Method and system of reducing charged particle beam write time

Abstract

To reduce charged particle beam write time by reducing an exposure amount required to expose a shot or a shape in a relatively low-density exposure area by casting an artificial background exposure amount in an exposure area.SOLUTION: A method for exposing a pattern in an area on a surface using a charged particle beam lithography is disclosed and includes inputting an original set of exposure information for the area. A backscatter is calculated for the area of the pattern based on the exposure information. An artificial background exposure amount is determined for the area. The artificial background exposure amount includes additional exposure information and is combined with the original set of exposure information to create a modified set of exposure information. A system for exposing a pattern in an area on a surface using a charged particle beam lithography is also disclosed.SELECTED DRAWING: Figure 9

Description

[Technical field]

[0001] Related Applications This application claims the benefit of U.S. Non-Provisional Patent Application No. 16 / 231,447, filed December 22, 2018, and entitled “Method and System of Reducing Charged Particle Beam Write Time,” which is incorporated by reference for all purposes.

[0002] This disclosure relates to lithography, and more particularly to the design and manufacture of a surface, which may be a reticle, wafer, or any other surface, using charged particle beam lithography. [Background technology]

[0003] The three common types of charged particle beam lithography are unshaped (Gaussian) beam lithography, shaped charged particle beam lithography, and multi-beam lithography. In all types of charged particle beam lithography, a charged particle beam deposits energy on a resist-coated surface to expose the resist.

[0004] Shaped charged particle beam lithography can be variable shaped beam (VSB) or partial projection exposure (CP), where shots of an electron beam are precisely shaped and steered to expose a resist-coated surface, such as the surface of a wafer or the surface of a reticle. In VSB, these shapes are simple shapes, usually limited to rectangles with certain minimum and maximum sizes and sides that are parallel to the axes of a Cartesian coordinate plane (i.e., "Manhattan" orientation), and 45-degree right triangles (i.e., triangles whose three interior angles are 45 degrees, 45 degrees, and 90 degrees) with certain minimum and maximum sizes. At a given location, electron exposures are applied to resist with these simple shapes. The total writing time for this type of system increases with the number of shots. In CP, the system has a stencil that has various apertures or characters in it, which can be complex shapes such as straight angle, linear at any angle, circular, nearly circular, annular, nearly annular, elliptical, nearly elliptical, partially circular, partially nearly circular, partially annular, partially nearly annular, partially nearly elliptical, or any curved shape, or a connected set of complex shapes, or a group of disassembled sets of connected sets of complex shapes. An electron beam can be projected onto the characters on the stencil to efficiently produce more complex patterns on the reticle. In theory, such a system can be faster than a VSB system because more complex shapes can be projected per shot, which takes more time. Thus, a shot of an E-shaped pattern with a VSB system requires four shots, whereas the same E-shaped pattern can be projected in one shot with a CP system.

[0005] Shaped charged particle beam lithography can use either a single shaped beam or multiple shaped beams that expose a surface simultaneously, with multiple shaped beams providing faster writing speeds than a single shaped beam. When multiple charged particle beams expose a surface simultaneously, charged particle beam lithography is often referred to as multi-beam lithography. Multi-beam lithography can be multiple beams of shaped or unshaped charged particle beam lithography.

[0006] In lithography, a lithographic mask, or reticle, contains geometric patterns that correspond to the circuit components to be integrated onto a substrate. Computer-aided design (CAD) software A CAD program may be utilized to generate the patterns used to manufacture the reticle. In designing the pattern, the CAD program may follow a predefined set of design rules to create the reticle. These rules are set by process, design, and end-use limitations. An example of an end-use limitation is that defining the shape of a transistor in a way that the transistor cannot operate satisfactorily at the required supply voltage. In particular, design rules may define the space tolerance between circuit devices or internal wiring. Design rules are used, for example, to ensure that circuit devices or wiring do not interact with each other in an undesirable way. For example, design rules are used to ensure that wiring does not get too close to each other in a way that could cause a short circuit. Design rule limitations reflect, among other things, the smallest dimensions that can be reliably manufactured. When referring to these small dimensions, the concept of critical dimensions is usually introduced. These are defined, for example, as the critical width or area of ​​a feature, or the critical space or critical space area between two features, and these dimensions require exquisite control.

[0007] In the production or manufacturing of semiconductor devices such as integrated circuits, optical lithography can be used to fabricate the semiconductor devices. Optical lithography is a printing process in which a lithographic mask or photomask or reticle is used to transfer a pattern onto a substrate such as a semiconductor or silicon wafer to create an integrated circuit (IC). Other substrates can include flat panel displays, holographic masks, or other reticles. Conventional optical lithography typically uses radiation with wavelengths of 193 nm or greater. Extreme ultraviolet (EUV) or X-ray lithography are also considered types of optical lithography, but use wavelengths much shorter than the 193 nm of conventional optical lithography. A reticle, or multiple reticles, can contain circuit patterns corresponding to individual layers of an integrated circuit, which can be imaged onto a specific area on a substrate coated with a layer of radiation-sensitive material known as photoresist or resist. Conventional optical lithography writers typically reduce the photomask pattern by up to a factor of four during the optical lithography process. Therefore, the pattern formed on the reticle or mask needs to be four times larger than the size of the desired pattern on the substrate or wafer.

[0008] In the production or manufacturing of semiconductor devices such as integrated circuits, non-optical methods can be used to transfer a pattern on a lithography mask to a substrate such as a silicon wafer. Nanoimprint lithography (NIL) is an example of a non-optical lithography process. In NIL, the lithography mask pattern is transferred to the substrate by contact between the lithography mask and the substrate. Lithography masks for NIL are typically manufactured as surface products using charged particle beam lithography.

[0009] In the production or manufacturing of semiconductor devices such as integrated circuits, maskless direct writing can also be used to manufacture semiconductor devices. Maskless direct writing is a printing process in which charged particle beam lithography is used to transfer a pattern to a substrate such as a semiconductor or silicon wafer to create an integrated circuit. Other substrates can also include flat panel displays, imprint masks for nanoimprinting, or reticles. The desired pattern of a layer is written directly onto a surface, which in this case is also a substrate. Once the patterned layer is transferred, it can undergo various other processes such as etching, ion implantation (doping), metallization, oxidation, and polishing. These processes are used to finish individual layers in the substrate. If several layers are required, the entire process or a variation thereof is repeated for each new layer. To produce the same substrate, some of these layers can be written using optical lithography and others can be written using maskless direct writing. Also, Some patterns of a given layer can be written using optical lithography, others using maskless direct writing. Eventually, a combination of multiple devices or integrated circuits will be present on the substrate. These integrated circuits are then separated from one another by dicing or sawing and packaged into individual packages. In the more general case, the pattern on the surface can be used to define artifacts such as display pixels, holograms, or magnetic recording heads.

[0010] One goal of integrated circuit manufacturing by optical lithography is to reproduce the original circuit design on a substrate through the use of a reticle, sometimes referred to as a mask or photomask, which is a surface that can be exposed using charged particle beam lithography. Integrated circuit manufacturers are always trying to use the physical material of semiconductor wafers as efficiently as possible. As engineers continue to reduce the size of circuits, integrated circuits contain more circuit elements, allowing them to use less power. As the size of the critical dimensions of integrated circuits decreases and their circuit density increases, the critical dimensions of the circuit pattern or physical design approach the resolution limit of the optical exposure tools used in conventional optical lithography. As the critical dimensions of the circuit pattern become smaller and approach this resolution value of the exposure tools, it becomes more difficult to accurately transfer the physical design into the actual circuit pattern developed on the resist layer. For the further use of optical lithography to transfer patterns with features smaller than the wavelength of light used in the optical lithography process, a process known as optical proximity correction (OPC) was developed. OPC modifies the physical design to compensate for distortions caused by effects such as optical diffraction and the optical interaction of features with nearby features. Resolution enhancement techniques implemented by reticles include OPC and inverse lithography techniques (ILT).

