Stepper (semiconductor technology)
In semiconductor technology, a stepper (also wafer stepper) is a type of system or a functional principle for the photolithographic structuring of a photoresist layer , one of the most important sub-processes in the complex production of integrated circuits , also known as microchips. The most important feature of steppers is the exposure of the wafer in several identical steps with a mask . Wafers are thin semiconductor wafers on the surface of which the integrated circuits are built. Further exposure methods are 1: 1 exposure and exposure using scanner systems.
background
The exposure systems used in semiconductor technology have the task of transferring the structures on a photomask ( reticle ) into an applied layer of photoresist ( resist ). This structured photoresist layer then serves as a contact mask for subsequent processes, such as the etching of the underlying material or a selective coating. Important criteria for this transfer are the highest possible structural fidelity, that is, how precisely the 2D geometry of the mask is transferred, and high positioning precision relative to previous structuring levels, cf. Overlay (semiconductor technology) .
In the early days of microelectronics up to the end of the 1970s, this structure transfer (the exposure of the photoresist) was carried out in a full-pane exposure. The photoresist layer on the entire wafer (at that time with a diameter of at most 100 mm) was exposed in one step. The photomask was as large as the wafer, and the structures on the mask were as large as the desired structures on the wafer (1: 1 projection exposure). With the constant downsizing of the structures on the wafer and the use of larger wafers, problems arose, for example the production of the structures on the mask and also the mask size.
For these reasons, new concepts with two production steps were developed by David W. Mann in Burlington USA as early as the early 1960s. GCA (Geophysical Corporation of America) buys David Mann in 1959 and becomes active in Kreuzlingen, Switzerland in 1965. The stepper principle became industrial in Europe after 1961 (first sale to Clevite Corp. with a NIKON optics) at IBM, Valvo, Philips, Siemens, Mullard, SGS Ates, GI, Texas Instruments and other semiconductor manufacturers as well as in research at the University of Aachen. A typical reduction in the David Mann Repeater is a factor of ten; ZEISS Objective now achieve better results. Before a David Mann photorepeater is used, a drawn mask on a 100x100mm glass plate is exposed to the emulsion in the David Mann Pattern Generator 3000. The accuracy of these early systems was based on two glass scales that were moved one above the other by a high-precision spindle. The translucent light signals generated interference impulses and, evaluated and interpolated with a DIGITAL PDP-8, enabled the positioning accuracy of 1µ to be achieved with the X and Y servomotors. Later models of the David Mann Pattern Generator 3600 had a larger (150 x 150 mm) even more precise XY table and were controlled with two HP laser measuring devices; a DIGITAL PDP-11 was required here. The GCA 4800 DSW (Direct Step on the Wafer) was used in industry in the USA, Japan and Europe from 1979.
Step-and-repeat principle
In contrast to the previously mentioned full-pane exposure, with the stepper the structures of the photo mask are not transferred to the entire wafer in one step. Instead, a certain section of the complete layout, for example the structures of a single chip or a few (2-8) chips, is successively transferred to different positions on the wafer. This transfer (Engl. In increments steps ) gave the procedure its name. Strictly speaking, it is based on the step-and-repeat principle.
The use of the stepper principle has a number of advantages compared to full-disk exposure: Reduction optics (often 5: 1) could now be used, because the maximum size of the masks that can still be handled remained the same. Larger structures on the masks also mean low demands on the mask itself, which is expressed, among other things, in lower production costs. In addition, defects caused by particles in the optical system or on the mask have become less critical, since most of the particles are not in the focus of the imaging system and are therefore not imaged in focus and they are also reduced in size. The use of light sources of ever shorter wavelengths (436 nm , 365 nm, 248 nm up to 193 nm in 2011) enables it, in combination with further improvements in the exposure systems (e.g. immersion lithography ) and resolution- improving techniques (RAT), structure widths up to down to 32 nm and less.
Step-and-scan principle
Exposure systems based on the step-and-scan principle work in a similar way to systems based on the step-and-repeat principle. In each step, only part of the entire wafer is exposed and the mask is imaged in a reduced size by the optical system (generally 4: 1). The difference between the two methods lies in the exposure of the section. In contrast to steppers with the step-and-repeat principle, the mask is only illuminated in a narrow strip and driven under this light strip, similar to what happens with line scanners or photocopiers . Systems that use this principle are often only referred to as scanners. Exposure systems with a 1: 1 projection exposure based on the scanner principle have not been used at the front end since the mid-1980s, but have been used again in the back end for several years due to the high throughputs in the 1-3 µm resolution range . The step-and-scan principle has been the preferred exposure principle in the manufacture of modern integrated circuits since the mid-1990s.
See also
literature
- Chris Mack: Fundamental Principles of Optical Lithography: The Science of Microfabrication . 1st edition. John Wiley & Sons, 2007, ISBN 978-0-470-01893-4 .