Method for maskless exposure with particle beams

By enhancing particle beam lithography with a target movement method that exceeds pixel exposure distances and using redundant beam groups, the method achieves high throughput and efficient writing speeds, addressing inefficiencies in existing technologies.

DE102008015305B4Active Publication Date: 2024-11-28IMS NANOFABTION
View PDF 9 Cites 0 Cited by

Patent Information

Application Number
DE102008015305
Authority / Receiving Office
DE · DE
Patent Type
Patents
Current Assignee / Owner
Filing Date
2008-03-20
Publication Date
2024-11-28
Estimated Expiration
2028-03-20

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

Method for irradiating a target with a beam of high-energy, electrically charged particles, comprising the following steps: - Providing a pattern definition device (102) with a plurality of apertures, - Illuminating the pattern definition means by means of the beam which traverses the pattern definition means through the apertures and thereby forms a shaped beam (pb) consisting of a corresponding plurality of partial beams, - Shapes of the shaped beam into an image on a location of the target (14, 40), wherein the image includes images of at least some of the plurality of apertures, wherein the images of the apertures have a first width (w), - Moving the target relative to the pattern-defining means along a path within an image plane in which the images of the apertures are formed, where, during pixel exposure times (Tu), the position of the image is moved along with the target, characterized by the fact that the pixel exposure time has a duration within which the relative movement of the target covers a distance, the "forward distance", which is greater than the first width (w) measured on the target, but less than or equal to the distance between adjacent partial rays along the direction of the relative movement, measured on the target, and after the pixel exposure time the position of the image is changed with respect to the target, wherein this change in position substantially compensates the movement of the position of the image during the pixel exposure time (Tu) with respect to the pattern definition means (102), wherein the relative movement of the target during successive pixel exposure times (Tu) is over a distance which exceeds the distance between adjacent partial beams along the direction of the relative movement.
Need to check novelty before this filing date? Find Prior Art

Description

Field of the invention and description of the prior art

[0001] The invention relates to an improved method for irradiating a target with a beam of energetic, electrically charged particles, wherein a pattern defining means having a plurality of apertures transparent to said particles is used, and by illuminating the pattern defining means with the beam passing through the pattern defining means through the apertures, a shaped beam is formed which consists of a corresponding plurality of sub-beams, and the shaped beam is formed into an image on the location of the target, the image comprising the images of at least a portion of the plurality of apertures;the aperture images have a first width, the target moving relative to the pattern defining means along a path within an image plane in which the images of the apertures are formed (or in a direction substantially perpendicular to the direction of propagation of the shaped beam), the location of the image moving with the target during pixel exposure times;

[0002] A method of this type and the associated particle beam device are disclosed in US Pat. No. 6,768,125 B2, which is hereby incorporated into the present disclosure as relevant prior art. This patent describes a method and apparatus for charged particle lithography and processing, designated PML2 (for 'Projection Mask-Less Lithography'), that implements a multi-beam direct writing concept and uses a programmable aperture plate system (APS) as a pattern definition (MD) device to shape a particle beam extracted from a single source of electrically charged particles.

[0003] Particle lithography and processing are used in semiconductor production and microstructure applications. In particular, direct structuring by ion beam irradiation is a promising future concept for the industrial fabrication of nanoscale devices with high resolutions, especially those using 32 nm and 22 nm technology and below. The MD device is preferably a programmable multi-aperture device. The implementation of an optical system for multi-beam projection based on a programmable multi-aperture plate allows for a significant improvement in achievable productivity compared to focused single-beam systems. (In this disclosure, "optical" always means "particle-optical").The reasons for the improved productivity are, first, the parallelization of the process using a large number of beams and, second, the increased current that can be imaged onto the substrate at the same resolution. Both are made possible by a significantly reduced Coulomb interaction in the beam. Furthermore, the moderate current density associated with the optical projection system results in an increased process rate when precursor gases are used for beam-induced chemical processes. Compared to a focused beam system, heat exposure due to extreme beam intensities can also be reduced or avoided.

[0004] In the PML2 setup according to US Pat. No. 6,768,125 B2 and related prior art, the position of the patterned beam on the target is fixed at the corresponding pixel position only for the time required for a pixel of the target to pass under the optical system. The patterned beam then jumps to the position of the next pixel adjacent to the previous one. In this way, each partial beam covers adjacent pixels on the target. This approach requires a fast loading rate of the pattern information, which can lead to an upper limit on the processing speed of the method.

