One of the main issues for the simulation of MOS transistors is the correct prediction of threshold voltages that depend on the active doping profiles in the channel under the gate oxide. Simulating a power MOS process we encountered a situation in which Sentaurus Process with default models failed to predict threshold voltages by as much as 3 V. An in-depth investigation revealed that the threshold voltage in our pMOS devices is determined by a very special distribution of the doping in the channel that involves both n-type and p-type doping which nearly compensate each other. As threshold voltages were found in the simulations to be particularly sensitive to boron segregation, silicon samples were implanted with boron and oxidized in several atmospheres for a variety of process times.The profiles were studied by advanced SIMS methods. Because of the limitations of the SIMS depth resolution, they had to be complemented by electrical measurements on MOS transistors. This combination finally allowed finding a new calibration for the segregation models which allows predicting the electrical characteristics of the transistors in a wide range of experimental conditions. Since the threshold voltage in our transistors turned out to be extremely sensitive to the boron segregation parameters, in contrast to technologies in which only one dopant type prevails, the newly achieved calibration should be superior to previous work. (C) 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Angular distributions of silicon atoms sputtered by gallium ions at grazing incidence were investigated experimentally and by simulation. The energies and the ion beam fluence studied are typical for using focused ion beam techniques for silicon micro-structuring. The angular distributions of the sputtered atoms at grazing ion beam incidence loose their cylindrical symmetry around the target surface normal due to the anisotropy of single collisions and of the collision cascades initiated by the energetic ions. The angular distributions were studied in this work both experimentally and using the simulation methods of Monte-Carlo and of Molecular Dynamics. To study the influence of the surface structure on the angular distributions of the sputtered atoms, different arrangements of silicon atoms on the surface of the targets were tested in simulations using Molecular Dynamics. Monte-Carlo simulations based on the theory of binary collisions were used to complement the experimental results on the angular distributions of sputtered silicon and the final results are presented as an analytical model.

For a realistic simulation of sputtering processes and the topography changes associated, we combined the 3-D topography simulator ANETCH with the Monte-Carlo ion implantation program MC_SIM. The coupling between the programs provides the possibility to study the results of physical sputtering processes for nearly arbitrary ion/target combinations without a priori knowledge about the respective yield from experiments. As a first application, simulations were carried out to optimize process parameters of sputtering experiments. In a second application, the topography of a trench after FIB preparation is compared to simulations. The side-wall evolution at an edge due to ion irradiation is studied as a third application.

Ion sputtering of germanium was studied within this work. Argon and nitrogen ions with energies of 20keV were used for the sputtering experiments. Grazing ion incidence angles in the range of 60–85° were examined experimentally and by Monte-Carlo simulations. The range of ion incidence angles was chosen in order to observe a contribution of primary knock-on atoms to the total angular distribution. Experiments were performed using the collector technique. The collectors, 300mm silicon wafers, were covered with a sub-monolayer concentration of germanium and were examined afterwards with spatial resolution using total X-ray reflection fluorescence (TXRF). Initially, simulation results significantly deviated from experiments. Therefore, the model of ion penetration into a target was modified, resulting in a better agreement between simulation and measurement.

We demonstrate the coupling of Monte Carlo sputter simulation with feature-scale simulation of profile evolution during sputter etching. With the Monte Carlo sputter simulation, the dependence of the sputter yield on the angle of incidence and on the energy of ions impinging onto the surface is determined. The yield curves obtained thereby are fed into a feature-scale etching profile simulator which predicts the local etch rates based on these sputter yield curves and on ion fluxes which are calculated for a substrate placed in a sputter reactor. For validating the simulations, a process sequence consisting of deposition and back etching (in an argon plasma) of silicon oxide between metal lines has been studied. Assuming an ion energy of 250 eV allows us to consistently reproduce profiles observed experimentally.

Angular distributions of sputtered germanium atoms during the grazing incidence of argon ions were investigated. Argon ion beams with angles of incidence of 75 degrees and 80 degrees and with ion energy of 20 keV were used in sputtering experiments. TXRF analysis showed to be an excellent method for measuring the angular distribution of sputtered atoms due to its low detection limit under total reflection conditions. The experimental data are compared to Monte-Carlo simulations. The observed differences are discussed, and suggestions for improving the Monte-Carlo simulation of ion sputtering are made. (C) 2008 Elsevier B.V. All rights reserved.

