Radiant energy – Invisible radiant energy responsive electric signalling – Infrared responsive
Reexamination Certificate
1999-06-01
2002-07-02
Anderson, Bruce (Department: 2881)
Radiant energy
Invisible radiant energy responsive electric signalling
Infrared responsive
C250S458100
Reexamination Certificate
active
06414313
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention generally relates to high-resolution electron beam systems and, more particularly, to electron beam projection lithography tools having variable beam current.
2. Description of the Prior Art
Many areas of technical endeavor have, in recent years, required increasing levels of dimensional accuracy and resolution. In particular, the field of integrated circuit manufactures has manifested increasing need for increased density of integration and reduced device (e.g. transistors, capacitors, interconnects and the like) dimensions. Reduced device dimensions and increased device proximity provides advantages in both economy of manufacture and level of performance and functionality of integrated circuit chips.
These advantages can readily be appreciated when it is considered that the reduction of size and spacing between devices on a chip allows more devices to be formed on a chip or wafer of given size by the same sequence of process steps having a substantially fixed cost. At the same time, increased proximity of devices reduces signal propagation time and the minimum cycle time of operation of the integrated circuit while reducing susceptibility to electrically coupled noise. Reduced device size also usually requires that the integrated circuit operate at a lower voltage which, while operating margins may be reduced, reduces power dissipation requirements for the chip. The reduced operating margins are of relatively less importance in view of the reduced susceptibility of noise coupling.
While semiconductor manufacturing processes have become very sophisticated and currently provide for production of structures at sizes much smaller than can be directly produced lithographically (e.g. gate sidewalls), at least one lithographic process is invariably required to define the device position and fundamental dimensions. As is well-understood in the art, a lithographic process includes the patterned exposure of a resist so that portions of the resist can be selectively removed to expose underlying areas for selective processing such as by etching, material deposition, implantation and the like. For current designs of integrated circuits, these dimensions have been reduced well below the resolution capability of lithographic processes utilizing radiant energy for selective exposure of the resist.
As an alternative to radiant energy (or, for example, X-Rays, which present other technical problems and/or limitations), charged particle beams have been used for high resolution lithographic resist exposure. In particular, electron beams are particularly favored since the low mass of electrons allows relatively accurate control of an electron beam at relatively low power and relatively high speed. (For this reason, all references to electron beams or electron beam systems hereinafter should be understood as references to a preferred charged particle beam arrangement without exclusion of or limitation as to any form of the latter, more general, class of systems.)
So-called probe-forming systems form a well-focussed spot at the target surface for exposure of the resist. “Gaussian beam” systems, as the name implies, use a spot of Gaussian cross-section and either vector-address or raster-scan the beam to directly write the circuit of interest. Alternatively, shaped-beam systems have higher throughput which is accomplished by parallel pixel exposure. A square shaping aperture is uniformly illuminated and imaged to another aperture, the size of which matches the image of the shaping aperture. The image of the shaping aperture is deflected onto the lower aperture and the compound image is then projected to the target (e.g. wafer). The Gaussian systems project one pixel at a time while the shaped beam systems can expose many pixels in parallel although the number of contiguous pixels concurrently exposed is relatively small.
For example, consider a rectangle of dimensions 0.1×2.0 micrometers. Using a Gaussian beam with a 0.05 micrometer feature size, the rectangle corresponds to forty pixels. A shaped beam system with a maximum spot size of 1.0 micrometer square can expose this rectangle in two exposures.
In general, a single exposure for a shaped-beam system is limited to a few hundred pixels, at most, while the full pattern required for a full integrated circuit may include hundreds of millions of pixels or more. Therefore, the throughput of probe-forming exposure tools, even of the shaped-beam type, is too low to be economically feasible for high density, large scale integrated circuits even though exposures can be made at relatively high speed.
To simultaneously expose much greater numbers of pixels with each exposure, electron beam projection lithography tools have been developed. These tools expose relatively large (e.g. 1 mm square) subfields formed on a reticle which may include an extremely complex pattern. Each subfield can include several million pixels or more and the number of sequential exposures which must be made for each chip is correspondingly reduced. Further, the sub-field image from the reticle is demagnified by the optics of the electron beam column to decrease the feature size and spacing below that produced on the reticle. However, this process and the resulting resist pattern is limited by the ultimate resolution of the electron beam system.
Specifically, a parameter of any charged particle beam column which relates to resolution is the numerical aperture or semi-angle of the particle trajectories of the beam. The beam is typically blanked onto an aperture, referred to as a beam-limiting aperture, at an image of the source cross-over. This aperture is usually smaller than the extent of the beam, causing the beam to be trimmed, and resulting in a beam particle density of Gaussian shape, with the tails truncated, hereinafter referred to as “truncated Gaussian” in shape.
As described in the prior art, electron beam projection lithography systems have a contrast aperture located at the front focal plane of the projection lens. Additionally, this is a plane where an image of the source cross-over occurs, conjugate to the cross-over at the blanking aperture. The width of this truncated Gaussian beam at the plane of the contrast aperture, divided by the focal length of the projector lens yields twice the numerical aperture. The numerical aperture is controlled by the diameter of the beam-limiting aperture.
The resolution of the beam is dependent on many factors including geometric aberrations and Coulomb interactions. Geometric aberrations include chromatic aberrations and spherical aberrations which vary as the numerical aperture to the first and third powers, respectively. The Coulomb interactions have three distinct components: space charge which can be corrected by refocussing the beam of electrons, the Boersch effect which adds energy spread to the beam and contributes to chromatic aberrations, and the stochastic trajectory displacement. The stochastic effect depends on beam energy, length of electron path and numerical aperture as well as beam current.
For a fixed beam column design (length) and beam energy, a trade-off exists between beam current and numerical aperture to balance the effects of the geometric aberrations and the stochastic Coulomb effect. All projection lithography systems to date use a single numerical aperture. This numerical aperture is generally optimized at the maximum target beam current so that the resolution is optimized as well as maximizing the throughput. However, a recent publication by W. Stickel (J. Vac. Sci. Technol., Vol.16(6), November/December, 1998, p. 3211) has analyzed the effect of the stochastic Coulomb interactions on resolution as a function of current, numerical aperture and other parameters. The current dependent loss of resolution varies in the range of about I
(0.6)
-I
(0.9)
and the numerical aperture dependent loss of resolution varies as 1/NA
(0.5)
-1/NA
(0.7)
, depending on the theoretical model and the assumptions used.
However, the target beam cu
Gordon Michael S.
Kendall Rodney A.
Pinckney David J.
Anderson Bruce
Whitham Curtis & Christofferson, P.C.
LandOfFree
Multiple numerical aperture electron beam projection... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Multiple numerical aperture electron beam projection..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Multiple numerical aperture electron beam projection... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2913994