[0005] The diagram of the Fig. Figure 19 illustrates the previously known writing procedure, using the example of a row of pixels to be exposed on a target, for example, a wafer covered with resist. The pixels of this single row are irradiated (exposed) by a number of sub-beams (each labeled with capital letters A, B, C, ...), and each sub-beam irradiates a sequence of adjacent pixels during successive time steps. The amount of exposure dose delivered to each pixel is symbolically shown by hatched and cross-hatched fields. The cross-hatched fields indicate pixels that are fully irradiated (according to a programmed pattern; consecutive exposures are symbolically shown as stacked on top of each other, while in fact they simply add up), whereas simply hatched fields indicate pixels during the course of an exposure.The dose for each pixel accumulates sequentially as the resist moves beneath the MD device that shapes the sub-beams (see the discussion of maskless multi-beam writing below with reference to ). Fig. 1). Each sub-beam has the same time interval in which it delivers a partial exposure dose to the substrate, and therefore delivers the same dose amount. Each sub-beam in a row is used to expose each pixel in that row. To switch the sub-beams on and off through the apertures in the MD device according to this exposure method, the MD apertures require pattern information at an appropriate data rate, and the image information must be distributed to each individual aperture on the APS.

[0006] The first two individual images of the Fig. Figure 19 illustrates the situation at the beginning and end of a pixel exposure time interval (per sub-beam), denoted by t = 0:0 and 0:1, respectively. It is clear from the illustration that the sub-beams are moved along with the target, so that their positions coincide with the (moving) positions of the pixels during the pixel exposure. When the pixel exposure is finished after t = 0:1, all sub-beams are repositioned to then align with the positions of the next pixels; the third sub-image, t = 1:0, shows the beginning of the exposure of the next pixel. The repositioning of the sub-beams occurs in a very short time, much shorter than the total time interval Tw between successive pixel exposures (Tw is equal to the time between the sub-images 0:0 and 1:0).) The distance between the partial beams is 5w, namely five times the width of the partial beams; but other integer values ​​are equally suitable for the previously known method, as discussed in US 6 768 125 B2.

[0007] The diagram of the Fig. 20 shows the timing of the previously known writing process of Fig. 19. At the top, the beam deflection x is shown as a function of time t; the beam deflection describes a sawtooth-like function of time. Also shown are the temporary oscillations until the deflection stabilizes after each repositioning. Below the beam deflection x, the states p are exemplary. A , ..., p E of five apertures of the APS (as examples from a large number of apertures, typically many thousands). Each partial beam A to E is projected onto the target only when the respective state p A to p Eis on low ('low'; no energy, partial beam switched on), as shown below with reference to Fig. 1. As can be seen, the states of the apertures do not change during each actual pixel exposure. Consequently, only 1-bit of information can be processed per aperture during a period of the sawtooth function.

[0008] The current approach makes it difficult to find a suitable design for on-chip data storage that is efficient and capable of delivering the required data rates. Currently, the use of shift registers is the most promising solution to this problem (see US 2005 / 0242303 A1).

[0009] From DE 10 2004 055 149 A1 a method for irradiating a target with a beam of high-energy, electrically charged particles with the features of the preamble of claim 1 is known.

[0010] Both US 6,835,937 B1 and US 6,137,113 A disclose a method in which optical beam tracking is performed for at least one entire image of a pattern definition means or an aperture plate on a substrate. In this case, the individual partial beams always write exclusively in a single corresponding cell. Once the entire subfield has been exposed, the entire beam is shifted by the width or height of the subfield, relative to the previous position of the beam. Therefore, writing methods are described in which each individual beam writes onto the target within an area assigned to it. Summary of the invention

[0011] It is the object of the invention to overcome the above-mentioned weaknesses and to provide a method that offers high throughput combined with an easily realizable design of the pattern definition (MD device).

[0012] This object is achieved by means of a target irradiation method having the features of claim 1 and a method for positioning a target using a beam of energetic, electrically charged particles having the features of claim 23, in which during pixel exposure times the location of the image is moved along with the target, the pixel exposure times having a duration within which the relative movement of the target covers a distance - the "forward distance" - which is greater than the first width, measured on the target, preferably at least a multiple of the first width (i.e. at least 2.0 times and more preferably 3 times the first width or more); and after the pixel exposure times the location of the image with respect to the target is changed (repositioned on the target) to substantially compensate for the movement of the location of the image during the pixel exposure times with respect to the pattern defining means.According to the invention, the relative movement of the target during successive pixel exposure times occurs over a distance that exceeds the distance between adjacent partial beams along the direction of the relative movement.

[0013] The invention provides a writing procedure that provides an optimal writing speed across the target and is suitable for an electronics design that can be realized using standard CMOS techniques and allows a simplified implementation of the blanking circuits in the pattern definition means.