Angular distributions of ion sputtered germanium and silicon atoms are investigated within this work. Experiments are performed for the case of grazing ion incidence angles, where the resulting angular distributions are asymmetrical with respect to the polar angle of the sputtered atoms. The performed experiments are compared to Monte-Carlo simulations from different pro-rams. We show here an improved model for the angular distribution, which has an additional dependence of the ion incidence angle.

Two monoclonal antibodies (MAb996 and MAb994) were produced by immunisation with a synthetic peptide with a sequence based upon that of the protein core of the gastrointestinal MUC2 mucin. The epitopes were identified as T G T Q for MAb996 and P T G T Q for MAb994. Antibody competition tests also confirmed the overlapping nature of the epitopes for the two antibodies. MAb994 and MAb996 were employed in immunoadsorbent columns for the fractionation of human colorectal carcinoma tissue extracts. While the two antibodies displayed only relatively minor differences in immunological specificity and affinity for the immunising synthetic MUC2 mucin core related peptide, they had the capacity to separate antigenically distinct molecules when used as immunoadsorbents. The findings indicated that subfractions of MUC2 antibody-defined mucins exist in human carcinomas and that these may be distinguished by the differential exposure of determinants in the mucin protein core. The results are in accord with the view that aberrant patterns of glycosylation of mucins in human intestinal tumours produces a spectrum of variably glycosylated macromolecules.

Monoclonal antibodies against the protein core of epithelial mucins have been found to react with the immunodominant sequence P D T R P A P (Burchell et al., 1989; Price et al., 1990a). Two immunoadsorbent matrices were prepared by linking the peptide A P D T R P A P G to CNBr-activated Sepharose and by linking the peptide C A P D T R P A P G to activated thiol-Sepharose, so that each immunoadsorbent contained the immunodominant motif. Anti-epithelial mucin antibodies (anti-breast carcinoma antibodies, anti-purified mucin antibodies and anti-human milk fat globule antibodies) were examined for reactivity with these preparations. The initial tests indicated that the substituted CNBr-activated Sepharose displayed lower non-specific antibody binding and this matrix was selected for further investigation. The anti-mucin antibodies were shown to react specifically with this affinity matrix and irrelevant antibodies failed to bind. A Sepharose-peptide immunoadsorbent column was examined for its capacity to purify several of these anti-mucin antibodies and it was determined that this procedure was highly efficient - purified IgG and IgM antibodies could be isolated from either hybridoma tissue culture supernatants or ascitic fluids. The capacity of the column was in excess of 40 mg antibody protein per ml of gel for the IgG3 antibody, C595 (anti-urinary mucin) and at least 10 mg antibody protein per ml of gel for the IgM antibody, NCRC-11 (anti-breast carcinoma).The procedure described permits the efficient purification of anti-mucin antibodies and provides a product which would be suitable for further investigations requiring highly immunoreactive antibodies (e.g., for radioimmunotherapy or immunoscintigraphy in patients with malignant disease).

Recombinant single-chain fragments (scFv) of the murine anti-MUC1 monoclonal antibody C595 have been produced using the original hybridoma cells as a source of variable heavy (V(H))- and variable light (V(L))-chain-encoding antibody genes. The use of the polymerase chain reaction (PCR), bacteriophage (phage) display technology and gene expression systems in E. coli has led to the production of soluble C595 scFv. The scFv has been purified from the bacterial supernatant by peptide epitope affinity chromatography, leading to the recovery of immunoreactive C595 scFv, which was similar in activity to the C595 parent antibody. Analysis by DNA sequencing, SDS-PAGE and Western blotting has demonstrated the integrity of the scFv, while ELISA, FACScan analysis, fluorescence quenching, quantitative immunoreactivity experiments and immunohistochemistry confirm that the activity of the scFv compares favourably with that of the parent antibody. The retention of binding activity to MUC1 antigen on human bladder and breast carcinoma tissue specimens illustrates the potential application of this novel product as an immunodiagnostic and immunotherapeutic reagent.

