COPS (crystal originated particle/pit), which are considered to be originated from grown-in defects formed during CZ crystal growth and degrade GOI (gate oxide integrity) of MOS ULSI devices, were studied with respect to the size and density distribution on a surface of CZ grown Si wafer by means of laser particle counter. Firstly we studied the relation between detection limit of COP size and various measurement parameters such as PMT (photomultiplier) gain and threshold level of laser particle counter (SFS6200) using AFM (atomic force microscope), OPP (optical precipitate profiler), and LST (infrared light scattering tomography). Then this method was applied to evaluation of as grown CZ Si wafer containing Ring-OSF region, which were revealed after annealing such as 1000 degrees Celsius for 1 hour in dry O₂, and the COP size and density distribution were compared to those of the grown-in defects measured by OPP and LST. It was found COP measured by laser particle counter had good correlation in the size and density distribution with the grown-in defects measured by OPP and LST, and the COP showed slightly higher density than the grown-in defects by OPP. This means COP measurement by laser particle counter is a more sensitive method to evaluate grown-in defects than OPP and LST.

This paper describes a methodology that has been incorporated into a fully integrated measurement system, the n&k Analyzer, that determines simultaneously the thickness, energy band gap, and n and k spectra (from 190 to 900 nm) of various forms of silicon, i.e., a-Si, poly-Si films, and mixtures of a-Si and poly-Si films. Additionally, the system also measures the average surface roughness. In turn, the n and k spectra of such films can be correlated to processing conditions, temperature being the most important one in LPCVD method. The n&k Analyzer can be used to identify the amorphous-polycrystalline transition regime and characterization of films produced in this regime.

InP is a very promising material for electronic and optoelectronic devices, especially for heterojunction devices employing InGaAs as the low energy gap material. However, the low metal-semiconductor barrier height of n-InP (phi_b less than or equal to 0.5 eV) and the lack of reliable oxide-insulator technology, limit the number of applications severely. Here, the need for an oxide-insulator layer is eliminated with a simple one step deposition technique. We report the enhancement of the Schottky barrier height in n-InP using a dry, room temperature indium tin oxide (ITO) film deposition technique in a controlled oxygen and argon partial pressure environment. This process appears to produce a stable reproducible surface. Resistivity values as low as 2 by 10^-5 (Omega) cm and optical transmission values greater than 90% in the visible and near-IR region of the spectrum were obtained. Barrier height values as high as phi_b equals 0.89 eV have been obtained from current-voltage (I-V) characteristics that were linear over several orders of magnitude of current with an interface trap time constant of 7.07 by 10^-6s obtained from conductance-frequency (G-f) measurements. The enhancement in the barrier height is explained by an analysis of the ITO/InP interface. This process is being investigated for fabricating a highly transparent ITO gate electrode to an InP optical field effect transistor (OPFET) structure.

The cathodoluminescence (CL) measurements of carbon doped and undoped aluminum nitride (AlN) thin films near the band- edge region were performed at temperatures of 300, 77 and 4.2 K. These films were grown on three different substrates sapphire, 6H-SiC and 4H-SiC. The typical thickness of these films was in the range of 0.35 - 0.4 micrometer. Two distinct peaks 'A' and 'B' were observed in undoped samples around 6.1 and 5.9 eV, respectively. Also, there is an expected peak within the low energy side of peak 'B.' The absorption spectra of different carbon doped AlN on sapphire substrate were carried out to study the Urbach tail's parameters which play an important role near the band-edge transitions.

The following paper presents the results of morphology (columnar structure) and properties investigations of 0.2 - 0.4 mkm thick amorphous hydrogenated carbon (C:H) films, prepared by radio frequency (rf) sputtering of graphite target in argon-hydrogen mixture.

We have constructed the automated complex oriented to the semiconductor wafer surface superlattice symmetry type evaluation. The basis of the method is the photo-e.m.f. signal registration as a function of the rotation angle of the upper transparent electrode. The main part of the apparatus is the molecular dipole probe being characterized by the value of the surface monolayer charge which in turn depends on the wafer surface micro curvature (atomic rows energetic relief). As the result of the present work the interdependence between the number of light director orientation axes and surface atoms coordination type which is, at the same time, the symmetry of the superlattice.

We have examined experimentally the parameters of the photo- e.m.f. generated in the system containing the monocrystalline semiconductor and thin layer of nematic liquid crystal (NLC). Registered maxima quantity perfectly correlates with the ordering axis number. The present paper shows one of the ways of biaxial molecular ordering in NLC not registered before. It was shown also that the proposed method is both applicable as to the semiconductor surface superlattice symmetry type analysis and also to the ordered LC molecules concentration evaluation.

