André Van Calster

Adhesion Strength of the Epoxy Polymer/Copper Interface for Use in Microelectronics

The adhesion of plated metals on top of chemically treated epoxy layers for build-up purposes was examined. Specifically, the influence of wet chemical pretreatments on the adhesion of plated copper to epoxy polymer is investigated. This adhesion is related to the surface roughness of the polymer and the chemical composition of its surface. The chemical composition of the surface was examined by X-ray photoelectron spectrscopy and placed in the context of the development of the interface during the wet chemical pretreatments and related to the theory developed in previous publications. Various combinations of pretreatments were followed by an identical electrochemical Cu deposition and peel strength measurement sequence. This allowed interpretation of the changes of the peel strength with pretreatments. Using this interpretation, the peel strength of build-up layers was maximized. We propose that the surface of the polymer layer develops into a fractal surface during wet chemical oxidation. Using this proposition, in combination with pore diffusion for the oxidizer, the evolution of peel strengths with chemical pretreatment times can be qualitatively understood. The peel strength of electrochemically deposited copper can be quantitatively related to the atomic force microscopy measurements for limited oxidation treatment times.

Kinetic study of wet chemical treatments on the surface roughness of epoxy polymer layers for buildup layers - II. Oxidative treatment of the surface

The adhesion of plated metal layers to epoxy polymer surfaces is of prime importance for the reliability of interconnections in microelectronics. An increase in the roughness of the polymer surface, caused by chemical treatment, plays an important part in the adhesion strength of plated metal layers by increasing the total area of interaction between both layers. This kinetic study of the influence of chemical treatments on the surface of the polymer consists of two parts. In Part I the kinetics of sweller treatments and of the solvent were examined in detail and a kinetic model for the influence of sweller and solvent was developed. In this second part the kinetics of the formation of roughness at the polymer surface due to oxidative liquid treatments is examined. A kinetic model is developed based on experimental studies, compromising the combined effects of sweller-oxidizer treatments and the influence of diffusion of the different components. A number of examples are given and verified by experiment. The roughness of the treated polymer samples is measured experimentally by atomic force microscopy. Time-dependant mass measurements were used to record mass changes due to diffusion and reaction.

Chemical modification of buildup epoxy surfaces for altering the adhesion of electrochemically deposited copper

The adhesion of electrochemically deposited copper on top of chemically surface modified epoxy layers for buildup purposes is examined. Using a wet-chemical surface synthesis reaction iminodiacetic acid, imidazole, and mercaptopyrimidine groups were chemically imprinted on the surface of buildup epoxy layers based on nucleofilic substitutions. The synthesis was followed by an identical electroless Cu deposition and peel strength measurement sequence. The identity of certain groups at the surface has pronounced influence on the adhesion strength of electrochemically deposited metals and is compared with traditional swell/ oxidation treatments of the surface of the polymer. By varying the identity of the chemical groups at the surface, the chemical part of the adhesion can be altered. The changes in roughness of the polymer surface were minimized by choosing solvents that exhibit minimum diffusion into the polymer and by limiting the reaction temperature. In this way the changes in the physical part of adhesion could be kept to a minimum. The chemical composition of the surface is examined by attenuated total reflection-infrared and X-ray photoelectron spectroscopy measurements. Weight changes were recorded after each separate wet-chemical treatment. Surface synthesis influences the peel strength negatively or positively and modifies the deposition rate and etching speed of electrochemically deposited copper.

Processing quality results for electroless/electroplating of high-aspect ratio plated through holes in industrially produced printed circuit boards

Abstract The plating quality results of high-aspect ratio through holes in industrially produced and sized printed circuit boards are reported in this paper. Test panels with 32,000 PTHs were processed on an existing industrial production line in order to test the possibilities of modifying the line for panels with higher aspect-ratio plated through holes. Depending on the process parameters different types of faults in the plating of these plated through holes were detected, varying from gas bubbles stuck inside the plated through hole, rough and non-plated glass fibers caused by the drilling of the holes, to parts of resist stuck inside the through holes. The type of error occurring is dependent on the processing used. Because each individual processing step can influence the following processing steps, an integrated quality approach has to be considered. Most of these problems could be solved by small changes in process parameters. Making a quantitative analysis of the type of errors resulted in a drastic reduction in the number of unplated through holes on industrial panels. This was done by identifying and classifying the type of error occurring. Aspect ratio’s of 10.7 could be plated with a high yield on an existing vertical electroless plating line.

