Patent Application
20150249041
A push for higher performance and smaller die size drives the semiconductor industry to reduce circuit chip area by approximately 50% every two years. The chip area reduction provides an economic benefit for migrating to newer technologies. The 50% chip area reduction is achieved by reducing the feature sizes between 25% and 30%. The reduction in feature size is enabled by improvements in manufacturing equipment and materials. For example, improvement in the lithographic process has enabled smaller feature sizes to be achieved, while improvement in chemical mechanical polishing (CMP) has in-part enabled a higher number of interconnect layers.
In the evolution of lithography, as the minimum feature size approached the wavelength of the light source used to expose the feature shapes, unintended interactions occurred between neighboring features. Today minimum feature sizes are approaching 45 nm (nanometers), while the wavelength of the light source used in the photolithography process remains at 193 nm. The difference between the minimum feature size and the wavelength of light used in the photolithography process is defined as the lithographic gap. As the lithographic gap grows, the resolution capability of the lithographic process decreases.
An interference pattern occurs as each shape on the mask interacts with the light. The interference patterns from neighboring shapes can create constructive or destructive interference. In the case of constructive interference, unwanted shapes may be inadvertently created. In the case of destructive interference, desired shapes may be inadvertently removed. In either case, a particular shape is printed in a different manner than intended, possibly causing a device failure. Correction methodologies, such as optical proximity correction (OPC), attempt to predict the impact from neighboring shapes and modify the mask such that the printed shape is fabricated as desired. The quality of the light interaction prediction is declining as process geometries shrink and as the light interactions become more complex.
In view of the foregoing, a solution is needed for managing lithographic gap issues as technology continues to progress toward smaller semiconductor device features sizes.
In one embodiment, an integrated circuit is disclosed to include a first gate electrode feature of a first gate electrode track. The first gate electrode feature of the first gate electrode track forms a first n-channel transistor as it crosses an n-diffusion region. The integrated circuit also includes a second gate electrode feature of the first gate electrode track. The second gate electrode feature of the first gate electrode track forms a first p-channel transistor as it crosses a p-diffusion region. The first and second gate electrode features of the first gate electrode track are separated by a first end-to-end spacing. The integrated circuit also includes a first gate electrode feature of a second gate electrode track. The first gate electrode feature of the second gate electrode track forms a second n-channel transistor as it crosses the n-diffusion region. The integrated circuit also includes a second gate electrode feature of the second gate electrode track. The second gate electrode feature of the second gate electrode track forms a second p-channel transistor as it crosses the p-diffusion region. The first and second gate electrode features of the second gate electrode track are separated by a second end-to-end spacing that is offset from the first end-to-end spacing such that a line of sight perpendicular to the first and second gate electrode tracks does not exist through the first and second end-to-end spacings.
In another embodiment, a cell of a semiconductor device is disclosed. The cell includes a substrate portion formed to include a plurality of diffusion regions. The plurality of diffusion regions respectively correspond to active areas of the substrate portion within which one or more processes are applied to modify one or more electrical characteristics of the active areas of the substrate portion. The plurality of diffusion regions are separated from each other by one or more non-active regions of the substrate portion.
Also in this embodiment, the cell includes a gate electrode level of the cell formed above the substrate portion. The gate electrode level includes a number of conductive features defined to extend in only a first parallel direction. Adjacent ones of the number of conductive features that share a common line of extent in the first parallel direction are fabricated from respective originating layout features that are separated from each other by an end-to-end spacing having a size measured in the first parallel direction. The size of each end-to-end spacing between originating layout features corresponding to adjacent ones of the number of conductive features within the gate electrode level of the cell is substantially equal and is minimized to an extent allowed by a semiconductor device manufacturing capability. The number of conductive features within the gate electrode level of the cell includes conductive features defined along at least four different virtual lines of extent in the first parallel direction across the gate electrode level of the cell.
A width size of the conductive features within the gate electrode level is measured perpendicular to the first parallel direction. The width size of the conductive features within a photolithographic interaction radius within the gate electrode level is less than a wavelength of light used in a photolithography process to fabricate the conductive features within the gate electrode level. The wavelength of light used in the photolithography process is less than or equal to 193 nanometers. The photolithographic interaction radius is five wavelengths of light used in the photolithography process.
Some of the number of conductive features within the gate electrode level of the cell are defined to include one or more gate electrode portions which extend over one or more of the active areas of the substrate portion corresponding to the plurality of diffusion regions. Each gate electrode portion and a corresponding active area of the substrate portion over which it extends together define a respective transistor device.
Also in this embodiment, the cell includes a number of interconnect levels formed above the gate electrode level of the cell. The substrate portion, the gate electrode level of the cell, and the number of interconnect levels are spatially aligned such that structures fabricated within each of the substrate portion, the gate electrode level of the cell, and the number of interconnect levels spatially relate to connect as required to form functional electronic devices within the semiconductor device.