[0014] The invention provides an electron beam writing device that enables writing at a speed limited only by the electron beam density and the transmittance of the APS. Power consumption and data rate to the APS can be managed within the UMC 0.25 µm process. Currently used architectures would fail by several orders of magnitude in both respects.

[0015] To ensure that the partial beams illuminate the entire surface of the target, the number of partial beams illuminating different pixels is chosen such that the first width multiplied by the number of these partial beams is equal to (or even greater than) the forward distance. In the usual case of groups of redundant partial beams (namely, the partial beams lying in the same line along the direction of relative motion of the target form groups, "redundancy groups," within which the partial beams are spaced apart by a distance that, measured on the target, is an integer multiple of the forward distance), the number of these redundancy groups is chosen such that the first width multiplied by the number of these groups is greater than or equal to the forward distance.

[0016] In a preferred embodiment of the invention, the relative movement of the target is a continuous linear movement over time intervals containing multiple pixel exposure times. Typically, it is the target that is moved while the MD device and the optical system are stationary. The target movement can preferably be a scanning movement along scanning sweeps in a main direction.

[0017] Furthermore, it is advantageous if, outside the pixel exposure times, the designed beam is masked out in such a way that it does not reach the target, but is absorbed by an absorber before reaching the target.

[0018] In a suitable choice of the relative motion of the target, the path length during a time period between the start times of successive pixel exposure times constitutes a second width that is an integer multiple of the first width. In this case, the distances between two partial beams lying in the same row along the direction of the relative motion of the target can be integer multiples of the first width, and at least some of these integer multiples can be chosen such that they are coprime to the integer multiple of the second width. A simple way to achieve this is for said integer multiples to differ by one.

[0019] In a preferred embodiment, the pattern-defining means comprises a plurality of electrostatic partial beam deflectors, each associated with an aperture of the pattern-defining means and configured to deflect the partial beam passing through the respective associated aperture by an amount dependent on an individual control signal, including an amount sufficient to deflect the partial beam from its desired path (effectively shutting off this partial beam so that it does not reach the target). In this case, the apertures of the pattern-defining means are controlled in a time-dependent manner according to a pattern to be formed on the target by means of a plurality of said control signals.

[0020] In this preferred embodiment, the apertures are preferably logically grouped into disjoint groups, and the control signals of the apertures of each group are supplied via a common control line in a time-staggered manner. Conveniently, the apertures of each group can be arranged in the pattern definition means along an orientation direction substantially perpendicular to the direction of relative movement of the target. The apertures of each group can extend over half the width of the area occupied by the apertures of the pattern definition means, measured along an orientation direction that is substantially (but not necessarily) perpendicular to the direction of relative movement of the target, with the control signals of this area being supplied by apertures from two opposite sides.

[0021] In this preferred embodiment, the pattern definition means may also comprise data storage means to which data for activating the control signals is fed and stored; allowing this data to be read out individually for corresponding activation of the control signals. The data storage means may be equipped with duplicate ports, wherein during each pixel exposure time, data for controlling the apertures is read out through a first port of the data storage means to the control signals, and data relating to the subsequent pixel exposure time is stored via a second port of the data storage means.

[0022] Another useful development of this preferred embodiment is to group the apertures of the pattern definition means into groups of mutually redundant apertures, wherein, for each group, the control signals associated with the apertures of the group are generated from a common signal that is copied sequentially to the control signals. In this case, the common signal can be copied to the control signals with respective time delays in accordance with the spatial offset of the respective apertures along the direction of relative movement of the target. In order to implement gray values ​​(cf.

[0023] US 2005 / 0242303 A1), the common signal of a group may contain information of a gray value to be implemented by means of the apertures belonging to the group, and then the gray value is expanded into a sequence of inactive and active control signals, in accordance with the gray value that defines the relative number of inactive control signals with respect to the total number of control signals in the group. The gray value expansion may occur before or after reading the data from the storage means, and the expansion method may advantageously include at least one of the following methods: expansion into an evenly spaced sequence, expansion into a pseudorandom number sequence, expansion using linear feedback shift registers, expansion using different shifts for the individual control signals; or a combination of these.Of course, the invention can also be combined with other known writing strategies, such as an additional global gating signal which asynchronously masks the outputs of the gating flip-flops (cf. WO 2007 / 112465).

[0024] Furthermore, the disk implementing the array of apertures may be duplicated and operated in parallel, so that the apertures are arranged in (at least) two parallel devices offset from each other by a distance along the beam.