CZ-Si has been subjected to different heating schedules in an effort to have a deeper insight about the appearance and disappearance of thermal donors, generation of new donors and thermal acceptors along with the amount of oxygen and carbon precipitated. Annealing of the oxygen rich silicon samples in the temperature range 465 - 540 degrees Celsius for 26 hours demonstrated the three OD stages -- growth region followed by annihilation of TDs and subsequently ND generation. Samples with low oxygen content pre-annealed at 455 degrees Celsius for 24 hrs and annealed at 470 degrees Celsius for varying annealing duration showed the complete absence of the first stage of rapid growth of TDs, but the pre-annealed samples with both low and high oxygen contents and subsequently annealed in the temperature range 430 - 630 degrees Celsius for short duration exhibited the dominance of thermal acceptors beyond the annealing temperature corresponding to maximum ND generation. The pattern of oxygen and carbon precipitation for such samples is also different. Activation energy for oxygen diffusion and self interstitials in silicon come out to be 0.6 eV and 0.4 eV respectively and the diffusion coefficient of oxygen in the pre-annealed samples and annealed at 470 degrees Celsius for different durations to be 6.18 by 10^-19 cm²s^-1. Oxygen reduction is governed by second order kinetics.

In this paper we demonstrate the feasibility of an in-line, non-destructive method, thermal wave imaging, to characterize the formation of titanium disilicide on narrow lines. The observation of thermal-wave signals and images allows us to distinguish the formation and structure of the metastable, high-resistivity, C49 phase and the equilibrium, low-resistivity C54 phase of the disilicide. Patterned wafers, having linewidths in the range of 1.5 to 0.4 micrometer were used to study the effect of rapid thermal anneal conditions and linewidth on disilicide formation. The transformation temperature of C49 to paper we demonstrate the feasibility of an in-line, non-destructive method, thermal wave imaging, to characterize C54 disilicide during a 20 second isothermal anneal was found to increase by more than 50 degrees Celsius when decreasing the linewidth to 0.4 micrometer.

By using molecular beam epitaxy (MBE) technology, one can now fabricate routinely ultra-pure GaAs and Si layers with the background residual impurity concentration lower than approximately 1 by 10¹⁴ cm^-3. We incorporated mass-separated impurities into these layers by means of high-energy ion-implantation (HE-I²) and low-energy ion beam impinging during MBE growth (CIBMBE). Low-temperature (2K) photoluminescence (PL) spectra from acceptor-doped GaAs layers prepared by HE-I² and CIBMBE methods revealed that multiple shallow emission levels are formed with increasing net hole concentration, N_A-N_D. These emissions were theoretically interpreted as the pairs between ground- and excited-states acceptors. The absence of these energy levels in the conventional specimens was attributed to a strong optical compensation effect induced by an extremely small amount of shallow donors that were accidentally introduced during sample preparation. The majority of these energy levels were found to present strong red or blue energy-shifting with increasing N_A-N_D. These PL features were confirmed to be used for the determination of N_A-N_D and compensation ratio totally by an optical method. High-dose incorporation of Fe atoms into Si substrates was carried out by HE-I² method for the synthesis of beta-FeSi₂. In 2K-PL spectra, one noticed strong edge-emission from beta- FeSi₂, indicating that HE-I² method is powerful to synthesize high quality silicides.

In the last several years, the semiconductor industry has come to the realization that the package into which a die is placed is at least as critical to the performance of the complete electronic system as the die itself. This realization has led to an explosive effort across the entire industry to advance the state-of-the-art in semiconductor packaging. To date, this effort has already produced semiconductor packaging options on the scale of the die (i.e., chip scale packaging -- CSP). While CSPs and other advanced packaging techniques provide improved electronic system performance, they also increase the quality control burden (despite the highly automated processes used to manufacture the packages, quality control remains, for the most part, a manual operation). This paper addresses the necessary requirements of automated visual inspection (AVI) for quality control of current and future semiconductor packaging. The necessary requirements of the station are subdivided into two categories: those pertaining to the hardware platform, and those pertaining to the software reasoning engine. Hardware requirements are discussed in terms of finding the best match between commercial, off-the- shelf, hardware components and a given inspection application. Components reviewed include: imagers, optics, illumination systems, auto-focus/alignment systems, material handlers, parallel image preprocessors, and host computers. Applications reviewed include: pin grid array (PGA), ball grid array (BGA), and flip-chip package inspection. Also discussed in the hardware section are options that may be used when commercial components are not adequate. Software requirements are discussed in terms of the functionality required to provide accurate, real-time characterization of package quality, to gain operator acceptance, and to produce meaningful statistics for use in process control.