Method for measuring the cell gap in liquid-crystal displays

A method for measuring the cell gap of liquid-crystal displays (LCDs) is presented. It is based on analyzing the interference spectrum of white light reflected by the LCD. It can be used before or after filling the LCD. The method is applicable for LCDs whose operation is based on the polarization or the scattering of light. Four methods for the calcu- lation of the cell gap are presented. The first method is based on the wavelengths corresponding with the peaks in the interference spectrum. Two other methods use analog and digital filtering techniques to filter out all the irrelevant information in the interference spectrum; the cell gap is then calculated with fast Fourier transform (FFT) or the inverse FFT. In the fourth method an empirical model is fitted to the measured interfer- ence spectrum.

New model for the characterization and simulation of TFTs in all operating regions

Abstract— A new, semi-empirical model that describes TFTs in all operating regions with only one dynamic and five static parameters is proposed. This model can easily be implemented in most simulation programs and used to predict TFT circuit behavior. The five static parameters can automatically be extracted from current measurements and form a very compact yet adequate representation of the TFT characteristics. This proves especially useful when storing and statistically processing measurement results for a large number of devices, such as the pixel TFTs in an active matrix.

Cell gap optimization and alignment effects in reflective PDLC microdisplays

In general this reduction of the cell gap improves the electro-optic properties of a polymer dispersed liquid crystal (PDLC) in reflective microdisplays. At the interface between the PDLC film and the silicon backplane or cover glass, the LC molecules have a different alignment from those in the droplets in the interior of the PDLC film. This is shown by microscopic observations and the temperature dependency of the brightness and capacitance of the displays. The influence of this alignment effect increases for smaller cell gaps and has an impact on the properties of the PDLC. During and after the filling of the displays, a compression and expansion of the cell gap takes place, respectively. If the curing of the PDLC takes place before the expansion of the cell gap has stopped, transparent areas in the PDLC film may occur some time after curing. This effect is caused by the expansion of the cell gap after curing resulting in the vertical alignment of LC molecules. This can be concluded from microscopic observations and from measurements of the refractive index and cell gap.

35.3: Fringe Field Effects in Microdisplays

Two-dimensional simulations of the liquid crystal behaviour are used to predict the influence of fringe fields on the optical performance of current and upcoming fast LCoS microdisplays based on the vertically aligned nematic mode. Results are demonstrated on LCoS devices using inorganic alignment layers.

34.1: Invited Paper: Microdisplays with High Pixel Counts

Recent developments in high-resolution microdisplays are discussed. The importance of resolution limiting factors such as diffraction of light, fringe fields, voltage requirements and stepper field size is shown. Difficulties related to stitching and large-area CMP are reviewed, Parallelism in the driver design is discussed and the first results of a 2560×2048 pixel microdisplay are presented.

Measurement methodology for vertically aligned nematic reflective displays

We developed a measurement method for the characteristics of microdisplays specifically aimed at vertically aligned nematic reflective cells. It allows determination of contrast ratio and cell gap, and gives good estimates for the pretilt angle and the elastic surface-coupling constant. The set-up consists of a laser source, high quality polarisers, a beamsplitter mirror, a quarter-wave plate and a sensitive photodiode. A model for the polarization changes in the light caused by each component allows the extraction of the initial phase retardation induced by the cell and gives a first estimate of the thickness. Simulation of the director configuration in liquid crystals is then used to enhance the accuracy by taking into account the properties of a real LC cell. Matching of the simulation and the measurements yields the required values together with a calibrated simulation model.