[0011] OPC can add sub-resolution lithography features to a mask pattern to reduce the difference between the original physical design pattern, i.e., between this design and the final circuit pattern transferred onto the substrate. The sub-resolution lithography features interact with the original pattern in the physical design and with each other to compensate for proximity effects and improve the final transferred circuit pattern. One feature that is added to improve the pattern transfer is called a "serif." A serif is a small feature that improves the accuracy or resilience to manufacturing variations in the printing of a particular feature. An example of a serif is a small feature that is positioned on a corner of a pattern to sharpen the corner in the final transferred image. The pattern that is intended to be printed on the substrate is called the main feature. The serif is a part of the main feature. It is common to discuss the OPC decorated pattern written on the reticle in terms of the main feature (i.e., the feature that reflects the design before OPC decoration) and the OPC features, which can include serifs, jogs, sub-resolution assist features (SRAFs), and negative features. OPC features must follow various design rules, including rules based on the minimum size of a feature that can be transferred to a wafer using optical lithography. Other design rules can come from the mask manufacturing process, or from the stencil manufacturing process if a lithographic charged particle beam writing system is used to form the pattern on the reticle.

[0012] SRAF features are used to improve the transfer of patterns. The precision or accuracy required in the surface manufacturing process for SRAF is often lower than that for the main feature. There are similar decorations introduced by OPC as negative features. Inside larger features there can be narrow spaces introduced to improve lithography performance. Because narrow spaces are harder to reliably write than similarly narrow lines, these negative features are often used to improve the final Generally, getting a good impression on the surface is often the most difficult part. Summary of the Invention [Means for solving the problem]

[0013] A method and system for exposing a pattern in a region on a surface using charged particle beam lithography includes inputting an original set of exposure information for the region, calculating backscatter for the region of the pattern based on the exposure information, and determining an artificial background exposure amount for the region, the artificial background exposure amount including additional exposure information and being combined with the original set of exposure information to create a modified set of exposure information. [Brief description of the drawings]

[0014] [Figure 1] 1 illustrates an example of a variable shaped beam (VSB) charged particle beam system known in the art. [Diagram 2] 1 shows an example of an electro-optical schematic of a multi-beam exposure system, as known in the art. [Diagram 3] A to D show various types of shots known in the art. [Figure 4] 1 illustrates an example of a multi-beam charged particle beam system known in the art. [Diagram 5] AC show the extent to which areas with both large and small features affect exposure dose after PEC, as known in the art. [Figure 6] AC show the extent to which areas with both large and small features affect exposure dose after PEC, as known in the art. [Figure 7] AG show rectangular shaped PEC iterations and resulting dose curves, as known in the art. [Figure 8A] Illustrated are examples of low density exposures versus high density exposures as known in the art. [Figure 8B] Illustrated are examples of low density exposures versus high density exposures as known in the art. [Figure 8C] Illustrated are examples of low density exposures versus high density exposures as known in the art. [Figure 8D] Illustrated are examples of low density exposures versus high density exposures as known in the art. [Figure 8E] Illustrated are examples of low density exposures versus high density exposures as known in the art. [Figure 8F] Illustrated are examples of low density exposures versus high density exposures as known in the art. [Figure 9] A-B show low density exposure areas with added artificial background exposure according to an embodiment of the present disclosure. [Figure 10] A-F show exemplary embodiments of artificial background exposure. [Figure 11] 1 illustrates a conceptual flow diagram for preparing a surface or reticle for use in manufacturing substrates such as integrated circuits on silicon wafers, as known in the art. [Figure 12] 1 illustrates an exemplary computing hardware device for use in embodiments of these methods. [Figure 13] 4 illustrates a flow in which artificial background exposure is calculated according to some embodiments. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] This disclosure describes a method for reducing charged particle beam writing time by reducing the exposure dose required to expose a shot or feature in a relatively low density exposure area by casting an artificial background exposure dose in the area.

[0016] The cost of charged particle beam lithography is directly related to the time required to expose a pattern onto a surface, such as a reticle or wafer. Traditionally, exposure time is related to the amount of light needed to write the pattern. For most complex integrated circuit designs, forming a set of layer patterns on either a set of reticles or a substrate is an expensive and time-consuming process. It would therefore be advantageous to be able to reduce the exposure light required to form these patterns, such as by reducing the time required to image these patterns on reticles and other surfaces.

[0017] Referring now to the drawings, where like numbers refer to like items, FIG. 1 illustrates one embodiment of a lithography system, such as a charged particle beam writer system, in this case an electron beam writer system 100, that uses a variably shaped beam (VSB) to produce a surface 112. The electron beam writer system 100 includes an electron beam source 114 that projects an electron beam 116 toward an aperture plate 118. The plate 118 includes an aperture 120 that is formed in the plate to allow the electron beam 116 to pass through. Once the electron beam 116 passes through the aperture 120, it is directed or deflected by a lens system (not shown) as an electron beam 122 toward another rectangular aperture plate or stencil mask 124. The stencil 124 has a plurality of openings or apertures 126 formed therein that define a variety of simple shapes, such as rectangles and triangles. Each aperture 126 formed in the stencil 124 can be used to form a pattern in a surface 112 of a substrate 134, such as a silicon wafer, reticle, or other substrate. The electron beam 130 exits one of the apertures 126 and passes through electromagnetic or electrostatic demagnification lenses 138, which reduce the size of the pattern exiting the aperture 126. In commonly available charged particle beam writer systems, the demagnification ratio is between 10 and 60. The demagnified electron beam 140 exits the demagnification lens 138 and is directed by a series of deflectors 142 onto the surface 112 as a pattern 128. The surface 112 is coated with a resist (not shown) that reacts with the electron beam 140. The electron beam 122 can be directed to overlap a variable portion of the aperture 126, which affects the size and shape of the pattern 128. A blanking plate (not shown) can be used to deflect beam 116 or shaped beam 122 to prevent the electron beam from reaching surface 112 during a period of time after each shot while the lenses directing beam 122 and deflector 142 are realigned for the next shot. By deflecting beam 116 onto non-apertured portions of stencil 124, stencil 124 can itself act as a deflector.Typically, the blanking plate is positioned to deflect the electron beam 116 to prevent it from illuminating the aperture 120. Conventionally, the blanking period can be a fixed length of time, or it can vary, for example, depending on the extent to which the deflector 142 needs to be realigned for the position of the next shot. It is noted that the VSB system can be considered a special (simple) case of the partial block exposure method, where the characters are only simple characters, usually rectangles or 45 degree right triangles. It is also possible to partially expose the characters. This can be done, for example, by blocking part of the particle beam. In this disclosure, partial projection is used to mean both partial block exposure and VSB projection.

[0018] In the electron beam writer system 100, the substrate 134 is mounted on a movable platform or stage 132. The stage 132 allows the substrate 134 to be repositioned so that a pattern larger than the maximum deflection capability or field size of the charged particle beam 140 can be written on the surface 112 in a series of subfields, each subfield being within the capability of a deflector 142 that deflects the beam 140. In one embodiment, the substrate 134 can be a reticle. In this embodiment, the reticle is exposed with a pattern and then undergoes various manufacturing steps that make it into a lithographic mask or photomask. This mask can then be used in an optical lithography tool to project an image of the reticle pattern 128, typically reduced in size, onto a silicon wafer to produce integrated circuits. More typically, the mask is used in another device to transfer the pattern 128 onto a substrate (not shown).

[0019] The shot exposure of a charged particle beam writer such as an electron beam writer system, whether a VSB, CP, or multi-beam writer, is a function of the intensity of the beam source 114 and, in this VSB example, the exposure time per shot. Typically, the beam intensity remains fixed and the exposure time is varied to obtain a variable shot exposure. In a process called proximity effect correction (PEC), the exposure time can be varied to compensate for various long-range effects such as medium-range effect correction (MEC), loading effect correction (LEC), and fogging effect correction (FEC), as well as backscatter. In this disclosure, the term PEC is used to describe MEC, LEC, FEC, and backscatter correction. An electron beam writer system allows for an overall exposure dose to be set, typically called the base exposure dose, which affects all shots in an exposure pass. Some electron beam writer systems perform exposure compensation calculations within the electron beam writer system itself and do not allow for the exposure dose of each shot to be individually assigned as part of the input shot list, so the input shots have unassigned shot exposures. In electron beam lithography systems such as these, prior to PEC, all shots have a base dose. Other electron beam lithography systems allow for dose allocation on a shot-by-shot basis. In electron beam lithography systems that allow for dose allocation on a shot-by-shot basis, the number of available dose levels may be 64-4096 or more, or there may be a relatively small number of available dose levels, such as 3-8 levels. Some embodiments of the present disclosure are directed to use with charged particle beam lithography systems that allow for dose level allocation.