We present our effort to construct an optoelectronic interconnection layer for Cray supercomputer board for board-level fast optical clock signal distribution. The optical interconnection layer under investigation employs optical channel waveguides with surface-normal waveguide couplers. The difficulties associated with the complicated 3-D multiple alignments are significantly reduced through the parallelism among the surface-normal fanout beams and the unique planar device feature. One-GHz optical clock signal is demonstrated experimentally with a guided-wave interconnection length of 45 cm. Some new techniques to fabricate multimode optical channel waveguides and 2-D array of surface-normal waveguide holographic gratings on board- level are also presented.

We report in this paper the surface-normal input and output grating couplers for 1-to-many fanouts aiming at optical clock signal distribution application. For wafer-scale interconnects, surface-relief polygonal gratings with a 1 micrometer period (0.5 micrometer feature size) were fabricated using reactive ion beam etching (RIE). Surface- normal input and output coupling schemes were carried out with a combined coupling efficiency of 65%. Employment of substrate modes in silicon greatly releases the required grating spacing for the demonstrated two-way surface-normal coupling. Seven and five-tenths GHz 1-to-4 clock signal distribution operating at 1.3 micrometer was demonstrated with a signal to noise ratio as high as 60 dB. Generalization of 1-to-many fanout can be realized by implementing a polygonal grating with an equivalent number of facets. For board level optical clock signal distribution system, a preliminary result using a parallelogramic grating to couple surface-normal input light into polyimide waveguide is also reported.

A novel double focus optical interferometer for scanning force microscope is presented in this paper. The deflection of the force sensing cantilever is measured by detecting the phase shift of two orthogonal polarized light beams reflected from the end of the cantilever and a reference plane with a hole in the center. We have tested some samples by using this instrument and gotten some high resolution pictures. The instrument provides 0.1 nm height resolution and 5 nm lateral resolution and is especially applicable in the profilometer of the integral circuit, fine grate, etc.

Spectroscopic reflectometry has been a mainstay of optical semiconductor metrology for many years, as has single wavelength ellipsometry performed at a fixed angle of incidence. More recently, spectroscopic ellipsometry has been employed to provide film thickness and optical constant measurements with greater precision and accuracy. In this work we demonstrate the combined use of spectroscopic ellipsometry (performed from 420 - 720 nm at 75 degree angle of incidence) and spectroscopic reflectometry (performed from 200 to 800 nm at normal incidence) for the characterization of thin films and multilayer stacks commonly encountered in semiconductor device fabrication. By employing sophisticated parametric dispersion models and simultaneously fitting both the reflectance and ellipsometric spectra, we are able to perform accurate and precise characterization of many structures that cannot be characterized with either technique alone or with combined reflectometry and single wavelength ellipsometry.

Application of electric field to heterostructure p-n junctions beyond the threshold produces lasing. On further increase of the electric field the peak of the intensity shifts towards red. The line width also increases. In this article it is shown that this shift as well as the change in the line width can be effectively utilized to distinguish the 3D to 0D structures. The shift in the bulk structure exhibits F^2/3 power law, where 'F' is the electric field, while 2D, 1D, and 0D samples give rise to a quadratic dependence as in single atoms. In addition elegant modifications in micro-fabricational techniques are suggested as to realize more efficient structures with ease.

The aim of this paper is the theoretical and experimental study of the mode structure laser radiation influence on the basic parameters of the sharp periodical submicron structures (grooves) system which is formed on the solid-state surface by means ofthe intracavity method. It is shown that the longitudinal resonator mode influence is one of the important factors determining the groove width. This factor may be the main one for wide-range infrared lasers with short cavities such as high-pressure carbon dioxide lasers. The comparison of the theoretical analysis with the experimental data enables to conclude that the treated surface damage looking like the sharp equidistant grooves system is formed only when the transverse mode TEM is used. The higher order transverse modes lead to the chaotically located microcraters producing. It is achieved that the grooves system contrast depends on the laser radiation spectrum width, the cavity length, the mirrors reflectivities and practically doesn't depend on the laser beam turn angle. Key words: Intracavity laser treatment of the surface, laser radiation mode structure, submicron periodical structures forming.

