|Publication number||USRE38900 E1|
|Application number||US 09/273,171|
|Publication date||Nov 29, 2005|
|Filing date||Mar 19, 1999|
|Priority date||Aug 8, 1995|
|Also published as||US5663076|
|Publication number||09273171, 273171, US RE38900 E1, US RE38900E1, US-E1-RE38900, USRE38900 E1, USRE38900E1|
|Inventors||Michael D. Rostoker, Nicholas F. Pasch, Ashok K. Kapoor|
|Original Assignee||Lsi Logic Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (15), Non-Patent Citations (2), Referenced by (8), Classifications (13), Legal Events (1)|
|External Links: USPTO, USPTO Assignment, Espacenet|
1. Field of the Invention
The present invention relates generally to systems for integrated circuit manufacture. More particularly, the present invention relates to a system and method for automating the photolithography procedure utilized in fabricating integrated circuits and for automating the manufacture of integrated circuits having an optimal feature size.
2. Description of the Related Art
The manufacture of integrated circuits is a complex process which combines the technologies of photolithography, physics and chemistry. In one common method of semiconductor manufacture, wafers of pure silicon are coated with thin layers of photo-resist. Each coated wafer is then exposed to a light source which is projected through an etched mask layer imaged adjacent to the wafer. The etched mask layer only passes light to selected regions of the wafers, resulting in the exposing of the resist at those selected regions. The exposed regions of the photo-resist are removed, opening small windows of silicon on the surface of the wafer. These windows enable doping impurities to be diffused or deposited onto the exposed wafer regions. Following exposure of the windowed regions of the wafer to the doping impurities, the resist is completely removed from the wafer and the process is repeated using additional layers of mask levels. Additional photo-resist layers may be subsequently used to selectively mask the wafer surface for further processing such as etching, formation of interconnect lines, and the like. Using this lithographic-diffusion technique, very complex systems of electronically active devices can be accurately produced. From start to finish it is not unusual for six to twenty mask levels to be used in the wafer manufacturing process.
Important advantages are achieved with semiconductor devices by making the individual electronic device features as small as possible. The smaller the individual devices are, the more devices that can be put on a single IC wafer. Higher densities, therefore, translate into lower materials cost for individual components. A further significant benefit of this reduced materials cost results from the fact that discrete defects on the silicon substrate randomly exist across the wafer. When the individual circuit dice consume smaller areas of substrate, the probability of silicon defects per unit die is decreased, thus resulting in higher yields and lower per unit cost. In addition to savings in manufacturing costs achievable with reduced device sizes, device speeds are increased and power is reduced per unit device as the devices become smaller.
Thus, what is needed, and one problem which is addressed by the present invention, is to design circuits having the smallest possible device feature sizes for an available manufacturing process.
A second problem that is being faced in the electronics industry, is the need for increasingly faster design and production cycles for semiconductor devices. The economies of integration, which can be achieved by utilizing customized semiconductor devices in electronic systems, makes it highly desirable to design application specific circuits where possible. One limitation to using customized circuits in electronic products is that the amount of time required to design and produce these integrated circuits makes them impractical for many commercial applications. Often, producers of electronic goods will design a product using conventional discrete or off-the-shelf components, and then over time begin to integrate and customize the circuitry as revisions to the product are produced. There may be a time in the not-too-distant future when a system will enable the circuit designer to enter a schematic of a required electronic device, and have a tabletop unit sitting next to the computer begin immediately generating finished integrated circuit products, much in the same way a printer reproduces a paper document. Today, however, the complexities of the device physics and chemistry require the use of complex lithography systems and highly controlled chemical process ovens.
What is further needed, therefore, is a system and method for automating photolithography procedures used in the fabrication of integrated circuits.