In one embodiment, a semiconductor chip includes a region including a plurality of transistors. Each of the plurality of transistors in the region forms part of circuitry associated with execution of one or more logic functions. The region includes at least ten conductive structures formed within the semiconductor chip. Some of the at least ten conductive structures form at least one transistor gate electrode. Each of the at least ten conductive structures respectively has a corresponding top surface. An entirety of a periphery of the corresponding top surface is defined by a corresponding first end, a corresponding second end, a corresponding first edge, and a corresponding second edge, such that a total distance along the entirety of the periphery of the corresponding top surface is equal to a sum of a total distance along the corresponding first edge and a total distance along the corresponding second edge and a total distance along the corresponding first end and a total distance along the corresponding second end. The total distance along the corresponding first edge is greater than two times the total distance along the corresponding first end. The total distance along the corresponding first edge is greater than two times the total distance along the corresponding second end. The total distance along the corresponding second edge is greater than two times the total distance along the corresponding first end. The total distance along the corresponding second edge is greater than two times the total distance along the corresponding second end. The corresponding first end extends from the corresponding first edge to the corresponding second edge and is located principally within a space between the corresponding first and second edges. The corresponding second end extends from the corresponding first edge to the corresponding second edge and is located principally within the space between the corresponding first and second edges. The top surfaces of the at least ten conductive structures are co-planar with each other. Each of the at least ten conductive structures has a corresponding lengthwise centerline oriented in a first direction along its top surface and extending from its first end to its second end. Each of the at least ten conductive structures has a length as measured along its lengthwise centerline from its first end to its second end. The first edge of each of the at least ten conductive structures is substantially straight. The second edge of each of the at least ten conductive structures is substantially straight. Each of the at least ten conductive structures has both its first edge and its second edge oriented substantially parallel to its lengthwise centerline. Each of the at least ten conductive structures has a width measured in a second direction perpendicular to the first direction at a midpoint of its lengthwise centerline. Each of the first direction and the second direction is oriented substantially parallel to the co-planar top surfaces of the at least ten conductive structures. The at least ten conductive structures are positioned in a side-by-side manner such that each of the at least ten conductive structures is positioned to have at least a portion of its length beside at least a portion of the length of another of the at least ten conductive structures. The width of each of the at least ten conductive structures is less than 45 nanometers. The region has a size of about 965 nanometers as measured in the second direction. Each of the at least ten conductive structures is positioned such that a distance as measured in the second direction between its lengthwise centerline and the lengthwise centerline of at least one other of the at least ten conductive structures is substantially equal to a first pitch that is less than or equal to about 193 nanometers. The region includes a first transistor of a first transistor type, a second transistor of the first transistor type, a first transistor of a second transistor type, and a second transistor of the second transistor type. Each transistor of the first transistor type that has its gate electrode formed by any of the at least ten conductive structures is included in a first collection of transistors, and each transistor of the second transistor type that has its gate electrode formed by any of the at least ten conductive structures is included in a second collection of transistors. The first and second collections of transistors are separated from each other by an inner sub-region of the region, wherein the inner sub-region does not include a source or a drain of any transistor. The first transistor of the first transistor type has a gate electrode formed by a portion of a first conductive structure of the at least ten conductive structures. Any transistor having its gate electrode formed by the first conductive structure is of the first transistor type. The second transistor of the first transistor type has a gate electrode formed by a portion of a second conductive structure of the at least ten conductive structures. The first transistor of the second transistor type has a gate electrode formed by another portion of the second conductive structure. The second conductive structure is positioned such that its lengthwise centerline is separated from the lengthwise centerline of the first conductive structure by the first pitch as measured in the second direction. The second transistor of the second transistor type has a gate electrode formed by one of the at least ten conductive structures. The conductive structure that forms the gate electrode of the second transistor of the second transistor type is positioned such that its lengthwise centerline is separated from the lengthwise centerline of the second conductive structure by the first pitch as measured in the second direction. Any transistor having its gate electrode formed by the conductive structure that forms the gate electrode of the second transistor of the second transistor type is of the second transistor type. The first conductive structure has a total length as measured in the first direction that is greater than one-half of a total length of the second conductive structure as measured in the first direction. The second conductive structure has at least one end substantially positioned at a first reference line oriented in the second direction. And, the conductive structure that forms the gate electrode of the second transistor of the second transistor type also has at least one end substantially positioned at the first reference line.
Other aspects and advantages of the invention will become more apparent from the following detailed description, taken in conjunction with the accompanying drawings, illustrating by way of example the present invention.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. It will be apparent, however, to one skilled in the art that the present invention may be practiced without some or all of these specific details. In other instances, well known process operations have not been described in detail in order not to unnecessarily obscure the present invention.
Generally speaking, a dynamic array architecture is provided to address semiconductor manufacturing process variability associated with a continually increasing lithographic gap. In the area of semiconductor manufacturing, lithographic gap is defined as the difference between the minimum size of a feature to be defined and the wavelength of light used to render the feature in the lithographic process, wherein the feature size is less than the wavelength of the light. Current lithographic processes utilize a light wavelength of 193 nm. However, current feature sizes are as small as 65 nm and are expected to soon approach sizes as small as 45 nm With a size of 65 nm, the shapes are three times smaller than the wavelength of the light used to define the shapes. Also, considering that the interaction radius of light is about five light wavelengths, it should be appreciated that shapes exposed with a 193 nm light source will influence the exposure of shapes approximately 5*193 nm (965 nm) away. When considering the 65 nm sized features with respect to 90 nm sized features, it should be appreciated that approximately two times as many 65 nm sizes features may be within the 965 nm interaction radius of the 193 nm light source as compared to the 90 nm sized features.
Due to the increased number of features within the interaction radius of the light source, the extent and complexity of light interference contributing to exposure of a given feature is significant. Additionally, the particular shapes associated with the features within the interaction radius of the light source weighs heavily on the type of light interactions that occur. Traditionally, designers were allowed to define essentially any two-dimensional topology of feature shapes so long as a set of design rules were satisfied. For example, in a given layer of the chip, i.e., in a given mask, the designer may have defined two-dimensionally varying features having bends that wrap around each other. When such two-dimensionally varying features are located in neighboring proximity to each other, the light used to expose the features will interact in a complex and generally unpredictable manner. The light interaction becomes increasingly more complex and unpredictable as the feature sizes and relative spacing become smaller.
Traditionally, if a designer follows the established set of design rules, the resulting product will be manufacturable with a specified probability associated with the set of design rules. Otherwise, for a design that violates the set of design rules, the probability of successful manufacture of the resulting product is unknown. To address the complex light interaction between neighboring two-dimensionally varying features, in the interest of successful product manufacturing, the set of design rules is expanded significantly to adequately address the possible combinations of two-dimensionally varying features. This expanded set of design rules quickly becomes so complicated and unwieldy that application of the expanded set of design rules becomes prohibitively time consuming, expensive, and prone to error. For example, the expanded set of design rules requires complex verification. Also, the expanded set of design rules may not be universally applied. Furthermore, manufacturing yield is not guaranteed even if all design rules are satisfied.