[0020] Charged particle beam systems can expose a surface with multiple individually controllable beams or beamlets. FIG. 2 shows an electro-optical schematic in which there are three charged particle beamlets 210. Associated with each beamlet 210 is a beam controller 220. Each beam controller 220 can, for example, allow its associated beamlet 210 to hit the surface 230, as well as prevent the beamlet 210 from hitting the surface 230. In some embodiments, the beam controller 220 can also control the beam blur, magnification, size and / or shape of the beamlet 210. In this disclosure, a charged particle beam system with multiple individually controllable beamlets is referred to as a multi-beam system. In some embodiments, charged particles from a single source can be subdivided to form multiple beamlets 210. In other embodiments, multiple sources can be used to generate multiple beamlets 210. In some embodiments, the beamlets 210 can be shaped by one or more apertures, while in other embodiments, there may be no apertures to shape the beamlets. Each beam controller 220 can individually control the exposure duration of its associated beamlet. Generally, the beamlets are reduced in size by one or more lenses (not shown) before striking a surface 230, which is typically coated with resist. In some embodiments, each beamlet can include a separate electro-optic lens, while in other embodiments, multiple beamlets, possibly including all beamlets, share an electro-optic lens.

[0021] 3A-3D show various types of shots. FIG. 3A shows an example of a rectangular shot 310. A VSB charged particle beam system can, for example, form rectangular shots with various x and y dimensions. FIG. 3B shows an example of a CP shot 320, which in this example is circular. A CP shot can represent any shape, for example an octagon, or a more complex shape such as the letter E. For the purposes of this disclosure, a shot is the exposure of a surface area over a period of time. The area may consist of multiple discontinuous small areas. A shot may consist of multiple other shots, which may or may not overlap and may or may not be exposed simultaneously. A shot may include a specified exposure amount, or may not have a specified exposure amount. The shots may use shaped beams, unshaped beams, or a combination of shaped and unshaped beams.

[0022] FIG. 3C shows an example of a shot 350 that is an array of circular patterns 352. Shot 350 can be formed in a variety of ways, including multiple shots of a single circular CP character, one or more shots of a CP character that is an array of circular apertures, and one or more multi-beam shots using a circular aperture. FIG. 3D shows an example of a shot 360 that is a sparse array of rectangular patterns 362 and 364. Shot 360 can be formed in a variety of ways, including multiple VSB shots, a CP shot, and one or more multi-beam shots using a rectangular aperture. In some multi-beam embodiments, shot 360 can include multiple interleaved groups of other multi-beam shots. For example, pattern 362 can be illuminated simultaneously, and then pattern 364 can be illuminated simultaneously at a different time than pattern 362.

[0023] FIG. 4 illustrates one embodiment of a charged particle beam exposure system 400. The charged particle beam system 400 is a multi-beam system in which multiple individually controllable shaped beams can simultaneously expose a surface. The multi-beam system 400 includes an electron beam source 402 that generates an electron beam 404. The electron beam 404 is directed to an aperture plate 408 by a capacitor 406, which may include electrostatic and / or magnetic elements. The aperture plate 408 includes multiple apertures 410 through which the electron beam 404 illuminates and through which the electron beam 404 passes to form multiple shaped beamlets 436. Thus, the multi-beam charged particle beam system 400 includes a charged particle beam source 402 and an aperture plate 408, which includes multiple apertures 410 through which the source of the charged particle beam 402 illuminates. In some embodiments, the multi-beam charged particle beam system includes a single aperture plate 408. In some embodiments, the aperture plate 408 can include hundreds or thousands of apertures 410. While FIG. 4 shows an embodiment with a single electron beam source 402, in other embodiments, the aperture 410 can be illuminated by electrons from multiple electron beam sources. As shown in FIG. 4, the aperture 410 can be circular, or of a different shape, for example, a rectangular aperture 120 as shown in FIG. 1. The set of beamlets 436 then illuminates a blanking controller plate 432. The blanking controller plate 432 includes multiple blanking controllers 434, each of which is aligned with a beamlet 436. Each blanking controller 434 can individually control its associated beamlet 436 to either allow the beamlet 436 to strike the surface 424 or prevent the beamlet 436 from striking the surface 424. The length of time the beam strikes the surface controls the total energy or "exposure" that the beamlet imparts. Typically, a multi-beam writer scans the entire area to be exposed.Thus, the writing time is constant regardless of the number of features or the complexity of the features.

[0024] The substrate 426 is positioned on a movable platform or stage 428, which can be repositioned using actuators 430. By moving the stage 428, the beam 440 can use multiple exposures or shots to expose an area larger than the dimensions of the largest size pattern formed by the beamlets 440. In some embodiments, the stage 428 remains stationary during an exposure and is then repositioned for the next exposure. In other embodiments, the stage 428 moves continuously and at a variable speed. In yet other embodiments, the stage 428 moves continuously but at a constant speed to allow for greater precision in the positioning of the stage. The stage 428 For those embodiments in which the beamlets move continuously, a set of deflectors (not shown) can be used to move the beams to match the direction and velocity of stage 428, allowing beamlet group 440 to remain stationary with respect to surface 424 during exposure. In yet other embodiments of a multi-beam system, individual beamlets within a beamlet group can be deflected across surface 424 independently of other beamlets in the beamlet group.

[0025] Other types of multi-beam systems can generate multiple unshaped beamlets 436, such as by using multiple charged particle beam sources to generate a Gaussian beamlet array. Although this disclosure is described with respect to charged particle beam lithography, the methods described can also be applied to laser lithography, and in particular to multi-beam laser lithography.

[0026] The exposure dose is controlled in all precision electron beam (eBeam) writers by controlling the exposure time, rather than dynamically varying the current density of the beam. It is understood that in practice the exposure time can be controlled much more precisely than the current density. For this reason, high dose exposures are achieved by increasing the time the beam is on compared to low dose exposures. It is contemplated that eBeam writers have variable current densities. The embodiments of the present disclosure are applicable where exposure dose control is achieved at least in part by controlling the exposure time. The exposure time for each beamlet 436 in a beamlet group can be individually controlled using the blanking controller 434. Thus, the multi-beam writing time is determined by the maximum exposure dose that needs to be applied. In writers where the stage stops for each beamlet group, such as a stripe, the beamlet with the highest exposure dose will influence the exposure time for the entire stripe. In writers with a constant velocity stage, the velocity may be constant across the entire design, which may be a reticle subset, the entire reticle or wafer, or any other surface, or may be partially constant for a period of time. Therefore, the beamlet with the highest exposure in the stripe can determine the stage speed and therefore the writing time for the entire design or surface. Even with a variable speed stage, the speed cannot be changed abruptly because the stage is heavy to provide stability and accelerating or decelerating consumes a lot of energy, causing vibration and blurring.

[0027] At advanced process nodes, features that need to be accurately written on the mask will be less than three times the size of the typical combined blur radius of forward scattering and resist blur for typical resists used to print the mask fast enough to be commercially worthwhile. Typical state-of-the-art masks write sub-resolution assist features (SRAFs) on the mask with widths less than 60 nm, with combined blurs in the 20 nm range or better. For masks for EUV lithography, it is generally expected that SRAFs of about 30 nm width will often be required on the mask. The optical proximity correction (OPC) or inverse lithography techniques (ILT) steps that generate the mask shape can further improve wafer performance if smaller shapes (such as SRAFs) can be printed accurately. However, it is also important that the mask shapes generated by OPC / ILT can be reliably generated on the mask across mask process variants. Establish mask design rules such as minimum allowable feature width and minimum allowable feature spacing on the mask, which are trade-offs. OPC / ILT must ensure that the output follows the mask design rules. Mask makers need to produce masks that precisely follow the mask design rules and across manufacturing variations to some agreed upon specification. The narrower the feature, the more difficult it is to write on the mask. Small square features are the most difficult to write. However, they have a smaller impact on wafer performance and are therefore important for cutting-edge mask processing, where narrow SRAFs must be written precisely.