In accordance with the present invention, a system enables the automated photolithography of semiconductor integrated circuits. The system is controlled as a processor which is coupled to a Rayleigh derator, a form factor generator, a logic synthesizer, a lithography module and a wafer process. A Rayleigh processor receives light source and numerical aperture information from the lithography module, and also receives manufacturing process data from the wafer process. The light source information, the numerical aperture data and the manufacturing information are combined together to derate the theoretical minimum resolvable feature size which may be manufactured using a wafer process. This derating consists of a combined consideration of both theoretical limitations produced by the lithography equipment, as well as measured results from manufacturing variations occurring in the wafer process. The Rayleigh processor communicates this minimum resolvable feature size to the form factor processor, which then uses standard sizing models to determine the minimum device size for each transistor in a circuit design net list. The form factor processor receives IDS (drain source current), VGS (gate to source voltage), and gate length information from the circuit net list and calculates a corresponding minimum manufacturable gate width which can be used to satisfy the design requirements. Once all transistors in the net list have been sized, a logic processor produces a physical design for production of a photolithographic wafer mask set. Following production of the wafer mask set, wafers are then manufactured in the wafer process. Manufacturing and yield data from the processed wafers is then collected and used to subsequently update the Rayleigh processor. In this way, the present invention is achieved: a direct coupling between the measurement of wafer process parameters and the automated sizing of semiconductor devices. Such invention enables the production of circuits having the smallest manufacturable device sizes available for the given photolithography equipment and wafer process.
Although the preferred embodiment relies on a single computer to control the automated system of the present invention, an alternative embodiment utilizes multiple computers or processing facilities to control various aspects of the system. For instance, in one alternative embodiment, Rayleigh derating is controlled using a first computer with the minimum feature size being stored. The minimum feature size is then conventionally communicated to a second computer which controls form factor generation and layout generation. A third computer controls the etching of wafer masks. A fourth computer monitors the wafer process and collects test results data for communication back to the Rayleigh derating performed by the first computer.
FIGS. 6(a) and 6(b) comprise a flow chart of the preferred method of the present invention for automating photolithography in the fabrication of integrated circuits.
Referring now to
The function of the IC design system 100 is to enable an integrated circuit designer to size and process integrated circuits as an automated function, beginning with the circuit net list and producing processed wafers as an end product. For purposes of the discussion below and the appended claims, the term “manufacture” is defined to generally comprise this process of designing and producing integrated circuit wafers. One of the several advantages of the present invention is the ability to automatically measure process parameters of the wafer process 125 and to use these process parameters to automatically calculate the minimum feature size which may be designed using the lithography module 138. Use of the automated system 100 is particularly well suited for devices having feature sizes (i.e. transistor gate widths) of less than approximately 0.3 microns, since the impact of even small process variations becomes significant at these reduced geometries. This minimum feature size, attainable in the lithography module 138, is then used to size transistors stored in a net list database 123, and then to synthesize and layout the logic necessary to produce the designed circuitry.
Rayleigh derator 128 is used to determine the minimum IC device feature size that can be practically manufactured using a specific lithographic light source 140 in combination with the various process tolerances of wafer process 125. The Rayleigh derator 128 begins operation by determining the theoretical minimum feature size available for a device, based on the light wavelength of light source 140 used in etching the wafer masks 137. The minimum theoretical device size that can be produced by a specific light source 140 is a direct function of the wavelength of the light source 140. The physics for this theoretical determination is well understood and is discussed in the text, Radiation and Optics: An Introduction to the Classical Theory, John M. Stone, McGraw-Hill Book Company, Inc. pp. 144, 180-182. The Rayleigh derator 128 then derates this minimum theoretical limit using image separability considerations produced as a function of the numerical aperture of the objective lens used in the lithography module 138. Once the theoretical minimum distance of line resolution is determined, this resolution is further derated by a manufacturing constant, m, which is empirically derived based on process tolerances extracted from the wafer process 125. Manufacturing factors which effect the derating of the minimum line resolution include conventional variations in mechanical steps used in the wafer process 125, accuracy of temperature controls, and chemical purities of the wafer process 125 materials. To a lesser extent the manufacturing constant, m, is also impacted by the mechanical and electrical stability and alignment of the lithography equipment used in lithography module 138. Other factors which are taken into account in generating the manufacturing constant m, include such effects as the Critical Resist Modulation Transfer Function (CMTF), proximity effects of the photolithographic process, and the modulation transfer function (MTF) of the lithography module 138. These issues are discussed in the text Silicon Processing For The VLSI Era, by Stanley Wolf and Richard Tauber, Lattice Press, pp. 412-413 and 464-467, which is incorporated herein by reference.