It should be appreciated that accurate prediction of all possible light interactions when rendering arbitrarily-shaped two-dimensional features is generally not feasible. Moreover, as an alternative to or in combination with expansion of the set of design rules, the set of design rules may also be modified to include increased margin to account for unpredictable light interaction between the neighboring two-dimensionally varying features. Because the design rules are established in an attempt to cover the random two-dimensional feature topology, the design rules may incorporate a significant amount of margin. While addition of margin in the set of design rules assists with the layout portions that include the neighboring two-dimensionally varying features, such global addition of margin causes other portions of the layout that do not include the neighboring two-dimensionally varying features to be overdesigned, thus leading to decreased optimization of chip area utilization and electrical performance.
In view of the foregoing, it should be appreciated that semiconductor product yield is reduced as a result of parametric failures that stem from variability introduced by design-dependent unconstrained feature topologies, i.e., arbitrary two-dimensionally varying features disposed in proximity to each other. By way of example, these parametric failures may result from failure to accurately print contacts and vias and from variability in fabrication processes. The variability in fabrication processes may include CMP dishing, layout feature shape distortion due to photolithography, gate distortion, oxide thickness variability, implant variability, and other fabrication related phenomena. The dynamic array architecture of the present invention is defined to address the above-mentioned semiconductor manufacturing process variability.
As illustrated in
A forbidden pitch, i.e., forbidden layout feature spacing, occurs when the neighboring layout features (101A-101C) are spaced such that peaks of the sinc function associated with one layout feature align with valleys of the sinc function associated with another layout feature, thus causing destructive interference of the light energy. The destructive interference of the light energy causes the light energy focused at a given location to be reduced. Therefore, to realize the beneficial constructive light interference associated with neighboring layout features, it is necessary to predict the layout feature spacing at which the constructive overlap of the sinc function peaks will occur. Predictable constructive overlap of the sinc function peaks and corresponding layout feature shape enhancement can be realized if the layout feature shapes are rectangular, near the same size, and are oriented in the same direction, as illustrated by the layout features (101A-101C) in
With regard to the definition of the underlying structure of the dynamic array, the dynamic array is built-up in a layered manner upon a base substrate 201, e.g., upon a silicon substrate, or silicon-on-insulator (SOI) substrate. Diffusion regions 203 are defined in the base substrate 201. The diffusion regions 203 represent selected regions of the base substrate 201 within which impurities are introduced for the purpose of modifying the electrical properties of the base substrate 201. Above the diffusion regions 203, diffusion contacts 205 are defined to enable connection between the diffusion regions 203 and conductor lines. For example, the diffusion contacts 205 are defined to enable connection between source and drain diffusion regions 203 and their respective conductor nets. Also, gate electrode features 207 are defined above the diffusion regions 203 to form transistor gates. Gate electrode contacts 209 are defined to enable connection between the gate electrode features 207 and conductor lines. For example, the gate electrode contacts 209 are defined to enable connection between transistor gates and their respective conductor nets.
Interconnect layers are defined above the diffusion contact 205 layer and the gate electrode contact layer 209. The interconnect layers include a first metal (metal 1) layer 211, a first via (via 1) layer 213, a second metal (metal 2) layer 215, a second via (via 2) layer 217, a third metal (metal 3) layer 219, a third via (via 3) layer 221, and a fourth metal (metal 4) layer 223. The metal and via layers enable definition of the desired circuit connectivity. For example, the metal and via layers enable electrical connection of the various diffusion contacts 205 and gate electrode contacts 209 such that the logic function of the circuitry is realized. It should be appreciated that the dynamic array architecture is not limited to a specific number of interconnect layers, i.e., metal and via layers. In one embodiment, the dynamic array may include additional interconnect layers 225, beyond the fourth metal (metal 4) layer 223. Alternatively, in another embodiment, the dynamic array may include less than four metal layers.
The dynamic array is defined such that layers (other than the diffusion region layer 203) are restricted with regard to layout feature shapes that can be defined therein. Specifically, in each layer other than the diffusion region layer 203, only linear-shaped layout features are allowed. A linear-shaped layout feature in a given layer is characterized as having a consistent vertical cross-section shape and extending in a single direction over the substrate. Thus, the linear-shaped layout features define structures that are one-dimensionally varying. The diffusion regions 203 are not required to be one-dimensionally varying, although they are allowed to be if necessary. Specifically, the diffusion regions 203 within the substrate can be defined to have any two-dimensionally varying shape with respect to a plane coincident with a top surface of the substrate. In one embodiment, the number of diffusion bend topologies is limited such that the interaction between the bend in diffusion and the conductive material, e.g., polysilicon, that forms the gate electrode of the transistor is predictable and can be accurately modeled. The linear-shaped layout features in a given layer are positioned to be parallel with respect to each other. Thus, the linear-shaped layout features in a given layer extend in a common direction over the substrate and parallel with the substrate.
The underlying layout methodology of the dynamic array uses constructive light interference of light waves in the lithographic process to reinforce exposure of neighboring shapes in a given layer. Therefore, the spacing of the parallel, linear-shaped layout features in a given layer is designed around the constructive light interference of the standing light waves such that lithographic correction (e.g., OPC/RET) is minimized or eliminated. Thus, in contrast to conventional OPC/RET-based lithographic processes, the dynamic array defined herein exploits the light interaction between neighboring features, rather than attempting to compensate for the light interaction between neighboring features.
Because the standing light wave for a given linear-shaped layout feature can be accurately modeled, it is possible to predict how the standing light waves associated with the neighboring linear-shaped layout features disposed in parallel in a given layer will interact. Therefore, it is possible to predict how the standing light wave used to expose one linear-shaped feature will contribute to the exposure of its neighboring linear-shaped features. Prediction of the light interaction between neighboring linear-shaped features enables the identification of an optimum feature-to-feature spacing such that light used to render a given shape will reinforce its neighboring shapes. The feature-to-feature spacing in a given layer is defined as the feature pitch, wherein the pitch is the center-to-center separation distance between adjacent linear-shaped features in a given layer.