[0028] That such small features are written on a mask in the presence of other, larger features on the same mask is typically achieved by dose adjustment. Since the overall mask writing time is one of the major drivers of mask cost, there is a trade-off between accuracy and speed / cost, choosing a resist that is sensitive enough to achieve the desired accuracy for the larger "main" features. However, the writing of SRAFs and other smaller features must also be done accurately, and a common technique is to boost (i.e. increase) or partially boost the exposure of the SRAFs to make them print better.

[0029] In the industry, the amount by which the standard exposure is boosted can generally vary from 1.2 times the standard exposure to 3 times the standard exposure, although the exposure can be any multiple of the standard exposure. A larger boost can lead to longer writing times, but can lead to more accurate printing of smaller features. Figures 5A-5B show the difference between a sufficiently large feature 502 and its exposure profile (i.e., exposure curve) 512 and a smaller SRAF feature 506 and its exposure profile 522, both of which are exposed with a standard exposure, measured along lines 504 and 508 through the shot contours of features 502 and 506, respectively, which have the same resist threshold as shown by line 514. Features 502 and 506 are broken down to show rectangular shapes that are much longer than their width, and each feature is exposed on the resist using the standard exposure at half the threshold of the standard exposure. Exposures below the resist threshold do not print.

[0030] In a VSB device, the shot contour is the shape of the eBeam projection directed to the writer. In a multi-beam device, the shot contour is the shape of the desired shape directed to the writer, resulting in rasterization to pixels and eBeam projection of different exposures depending on the pixel used to write the shape. For ease of understanding, in this disclosure, the exposure profile is depicted and described as if the device were a VSB, but the embodiments can also be applied to multi-beam. In a multi-beam device, rasterization to pixels further complicates the exposure profile, which is dependent on the position of the shape relative to the pixel grid. For understanding of the concepts of this disclosure, these additional complexities of multi-beam writing are not relevant.

[0031] For a large enough feature, dose profile 512 reaches a plateau at the standard dose. This plateau is the same dose even if the width of feature 502 is large. For a smaller feature, shown by feature 506, there is not enough energy to make the peak of the dose profile reach the standard dose. The feature in dose profile 522 does not have a plateau. Furthermore, dose profile 522 does not cross resist threshold 514 at the same location as the width of feature 506. Because there is not enough energy, the exposed feature on the resist is narrower than desired.

[0032] The exposure dose margins on the left edges of features 502 and 506 are indicated by edge slopes 516 and 528, respectively. Slope 528 is shallower than slope 516. The narrower the feature 506, the lower the exposure dose profile and the shallower the slope 528. Once feature 502 is wide enough that the exposure dose profile 512 reaches a plateau, the wider the feature, the more the slope 516 does not change. The shallower the slope, the less adequate the exposure dose margin. That is, the critical dimension (CD), which is the width of the feature, experiences more variation given a particular exposure dose variation. As used herein, the term exposure dose margin refers to the tolerance of a pattern defined by a charged particle beam shot set against all types of manufacturing variations, including those related to exposure dose. A better exposure dose margin indicates a higher tolerance. Those skilled in the art will appreciate that resilience to exposure dose variations is a fundamental property of many sources of manufacturing variation. It is generally understood that the dose margin is a good indicator of the amount of exposure that can be achieved. To improve the resilience to manufacturing variations as indicated by the dose margin, one conventional method is to use a base dose higher than standard for smaller features, as indicated by the dose profile 532 in FIG. 5C. In this example, a base dose greater than 1.0 is used, where the dose is selected such that the CD, as indicated by the distance between the points of the dose profile 532 above the resist threshold 514, is the desired width of the target feature 506. Many combinations of dose adjustment and shape adjustment are used in the industry. FIG. 5C is an example where only dose adjustment is deployed. In this example, once the dose is calculated for the feature 506, the dose margin can be measured by calculating the slope 538 of the dose profile 532 above the threshold 514. It is noted that edge slope 538 may be more suitable (i.e., have a steeper slope) than edge slope 528, but still less suitable than edge slope 516.

[0033] Dose margin is also important for printability of small features because of contrast. Contrast is the difference in the amount of energy applied to the resist in the inner border of the exposure area versus the amount of energy applied to the resist in the outer border of the exposure area. The border in semiconductor mask processing can be a few nanometers to tens of nanometers. Since the exposure profile is a continuous function on the length scale of interest, dose margin and contrast are highly correlated to each other. Insufficient contrast causes features to be unresolvable. Reducing the minimum size of features that can be reliably resolved in a given resist is important to the economics of mask making.

[0034] There is another factor that affects the dose that the eBeam writer needs to cast directly to print a feature on a surface. This is called proximity effect correction (PEC) and corrects for backscattering of electrons cast by the eBeam "shot" around the area of ​​interest. Figures 6A-6B show the resulting doses corresponding to the shots of Figures 5A-5B after PEC. In Figure 6A, the backscattering of electrons cast by the shot around feature 502 and its surroundings, which are typically within a distance from the shot on the order of 10 μm, is shown by the dashed line representing backscattering 618. If the surrounding area is densely dosed with exposure, backscattering 618 will be high. If the surrounding area is sparsely dosed with exposure, backscattering 618 will be low, perhaps close to zero. PEC, as is commonly implemented, iteratively optimizes the dose applied to a large enough feature 502 so that the dose applied to illuminate feature 502 causes the width of feature 502 to be printed accurately despite the additional dose contributed by backscattering. The corrected dose profile 612 of the exposure (excluding backscatter) has less exposure than dose profile 512. Edge slope 616 is shallower than edge slope 516. The more backscatter received from adjacent regions, the less adequate the dose margin. FIG. 6B shows the same PEC applied to smaller features, such as for SRAF. Since one would typically apply PEC to all features of all sizes in the same way, dose profile 622 (not including the base dose due to backscattering, shown by the horizontal dashed line for backscatter 618) would be even smaller in magnitude compared to dose profile 522. Thus, edge slope 628 is even shallower than edge slope 528, and substantially shallower than edge slope 616, resulting in less adequate dose margin and making these features much more susceptible to manufacturing variations. FIG. 6C shows the dose profile after PEC for the shot described in FIG. 5C, where the dose has been increased to meet the desired critical dimension (CD). The edge slope 638 is shallower (bad) than the pre-PEC edge slope 538, but is better than the post-PEC edge slope 628 at standard exposure.

[0035] In some embodiments, providing sufficient exposure margin involves increasing the exposure so that the exposure after PEC is within pre-set limits (such as 0.7 to 1.3 times the standard exposure) and performing linearity correction to ensure that the size of the features is accurate. This includes:

[0036] Backscattering is caused by charged particles such as electrons that "bounce" back after colliding with the resist and the material underneath it. The effective radius of backscattering may be, for example, 10 microns, which is much larger than the size of the shot. Thus, all backscattering from nearby shots within that effective radius of the shot adds exposure dose at the location of the shot. If not corrected, backscattering can cause the pattern to register more widely by adding more exposure dose to the shot than the intended exposure dose of the shot (especially in areas of high exposure density).

[0037] 7A-7G show an iterative sequence applied by PEC, which corrects the exposure dose used to irradiate each feature, as known in the art, to correct for the effects of backscattering. FIG. 7A shows an example of a rectangular shot 702 contour. The interruptions indicate that shot 702 can be arbitrarily long in the vertical direction. The width of shot 702, the distance between "g" and "h", is assumed to be between 100 nm and 300 nm in this example for illustration purposes, which assumes a commonly used mask manufacturing process and is significantly smaller than the assumed backscattering range of about 10 μm. Shot sizes for multi-beam devices are typically 5-20 nm. For multi-beam devices, a collection of shots that are combined irradiate a feature such as shot 702. Backscattering is applied to a single shot in the same way as the collection of shots, and PEC correction is applied in the same way. FIG. 7B shows an example of an exposure dose graph 710 showing the exposure dose along line 704 through shot contour 702 at a standard shot exposure, with little backscattering. Other long distance effects are also assumed to have no contribution to the background exposure in FIG. 7B, leading to a background exposure level of near zero. The total exposure delivered to the resist is shown on the y-axis and is 1.0 times the standard exposure. Due to the near zero background exposure, the total exposure and the shot exposure are nearly the same. The exposure graph 710 also shows the resist threshold 714 at half the standard exposure. The CD variation of the feature represented by the exposure graph 710 in the x-direction is inversely proportional to the slope of the exposure curve (i.e., exposure profile) 712 at the x-coordinates "g" and "h" where it crosses the resist threshold 714. The edge slope 716 of the exposure curve 712 at the threshold 714 is shown as the hypotenuse of a right triangle.

