Y. C. Chan

Carrier transport in thin films of organic electroluminescent materials

Abstract Using the time of flight technique, the carrier mobilities of organic electroluminescent materials, N , N ′-bis-(1-naphhyl)- N , N ′-diphenyl-1,1′-biphenyl-4,4′-diamine (NPB), tris(8-hydroxyquinoline) aluminum (Alq), and Tb(AcA) 3 phen (AcA: acetylacetone, phen: phenanthroline) (Tb-complex), were measured. At room temperature and an electric field of 2.3×10 5 V/cm, the effective mobility of holes in NPB, electrons in Alq and Tb-complex were (8.8±2)×10 −4 , (6.7±2)×10 −5 and (3±2)×10 −4 cm 2 V −1 S −1 , respectively. The hole mobility of NPB shows an exponential root field dependence as commonly seen in some organic materials.

Comparative study of micro-BGA reliability under bending stress

The micro-ball grid array (/spl mu/BGA), a form of chip scale package (CSP), was developed as one of the most advanced surface mount devices, which may be assembled by ordinary surface mount technology. In the latest /spl mu/BGA type, eutectic tin-lead solder ball bumps are used instead of plated nickel and gold (Ni/Au) bumps. Assembly and reliability of the /spl mu/BGAs PCB, which is soldered by conventional surface mount technology, has been studied in this paper. The bending cycle test (1000 /spl mu//spl epsi/ to -1000 /spl mu//spl epsi/), is used to investigate the fatigue failure of solder joints of /spl mu/BGA, PBGA, and CBGA packages reflowed with different heating factors (Q/sub /spl eta//), defined as the integral of the measured temperature over the dwell time above liquidus (183/spl deg/C). The fatigue lifetime of the /spl mu/BGA assemblies firstly increases and then decreases with increasing heating factor. The greatest lifetime happens while Q/sub /spl eta// is near 500 second-degree. The optimal Q/sub n/ range is between 300 and 750 s/spl deg/C. In this range, the lifetime of the /spl mu/BGA assembly is greater than 4500 cycles if the assemblies are reflowed in nitrogen ambient. SEM micrographs reveal that both /spl mu/ & P-BGA assemblies fail in the solder joint at all heating factors. All fractures are near and parallel to the PCB pad. In the /spl mu/BGA assemblies cracks always initiate at the point of the acute angle where the solder joint joins the PCB pad, and then propagate in the section between the Ni/sub 3/Sn/sub 4/ intermetallic compound (IMC) layer and the bulk solder. In the CBGA assembly reliability test, the failures are in the form of delamination, at the interface between the ceramic base and metallization pad.

Study of Ni3P growth due to solder reaction-assisted crystallization of electroless Ni-P metallization