To provide the desired exposure reinforcement between neighboring features, the linear-shaped layout features in a given layer are spaced such that constructive and destructive interference of the light from neighboring features will be optimized to produce the best rendering of all features in the neighborhood. The feature-to-feature spacing in a given layer is proportional to the wavelength of the light used to expose the features. The light used to expose each feature within about a five light wavelength distance from a given feature will serve to enhance the exposure of the given feature to some extent. The exploitation of constructive interference of the standing light waves used to expose neighboring features enables the manufacturing equipment capability to be maximized and not be limited by concerns regarding light interactions during the lithography process.
As discussed above, the dynamic array incorporates a restricted topology in which the features within each layer (other than diffusion) are required to be linear-shaped features that are oriented in a parallel manner to traverse over the substrate in a common direction. With the restricted topology of the dynamic array, the light interaction in the photolithography process can be optimized such that the printed image on the mask is essentially identical to the drawn shape in the layout, i.e., essentially a 100% accurate transfer of the layout onto the resist is achieved.
In the exemplary embodiment of
The base grid is defined with consideration for the light interaction function, i.e., the sinc function, and the manufacturing capability, wherein the manufacturing capability is defined by the manufacturing equipment and processes to be utilized in fabricating the dynamic array. With regard to the light interaction function, the base grid is defined such that the spacing between gridpoints enables alignment of peaks in the sinc functions describing the light energy projected upon neighboring gridpoints. Therefore, linear-shaped features optimized for lithographic reinforcement can be specified by drawing a line from a first gridpoint to a second gridpoint, wherein the line represents a rectangular structure of a given width. It should be appreciated that the various linear-shaped features in each layer can be specified according to their endpoint locations on the base grid and their width.
Although
The layout architecture of the dynamic array follows the base grid pattern. Thus, it is possible to use grid points to represent where changes in direction occur in diffusion, wherein gate electrode and metal linear-shaped features are placed, where contacts are placed, where opens are in the linear-shaped gate electrode and metal features, etc. The pitch of the gridpoints, i.e., the gridpoint-to-gridpoint spacing, should be set for a given feature line width, e.g., width 303 in
The various layers of the dynamic array are defined such that the linear-shaped features in adjacent layers extend in a crosswise manner with respect to each other. For example, the linear-shaped features of adjacent layers may extend orthogonally, i.e., perpendicularly with respect to each other. Also, the linear-shaped features of one layer may extend across the linear-shaped features of an adjacent layer at an angle, e.g., at about 45 degrees. For example, in one embodiment the linear-shaped feature of one layer extend in the first reference direction (x) and the linear-shaped features of the adjacent layer extend diagonally with respect to the first (x) and second (y) reference directions. It should be appreciated that to route a design in the dynamic array having the linear-shaped features positioned in the crosswise manner in adjacent layers, opens can be defined in the linear-shaped features, and contacts and vias can be defined as necessary.
The dynamic array minimizes the use of bends in layout shapes to eliminate unpredictable lithographic interactions. Specifically, prior to OPC or other RET processing, the dynamic array allows bends in the diffusion layer to enable control of device sizes, but does not allow bends in layers above the diffusion layer. The layout features in each layer above the diffusion layer are linear in shape, e.g.,
An exemplary buildup of dynamic array layers from diffusion through metal 2 are described with respect to
The gate electrode features 501 form n-channel and p-channel transistors as they cross the diffusion regions 403 and 401, respectively. Optimal gate electrode feature 501 printing is achieved by drawing gate electrode features 501 at every grid location, even though no diffusion region may be present at some grid locations. Also, long continuous gate electrode features 501 tend to improve line end shortening effects at the ends of gate electrode features within the interior of the dynamic array. Additionally, gate electrode printing is significantly improved when all bends are removed from the gate electrode features 501.
Each of the gate electrode tracks may be interrupted, i.e., broken, any number of times in linearly traversing across the dynamic array in order to provide required electrical connectivity for a particular logic function to be implemented. When a given gate electrode track is required to be interrupted, the separation between ends of the gate electrode track segments at the point of interruption is minimized to the extent possible taking into consideration the manufacturing capability and electrical effects. In one embodiment, optimal manufacturability is achieved when a common end-to-end spacing is used between features within a particular layer.
Minimizing the separation between ends of the gate electrode track segments at the points of interruption serves to maximize the lithographic reinforcement, and uniformity thereof, provided from neighboring gate electrode tracks. Also, in one embodiment, if adjacent gate electrode tracks need to be interrupted, the interruptions of the adjacent gate electrode tracks are made such that the respective points of interruption are offset from each other so as to avoid, to the extent possible, an occurrence of neighboring points of interruption. More specifically, points of interruption within adjacent gate electrode tracks are respectively positioned such that a line of sight does not exist through the points of interruption, wherein the line of sight is considered to extend perpendicularly to the direction in which the gate electrode tracks extend over the substrate. Additionally, in one embodiment, the gate electrodes may extend through the boundaries at the top and bottom of the cells, i.e., the PMOS or NMOS cells. This embodiment would enable bridging of neighboring cells.
With further regard to
The gate electrode features 501 and diffusion contacts 503 share a common grid spacing. More specifically, the gate electrode feature 501 placement is offset by one-half the grid spacing relative to the diffusion contacts 503. For example, if the gate electrode features 501 and diffusion contact 503 grid spacing is 0.36 μm, then the diffusion contacts are placed such that the x-coordinate of their center falls on an integer multiple of 0.36 μm, while the x-coordinate of the center of each gate electrode feature 501 minus 0.18 μm should be an integer multiple of 0.36 μm. In the present example, the x-coordinates are represented by the following:
Diffusion contact center x-coordinate=I*0.36 μm, where I is the grid number;
Gate electrode feature center x-coordinate=0.18 μm+I*0.36 μm, where I is the grid number.
The grid based system of the dynamic array ensures that all contacts (diffusion and gate electrode) will land on a horizontal grid that is equal to a multiple of one-half of the diffusion contact grid and a vertical grid that is set by the metal 1 pitch. In the example above, the gate electrode feature and diffusion contact grid is 0.36 μm. The diffusion contacts and gate electrode contacts will land on a horizontal grid that is a multiple of 0.18 μm. Also, the vertical grid for 90 nm process technologies is about 0.24 μm.