Referring now to
The light source library 205 contains a listing of information relating to the wavelengths of various light sources 140 used in the lithography module 138. Typical wavelength values of preferred light sources 140 include enhanced mercury (Hg+) arc sources which have a wavelength of 2560 Angstroms, and krypton fluoride (KrF) lasers which have a wavelength of approximately 2200 Angstroms.
The numerical aperture library 207 contains a listing of the various values for numerical apertures used in the lithography module 138. This numerical aperture is a value assigned to the lithography module 138 objective lens used in focusing the laser beam 133 onto the wafer mask 137 during the mask etching process. This numerical aperture is a function of the specific lithography equipment used in mask production, and varies between light sources 140 and lithography modules 138. Typical values for numerical apertures range from 0.3 to 0.7. In the preferred embodiment, a numerical aperture of 0.5 is typical.
Also attached to the Rayleigh processor 201 is the k Constant register 209. The k constant register 209 contains a derating value which is theoretically determined and represents minimum separability. This number is often referred to as the Rayleigh constant, k, as described in Radiation and Optics: An Introduction to the Classical Theory, supra.
The Rayleigh processor 201 combines the data stored in the m database 211 with the appropriate wavelength value stored in the light source library 205, the numerical aperture value stored in the aperture library 207, and the Rayleigh constant k stored in the constant register 209, to produce a number which represents the minimum resolvable distance (minimum feature size) which can be achieved using the lithography module 138 in combination with the wafer process 125. This minimum resolvable distance is processed according to the equation:
where λ is equal to the wavelength of the light source 140, NA is equal to the numerical aperture, k is equal to the Rayleigh constant, and m is equal to the manufacturing process constant. This minimum resolvable distance represents the minimum feature size that can be reproduced accurately in a semiconductor circuit using the lithography module 138 in combination with the wafer process 125. In MOS circuit manufacturing, this minimum feature size defines the minimum gate length that can be realized for transistor devices. Since speed, power dissipation and manufacturing costs are all a function of this minimum feature size, optimal design requires that, where possible, transistors having the smallest possible gate lengths be used. This minimum feature size for the gate length of an MOS transistor therefore defines a form factor which is used in the design of MOS semiconductors. Using an enhanced mercury (Hg+) arc light source 140 having a light wavelength of 2560 Angstroms, in combination with a reasonably stable wafer process 125, 0.35 micron transistor lengths are achievable. Use of the krypton fluoride laser light source 140 (shown in light source library 205) having a light wavelength of approximately 2200 Angstroms, with highly controlled manufacturing processes, enables the production of 0.25 micron gate length devices.
Following calculation of the minimum feature size in the Rayleigh derator 128, processor 116 (
Referring now to
When the net list is stored in system 100, connectivity information defining each node connection for the devices M1 through Mn is known and stored. However, no device size information is generally available with the net list until processing by the form factor generator 132 occurs. Following determination of the derated minimum feature size and storage of the minimum feature size in feature size memory 130, the form factor generator 132 is used to determine the appropriate transistor size for each of the devices shown in FIG. 3.
Referring now to
For each transistor of the electrical circuit design, form factor processor 401 receives a value for the drain-source current, IDS, and gate-to-source voltage, VGS, from the net list database 123. The form factor processor 401 also reads the minimum feature size from the feature size memory 130, and use this minimum feature size as the default gate length value L. Alternatively, if a non-minimum size L is desired to be used in the electrical circuit design, a preferred value for gate length can be transferred from the net list database 123 to the form factor processor 401 along with the IDS and VGS values for that transistor. Form factor processor 401 calculates transistor gate widths using the conventional sizing formula:
where IDS is defined as the drain-source current, Cgox represents the gate oxide capacitance, W represents the gate width, L represents the gate length, VGS represents the gate-to-source voltage and VT represents the gate threshold voltage. The value for gate threshold voltage VT is read from the VT library 403. Cgox library 405 contains the value used by the form factor processor 401 for gate oxide capacitance. Form factor processor 401 receives the values for IDS, VGS, and L; the processor 401 then reads values of VT, Cgox, and then calculates an aspect ratio (W/L) for each of the transistors identified in the net list database 123. Each value of W which is determined by form factor processor 401 is then stored in the appropriate W column in net list database 123.
