The subject matter of this application is related to the subject matter in a co-pending non-provisional application by the same inventors as the instant application and filed on the same day as the instant application entitled, “CALCULATING ETCH PROXIMITY-CORRECTION USING IMAGE-PRECISION TECHNIQUES,” having Ser. No. 10/955,189 and filing date 29 Sep. 2004.
1. Field of the Invention
This invention relates to the process of fabricating semiconductor chips. More specifically, the invention relates to a method and apparatus for calculating etch proximity-correction during the OPC (Optical Proximity Correction) process.
2. Related Art
The relentless miniaturization of integrated circuits has been a key driving force behind recent advances in computer technology. As this miniaturization process causes features on integrated circuits to become progressively smaller, post-lithographic process effects are accounting for an ever-increasing portion of the CD (Critical Dimension) error budget. As a result, accurate modeling of these post-lithographic process effects is becoming increasingly more important during the OPC (Optical Proximity Correction) process.
It is well known that for sub-90 nm processes, resist and etch effects can no longer be treated as a small perturbation on a purely optical OPC model. Hence, OPC models must account for such etch proximity-effects that occur due to the main-etch step and any additional etch steps, such as resist trim, that follow the lithography step.
Etch proximity-effects are determined by the complex physical, transport, and chemical interactions in an etch chamber. Moreover, etch proximity-effects are heavily influenced by the actual layout of the integrated circuit. For example, one important source of etch proximity-effects is the deposition of passivant molecules from the gas phase during etch processing. These passivant molecules move in straight lines through the gas phase and deposit on sidewalls of the features of the integrated circuit. Note that, since these passivant molecules move in straight lines, the geometry of the layout plays a critical role in determining the deposition of the passivant molecules.
Unfortunately, existing OPC models do not calculate etch proximity-correction accurately. These OPC models typically use a function that is empirically-fit to model etch proximity-effects, and they use a linear convolution technique to calculate the etch proximity-correction at the target point. Linear convolution techniques linearly superimpose (add together) the contribution of each polygon to the overall proximity effect. This cannot capture the passivation effect accurately because polygons that are not visible from the target point do not contribute to the proximity effect. Linear convolution techniques cannot distinguish between visible and occluded polygons in every case. As a result, in existing OPC models, the calculated etch proximity-correction lack much of the polygon orientation and the relative placement information needed for modeling of etch proximity. This is the main reason why existing OPC models do not calculate etch proximity-correction accurately.
Hence, what is needed is a method and apparatus for accurately calculating etch proximity-correction by taking into account the orientation and relative placement of the features in the layout.
One embodiment of the present invention provides a system that calculates etch proximity-correction during an OPC (Optical Proximity Correction) process. During operation, the system receives a layout for an integrated circuit. Next, the system selects a target point on an edge in the layout. The system then creates a list of edges within an ambit of the target point. Next, the system removes edges from the list that are not visible from the target point. The system then computes a line integral of a statistical function over the remaining edges in the list, wherein the line integral of the statistical function models the etch proximity effects correlated with positions of the edges visible from the target point. Next, the calculated etch proximity-correction is applied to an area in proximity to the target point.
In a variation on this embodiment, the system removes back-facing edges from the list by computing the dot product of the surface normal of a candidate edge with a vector that points from the target point to the candidate edge's midpoint. Next, the system determines that the candidate edge is back-facing if the computed dot product is positive. If the candidate edge is determined to be back-facing, the system removes the candidate edge from the list.
In a variation on this embodiment, the system identifies and removes occluded edges from the list by performing a visual surface calculation.
In a variation on this embodiment, the system selects the midpoint of the edge to be the target point.
In a variation on this embodiment, the system creates the list of edges by selecting all edges that are within a circular area that is centered on the target point.
In a variation on this embodiment, the system computes the line integral by computing the line integral only on the visible portion of an edge, which is defined as the portion of the edge that is visible from the target point.
In a variation on this embodiment, if the statistical function is Gaussian, the system computes a line integral of the Gaussian over an edge, where an integration variable s parameterizes the edge, such that if L is the length of the edge, then s=0 is one end of the edge and s=L is the other end of the edge.
In a variation on this embodiment, if the statistical function is Gaussian, the line integral can be efficiently calculated as the difference of two error functions multiplied by an exponential. For example, if the statistical function is K(x, y)=e−(x
Note that, in the above expression, erf(x) represents the error function.