[0038] The condition of FIG. 7B of near-zero background exposure is not reflective of a real design. A real design will usually have many other shots within the backscatter distance of shot 702. FIG. 7C shows an example of an exposure dose graph 720 of a shot with standard exposure with backscatter 728 due to 50% exposure density. In the exposure dose graph 720, the exposure dose curve 722 shows the cross-sectional exposure dose of shot 702 in addition to the background exposure (backscatter 728). The edge slope 726 of the exposure dose curve 722 at threshold 714 is shown as the hypotenuse of a right triangle. The CD variation of curve 722 is larger (worse) than that of curve 712, as shown by the shallower edge slope 726 where curve 722 crosses resist threshold 714 beyond points "g" and "h" compared to edge slope 716. The shallower edge slope 726 is due to the resist threshold intersecting the lower part of the exposure curve where the exposure margin is inadequate due to background exposure caused by backscatter 728. The intersection of the exposure curve 722 at the resist threshold 714 is wider than points "g" and "h", indicating that the printed CD is larger than desired. Backscatter 728 "pre-exposes" the resist, so irradiating the exposure curve 722 with a standard exposure in the presence of backscatter 728 causes the CD to print larger than desired.

[0039] FIG. 7D shows the new dose curve 732 after PEC. The intermediate dose graph 725 still shows 50% backscatter 728, but with a reduced dose curve 732. This reduced dose curve, along with the current backscatter, is now consistent with the dose curve 732 and the resist threshold 71. 7. The intersection with 4 was calculated to be exactly measured at "g" and "h". This is the result of the first iteration of PEC, which corrects for the overexposure due to the presence of backscatter. In this iteration, the amount of backscatter is assumed to be the same as backscatter 728, since there is no way to know what the new backscatter will be after PEC has been run on all the surrounding shots. The PEC calculation reduces the exposure dose applied to shot 702 so that resist threshold 714 intersects with exposure profile 732 at "g" and "h". The edge slope 736 of exposure curve 732 at threshold 714 is shown as the hypotenuse of a right triangle. Note that slope 736 is shallower than slope 716, which would cause PEC to make the exposure margin inadequate.

[0040] The exposure dose graph 730 in FIG. 7E shows why PEC requires multiple iterations. The PEC calculation performed as shown in FIG. 7D will arrive at "g" and "h" accurately if and only if backscatter 738 is identical to backscatter 728. However, since PEC is performed for every shot, applying PEC to all shots around this shot will result in a reduction in backscatter 738 and a reduction in the exposure dose for each shot (because the exposure dose for this shot has been reduced from curve 722 to curve 732). The new backscatter is shown in FIG. 7E as backscatter 738. Backscatter 738 is lower than backscatter 728. As can be seen, the reduction in backscatter results in a recorded pattern size that is smaller than the desired size of the original shot 702.

[0041] The second iteration of PEC corrects for this by recalculating the PEC with this new (reduced) amount of backscatter (Backscatter 738). Now that the backscatter is lower, the dose profile 732 needs to be increased to reach the target "g" and "h" and print a feature such as shot 702 with the corrected CD.

[0042] As shown in dose graph 740 of FIG. 7F, increasing the dose for PEC on all features of the design results in a dose profile 742 that is higher than dose curve 732. Then, after all other neighboring features have been adjusted, there is backscatter 748 that is higher than backscatter 738. A third iteration of PEC corrects for this by again recalculating the PEC with this new (increased) amount of backscatter. These iterations continue until they converge to within a predefined tolerance of the amount of backscatter, as shown by dose level 758 in dose graph 750 of FIG. 7G, otherwise the dose profile 752 produces exactly the desired size of the original shot 702 at threshold 714. The converged dose is then used to irradiate each individual feature on the mask, resulting in a backscatter corrected set of doses for all features as the PEC result.

[0043] 8A-8B show an example of how PEC affects isolated features. In FIG. 8A, features 810 and 815 are isolated patterns in an area that is otherwise unoccupied by other exposures. The initial exposure can be calculated at 1.0 times the standard exposure for both features. FIG. 8B shows that the calculated exposure after PEC is still approximately 1.0 of the standard exposure, and is substantially unaffected as a result of there being no surrounding backscatter other than that contributed by features 810 and 815.

[0044] 8C-8D show an example of how PEC affects features in a 50% exposure density region (i.e., 50% of the region is covered by the pattern). In FIG. 8C, features 820, 822, 824, 826, and 828 are shown in a repeated line-and-space pattern, where the exposed features and spaces between them are equal in width, and the line-and-space pattern is repeated in all directions with an effective radius of 10 μm. The initial exposure dose per feature can be calculated at 1.0 times the standard exposure dose. FIG. 8D shows, as an example, that the exposure dose is reduced to 0.67 of the standard exposure dose for all features in the repeated pattern during PEC due to ambient backscattering from neighboring features.

[0045] 8E-8F show an example of how PEC affects features in a high density region. In FIG. 8E, features 830, 832, and 834 are shown, with feature 832 being the feature of interest. Features 830 and 834 are wide exposed areas with widths greater than the effective radius of 10 μm, and the pattern extends up and down to the effective radius of 10 μm. The initial exposure dose for each feature can be calculated at 1.0 times the standard exposure dose. FIG. 8F shows that the exposure dose is reduced, for example, to 0.5 of the standard exposure dose for feature 832, due to ambient backscattering from neighboring features in the PEC. The PEC reduces the exposure dose for each exposure by a computationally appropriate amount to make the CD equal to what it would be if there was no backscattering for a large enough feature. Thus, the PEC effectively annihilates the effect of backscattering. Backscattering is always a positive value. Thus, the PEC always reduces the exposure dose to compensate.

[0046] Because PEC works by a "flattened" Gaussian, the dose required to expose a feature in an area with high backscattering is significantly less. The difference in the dose required between an area with high backscattering (where the feature is surrounded by a high dose density) and an area with no backscattering (where the feature is surrounded by a low dose density) can be as much as or more than a factor of two. Modified Exposure D for a Pixel or Shot with a Standard Exposure for the First Iteration of the PEC Calculation pec is given by the following equation:

[0047]

number

[0048] At the resist threshold of 0.5 standard exposure, T emp is the forward scattering fraction at the resist threshold, and eta (η) is a normalization constant. T emp , eta (η) of 0.5, and exposure density of 100%, D pecis calculated to be 0.5. The SRAF being illuminated is assigned twice the standard exposure (2.0x), but if the SRAF is in a high density area with high backscattering, the pixel or shot will use approximately 1.0x the standard exposure, 0.5*(2.0)=1.0, after the first iteration of the PEC calculation.

[0049] This is because that area is exposed primarily by energy from backscattering, i.e. only a small fraction of the energy from this pixel or shot is needed to cast enough energy to reach the threshold for exposing the resist.

[0050] Areas with near-zero backscatter require the most eBeam exposure per shot. Therefore, for constant write-time multi-beam writers, the write time for the entire device is paced by writing these isolated patterns with the highest boosted exposure, usually for narrow features such as SRAF. Typical test masks have a mix of patterns in high density areas and patterns in low density areas. Typical production masks have much less variability. Some masks have patterns that are all high density (e.g., 70% average exposure density). Other masks have patterns that are all sparse (e.g., 25% average exposure density). However, many production masks combine several test patterns among them and therefore cannot guarantee low variability. In either case, the multi-beam writing speed, especially for constant write-time writers, is determined by the pixel combining the boosted high exposure into an environment with less backscatter cast. To a lesser extent, but still significantly, the writing speed of VSB writers and possibly variable writing time writers such as multi-beam writers, is paced by shots that combine enhanced high exposure into an environment with low backscatter.

[0051] Artificial Background Exposure In some embodiments, the introduction of artificial background exposure in areas that would otherwise have low backscattering reduces pixel or shot exposure and therefore reduces the overall writing time, resulting in lower pixel or shot exposure after PEC, a reduced maximum pixel or shot exposure for a mask or mask section, and an increased stage travel speed, significantly reducing the overall writing time.