In one embodiment, the gate electrode contact 601 extension 701 beyond the gate electrode feature 501 is set such that maximum overlap is achieved between the gate electrode contact 601 and the gate electrode feature 501. The extension 701 is defined to accommodate line end shortening of the gate electrode contact 601, and misalignment between the gate electrode contact layer and gate electrode feature layer. The length of the gate electrode contact 601 is defined to ensure maximum surface area contact between the gate electrode contact 601 and the gate electrode feature 501, wherein the maximum surface area contact is defined by the width of the gate electrode feature 501.
Each of the metal 1 tracks 801-821 may be interrupted, i.e., broken, any number of times in linearly traversing across the dynamic array in order to provide required electrical connectivity for a particular logic function to be implemented. When a given metal 1 track 801-821 is required to be interrupted, the separation between ends of the metal 1 track segments at the point of interruption is minimized to the extent possible taking into consideration manufacturing capability and electrical effects. Minimizing the separation between ends of the metal 1 track segments at the points of interruption serves to maximize the lithographic reinforcement, and uniformity thereof, provided from neighboring metal 1 tracks. Also, in one embodiment, if adjacent metal 1 tracks need to be interrupted, the interruptions of the adjacent metal 1 tracks are made such that the respective points of interruption are offset from each other so as to avoid, to the extent possible, an occurrence of neighboring points of interruption. More specifically, points of interruption within adjacent metal 1 tracks are respectively positioned such that a line of sight does not exist through the points of interruption, wherein the line of sight is considered to extend perpendicularly to the direction in which the metal 1 tracks extend over the substrate.
In the example of
The metal 1 track pattern is optimally configured to optimize the use of “white space” (space not occupied by transistors). The example of
The pitch (center-to-center spacing) of the metal 2 tracks 1001 is minimized while ensuring optimization of lithographic reinforcement, i.e., resonant imaging, provided by neighboring metal 2 tracks. It should be appreciated that regularity can be maintained on higher level interconnect layers in the same manner as implemented in the gate electrode and metal 1 layers. In one embodiment, the gate electrode feature 501 pitch and the metal 2 track pitch is the same. In another embodiment, the contacted gate electrode pitch (e.g., polysilicon-to-polysilicon space with a diffusion contact in between) is greater than the metal 2 track pitch. In this embodiment, the metal 2 track pitch is optimally set to be ⅔ or ¾ of the contacted gate electrode pitch. Thus, in this embodiment, the gate electrode track and metal 2 track align at every two gate electrode track pitches and every three metal 2 track pitches. For example, in a 90 nm process technology, the optimum contacted gate electrode track pitch is 0.36 μm, and the optimum metal 2 track pitch is 0.24 μm. In another embodiment, the gate electrode track and the metal 2 track align at every three gate electrode pitches and every four metal 2 pitches. For example, in a 90 nm process technology, the optimum contacted gate electrode track pitch is 0.36 μm, and the optimum metal 2 track pitch is 0.27 μm.
Each of the metal 2 tracks 1001 may be interrupted, i.e., broken, any number of times in linearly traversing across the dynamic array in order to provide required electrical connectivity for a particular logic function to be implemented. When a given metal 2 track 1001 is required to be interrupted, the separation between ends of the metal 2 track segments at the point of interruption is minimized to the extent possible taking into consideration manufacturing and electrical effects. Minimizing the separation between ends of the metal 2 track segments at the points of interruption serves to maximize the lithographic reinforcement, and uniformity thereof, provided from neighboring metal 2 tracks. Also, in one embodiment, if adjacent metal 2 tracks need to be interrupted, the interruptions of the adjacent metal 2 tracks are made such that the respective points of interruption are offset from each other so as to avoid, to the extent possible, an occurrence of neighboring points of interruption. More specifically, points of interruption within adjacent metal 2 tracks are respectively positioned such that a line of sight does not exist through the points of interruption, wherein the line of sight is considered to extend perpendicularly to the direction in which the metal 2 tracks extend over the substrate.
As discussed above, the conduction lines in a given metal layer above the gate electrode layer may traverse the dynamic array in a direction coincident with either the first reference direction (x) or the second reference direction (y). It should be further appreciated that the conduction lines in a given metal layer above the gate electrode layer may traverse the dynamic array in a diagonal direction relative to the first and second reference directions (x) and (y).
As with the metal 1 and metal 2 tracks discussed above, the diagonal traversing conductor tracks 1101 and 1201 of
An optimal layout density within the dynamic array is achieved by implementing the following design rules:
at least two metal 1 tracks be provided across the n-channel device area;
at least two metal 1 tracks be provided across the p-channel device area;
at least two gate electrode tracks be provided for the n-channel device; and
at least two gate electrode tracks be provided for the p-channel device.
Contacts and vias are becoming the most difficult mask from a lithographic point of view. This is because the contacts and vias are getting smaller, more closely spaced, and are randomly distributed. The spacing and density of the cuts (contact or vias) makes it extremely difficult to reliably print the shapes. For example, cut shapes may be printed improperly due to destructive interference patterns from neighboring shapes or lack of energy on lone shapes. If a cut is properly printed, the manufacturing yield of the associated contact or via is extremely high. Sub-resolution contacts can be provided to reinforce the exposure of the actual contacts, so long as the sub-resolution contacts do not resolve. Also, the sub-resolution contacts can be of any shape so long as they are smaller than the resolution capability of the lithographic process.
Grid location 1303 in
As feature sizes decrease, semiconductor dies are capable of including more gates. As more gates are included, however, the density of the interconnect layers begins to dictate the die size. This increasing demand on the interconnect layers drives higher levels of interconnect layers. However, the stacking of interconnect layers is limited in part by the topology of the underlying layers. For example, as interconnect layers are built up, islands, ridges, and troughs can occur. These islands, ridges, and troughs can cause breaks in the interconnect lines that cross them.
To mitigate these islands and troughs, the semiconductor manufacturing process utilizes a chemical mechanical polishing (CMP) procedure to mechanically and chemically polish the surface of the semiconductor wafer such that each subsequent interconnect layer is deposited on a substantially flat surface. Like the photolithography process the quality of the CMP process is layout pattern dependent. Specifically, an uneven distribution of a layout features across a die or a wafer can cause too much material to be removed in some places and not enough material to be removed in other places, thus causing variations in the interconnect thickness and unacceptable variations in the capacitance and resistance of the interconnect layer. The capacitance and resistance variation within the interconnect layer may alter the timing of a critical net causing design failure.