Referring now to
Referring now to FIGS. 6(a) and 6(b), a flow chart illustrates the preferred method of automated photolithography in the fabrication of integrated circuits. Beginning in step 604, processor 116 selects a light source 140 from an appropriate lithography module 138. In the IC design system 100, it is contemplated that more than one lithography module 138 may be available. Once the light source is selected 604, the processor 116 transfers information relating to the selected light source 140 to the Rayleigh processor 201 located in the Rayleigh derator 128. The Rayleigh processor 201 then accesses light source library 205 to identify the wavelength associated with the selected light source 140. Once the lithography module 138 is identified, the Rayleigh processor 201 also looks up the value of the numerical aperture used within module 138 in the aperture library 207.
Measurements relating to the manufacturing parameters of wafer process 125 are transferred to the m database 211 in step 602. The measurement of manufacturing parameters 602 may occur just prior to lithography of the wafer masks 137, or the measurements may take place as an ongoing process, with feedback updating measured wafer process 125 data as a function of wafer yield. The Rayleigh processor 201 then reads a Rayleigh constant, k, from the constant register 209 and combines this constant with the manufacturing constant m stored in m database 211. The Rayleigh processor 201 subsequently determines 608 minimum feature size R at wafer process 125. Determination 608 of the minimum feature size is processed according to the equation:
where λ is equal to the wavelength of the light source 140, NA is equal to the numerical aperture, k is equal to the Rayleigh constant, and m is equal to the manufacturing process constant This minimum feature size R is then stored in feature size memory 130, and used by the form factor generator 132 to complete transistor sizing of the circuit design stored in the net list database 123.
Processor 116 reads 610 the net list stored in net list database 123 for the circuit design being processed. Processor 116 then transfers drain-source current values (IDS) and gate-to-source voltages (VGS) to the form factor processor 401. The form factor processor 401 also receives gate length information for each transistor device stored in the net list database 123, not having minimum feature size gate lengths as stored in feature size memory 130. If no gate length information is stored in the net list database 123 for a specific transistor device, then the processor 116 reads the minimum feature size R from the feature size memory 130 and transfers this value to the form factor processor 401 for generation of a device gate width. Along with the gate length, form factor processor 401 reads the threshold voltage (VT) from the VT library 403, the gate oxide capacitance (Cgox) from the Ccox library 405, and reads the mobility constant from the μ memory 407. The form factor processor 401 then generates 612 the appropriate device gate width according to the conventional sizing equation:
where VT is defined as the threshold voltage, Cgox is the gate oxide capacitance, μ is the mobility constant, IDS is the drain-source current, VGS is the gate-to-source voltage, W is the gate width, and L is the gate length.
After the form factor processor 401 generates 612 the transistor's gate width, the processor 401 tests the net list database 123 in step 614 to determine whether additional devices are present in the net list database 123. If additional devices are present, the form factor processor 401 reads 610 the net list database 123 for additional transistor IDS, VGS, and L values to continue the generation 612 of additional gate widths. If no additional devices require processing, the method continues in Step 620 with the generation and storing of the physical design by the logic processor 501. The logic processor 501 reads net list data stored in net list database 123, and in conjunction with user defined parameters stored in user defined constraints library 505, produces a physical design for storage in physical design storage 507. This physical design preferably includes a circuit routing layout which specifies device geometry orientation and interconnect routing of the physical integrated circuit on a silicon substrate. Once the physical design is completed, in Step 622, testing and simulation of the physical design is performed by test and simulation module 509.
The tested physical design is then used to generate a wafer mask 137 set in Step 626. The wafer mask 137 set is generated by processor 116 transferring the physical design from physical design storage 507 to the lithography module 138. This physical design consists of a plurality of mask layers, where each mask layer represents a manufacturing step used in the wafer process 125. Each wafer mask 137 is conventionally etched 626 in the lithography module 138 by the light source 140. Processor 116 drives the lithography module 138 such that the light source 140 etches the wafer mask 137 patterns to correspond to the various layers of the physical design. The wafer masks 137 are then used by wafer process 125 to process semiconductor wafers 628 in a conventional manner. Subsequent to the processing of the semiconductor wafer 628 test measurements are made 630 of the wafers to enable the adjustment of manufacturing parameters 632 for use in improving the derating process of Step 606.