[0052] 9A-B show how the introduction of artificial background exposure according to some embodiments can mimic 50% density similar to FIGS. 8C and 8D. FIG. 9A shows isolated shapes 910 and 915 similar to FIG. 8A, but the added artificial background exposure is in the form of shape 913. The initial exposure is set to 1.0 times the standard exposure for shapes 910 and 915 as before, and the artificial background exposure is 0.30 standard exposure for shape 913 in this example. It is important to note that the exposure applied to this new shape 913 is below the printing threshold. FIG. 9B shows the exposure reduction at 0.67 standard exposure after PEC, similar to FIG. 8D, reducing the overall exposure required to print isolated shapes 910 and 915 up front. In some embodiments, shape 913 is also reduced to 0.20 standard exposure. In this simplified example, shape 913 is assumed to extend beyond the backscatter radius away from shapes 910 and 915. In most cases, shape 913 is significantly reduced in size such that it adds artificial background exposure only to the area immediately surrounding shapes 910 and 915. Other such artificial background exposure to other shapes surrounding shapes 910, 913, and 915 contributes actual backscattering to those shapes. In embodiments of the invention, all such energy contributions are calculated using PEC when the actual post-PEC exposure of shapes 910 and 915 is calculated. In some embodiments, the exposure of shape 913 is not altered by every iteration of PEC. In some embodiments, the exposure of shape 913 is altered by every iteration of PEC.

[0053] Artificial background exposure can be added in areas where backscatter is low enough and the exposure margin after PEC is sufficiently better than "good enough" according to some parameters set in practice. Figures 10A-10F show some examples of artificial background exposure, where the added artificial background exposure is in the form of an additional pattern (a pattern with hashed fill lines) that is irradiated in combination with the desired pattern. The artificial background exposure is a sub-threshold exposure, i.e., an exposure below the resist threshold required to print. Figure 10A shows how an artificial background exposure shape 1001 can be cast over an area containing a desired pattern shape 1000 to be printed on a surface, completely covering the pattern in that area. Figure 10B shows an artificial background exposure shape 1002 in that area, excluding the area where the desired pattern shape 1000 is already present, so as not to cover the pattern. Figure 10C shows an artificial background exposure shape 1003 in an area surrounding the desired pattern shape 1000 with some margin, e.g., up to 3 sigma away from the edge of the pattern. For sufficiently large patterns, such as the pattern shape 1010 of Figure 10D, the artificial background exposure shapes 1011 and 1012 are only applied beyond a predetermined distance from the edge of the pattern shape 1010 printed on the surface. By creating margins around features such as edges, line ends, or SRAFs, the artificial background exposure applied in this manner may in some cases be more appropriate than just increasing the background exposure everywhere. Alternatively, the artificial background exposure can be added everywhere in the area, with the artificial background exposure shapes 1004 and 1005, respectively, partially covering the desired pattern shape 1000, as shown, for example, in Figures 10E and 10F, respectively.

[0054] In some embodiments, a predetermined artificial background exposure is defined, such as within the range of 20% to 30%, and applied to every location. A minimum target backscatter amount, or a predetermined threshold, for example 30%, is defined. In some embodiments, if there is not enough backscatter naturally for the PEC to reduce the exposure, the ambient exposure is increased by adding artificial background exposure to reach a minimum backscatter amount. In some embodiments, the additional backscatter contributed by the artificial background exposure may or may not be taken into account in the artificial background exposure amount added, by pre-calculating the additional backscatter contributed by the artificial background exposure. The embodiments described herein are examples, and other variations in adding artificial background exposure are possible. In some embodiments, the minimum backscatter amount has a built-in margin that accounts for the additional backscatter resulting from the artificial background exposure. Since the backscatter amount is mathematically linear in behavior, the amount of backscatter added by the artificial background exposure can be calculated independently of the exposure amount of the shot in any given area. In some embodiments, the total mask area is subdivided into partitions of a size or sizes, and the artificial background exposure is determined for each partition, with each partition having one artificial background exposure within the partition, but different partitions may have different artificial background exposures. Typically, the PEC is calculated on a coarse grid, such as a 50 nm grid or a 300 nm grid. For each grid or partition, a PEC adjustment is calculated. In some implementations, the PEC adjustment for any position within a partition is interpolated based on the PEC adjustments calculated for adjacent partitions. In some embodiments, the calculation of the artificial background exposure is done on the same grid as the grid used for the PEC. In such implementations, the amount of backscatter received during each PEC grid is compared to a specified minimum target backscatter to determine the amount of artificial background exposure to impart to the area covered by the PEC grid.In some embodiments, the artificial background exposure for any location in an individual partition is interpolated across the partition based on the artificial background exposure for neighboring partitions, etc. The subsequent PEC step reduces the exposure for that shot / pixel, reducing the writing time for that shot / pixel. Doing this for all shots / pixels on the mask or mask section reduces the exposure and therefore the writing time for that mask section. The subsequent PEC step accounts for the addition of the artificial background exposure and any additional backscattering that the artificial background exposure contributes. PEC adjusts the shot exposure for all shots on the mask to ensure that all CDs are on target in exactly the same way that PEC has always worked for natural backscattering. The adjusted shot exposure margin is less adequate than before the addition of the artificial background exposure. However, the parameters adjusted by the user to minimize backscattering allow the exposure margin to remain within the acceptable exposure margin determined for the particular mask process.

[0055] Casting any kind of exposure dose (but small doses) with a conventional VSB system requires separate shots and blanking times to separate the shots. Since the blanking times are usually about the same time as the exposure time at standard exposure, a reasonable first order approximation of the shot time may be taken as taking 2 time units at standard exposure. With this approximation, a shot at 10% exposure takes 1.1 time units. Since writing time is paramount to optimizing both mask cost and yield, and since complex state-of-the-art masks have already proven to take too long to write, it is not commercially feasible to add artificial background exposures at every location where there would otherwise be no shots. Nevertheless, VSB mask writers typically have a heavy mask stage that is variable speed, and the mask stage can only change speed gradually, so reducing the peak exposure density is not feasible for VSB systems. This can help reduce the drawing time.

[0056] Specifically for VSB devices, some embodiments of the present disclosure include intentionally leaking some eBeam energy during blanking times to generate artificial background exposure. The exact location where the leaking eBeam is cast does not need to be precisely controlled. Since backscattering is a large scale effect within a 10 μm radius, nm level control of the location is not important. The amount of leakage and duration of leakage during blanking times can be precisely calculated. Such calculations allow the path the eBeam is traveling to be calculated and to avoid casting leakage near any pattern edges (i.e., within the combined forward blur of any pattern edges), which requires precise control.

[0057] Surface writing for leading edge technology nodes typically involves multiple passes of charged particle beam writing to write and overwrite a given shape on a reticle by a process called multi-pass exposure. Typically, two to four passes are used to write the reticle, averaging out accuracy errors in the charged particle beam writer, resulting in a more accurate photomask. Also, typically, the list of shots, including the exposure dose, is the same for all passes. In one variation of multi-pass exposure, the list of shots may differ between exposure passes, but the combination of shots in any one exposure pass covers the same area. Multi-pass writing can reduce overheating of the resist coating the surface. Also, multi-pass writing averages out random errors in the charged particle beam writer. Multi-pass writing, using different shot lists for different exposure passes, can also reduce the effect of certain systematic errors during the writing process. In some embodiments, in VSB and multi-pass writing, only one or some of the writing passes cast an artificial background exposure dose. In other words, in some embodiments, the original exposure information set contains information for multiple exposure passes, and the artificial background exposure is added only during an exposure pass. This is sufficient since accuracy is not important for the artificial background exposure, and by not affecting all passes, it saves writing time. The other passes have a reduced peak exposure of a pixel, or a reduced peak exposure density of an area, without the additional writing time required for the artificial background exposure, as in the embodiment shown in FIG. 10A.

[0058] In some embodiments, relatively isolated patterns are intentionally surrounded by artificial background exposure that results in increased backscattering. Correction for backscattering, PEC, is achieved by reducing the exposure dose of a pixel or shot, with less reduction at low exposure densities and more reduction at high exposure densities. However, reducing the exposure dose results in inadequate exposure margins.

