for ultrasonic wirebonding applications

Lead-free piezoelectric ceramic 0.90(Bi1/2Na1/2)TiO3–0.05(Bi1/2K1/2)TiO3–0.05BaTiO3 (BNKBT-5) rings (OD=12.7 mm, ID=5.1 mm and 2.3-mm thick) were fabricated and characterized. Four ceramic rings were used as the driving element in an ultrasonic wirebonding transducer and the performance of the transducer was characterized. The lead-free transducer was found to have comparable voltage rise and fall times as a lead-based PZT transducer and has a relatively large vibration amplitude, thus showing that BNKBT-5 has the potential to be used in fabricating lead-free ultrasonic wirebonding transducers.

A microslip model of the bonding process in ultrasonic wire bonders Part I: Transient response

Ultrasonic wire bonders with precision capillary tips are widely used for bonding electrical wires to IC chips and circuits. The industry is driving wire-bonding technology towards increased yields, decreased pitch, and ever decreasing cost. However, many technical and material issues will be involved in achieving these goals, such as bond quality monitoring, new failure modes, reliability problems in new plastic mold compounds, and increasing wire sweep problems, in particular lack of the quantitative understanding and validated mathematical model of the bonding process. In this paper, a microslip model is proposed for the bonding process based on the bonding pattern observed, i.e. the wire is bonded at the periphery while the centre is left unbonded. A bilinear hysteresis restoring force is assumed in the microslip model because of its simplicity. The transient displacement response is calculated, and its pattern shows very good agreement with the experiment.

A microslip model of the bonding process in ultrasonic wire bonders part II: steady teady state response

Ultrasonic wire bonders with precision capillary tips are widely used for bonding electrical wires to IC chips and circuits. The industry is driving wire-bonding technology towards increased yields, decreased pitch, and ever decreasing cost. However, many technical and material issues will be involved in achieving these goals, such as bond quality monitoring, new failure modes, reliability problems in new plastic mold compounds, and increasing wire sweep problems, in particular lack of the quantitative understanding and validated mathematical model of the bonding process. In a previous paper [Hu CM, Guo N, Du H, Li WH (in press) A microslip model of the bonding process in ultrasonic wire bonders - Part I: transient response. Int J Adv Manuf Technol], a microslip model has been proposed for bonding process based on the bonding pattern observed, i.e. the wire is bonded at the periphery while the centre is left unbonded. A bilinear hysteresis restoring force is assumed in a microslip model because of its simplicity. The predicted transient responses show good agreement with the experiment. In this paper, the analysis is extended to steady- state response, and discussion and comparison are made with respect to the transient response.

Study of 1-3 composite transducer for ultrasonic wirebonding application

Lead zirconate titanate (PZT)/epoxy 1-3 composites consisted of PZT rods embedded in an epoxy matrix were used as the driving elements in an ultrasonic wirebonding transducer for microelectronic packaging. The performance of the 1-3 composite PZT transducer with 0.91 volume fraction of PZT has been studied. The composite transducer has fewer lateral vibrations and higher damping. It has the potential to be used in new generation high-speed and fine-pitch ultrasonic wirebonding machines.