[0025] Another aspect of the invention relates to the positioning of a target using a beam of energetic, electrically charged particles, in particular in connection with a target irradiation method as dealt with here, wherein during the positioning method the target is moved relative to the pattern defining means along a path within an image plane in which the images of the apertures are formed, the relative movement of the target covering a "forward distance" which is greater than the first width measured on the target, while at least a part of the beam impinges on markings provided on the target which then generate secondary radiation, and the secondary radiation is measured in a detection system with a sufficiently high sampling rate in order to adjust the target position in such a way that the secondary radiation thus measured reaches an extreme value (maximum or minimum).According to the invention, the relative movement of the target during successive pixel exposure times occurs over a distance that exceeds the distance between adjacent partial beams along the direction of the relative movement.

[0026] In this method, preferably the substrate guide moves the target at a uniform speed equal to that at which the target moves during a subsequent exposure process according to the exposure method described above.

[0027] A suitable implementation of the markings includes structures that are part of a multilayer structure, and these structures have been formed on the target in at least one preceding process step. In particular, the markings may include components for generating secondary radiation containing Auger electron radiation, and the detection system may employ an energy-selective detector configured to detect this Auger radiation. In another implementation, the markings may include components with a characteristic backscattering yield.

[0028] A particular advantage is that the target positioning process and a subsequent particle beam exposure / irradiation process performed on the target can use the same pattern definition means, with the target moving relative to the pattern definition means in a corresponding manner. However, the forward distance of the target positioning process can be greater than that during the particle beam exposure / irradiation process. Brief description of the drawings

[0029] In the following, the invention is described in more detail with reference to the drawings, which show: Fig. 1 shows a schematic overview of a particle beam exposure device suitable for the invention in a longitudinal section; Fig. 2 shows the basic procedure of writing according to the invention, in which the partial beams “trot” over the target; Fig. Figure 3 shows a time diagram of the beam deflection in the “trot” approach; Fig. 4 illustrates redundant writing as an extension of the “trotting” approach; Fig. 5 illustrates the writing process in a combination of the concepts of Fig. 3 and Fig. 4; Fig. Figure 6 is a block diagram of the electronics of a blanking disk in an MD device for implementing the “trotting” approach; Fig. Figure 7 is a block diagram of a RAM memory in the electronics of the Fig. 6; Fig. Figure 8 is a block diagram of the circuit layout of the data supply of an aperture half-slit in the blanking plate in the electronics of the Fig. 6; Fig. 9 shows a timing chart of the control flow and loading sequence of the pattern data; Fig. 10 shows a modification of the structure Fig. 8 for redundant column information; Fig. Figure 11 shows an MD device (longitudinal section detail) with a redundant design with two blanking plate devices operating in parallel; Fig. Figure 12 shows a plan view of the geometric design of an aperture arrangement designed to implement the “trotting” approach; Fig. 13 shows a variant of the design of the Fig. 12 with a redundancy of two; Fig. 14 shows a section of the Fig. 13, with a unit cell highlighted; Fig. 15, Fig. 15a and Fig. 16 illustrate the use of markers on the target for target positioning, namely Fig. 15 a first stage, the initial positioning, and Fig. 16 a second stage, fine positioning; Fig. 15a shows a variant with embedded markers; Fig. Figure 17 shows a state-of-the-art MD device (longitudinal section detail), which, however, can also be used in the device of the Fig. 1 is usable; Fig. 18 shows a plan view with the geometric design of the apertures in a previously known MD device; Fig. 19 and Fig. 20 show the procedure of writing according to the state of the art in contrast to the analogue Fig. 2 or 3. Detailed description of the invention

[0030] The preferred embodiment of the invention discussed below is a further development of the PML2-type particle beam exposure apparatus with a pattern definition device (MD device) as described in the applicant's US Pat. No. 6,768,125 B2 (= GB 2,389,454 A) and with a high-reduction projection system. In the following, the technical background of the apparatus is first presented, insofar as relevant to the invention, followed by detailed descriptions of embodiments of the invention. It should be emphasized that the invention is not limited to the following embodiments or to the particular design of the MD system, which represents merely one possible implementation of the invention; rather, the invention is equally suitable for other types of processing systems that employ a particle beam with projection stages. PML2 system

[0031] A schematic overview of a maskless particle beam processing apparatus PML2 employing the invention is shown in Fig. 1. In the following, only those details are given which are necessary to disclose the invention; for the sake of clarity, the components in Fig. 1 is not shown true to size, especially the lateral width of the particle beam is exaggerated. For further details, the reader is referred to US Pat. No. 6,768,125 B2.