The CMP process requires that dummy fill be added in the areas without interconnect shapes so that a substantially uniform wafer topology is provided to avoid dishing and improve center-to-edge uniformity. Traditionally, dummy fill is placed post-design. Thus, in the traditional approach the designer is not aware of the dummy fill characteristics. Consequently, the dummy fill placed post-design may adversely influence the design performance in a manner that has not been evaluated by the designer. Also, because the conventional topology prior to the dummy fill is unconstrained, i.e., non-uniform, the post-design dummy fill will not be uniform and predictable. Therefore, in the conventional process, the capacitive coupling between the dummy fill regions and the neighboring active nets cannot be predicted by the designer.
As previously discussed, the dynamic array disclosed herein provides optimal regularity by maximally filling all interconnect tracks from gate electrode layer upward. If multiple nets are required in a single interconnect track, the interconnect track is split with a minimally spaced gap. For example, track 809 representing the metal 1 conduction line in
Because the dynamic array sets the size and spacing of the linearly shaped features, i.e., tracks and contacts, in each mask layer, the design of the dynamic array can be optimized for the maximum capability of the manufacturing equipment and processes. That is to say, because the dynamic array is restricted to the regular architecture for each layer above diffusion, the manufacturer is capable of optimizing the manufacturing process for the specific characteristics of the regular architecture. It should be appreciated that with the dynamic array, the manufacturer does not have to be concerned with accommodating the manufacture of a widely varying set of arbitrarily-shaped layout features as is present in conventional unconstrained layouts.
An example of how the capability of manufacturing equipment can be optimized is provided as follows. Consider that a 90 nm process has a metal 2 pitch of 280 nm. This metal 2 pitch of 280 nm is not set by the maximum capability of equipment. Rather, this metal 2 pitch of 280 nm is set by the lithography of the vias. With the via lithography issues removed, the maximum capability of the equipment allows for a metal 2 pitch of about 220 nm. Thus, the design rules for metal 2 pitch include about 25% margin to account for the light interaction unpredictability in the via lithography.
The regular architecture implemented within the dynamic array allows the light interaction unpredictability in the via lithography to be removed, thus allowing for a reduction in the metal 2 pitch margin. Such a reduction in the metal 2 pitch margin allows for a more dense design, i.e., allows for optimization of chip area utilization. Additionally, with the restricted, i.e., regular, topology afforded by the dynamic array, the margin in the design rules can be reduced. Moreover, not only can the excess margin beyond the capability of the process be reduced, the restricted topology afforded by the dynamic array also allows the number of required design rules to be substantially reduced. For example, a typical design rule set for an unconstrained topology could have more than 600 design rules. A design rule set for use with the dynamic array may have about 45 design rules. Therefore, the effort required to analyze and verify the design against the design rules is decreased by more than a factor of ten with the restricted topology of the dynamic array.
When dealing with line end-to-line end gaps (i.e., track segment-to-track segment gaps) in a given track of a mask layer in the dynamic array, a limited number of light interactions exist. This limited number of light interactions can be identified, predicted, and accurately compensated for ahead of time, dramatically reducing or completely eliminating the requirement for OPC/RET. The compensation for light interactions at line end-to-line end gaps represents a lithographic modification of the as-drawn feature, as opposed to a correction based on modeling of interactions, e.g., OPC/RET, associated with the as-drawn feature.
Also, with the dynamic array, changes to the as-drawn layout are only made where needed. In contrast, OPC is performed over an entire layout in a conventional design flow. In one embodiment, a correction model can be implemented as part of the layout generation for the dynamic array. For example, due to the limited number of possible line end gap interactions, a router can be programmed to insert a line break having characteristics defined as a function of its surroundings, i.e., as a function of its particular line end gap light interactions. It should be further appreciated that the regular architecture of the dynamic array allows the line ends to be adjusted by changing vertices rather than by adding vertices. Thus, in contrast with unconstrained topologies that rely on the OPC process, the dynamic array significantly reduces the cost and risk of mask production. Also, because the line end gap interactions in the dynamic array can be accurately predicted in the design phase, compensation for the predicted line end gap interactions during the design phase does not increase risk of design failure.
In conventional unconstrained topologies, designers are required to have knowledge of the physics associated with the manufacturing process due to the presence of design dependent failures. With the grid-based system of the dynamic array as disclosed herein, the logical design can be separated from the physical design. More specifically, with the regular architecture of the dynamic array, the limited number of light interactions to be evaluated within the dynamic array, and the design independent nature of the dynamic array, designs can be represented using a grid point based netlist, as opposed to a physical netlist.
With the dynamic array, the design is not required to be represented in terms of physical information. Rather, the design can be represented as a symbolic layout. Thus, the designer can represent the design from a pure logic perspective without having to represent physical characteristics, e.g., sizes, of the design. It should be understood that the grid-based netlist, when translated to physical, matches the optimum design rules exactly for the dynamic array platform. When the grid-based dynamic array moves to a new technology, e.g., smaller technology, a grid-based netlist can be moved directly to the new technology because there is no physical data in the design representation. In one embodiment, the grid-based dynamic array system includes a rules database, a grid-based (symbolic) netlist, and the dynamic array architecture.
It should be appreciated that the grid-based dynamic array eliminates topology related failures associated with conventional unconstrained architectures. Also, because the manufacturability of the grid-based dynamic array is design independent, the yield of the design implemented on the dynamic array is independent of the design. Therefore, because the validity and yield of the dynamic array is preverified, the grid-based netlist can be implemented on the dynamic array with preverified yield performance.