This paper first describes how growth rate of ultra-thin thermal oxide and oxide charges are influenced by an incorporation of Al and/or Fe into SiO₂. Al and Fe in SiO₂ enhanced the oxide growth rate at an early stage of thermal oxidation, while they reduced it around 10 percent in several hundreds minutes at 1123 K. In thermally oxidized Al- or Fe-contaminated p-type Si, observed ac surface photovoltages (SPVs) showed that the positive fixed oxide charge (Q_f) was compensated by the induction of a negative charge which was already proposed by the conceptual model positing that (AlOSi)^- and (FeOSi)^- networks are formed in SiO₂. During thermal oxidation, the Al and Fe existing in the native oxide were left behind at the very top of the SiO₂ causing a metal-induced negative charge. In the course of thermal oxidation of Al- contaminated Si wafers, the net oxide charge has diminished in n-type Si, while it has increased in p-type Si. This is possibly because the Al suboxides, for example, in the thinnest oxides changed into Al₂O₃ in thicker oxide, resulting in losing the negative charge. The Al-induced negative charge in SiO₂ might enhance oxide growth at an early stage of oxidation by compensating the positive Q_f. Furthermore, in oxidized n-type Si, a new idea is proposed that most part of positive Q_f originating in Si dangling bond may be effectively neutralized by abundant electrons in n-type Si.

The acquisition of surface photo-voltage (SPV) instrument is only one step in the implementation of a metallic contamination reduction program. In order to take manufacturing decisions based on surface photo-voltage data, one must understand the strengths and limitations of the technique and avoid its pitfalls. We review the steps taken to implement SPV in our fab.

Influence of an inner electric field on such emission phenomena like: secondary emission, photoemission and field emission has been investigated. The applied sample-emitter was a glass wafer (thickness 0.2 mm) covered on both sides by semiconducting films In₂O₃:Sn. A voltage (in the interval -2000V divided by 0V) generating transverse electric field was applied to one of the films. This film had a thickness of about 200 nm. The second film (emitting electrons) had a thickness 100 nm or 10 nm. The secondary emission measurements were made by the retarding field method using four grid retarding potential analyzer. It was found that the secondary emission coefficient changes non- monotonically with increasing field intensity. Electron emission measurements without using a primary electron beam were made with the electron multiplier cooperating with a multichannel pulse amplitude analyzer. The measurements were performed in the vacuum of about 2 multiplied by 10^-6 Pa. Influence of film thickness on the intensity of field controlled emission and field controlled photoemission was also studied. It was also found that the frequency of counts (generated by electrons in the electron multiplier) depends on the polarizing voltage approximately in an exponential way. Some departures from this dependence can be observed at higher U_pol voltages (above 1000 V). Thus, at an appropriate high voltage U_pol conditions for a cascade emission are created. At lower voltages the conditions correspond to a semiconductor with a negative electron affinity.

Recent developments in SiO₂ diagnostics have opened a possibility for measuring the mobile charge in oxide without the expensive and time consuming preparation and analysis of MOS test structures. A key element of the new approach is that corona charge is deposited directly onto the SiO₂, over the entire wafer surface, to create an electric field needed for ion drift. Mobile charge density is determined by mapping of the oxide voltage change after temperature stress. As a result, whole-wafer mobile charge maps with over 6000 points may be acquired in about ten minutes at a sensitivity below 10¹⁰ c^-2 for 100 nm oxides. A distinctive feature of the new technique is that the measurement of oxide voltage shift is done under strong inversion or accumulation and is insensitive to interface traps and oxide fixed charge. In other methods based on the flat band voltage shift these charges near the SiO₂-Si interface are a major limitation.

We developed a robust measurement recipe for six layer SOI film stack. Both spectrometer and BPR were combined to characterize the plate and storage polysilicons. A new global optimization method was developed to find the best solution in parameter spaces with up to 12 parameters. Such a recipe was applied to production wafers with over 50 site die mapping. The 5 day repeatability shows the measurements were stable and robust.

Industrial Engineering is fairly new to the semiconductor industry, though the awareness to its importance has increased in recent years. The US semiconductor industry in particular has come to the realization that in order to remain competitive in the global market it must take the lead not only in product development but also in manufacturing. Industrial engineering techniques offer one ofthe most effective strategies for achieving manufacturing excellence. Industrial engineers play an important role in the success of the manufacturing facility. This paper defines the Industrial engineers role in the IC facility, set the visions of excellence in semiconductor manufacturing and highlights 10 roadblocks on the journey towards manufacturing excellence.