Kiichi Takamoto

Design method for an electrostatic einzel lens having an asymmetric structure

The guiding principle for designing an asymmetric einzel lens is obtained on the basis of an examination of the axial potential distribution in symmetric einzel lenses with lower aberrations. The asymmetric axial potential distribution resulting in lower aberrations is expressed by four quadratic curves. After the asymmetric distribution is optimized for the required optical properties, the lens structure to achieve it is determined with only a few calculation cycles. As an example, an asymmetric einzel lens is designed under the condition that the aperture angle is 5 mrad, the beam energy spread is 1×10−4, and the working distance is 10 mm. The lens has a total axial aberration as small as 0.038 μm.

Electron ray tracing with high accuracy

An electron ray tracing program is developed to investigate the overall geometrical and chromatic aberrations in electron optical systems. The program also computes aberrations due to manufacturing errors in lenses and deflectors. Computation accuracy is improved by (1) calculating electrostatic and magnetic scalar potentials using the finite element method with third‐order isoparametric elements, and (2) solving the modified ray equation which the aberrations satisfy. Computation accuracy of 4 nm is achieved for calculating optical properties of the system with an electrostatic lens.

Pattern Fidelity in Submicron Lithography with a Rectangular Electron Beam

The energy density profiles deposited in a PMMA resist film by a 0.5 µm square electron beam are calculated by Monte Carlo simulation. The influence of a combination of beam accelerating voltages and beam edge widths on the pattern fidelity is studied by examining some characteristic parameters. Simulation shows that one of these parameters, the energy efficiency, has a tendency to saturate above 50 kV, and that larger beam edge widths are allowable for higher voltages in obtaining the desired values for the characteristic parameters, which are exemplified by a normalized energy edge slope and the roundness of the square pattern. A combination of voltages of 30–50 kV and a beam edge width of <0.2 µm is desirable in 0.5 µm lithography. High-fidelity 0.5 µm patterns are obtained experimentally with an electron beam system employing writing strategies based on this simulation.

Distortion correction and overlay accuracies achieved by the registration method using two‐stage standard mark system

A registration method using the two‐stage standard mark system is developed for high accuracy overlaying of patterns on a deformed wafer in an electron beam lithographic system used for direct wafer writing of 0.5 μm very large‐scale integrated (VLSI) patterns. Deflection distortions are corrected according to height z of a pattern writing field on a wafer. In this paper, deflection distortion correction and overlay accuracies are examined experimentally. Main field (2.6×2.6 mm) distortions are corrected using the third‐order power function of the main field coordinates (X, Y). Correction coefficients related to shift, deflection gain, and deflection rotation are shown to be a linear function of z, in spite of varying the beam focusing level with z. Each distortion expressed in terms up to the third‐order power of (X, Y) can be corrected with an accuracy within 0.02 μm. Overlay accuracies within 0.12 μm (3σ) are yielded when two‐layer test patterns are exposed on a wafer deformed about 75 μm in the Z dire...

Feasibility of Reflection Electron Energy Loss Spectroscopy with a Small Scattering Angle for Research in Surface Science

A reflection electron energy loss spectroscopy (R-EELS) system has been developed to investigate local surface atomic structures around light atoms such as C, N, and O. In this system, the primary electrons are incident on a surface at a grazing angle, and the electrons scattered inelastically at a small scattering angle are energy-analyzed to make the momentum transfer from the primary electrons to the scattered ones as small as possible. With this system, one can measure an energy loss near edge structure (ELNES) and an extended energy loss fine structure (EXELFS), compared with an X-ray absorption near edge structure (XANES) and an extended X-ray absorption fine structure (EXAFS). The feasibility of the system in surface science is examined by measurements of ELNES and EXELFS for oxygen-adsorbed Ni(100) surfaces.

Dependence of minimum linewidth on electron‐beam properties in submicron lithography using a rectangular beam

The relation between beam properties and minimum linewidth for a rectangular electron beam for submicron lithography is obtained through experiments and electron scattering simulation. Patterns are exposed in a 0.5‐μm‐thick PMMA resist film on a silicon wafer. The result gives a guiding principle in designing an electron‐optical column of a rectangular electron‐beam exposure system. A minimum linewidth of 0.5 μm can be obtained using a beam with a beam edge width δ=0. 18 μm at a beam voltage Va=25 kV and 0.2‐μm minimum linewidth using a beam with δ=0.08 μm at Va=30 kV.

Imaging in an optical projection system with a light source composed of point sources

The effects of a points source on imaging characteristics are examined using a one-dimensional model. Intensity distributions for a line-and-space pattern with 5 lines are calculated and the imaging characteristics are evaluated, compared with those obtained using a plane-source system. In a system with a light source composed of 2 point-sources, the focal tolerance determined from linewidth and contrast becomes maximum. Using a light source composed of 3 point-sources, the line-and-space pattern is imaged with precise linewidths while maintaining a focal tolerance slightly larger than that for the plane-source system.

Imaging in an Optical Projection System with a Laser Light Source. II: Multi-Transverse-Mode Laser Light Illumination

Imaging characteristics are analyzed using a one-dimensional model. The second source plane in an illumination system is illuminated by a laser light oscillating in a multitude of transverse modes. Intensity distributions for a line-and-space pattern with 5 lines are calculated. Focal tolerance is examined for intensity distributions due to the light oscillating in only even transverse modes. No linear relationship between the number of transverse modes and focal tolerance is observed. When a phase of the light field for each even transverse mode is randomly given, contrasts of intensity distributions do not deteriorate with defocus.

Stitching accuracy in a variable‐shaped electron‐beam exposure system

Strategies for stitching rectangular electron beams or deflection fields, and measurement results of field stitching accuracy are described. A beam shaping aperture is newly developed to stitch rectangular beams with high accuracy. The aperture edges are defined by fine gold wires under tensile stress. Rectangular beams projected using this aperture are estimated to have an edge roughness within 0.01 μm and an orthogonality within 1 mrad. An x‐y table system is fabricated so as to reduce each of the beam positioning errors due to table movement to 0.02 μm or less. Deflection distortion is measured, using a two‐stage standard mark system mounted on the x‐y table, and corrected. A test pattern is then written and field stitching accuracy evaluated. About 90% of the measured stitching accuracies are below 0.05 μm for a 1.5×1.5 mm field and about 50% for a 2.5×2.5 mm field.

Optimization of system parameters for throughput in a variable rectangular electron beam exposure system

Optimization of system parameters for throughput is investigated under the condition that performances of an electron optical system and an X–Y table system, and large‐scale integrated (LSI) pattern properties are given. The beam is deflected by a vector scanning method. The table is operated in a step and repeat mode. Exposure time is calculated from resist exposure time, settling time of beam deflection digital–analog converters, and table movement time. It is shown that there exist field sizes, aperture angles, and maximum beam sizes which minimize the exposure time. Optimum system parameters depend on LSI pattern properties. The concept of optimizing system parameters for throughput can be applied to system parameter determination in the development of a high throughput exposure system.