The conductive lines 1403A-1403G are arranged to extend over the substrate 1405 in a common direction 1407. It should also be appreciated that each of the number of conductive lines 1403A-1403G are restricted to extending over the diffusion region 1401 in the common direction 1407. In one embodiment, the conductive lines 1403A-1403G defined immediately over the substrate 1405 are polysilicon lines. In one embodiment, each of the conductive lines 1403A-1403G is defined to have essentially the same width 1409 in a direction perpendicular to the common direction 1407 of extension. In another embodiment, some of the conductive lines 1403A-1403G are defined to have different widths relative to the other conductive lines. However, regardless of the width of the conductive lines 1403A-1403G, each of the conductive lines 1403A-1403G is spaced apart from adjacent conductive lines according to essentially the same center-to-center pitch 1411.
As shown in
It should be appreciated that the semiconductor chip structure 1400 represents a portion of the dynamic array described above with respect to
The concept of the necessary active portion 1415 and the uniformity extending portions 1417 also applies to higher level interconnect layers. As previously described with regard to the dynamic array architecture, adjacent interconnect layers traverse over the substrate in transverse directions, e.g., perpendicular or diagonal directions, to enable routing/connectivity required by the logic device implemented within the dynamic array. As with the conductive lines 1403A-1403G, each of the conductive lines within an interconnect layer may include a required portion (necessary active portion) to enable required routing/connectivity, and a non-required portion (uniformity extending portion) to provide lithographic reinforcement to neighboring conductive lines. Also, as with the conductive lines 1403A-1403G, the conductive lines within an interconnect layer extend in a common direction over the substrate, have essentially the same width, and are spaced apart from each other according to an essentially constant pitch.
In one embodiment, conductive lines within an interconnect layer follow essentially the same ratio between line width and line spacing. For example, at 90 nm the metal 4 pitch is 280 nm with a line width and line spacing equal to 140 nm. Larger conductive lines can be printed on a larger line pitch if the line width is equal to the line spacing.
The invention described herein can be embodied as computer readable code on a computer readable medium. The computer readable medium is any data storage device that can store data which can be thereafter be read by a computer system. Examples of the computer readable medium include hard drives, network attached storage (NAS), read-only memory, random-access memory, CD-ROMs, CD-Rs, CD-RWs, magnetic tapes, and other optical and non-optical data storage devices. The computer readable medium can also be distributed over a network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion. Additionally, a graphical user interface (GUI) implemented as computer readable code on a computer readable medium can be developed to provide a user interface for performing any embodiment of the present invention.
While this invention has been described in terms of several embodiments, it will be appreciated that those skilled in the art upon reading the preceding specifications and studying the drawings will realize various alterations, additions, permutations and equivalents thereof. Therefore, it is intended that the present invention includes all such alterations, additions, permutations, and equivalents as fall within the true spirit and scope of the invention.
This application is a continuation application under 35 U.S.C. 120 of prior U.S. application Ser. No. 13/774,919, filed on Feb. 22, 2013, which is a continuation application under 35 U.S.C. 120 of prior U.S. application Ser. No. 12/572,225, filed on Oct. 1, 2009, issued as U.S. Pat. No. 8,436,400, on May 7, 2013, which is a continuation application under 35 U.S.C. 120 of prior U.S. application Ser. No. 12/212,562, filed Sep. 17, 2008, issued as U.S. Pat. No. 7,842,975, on Nov. 30, 2010, which is a continuation application under 35 U.S.C. 120 of prior U.S. application Ser. No. 11/683,402, filed Mar. 7, 2007, issued as U.S. Pat. No. 7,446,352, on Nov. 4, 2008, which claims priority under 35 U.S.C. 119(e) to U.S. Provisional Patent Application No. 60/781,288, filed Mar. 9, 2006. Each of the above-identified applications is incorporated herein by reference in its entirety. This application is related to each application identified in the table below. The disclosure of each application identified in the table below is incorporated herein by reference in its entirety. Attorney Docket No.Application No.Filing DateTELAP004AC212/561,207Sep. 16, 2009TELAP004AC312/561,216Sep. 16, 2009TELAP004AC412/561,220Sep. 16, 2009TELAP004AC512/561,224Sep. 16, 2009TELAP004AC612/561,229Sep. 16, 2009TELAP004AC712/561,234Sep. 