Optical Proximity Correction (OPC)
The EDA software design step 110, in turn, includes a number of sub-steps, namely, system design (step 112), logic design and function verification (step 114), synthesis and design for test (step 116), design planning (step 118), netlist verification (step 120), physical implementation (step 122), analysis and extraction (step 124), physical verification (step 126), resolution enhancement (step 128), and mask data preparation (step 130).
Optical Proximity Correction (OPC) takes place within the resolution enhancement step 128, which involves geometric manipulations of the layout to improve manufacturability of the design. Contrary to its name, OPC generally refers to proximity corrections for a variety of factors, which include, optical, micro-loading, etch, resist, phase shift mask (PSM), etc. Typically, OPC starts by dissecting the layout into a number of segments (or edges) to identify the critical features. Next, a model-based simulation is performed on each segment (or edge) to predict the pattern that would be generated by the layout. Finally, the required correction is computed for each segment (or edge) by comparing the predicted pattern with the target pattern. Note that, during the OPC process, edges are typically divided into segments. The etch proximity-correction technique discussed in the present application can be applied to edges or segments. Hence, in the present application, the term “edge” is used interchangeably with the term “segment”. Furthermore, the term “edge” is used in this application in the broadest possible meaning: an edge represents an arbitrary portion of a layout.
Due to the relentless miniaturization of integrated circuits, OPC has become an indispensable step in the design and fabrication of integrated circuits. Moreover, as miniaturization continues, accurate modeling of post-lithographic process effects, such as etch proximity-effects, is becoming more and more important in OPC.
Etch Proximity Correction
Etch proximity-effects are determined by the complex physical, transport, and chemical interactions in an etch chamber. Moreover, etch proximity-effects are heavily influenced by the actual layout of the integrated circuit.
One of the important sources for etch proximity-effects is the deposition of passivant molecules 202 and 204 from the gas phase during etch processing. (Note that, the dotted arrows in
Furthermore, passivant molecules typically have a high sticking coefficient. As a result, they deposit preferentially on sidewalls adjacent to wide trench areas, such as the sidewall of the isolated line 208. In contrast, the transport of passivant molecules, such as molecule 202, into dense features 206 is limited. It is evident that, since these passivant molecules move in straight lines, the geometry of the layout plays a critical role in determining the deposition of the passivant molecules.
Note that, it is impossible to exactly simulate the complex physical, transport, and chemical interactions in an etch chamber. Moreover, the trench height, which affects the amount of deposition, is also not readily available. Hence, such factors must be modeled using a statistical function.
It will be apparent to one skilled in the art that a variety of statistical functions can be used for modeling such factors. Specifically, in one embodiment of the present invention, the statistical function is chosen so that a line integral of the statistical function models the amount of deposition of the passivant molecules. Furthermore, in another embodiment of the present invention, the statistical function is chosen so that a line integral of the statistical function models the etch proximity effects correlated with positions of the edges visible from a target point.
Furthermore, the present invention is not limited to modeling the effect of passivant deposition on etch bias. It will be apparent to one skilled in the art that the present invention can be used to model any etch processing mechanism which depends on the placement and orientation of surfaces visible from the target point.
The deposition of passivant molecules, such as passivant molecules 202 and 204, can inhibit subsequent chemical etching, which can cause the actual pattern to be different from the target pattern. This can deteriorate the manufacturability of the design. Hence, the target pattern needs to be corrected for etch proximity-effects, so that the actual pattern is close to the target pattern.
In one embodiment of the present invention, the etch-proximity-effect model is separate from other proximity-effect models. In another embodiment of the present invention, the etch-proximity-effect model is lumped into a single proximity-effect model that additionally contains other proximity-effect models. Furthermore, applying proximity-correction to a pattern can be an iterative process, wherein in each iteration the pattern is geometrically manipulated using a proximity-effect model.
Calculating Etch Proximity-correction Using Object-precision Techniques
Etch proximity-correction is computed for an edge of a polygon, such as polygon 402. More specifically, etch proximity-correction is computed for a target point, such as target point 412, on an edge of a polygon, such as polygon 402.
In one embodiment of the present invention, an edge is divided into a plurality of segments, and etch proximity-correction is calculated for a target point on each segment. Furthermore, in one embodiment of the present invention, etch proximity-correction is calculated for multiple target points on an edge of a polygon.
The process begins with receiving the layout (step 552). The layout describes the orientation and relative placement of the various polygons, such as polygons 402, 404, 406, 408, and 410, which belong to the integrated circuit.