Nickel (Ni) has been widely used for advanced electronic applications to replace Cu or act as a diffusionbarrier layer in the conventional Au/Cu metallization because of a slower intermetallic compound (IMC) formation rate [1] and a relatively low diffusion rate through Au and Cu [2, 3]. Amorphous electroless nickel-phosphorus (Ni-P) has attracted much interest and widespread uses in the PCB fabrication and UBM for flip chip technology because it offers a lower cost alternative to a more expensive physical Ni deposition methods, good corrosion resistivity, strong adhesion, and good solder wetting [4–6]. It has been found that the amorphous Ni-P alloy will only undergo a selfcrystalline transformation to Ni and Ni3P at temperature above 300 ◦C [7]. Recently, Jang et al. [5] have found that the solder reaction will assist crystallization of electroless Ni-P UBM in flip chip packages even if the reflow temperature is well below the self-crystallization temperature. However, the details of the Ni3P growth in electroless Ni-P are still unclear. In this letter, some results of Ni3P growth in electroless Ni-P with chemical composition of Ni77P23 are presented and then compared with those results obtained in Ref. [5] in which the electroless Ni-P has a composition of Ni85P15. Electroless Ni-P was deposited on Cu pad substrate and a gold (Au) flash was deposited on the top of that electroless Ni-P in order to avoid oxidation of nickel surface. Eutectic Pb-Sn solder balls were placed on the prefluxed Au/Ni-P/Cu pad substrates and reflowed at temperature of 220 ◦C for different annealing time from 0.5 to 90 min. For interfacial microstructural examination, samples were mounted in epoxy and then sectioned using a slow speed diamond saw. The crosssectioned samples were ground, polished and etched with 10% conc. HCl+ 90% water. The chemical composition of electroless Ni-P was found to be Ni77P23 by using Philips XL 40 FEG scanning electron microscope (SEM) equipped with energy dispersive X-ray analysis (EDX). The structure of electroless Ni-P was confirmed to be amorphous by X-ray diffraction (XRD). Fig. 1 shows the SEM image of cross-sectioned sample after reflow at 220 ◦C for 0.5 min. The IMC layer is confirmed to be Ni3Sn4 by XRD and EDX analyses. It is found that both chunky-type and needle-type Ni3Sn4 IMCs are present. It is interesting to note that there is a thin dark layer between the Ni3Sn4 IMC and the electroless Ni-P. This dark layer is confirmed to be Ni3P, which is consistent with the findings of Jang et al. [5]. Fig. 2 shows the relationship between the thickness of Ni3P and the annealing time at 220 ◦C. During the first 15 min of annealing, the Ni3P grows rapidly due to the solder reaction-assisted crystallization of electroless Ni-P. It is found that the thickness of Ni3P is linearly proportional to the square root of the annealing time during the first 15 min. This means that the Ni3P growth is controlled by a diffusion process during the first 15 min of annealing. However, it is interesting to note that the growth of Ni3P is stopped or even dropped after the first 15 min of annealing. For the sake of comparison, Fig. 2 also shows the results of Ni3P thickness growth in electroless Ni85P15 [5]. It shows that the Ni3P layer in electroless Ni85P15 continued to grow throughout the entire annealing range from 0 to 40 min. If we assume all the Ni and P elements in electroless Ni-P react into Ni3P and Ni3Sn4 without any loss into the solder, a general solder reaction-assisted crystallization equation in electroless NiaPb will be as follows:

Impact properties of PBGA assemblies reflowed in nitrogen ambient and compressed air

In this paper, the solder joint brittleness of plastic ball grid array (PBGA) assemblies has been studied by means of impact testing. The results show that the maximum impact force and the impact energy for PBGA assemblies reflowed in nitrogen ambient is higher than for those reflowed in compressed air. Reduction of the oxygen content from 1000 ppm to 50 ppm, reduces the initiation energy of impact. The impact fractured surfaces occur between the printed circuit board (PCB) pads and solder joints, i.e., in the brittle intermetallics region. The pores and impurities are mainly distributed within a ring at perimeter of the interface, with greater density at the edges than at the center. Reflow in a nitrogen ambient reduces the defect density, with reduction in oxygen content further reducing the density. Impact cracks begin from the defects of the Cu/solder interface, and propagate along the intermetallics.

Metallurgical reaction and mechanical strength of electroless Ni-P solder joints for advanced packaging applications

We have studied the metallurgical reaction and mechanical strength of the electroless Ni-P solder joints as a function of reflow time at 220 °C. It is found that both Ni3Sn4 intermetallics and Ni3P are formed due to the solder reaction-assisted crystallization. However, after the first 15 min of reflow, an unusual depression of Ni3Sn4 growth has been observed. A detailed description of the diffusion mechanism has been presented to explain the prohibition of the Ni3Sn4 growth. It is found that the growth of Ni3Sn4 and Ni3P may have a mutual effect on each other during the solder reaction since there is a direct correlation between the depression of the Ni3Sn4 growth and the ending of Ni3P growth. The characteristic of the mechanical strength of electroless Ni-P solder joints has been demonstrated. A correlation between the mechanical strength and the interfacial metallurgical reaction has been discussed. Also, it is found that different reflow times will result in different fracture interfaces of the sheared electroless Ni-P solder joints. The detailed explanation of the fracture surface morphology has been explored.