[0059] The acceptable exposure dose margin can be determined by the amount of size variation caused by a reduction in exposure dose. In one embodiment, an isolated feature exposed at a standard exposure dose can be irradiated using a larger pattern and a reduced exposure dose to an acceptable level without adding artificial background exposure dose.

[0060] For example, when charged particle beam lithography is used to expose a repeated pattern onto a surface, the size of each pattern instance measured on the final manufactured surface will vary slightly due to manufacturing variations. The amount of size variation is an important manufacturing optimization criterion. Too much variation, especially for the smallest size features, can cause the feature to not print at all and the resulting circuit to not work properly. Manufacturing variations such as line edge roughness and corner rounding are also present in the actual pattern on the surface. Furthermore, greater size variation results in greater variation in circuit performance, which translates into greater design margins being required, resulting in faster and lower power integrated circuits. It becomes increasingly difficult to design for a mask that is too large to be designed with. This variation is referred to as critical dimension (CD) variation. Low CD variation, especially uniformly low CD variation across all features across the mask, is desirable, indicating that manufacturing variations result in relatively small size variations on the final manufactured surface. On a smaller scale, the effect of high CD variation may be observed as line edge roughness (LER). Each part of the line edge is manufactured slightly differently, causing LER, which causes some waviness in lines that are intended to have straight edges. CD variation is inversely proportional to the slope of the exposure dose curve at resist threshold, referred to as edge slope. Thus, edge slope and its inverse proportionality, which is the exposure dose margin, are important optimization factors for particle beam writing of surfaces. In this disclosure, edge slope and exposure dose margin are terms used interchangeably.

[0061] In some embodiments, the method includes suggesting an appropriate amount of artificial background exposure through the use of edge slope that reaches a target level (i.e., "good enough" edge slope). Since resilience to manufacturing variations is a statistical concept, the meaning of "good enough" edge slope is not an exact expression of a strict inequality. Some embodiments of the method of the present invention determine an artificial background exposure amount where the edge slope is above the "good enough" level. For example, a particular mask manufacturing process can be adjusted to produce reliable manufacturing results for a 100 nm x 2 μm line written in a 75% exposure density area, with a relatively shallow edge slope to generate more backscatter and minimize the exposure of the shot / pixel after PEC. In some embodiments, the edge slope of that line after PEC is indicated as "good enough". There are other geometries, such as a 40 nm x 200 nm space written in a 75% density area, where the edge slope is inadequate, and the mask manufacturing process can determine that the edge slope is not "good enough" but is still manufacturable with an overall tradeoff of economy, time, and manufacturing reliability. The function of "good enough" edge slope for embodiments of the present invention is that being even better than "good enough" indicates that it takes too long to cast that exposure dose compared to the incremental benefit of resilience to manufacturing variations. Since the conventional teaching is to maximize edge slope, it is counterintuitive to reduce the edge slope below an optimal level, i.e., below a target level.

[0062] In some embodiments, the artificial background exposure is calculated automatically. In one embodiment, the artificial background exposure can be calculated to achieve an acceptable level of contrast (drop in contrast). In another embodiment, the artificial background exposure can be calculated to achieve an exposure margin with an edge slope above a target minimum. The minimum acceptable edge slope can be determined by calculating the edge slope at a given edge location of a given pattern in a given backscatter region.

[0063] In some embodiments, an artificial background exposure is added before PEC to reduce the exposure during PEC, allowing the edge slope with the reduced exposure to fall below the target level, thereby reducing the writing time. In some embodiments, the target or "good enough" level can be calculated by simulating the edge slope of a feature that is known to be performing well enough in the manufacturing process to be used for manufacturing purposes. For example, for state-of-the-art photomask manufacturing, a repeating pattern of 100 nm wide lines separated by 100 nm wide spaces is made to be stable when exposed with a standard exposure before PEC. In the context of a 100 nm line and space pattern, a simulation of the edge slope for a 100 nm wide line can be considered a "good enough" exposure margin. To calculate the artificial background exposure, the 100 nm line and space pattern has a 50% exposure density in the surrounding area. The natural backscatter due to the 50% exposure density is calculated as the minimum backscatter amount. The sum of the natural backscatter and the artificial background exposure already present in the area is calculated to be the total backscatter amount for the entire mask. Evenly distributed to contribute to the overall background exposure up to the minimum backscatter that is uniform in size. Although a 50% exposure density may not produce the most adequate acceptable exposure margin after PEC, it can be described as a good practical target to achieve for the purposes of calculating the minimum acceptable exposure from which the artificial background exposure is calculated.

[0064] FIG. 11 is a conceptual flow diagram 1100 for preparing a surface, such as a reticle, or other surface, using charged particle beam lithography, as known in the art. In a first step 1102, a physical design, such as a physical design of an integrated circuit, is created. This can include determining logic gates, transistors, metal layers, and other articles that are required to be found in a physical design, such as a physical design of an integrated circuit. Next, in step 1104, optical proximity corrections (OPC) are determined for the physical design of step 1102, or for the portion of the physical design that is desired on a wafer to create a mask design 1106. OPC calculates the mask shape required to optimally produce the desired wafer shape across manufacturing variations. OPC creates the mask design 1106 by modifying the physical design to compensate for distortions caused by effects such as optical diffraction and optical interactions of features with nearby features. In step 1108, mask data preparation (MDP) operations can include fracturing operations, shot placement operations, dose allocation operations, or shot sequence optimization. In some embodiments for multi-beam mask writing, MDP step 1108 can include generating a plurality of multi-beam shots, each multi-beam shot including one or more beamlets, and a dose region is determined and a dose is assigned to the beamlets within each dose region. In some embodiments, the dose of the beamlets can be varied within the dose region.

[0065] In step 1118, a proximity effect correction (PEC) refinement is performed to adjust the exposure dose to account for backscattering, fogging, and loading effects, and in step 1120, exposure information is created for the adjusted exposure dose. The adjusted exposure dose in the exposure information of step 1120 is used to generate a surface in a mask writing step 1122 using a charged particle beam writer, such as an electron beam writer system. Depending on the type of charged particle beam writer used, the PEC refinement 1118 can be performed by the charged particle beam writer. The mask writing step 1122 can include a single exposure pass, or multiple exposure passes. The electron beam writer system projects an electron beam through a stencil or aperture plate onto the surface to form a mask image 1124 that includes the pattern on the surface. The completed surface, such as a reticle, can then be used in an optical lithography system, shown in step 1126.

[0066] FIG. 13 shows a flow 1300 illustrating a method of exposing a pattern in an area on a surface where an artificial background exposure dose is calculated and applied, according to some embodiments. Step 1305 inputs an original exposure information set for the pattern. In some embodiments, the original exposure information set is a collection of pixels and their exposure doses. In another embodiment, the original exposure information set is a shape set that may include curves or completely non-polygonal descriptions (such as circles or splines). In yet another embodiment, the original exposure information set is a shot set (e.g., VSB) to which a pre-PEC exposure dose is assigned or implied, or a shape set (e.g., multi-beam writing) to which a pre-PEC exposure dose is assigned or implied. In general, the input for step 1305 is something else that describes the shape or exposure dose.

[0067] In step 1310, backscatter is calculated based on the original set of exposure information. In one embodiment, the backscatter calculation in step 1310 is an initial determination of PEC that creates an initial map of backscatter. PEC is an iterative process that adjusts the exposure dose to print the desired size in the presence of backscatter. The initial backscatter from the initial PEC determination The map is a good relative indicator of higher and lower dose density regions.In some embodiments, step 1312 identifies regions where the calculated backscatter is below a predefined threshold.