16, 2009TELAP004AC812/561,238Sep. 16, 2009TELAP004AC912/561,243Sep. 16, 2009TELAP004AC1012/561,246Sep. 16, 2009TELAP004AC1112/561,247Sep. 16, 2009TELAP004AC1212/563,031Sep. 18, 2009TELAP004AC1312/563,042Sep. 18, 2009TELAP004AC1412/563,051Sep. 18, 2009TELAP004AC1512/563,056Sep. 18, 2009TELAP004AC1612/563,061Sep. 18, 2009TELAP004AC1712/563,063Sep. 18, 2009TELAP004AC1812/563,066Sep. 18, 2009TELAP004AC1912/563,074Sep. 18, 2009TELAP004AC2012/563,076Sep. 18, 2009TELAP004AC2112/563,077Sep. 18, 2009TELAP004AC2212/567,528Sep. 25, 2009TELAP004AC2312/567,542Sep. 25, 2009TELAP004AC2412/567,555Sep. 25, 2009TELAP004AC2512/567,565Sep. 25, 2009TELAP004AC2612/567,574Sep. 25, 2009TELAP004AC2712/567,586Sep. 25, 2009TELAP004AC2812/567,597Sep. 25, 2009TELAP004AC2912/567,602Sep. 25, 2009TELAP004AC3012/567,609Sep. 25, 2009TELAP004AC3112/567,616Sep. 25, 2009TELAP004AC3212/567,623Sep. 25, 2009TELAP004AC3312/567,630Sep. 25, 2009TELAP004AC3412/567,634Sep. 25, 2009TELAP004AC3512/567,641Sep. 25, 2009TELAP004AC3612/567,648Sep. 25, 2009TELAP004AC3712/567,654Sep. 25, 2009TELAP004AC3812/571,343Sep. 30, 2009TELAP004AC3912/571,351Sep. 30, 2009TELAP004AC4012/571,357Sep. 30, 2009TELAP004AC4112/571,998Oct. 1, 2009TELAP004AC4212/572,011Oct. 1, 2009TELAP004AC4312/572,022Oct. 1, 2009TELAP004AC4412/572,046Oct. 1, 2009TELAP004AC4512/572,055Oct. 1, 2009TELAP004AC4612/572,061Oct. 1, 2009TELAP004AC4712/572,068Oct. 1, 2009TELAP004AC4812/572,077Oct. 1, 2009TELAP004AC4912/572,091Oct. 1, 2009TELAP004AC5012/572,194Oct. 1, 2009TELAP004AC5112/572,201Oct. 1, 2009TELAP004AC5212/572,212Oct. 1, 2009TELAP004AC5312/572,218Oct. 1, 2009TELAP004AC5412/572,221Oct. 1, 2009TELAP004AC5512/572,225Oct. 1, 2009TELAP004AC5612/572,228Oct. 1, 2009TELAP004AC5712/572,229Oct. 1, 2009TELAP004AC5812/572,232Oct. 1, 2009TELAP004AC5912/572,237Oct. 1, 2009TELAP004AC6012/572,239Oct. 1, 2009TELAP004AC6112/572,243Oct. 1, 2009TELAP004AC6213/774,919Feb. 22, 2013TELAP004AC6313/774,940Feb. 22, 2013TELAP004AC6413/774,954Feb. 22, 2013TELAP004AC6513/774,970Feb. 22, 2013TELAP004AC6613/827,615Mar. 14, 2013TELAP004AC6713/827,755Mar. 14, 2013TELAP004AC6813/834,302Mar. 15, 2013TELAP004AC6913/837,123Mar. 15, 2013TELAP004AC7514/304,778Jun. 13, 2014BECKP004B12/013,342Jan. 11, 2008BECKP004B.C113/073,994Mar. 28, 2011BECKP004B.C214/276,528May 13, 2014BECKP004C12/013,356Jan. 11, 2008BECKP004C.C113/047,474Mar. 14, 2011BECKP004D12/013,366Jan. 11, 2008BECKP004D.C113/007,582Jan. 14, 2011BECKP004D.C213/007,584Jan. 14, 2011TELAP01412/363,705Jan. 30, 2009TELAP015A12/402,465Mar. 11, 2009TELAP015AC112/753,711Apr. 2, 2010TELAP015AC212/753,727Apr. 2, 2010TELAP015AC312/753,733Apr. 2, 2010TELAP015AC412/753,740Apr. 2, 2010TELAP015AC512/753,753Apr. 2, 2010TELAP015AC612/753,758Apr. 2, 2010TELAP015AC6A13/741,298Jan. 14, 2013TELAP015AC712/753,766Apr. 2, 2010TELAP015AC7A13/589,028Aug. 17, 2012TELAP015AC812/753,776Apr. 2, 2010TELAP015AC912/753,789Apr. 2, 2010TELAP015AC1012/753,793Apr. 2, 2010TELAP015AC1112/753,795Apr. 2, 2010TELAP015AC1212/753,798Apr. 2, 2010TELAP015AC12A13/741,305Jan. 14, 2013TELAP015AC1312/753,805Apr. 2, 2010TELAP015AC1412/753,810Apr. 2, 2010TELAP015AC1512/753,817Apr. 2, 2010TELAP015AC1612/754,050Apr. 5, 2010TELAP015AC1712/754,061Apr. 5, 2010TELAP015AC1812/754,078Apr. 5, 2010TELAP015AC1912/754,091Apr. 5, 2010TELAP015AC2012/754,103Apr. 5, 2010TELAP015AC2112/754,114Apr. 5, 2010TELAP015AC2212/754,129Apr. 5, 2010TELAP015AC2312/754,147Apr. 5, 2010TELAP015AC2412/754,168Apr. 5, 2010TELAP015AC2512/754,215Apr. 5, 2010TELAP015AC2612/754,233Apr. 5, 2010TELAP015AC2712/754,351Apr. 5, 2010TELAP015AC27A13/591,141Aug. 21, 2012TELAP015AC2812/754,384Apr. 5, 2010TELAP015AC2912/754,563Apr. 5, 2010TELAP015AC3012/754,566Apr. 5, 2010TELAP015AC3113/831,530Mar. 14, 2013TELAP015AC3213/831,605Mar. 15, 2013TELAP015AC3313/831,636Mar. 15, 2013TELAP015AC3413/831,664Mar. 15, 2013TELAP015AC3513/831,717Mar. 15, 2013TELAP015AC3613/831,742Mar. 15, 2013TELAP015AC3713/831,811Mar. 15, 2013TELAP015AC3813/831,832Mar. 15, 2013TELAP015AC4014/242,308Apr. 1, 2014TELAP015AC4514/273,483May 8, 2014TELAP015AC4614/303,587Jun. 12, 2014TELAP015AC4714/476,511Sep. 3, 2014TELAP015AC4814/642,633Mar. 9, 2015TELAP01612/399,948Mar. 7, 2009TELAP01712/411,249Mar. 25, 2009TELAP017.D13/085,447Apr. 12, 2011TELAP01812/484,130Jun. 12, 2009TELAP01912/479,674Jun. 5, 2009TELAP02012/481,445Jun. 9, 2009TELAP02112/497,052Jul. 2, 2009TELAP021C113/540,529Jul. 2, 2012TELAP022A12/466,335May 14, 2009TELAP022B12/466,341May 14, 2009TELAP02312/512,932Jul. 30, 2009TELAP04812/435,672May 5, 2009TELAP04912/775,429May 6, 2010TELAP05112/904,134Oct. 13, 2010TELAP05313/373,470Jan. 14, 2011TELAP05413/312,673Dec. 6, 2011TELAP05513/473,439May 16, 2012TELAP05613/740,191Jan. 12, 2013
| Number | Date | Country | |
|---|---|---|---|
| 60781288 | Mar 2006 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 13774919 | Feb 2013 | US |
| Child | 14711731 | US | |
| Parent | 12572225 | Oct 2009 | US |
| Child | 13774919 | US | |
| Parent | 12212562 | Sep 2008 | US |
| Child | 12572225 | US | |
| Parent | 11683402 | Mar 2007 | US |
| Child | 12212562 | US |