This report is an analysis of data relating to the epitopes of 28 murine monoclonal antibodies reactive with the protein core of human carcinoma-associated MUC1 mucins. All anti-MUC1 antibodies define epitopes of linear sequences of 3, 4 or 5 amino acids within the hydrophilic domain, APDTRPAP, which is expressed multiple times in a highly conserved 20 amino acid repeat sequence of the MUC1 core. The R residue is present in the epitopes defined by all of the 28 anti-MUC1 monoclonal antibodies. Epitopes of antibodies originally prepared against immunogens containing human milk fat globule membranes include the motif DTR in over 90% of the examples studied.

Monoclonal antibodies have been prepared against a synthetic peptide with a sequence corresponding to a repeated hydrophilic region of the protein core of the human MUC-2 gastrointestinal mucin. Peptide conjugates, prepared by glutaraldehyde cross-linking with keyhole limpet haemocyanin (KLH) and bovine serum albumin (BSA), were employed as the immunogen and target antigen (for screening by ELISA), respectively. However, for the measurement of antibody binding to peptide by an ELISA procedure, an alternative strategy was developed and is described in this report: peptides were conjugated directly to BSA immobilized by physical adsorption to the surface of microtitre plate wells. This procedure permits peptides to be tested as target antigens by ELISA without prior preparation of peptide-carrier conjugates.

A strategy for directing and enhancing B cell immune responses against synthetic peptide determinants has been developed in order to produce antibodies specifically against protein epitopes of clinical relevance. A peptide sequence based upon the MUC-1 mucin protein core was selected for this purpose since anti-MUC-1 antibodies have proven diagnostic application and therapeutic potential in human breast and ovarian cancer. Peptide constructs were synthesised co-linearly linking the immunodominant B cell determinant region, PDTRPAP, in the protein core of the MUC-1 mucin, to sequence 111-120 of influenza haemagglutinin A/X-31, a determinant recognised by T helper cells through association with MHC class II molecules. Induction of anti-MUC-1 antibodies to the B cell determinant region by immunisation with peptide was shown to be dependent upon both the presence and the position of the T cell determinant. In addition, haplotype mismatching with respect to the T cell determinant resulted in a significant lowering of the anti-MUC-1 antibody response in peptide construct immunised mice. These findings are relevant to the design of immunogens to produce antibodies against peptide epitopes of tumour associated proteins and glycoproteins.

Human polymorphic epithelial mucin is a high-molecular-mass glycoprotein that associates to provide protection to the epithelial-cell surface and may afford the malignant cell a selective advantage for growth. The scanning-tunnelling-microscopy micrographs obtained in the present study identify the purified human ovarian-carcinoma polymorphic epithelial mucin glycoproteins as rod-shaped molecules of mixed length. The dimensions of the individual molecules range from 25 to 45 nm in length and are 3-4 nm in width. The images further suggest that lateral association of the rods occurs.

A murine anti-(human gastric carcinoma) monoclonal antibody, GL-013 (IgG1), which reacts with a high-molecular-mass glycoprotein from colorectal tumour tissue [Yang and Price (1989) Anticancer Res 9: 1707], was examined for reactivity against a panel of purified and partially purified antigens associated with tumours of the gastrointestinal tract. These included carcinoembryonic antigen (CEA), normal cross-reacting antigen, Y-hapten glycoproteins, and perchloric acid extracts and glycolipid preparations from colorectal tumours. While the GL-013 antibody failed to bind to these antigens, it was found to react strongly with synthetic peptides with sequences based upon that reported for the protein core of a human gastrointestinal mucin [Barnd et al. (1989) Proc Natl Acad Sci USA 86: 7159; Gum et al. (1989) J Biol Chem 264: 6480]. In control tests, a series of other anti-(colorectal tumour) antibodies (IgG1 and IgG3), with broad reactivity towards gastrointestinal carcinomas, as well as an anti-CEA antibody, (IgG1) failed to react with the synthetic peptides. It is concluded that the anti-(gastric carcinoma) monoclonal antibody GL-013 binds to a threonine-rich peptide epitope expressed within the protein core of gastrointestinal mucins.