Next, the system selects a target point (step 554). The target point 412 is located on an edge of a polygon 402 for which etch proximity-correction is to be calculated. In one embodiment of the present invention, the midpoint of an edge of a polygon is selected to be the target point.
The system then creates a list of edges within an ambit of the target point 412 (step 556). For example, in
Next, the system removes edges that are not visible from the target point 412 (step 558).
It will be apparent to one skilled in the art that, a variety of visual surface determination techniques can be used to identify and remove edges in the list that are not visible from the target point 412.
Specifically, in one embodiment of the present invention, the system uses a process called back-face culling to remove back-facing edges.
In back-face culling, the system computes the dot product of the surface normal 488 of a candidate edge 476 with a vector 490 that points from the target point 412 to the candidate edge's midpoint 492. Note that the surface normal 488 is a perpendicular bisector pointing away from the polygon fill 410.
If the dot product is positive, the candidate edge is identified as a back-facing edge, and the system removes the candidate edge from the list. For example, the dot product of the surface normal 488 and the vector 490 that points from the target point 412 to the candidate edge's midpoint 492 is negative. Hence, candidate edge 476 is not removed from the list. On the other hand, the dot product for the following candidate edges is positive, and hence they are removed from the list: edges 452, 454, 456, 462, 464, 466, 468, 472, 474, 482, and 484.
Finally, the system calculates the etch proximity-correction by computing a line integral of a statistical fitting function over the remaining edges in the list (step 560).
In one embodiment of the present invention, the line integral is calculated over only on the visible portion of the remaining edges, which is defined as the portions of the edges that are visible from the target point. For example, in
Furthermore, in one embodiment of the present invention, if the statistical function is Gaussian, the line integral of this Gaussian over an edge with an integration variable s, where s parameterizes the edge, such that if L is the length of the edge, then s=0 is one end of the edge and s=L is the other end of the edge.
Specifically, in one embodiment of the present invention, the line integral can be efficiently calculated as the difference of two error functions multiplied by an exponential. For example, if the statistical function is K(x, y)=e−(x
Note that, in the above expression, erf(x) represents the error function.
Typically, a proximity-correction process starts by dissecting the layout into a number of segments (or edges) to identify the critical features. Next, a model-based simulation is performed on each segment (or edge) to predict the pattern that would be generated by the layout. Finally, the required correction is computed for each segment (or edge) by comparing the predicted pattern with the target pattern. Note that the proximity-correction process can be an iterative process, wherein in each iteration the pattern is geometrically manipulated using the proximity-effect model.
The foregoing description is presented to enable one to make and use the invention, and is provided in the context of a particular application and its requirements. It is not intended to be exhaustive or to limit the invention to the forms disclosed. Various modifications to the disclosed embodiments will be readily apparent, and the general principles defined herein may be applied to other embodiments and applications without departing from the spirit and scope of the invention. Thus, the invention is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein. Accordingly, many modifications and variations will be apparent. The scope of the invention is defined by the appended claims.
The data structures and code described in this detailed description can be stored on a computer readable storage medium, which may be any device or medium that can store code andlor data for use by a computer system. This includes, but is not limited to, magnetic and optical storage devices such as disk drives, magnetic tape, CDs (compact discs), and DVDs (digital versatile discs or digital video discs).
Number | Name | Date | Kind |
---|---|---|---|
5051598 | Ashton et al. | Sep 1991 | A |
5451487 | Abe et al. | Sep 1995 | A |
5663893 | Wampler et al. | Sep 1997 | A |
5863683 | Ohta et al. | Jan 1999 | A |
6243855 | Kobayashi et al. | Jun 2001 | B1 |
6289499 | Rieger et al. | Sep 2001 | B1 |
6453457 | Pierrat et al. | Sep 2002 | B1 |
6467076 | Cobb | Oct 2002 | B1 |
6501083 | Kamijo | Dec 2002 | B1 |
6684382 | Liu | Jan 2004 | B2 |
6753115 | Zhang et al. | Jun 2004 | B2 |
6792590 | Pierrat et al. | Sep 2004 | B1 |
6918104 | Pierrat et al. | Jul 2005 | B2 |
7003757 | Pierrat et al. | Feb 2006 | B2 |
7079223 | Rosenbluth et al. | Jul 2006 | B2 |
20050273753 | Sezginer | Dec 2005 | A1 |
Number | Date | Country | |
---|---|---|---|
20060075380 A1 | Apr 2006 | US |