The invention has now been explained with reference to specific embodiments. Other embodiments will be apparent to those of ordinary skill in the art in light of this disclosure. Therefore it is not intended that this invention be limited, except as indicated by the appended claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US4948458 *||Aug 14, 1989||Aug 14, 1990||Lam Research Corporation||Method and apparatus for producing magnetically-coupled planar plasma|
|US5129009||Jun 4, 1990||Jul 7, 1992||Motorola, Inc.||Method for automatic semiconductor wafer inspection|
|US5329334 *||Mar 2, 1993||Jul 12, 1994||Lsi Logic Corporation||Integrated circuit test reticle and alignment mark optimization method|
|US5414636 *||Aug 29, 1994||May 9, 1995||Fujitsu Limited||Semiconductor memory device having a memory cell capacitor and a fabrication process thereof|
|US5426375 *||Feb 26, 1993||Jun 20, 1995||Hitachi Micro Systems, Inc.||Method and apparatus for optimizing high speed performance and hot carrier lifetime in a MOS integrated circuit|
|US5468681||Jun 15, 1994||Nov 21, 1995||Lsi Logic Corporation||Process for interconnecting conductive substrates using an interposer having conductive plastic filled vias|
|US5477545||Feb 9, 1993||Dec 19, 1995||Lsi Logic Corporation||Method and apparatus for testing of core-cell based integrated circuits|
|US5485243||Apr 30, 1993||Jan 16, 1996||Lsi Logic Corporation||Afocal concentrator for low wavelength lithography, particularly for semiconductor lithography|
|US5495419||Apr 19, 1994||Feb 27, 1996||Lsi Logic Corporation||Integrated circuit physical design automation system utilizing optimization process decomposition and parallel processing|
|US5498579 *||Jun 8, 1994||Mar 12, 1996||Intel Corporation||Method of producing semiconductor device layer layout|
|US5526517||May 15, 1992||Jun 11, 1996||Lsi Logic Corporation||Concurrently operating design tools in an electronic computer aided design system|
|US5566078||May 26, 1993||Oct 15, 1996||Lsi Logic Corporation||Integrated circuit cell placement using optimization-driven clustering|
|US5663076||Aug 8, 1995||Sep 2, 1997||Lsi Logic Corporation||Automating photolithography in the fabrication of integrated circuits|
|US5691949||Jan 17, 1996||Nov 25, 1997||Tactical Fabs, Inc.||Very high density wafer scale device architecture|
|US5821624||Jun 5, 1995||Oct 13, 1998||Lsi Logic Corporation||Semiconductor device assembly techniques using preformed planar structures|
|1||*||J. B. Carter et al., "Transformer Coupled Plasma Etch Technology for the Fabrication of Sub-half Micron Structures," 1992.|
|2||U.S. Appl. No. 09/064,802, filed Apr. 22, 1998, Michael D. Rostoker et al., Automating Photolithography in the Fabrication of Printed Circuits.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US7533359 *||May 19, 2006||May 12, 2009||Cadence Design Systems, Inc.||Method and system for chip design using physically appropriate component models and extraction|
|US7745301||Aug 21, 2006||Jun 29, 2010||Terapede, Llc||Methods and apparatus for high-density chip connectivity|
|US8312413 *||Nov 13, 2012||International Business Machines Corporation||Navigating analytical tools using layout software|
|US8635582||Sep 12, 2012||Jan 21, 2014||International Business Machines Corporation||Navigating analytical tools using layout software|
|US8721121 *||Dec 21, 2010||May 13, 2014||Roger Daniel Briles||Decorative light string with blinking lights|
|US8957511||Aug 21, 2006||Feb 17, 2015||Madhukar B. Vora||Apparatus and methods for high-density chip connectivity|
|US20060265680 *||May 19, 2006||Nov 23, 2006||Cadence Design Systems, Inc.||Method and system for chip design using physically appropriate component models and extraction|
|US20110185325 *||Jan 22, 2010||Jul 28, 2011||International Business Machines Corporation||Navigating Analytical Tools Using Layout Software|
|U.S. Classification||438/14, 430/311, 438/16, 430/30, 700/121, 716/54|
|Cooperative Classification||G03F7/20, G03F7/70433, G03F7/705|
|European Classification||G03F7/70J2, G03F7/70L2B, G03F7/20|