[0068] The calculations performed in step 1314 determine an artificial background exposure, where the artificial background exposure includes additional exposure information. In some embodiments, step 1314 is performed in the region where the calculated backscatter is below a predetermined threshold determined in step 1312. In some embodiments, step 1314 adds an artificial background exposure only if the calculated backscatter is below a predetermined threshold. That is, the artificial background exposure is the difference between the predetermined threshold and the calculated backscatter. In some embodiments, in step 1314, an exposure margin or edge slope is calculated for the desired pattern in the region to maintain a "good enough" edge slope or contrast for the desired pattern in the region, and an artificial background exposure is determined. In some embodiments, step 1314 adds an artificial background exposure to maintain a "good enough" edge slope for the pattern in the region that reaches a certain criterion, such as a minimum object size for a "good enough" edge slope. For example, features that are smaller in size for SRAF may be excluded from the "good enough" edge slope. As another example, smaller sized features for SRAF may have different "good enough" edge slopes to achieve when determining the artificial background exposure to be added. In some embodiments, these calculations may result in the creation of an artificial background exposure that includes an additional pattern with a sub-threshold exposure so that the additional pattern is not printed. In some embodiments, the calculation and creation of the additional pattern occurs after the MDP step 1108 of the flow 1100 of FIG. 11, and the additional pattern overlaps the desired pattern in a geometric shape in the mask design 1106 of the flow 1100 of FIG. 11. A proximity effect correction (PEC) refinement step 1118 of the flow 1100 of FIG. 11 may retrieve the modified exposure information set of step 1316 that includes the original exposure information set combined with the additional exposure information (i.e., the additional pattern), and step 1120 generates exposure information that includes the artificial background exposure and may also include backscattering resulting from the artificial background exposure.In some embodiments, the modified exposure information set reduces the overall writing time compared to the original exposure information set. In some embodiments, the modified exposure information set is refined by the PEC resulting in an adjusted exposure dose for the modified exposure information set. The exposure information is then used to expose the surface at 1124 of FIG. 11. In some embodiments, the exposure dose of a pixel or shot is modified by adding an artificial background exposure. In some embodiments, a pixel or shot casts an artificial background exposure in an otherwise empty area (a pixel that starts with zero exposure) and the exposure dose is relatively similar for example 3×3 pixels or more than the number of neighboring pixels. In some embodiments, some or all of steps 1310, 1312, 1314 and 1316 of FIG. 13 followed by steps 1118 and 1120 of FIG. 11 are performed simultaneously with mask writing step 1122 of flow 1100 of FIG. 11, with in-line processing performing data manipulation while the mask is being exposed. That is, in some embodiments, the mask exposure is performed in-line by one or more steps selected from the group consisting of calculating backscattering, determining areas where the calculated backscattering is below a predetermined threshold, determining an artificial background exposure amount, and creating a modified exposure information set.

[0069] The preparation of mask data, calculation of artificial background exposure, and proximity effect correction described in this disclosure can be performed using a general purpose computer using a central processing unit (CPU) with appropriate computer software as a computing device. Also, due to the large amount of calculations required, multiple computer or CPU processor cores can be used in parallel. In one embodiment, the calculations are performed using a floating-point filter to support parallel processing. For one or more computationally intensive steps within a row, the row may be subdivided into multiple two-dimensional geometric regions. In another embodiment, a dedicated hardware device, used either alone or in multiples, may be used to perform the computation of one or more steps faster than using a general purpose computer or processor core. The specialized computing hardware device or processor may include, for example, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), or a digital signal processor (DSP) chip. In one embodiment, the dedicated hardware device may be a graphics processing unit (GPU). In another embodiment, the optimization and computation process described in this disclosure may include an iterative process of modifying and recomputing possible solutions to minimize the total charged particle beam writing time, or some other parameter. In yet another embodiment, the process may be a deterministic computation without iterations.

[0070] FIG. 12 illustrates an example of a computing hardware device 1200 that can be used to perform the computations described in this disclosure. The computing hardware device 1200 includes a central processing unit (CPU) 1202 to which a main memory 1204 is connected. The CPU can include, for example, eight processing cores, thereby improving the performance of any portion of computer software that is multithreaded. The size of the main memory 1204 can be, for example, 64 GBytes. The CPU 1202 is connected to a Peripheral Component Interconnect Express (PCIe) bus 1220. A graphics processing unit (GPU) 1214 is also connected to the PCIe bus. In the computing hardware device 1200, the GPU 1214 may or may not be connected to a graphics output device, such as a video monitor. When the GPU 1214 is not connected to a graphics output device, it can be used purely as a high-speed parallel computation engine. Computing software can obtain significantly higher performance by using the GPU for part of the computation compared to using the CPU 1202 for all computations. CPU 1202 communicates with GPU 1214 via PCIe bus 1220. In other embodiments (not shown), GPU 1214 may be integrated with CPU 1202 rather than connected to PCIe bus 1220. Disk controller 1208 may also be connected to the PCIe bus, for example with two disks 1210 connected to disk controller 1208. Finally, local area network (LAN) controller 1212 may also be connected to the PCIe bus and provide a Gigabit Ethernet (GbE) connection to other computers. In some embodiments, computer software and / or design data are stored on disk 1210.In other embodiments, either the computer programs or the design data, or both the computer programs and the design data, may be accessible from other computers or file serving hardware via Gigabit Ethernet (GbE Ethernet).

[0071] Although the present specification has been described in detail with respect to certain embodiments, it is to be understood that those skilled in the art, upon gaining an understanding of the foregoing, may readily conceive modifications, variations, and equivalents to these embodiments. These and other modifications and variations to the mask data preparation method and proximity effect correction method of the present invention may be implemented by those skilled in the art without departing from the scope of the inventive subject matter as more particularly set forth in the appended claims. Moreover, those skilled in the art will appreciate that the foregoing description is for illustrative purposes only and is not intended to be limiting. Steps may be added to, subtracted from, or altered from the steps herein without departing from the scope of the present invention. In general, any flowcharts presented are intended to illustrate one possible sequence of basic operations for implementing a function. Only one embodiment is intended, and many variations are possible. Thus, it is intended that the present subject matter cover such modifications and variations as come within the scope of the appended claims and their equivalents.

Claims

1. 1. A method of exposing a pattern in an area on a surface using charged particle beam lithography, comprising: inputting a set of original exposure information for said region; calculating backscatter for the region based on the original set of exposure information; determining an artificial background exposure amount for the region, the artificial background exposure amount including additional exposure information that is combined with the original set of exposure information to create a modified set of exposure information; The method comprising:

2. The method of claim 1 , wherein the modified set of exposure information reduces an overall writing time compared to the original set of exposure information.

3. The method of claim 1 , wherein the modified set of exposure information is refined by a PEC to result in an adjusted exposure dose for the modified set of exposure information.

4. The method of claim 1 , wherein the artificial background exposure comprises additional patterns with below-threshold exposure.

5. The method of claim 1 , wherein the artificial background exposure dose is only applied above a predetermined distance from an edge of the pattern on the surface.

6. The method of claim 1 , wherein the original set of exposure information includes information for multiple exposure passes, and wherein only the artificial background exposure amount is added to an exposure pass.

7. The method of claim 1 , wherein in the regions where the calculated backscatter is below a predetermined threshold, the artificial background exposure is the difference between the predetermined threshold and the calculated backscatter.

8. The method of claim 1 , wherein the region is subdivided into a plurality of partitions and the amount of artificial background exposure is determined for each partition.

9. The method of claim 8 , wherein the artificial background exposure for any location within an individual partition is interpolated across said partitions.

10. The method of claim 9 , wherein the interpolation is based on the artificial background exposure for adjacent partitions.

11. The method of claim 1 , further comprising calculating an edge slope for the pattern to be exposed.

12. The artificial background exposure dose that achieves an edge slope of the exposed pattern above a target minimum is determined; The method of claim 11 , wherein the target minimum is determined by calculating the edge slope at a given edge location of a given pattern within a given backscatter region.

13. The mask exposure may include one or more steps selected from the group consisting of: calculating the backscattering; determining the areas where the calculated backscattering is below a predetermined threshold; determining the artificial background exposure amount; and creating the modified exposure information set. The method of claim 1 , wherein the method is performed in-line by steps.

14. The method of claim 1 , further comprising exposing the surface with the modified set of exposure information.

15. 1. A system for exposing a pattern in an area on a surface using charged particle beam lithography, comprising: a device configured to input a set of original exposure information for said region; a device configured to calculate backscattering for the region of the pattern based on the exposure information; a device configured to determine an artificial background exposure amount for the region, the artificial background exposure amount including additional exposure information that is combined with the original set of exposure information to create a modified set of exposure information; The system comprising:

16. The system of claim 15 , wherein the modified set of exposure information reduces an overall writing time compared to the original set of exposure information.

17. 16. The system of claim 15, wherein the modified set of exposure information is refined by a PEC to result in an adjusted exposure dose for the modified set of exposure information.

18. The system of claim 15 , wherein the artificial background exposure comprises additional patterns with below-threshold exposure.

Images