Effect of pinhole Au/Ni/Cu substrate on self-alignment of advanced packages

Abstract The self-alignment of advanced packages (μBGA) on both non-pinhole and pinhole Au/Ni/Cu pads has been discussed. It is found that a slight reduction of self-alignment of the packages using pinhole pads occurs. Rutherford backscattering spectrometer (RBS) results suggest that this reduction should not be attributed to the oxide formation of the surface or interface layer in the Au/Ni/Cu pads. The solder wetting experiments show that slow spreading of molten solder on pinhole pads may result in a reduction of effective board pad surface area that can be wetted. This will reduce the restoring force of the solder joints, and thus causing a less better self-alignment of the packages using pinhole pads. Oxidation of nickel at the exposed area and Au/Ni interface is observed to occur by direct exposure of substrate pads through pinholes during aging. The solder wetting of the aged pads has been described. For flux reflow soldering, the aging of the pads seems to have no serious effect on the self-alignment of the package. However, it is found from the peel-off test that a few solder joints of the samples after reflow have weak adhesion strength at the solder and aged pinhole pad interface. The mechanism for this weak adhesion strength has been proposed.

Nondestructive defect detection in multilayer ceramic capacitors using an improved digital speckle correlation method with wavelet packet noise reduction processing

The nondestructive detection of defects in multilayer ceramic capacitors (MLCs) in-surface mount printed circuit board assemblies has been demonstrated by using an improved digital speckle correlation method (DSCM). The internal cracks in MLCs that contribute to the thermal displacements on the MLC surface after dc electrical loading may be uniquely identified using this improved DSCM combined with double lens optical arrangement. However, it is found that Joule heating of the MLC sample takes time, and therefore the thermal displacements on the MLC surface are not obvious at the beginning of the dc electrical loading. In order to shorten the detection time and increase the resolution of the DSCM, a wavelet packet noise reduction process is introduced into the DSCM technique. This new algorithm is used to reduce the background noise in the signal so as to improve the accuracy of detection of defect locations and reduce the detection time. By introducing wavelet packet noise reduction processing, the DSCM is found to be more sensitive to and faster at detecting defects in MLC samples. Furthermore, the DSCM with wavelet packet noise reduction process can cope with the problems of edge effect, rough and warped surface, which are the limitations of the scanning acoustic microscope (SAM).

Digital speckle correlation method based on wavelet-packet noise-reduction processing

Despite the advantages of being highly sensitive and nondestructive, the digital speckle correlation method (DSCM) may have difficulties in detecting tiny defects such as delaminations in multilayer ceramic capacitors. This is because the presence of background noise always complicates the data processing. We present a new algorithm, which employs the wavelet-packet noise-reduction process together with the improved DSCM, to improve data processing. Both the computational error and the noise are shown to be reduced successfully by this new algorithm. The accuracy (or precision) of the improved DSCM is increased after operation of the wavelet-packet noise-reduction process. The most important feature of this new algorithm is that it can extract a small hillock signal from a large noisy background in a DSCM deformation result. This helps to save time in the detection of tiny defects, such as delamination, in a miniaturized electronic component.

Gap states and stability of rapidly deposited hydrogenated amorphous silicon films

Abstract The sub-band gap absorption spectrum α(hv) in a-Si:H films prepared at high deposition rates was measured by the constant photo-current method (CPM) for photon energy ranging from 0.8 to 1.7 eV in a thermally-annealed state and light-soaked state. The Simmons–Taylor theory and occupation statistics of correlated defects are used to model the distribution of band tail and gap states. It is found that the density of gap states increases after light soaking, however, there is no evident change found in the density and distribution of band tail states. Measurements of the light-induced changes find that the photoconductivity decreases by less than one order of magnitude after long time of light illumination for the high rate deposited a-Si:H. This demonstrates that the high rate deposited samples have relatively high stability compared with conventionally deposited a-Si:H.