[0032] As already mentioned, the PML2 system uses a particle beam generated by a particle source. An illumination optics system shapes the beam into a wide beam, which illuminates an MD device with a regular array of apertures to define a beam pattern to be projected onto a target surface. Each aperture defines a small beam, and the passage of each beam through an aperture can be controlled to allow particles of the beam to pass through the respective aperture toward the target ("switching on") or deactivate them ("switching off"). The beam passing through the apertures forms a shaped beam that carries the pattern information represented in the spatial arrangement of the apertures. The shaped beam is then projected onto the target (e.g.,A semiconductor substrate is projected onto the target, forming an image of the apertures to modify the target at the illuminated areas. The image generated by the beam is scanned along a straight path across each chip array; additional scanning movements of the beam perpendicular to the line-by-line scanning direction are not required (except as necessary to compensate for lateral errors in the target station's working motion).

[0033] The main components of the device 100 are - in the order of the direction of the lithography beam lb, pb, which in this example is directed vertically downwards in Fig. 1—an illumination system 101, an MD system 102, a projection system 103, and a target station 104 with the target or substrate 14. The particle-optical systems 101, 103 are implemented using electrostatic or electromagnetic lenses. The electro-optical parts 101, 102, 103 of the device 100 are located in a vacuum housing (not shown) maintained under high vacuum to ensure undisturbed propagation of the beam 1b, pb along the optical axis of the device.

[0034] The illumination system 101 comprises, for example, an ion source 11, an extractor assembly 11a that defines the location of the virtual source, a particle filter / general blanker 12, and illumination optics in the form of a condenser lens system 13. The ions used can be, for example, hydrogen ions or heavy ions; in the context of this disclosure, heavy ions refer to ions of elements heavier than C, such as O, N, or the noble gases Ne, Ar, Kr, Xe. In addition to ions, the particles can be electrons (emitted by an electron gun), or, in general, other electrically charged particles can also be used.

[0035] The ion source 11 emits energetic ions mainly of a specific species, such as Ar +Ions with a defined (kinetic) energy of typically a few keV (e.g., 5 keV at the MD system 102) with a comparatively narrow energy range of, for example, ΔE = 1 eV; however, it is generally unavoidable that the source also emits electrical particles of other species. A velocity / energy-dependent filter 12 serves to filter out such unwanted particle species; this device can also be used to "blank" the beam as a whole during the repositioning of the partial beams. With the aid of an electro-optical condenser lens system 13, the ions emitted by the source 11 are formed into a broad-area, essentially telecentric ion beam, which serves as the lithography beam 1b.The telecentricity of the beam lies within a deviation range of ±25 µrad from the optical axis at the MD device location, resulting in a telecentricity range of ±5 mrad from the optical axis at the substrate location, assuming a 200x reduction system and equal particle energies at the MD device and substrate.

[0036] The lithography beam 1b then irradiates an MD device, which, together with the devices used to hold it in place, forms the MD system 102. The MD device is held at a specific position in the path of the lithography beam 1b, which thus irradiates an aperture pattern formed by a plurality of apertures 21. As already mentioned, each of the apertures can be "switched on" or "open," allowing the partial beam passing through the respective aperture to reach the target; in this case, the aperture is said to be transparent to the incident beam. Otherwise, the aperture is "switched off" or "closed," in which case the beam path of the respective partial beam is influenced such that it is absorbed or otherwise removed from the beam path before it can reach the target; thus, the aperture is, as a result, non-transparent or opaque to the beam.The pattern of the switched-on apertures is chosen according to the pattern to be illuminated on the substrate, since these apertures are the only areas of the MD device that are transparent to the beam lb; the beam is thus directed into one of the apertures (ie into . Fig. 1 under the MD system 102) is formed. The architecture and operation of the MD device, particularly with regard to its blanking plate, are described in detail below. In Fig. 1, only five partial beams (out of a large number) are shown in the designed beam pb, of which the second partial beam from the left is switched off because it is absorbed on an absorber plate 17; the other, switched-on partial beams pass through a central opening of the plate 17 and are thus projected onto the target.

[0037] The pattern represented by the shaped beam pb is then projected onto the substrate 14 by an electro-optical projection system 103, where it forms an image of the switched-on mask apertures. The projection system 103 realizes a reduction of, for example, 200x. The substrate 14 is, for example, a silicon wafer covered with a photoresist. The wafer 14 is held and positioned by a wafer stage (not shown) of the target station 104. A secondary radiation detector 15 can be used to detect the proper positioning of the substrate with respect to the beam; this is discussed in more detail below ("Mark Reading").

[0038] The projection system 103 is preferably composed of two consecutive electro-optical projector stages with a crossover c1 and c2, respectively. The electrostatic lenses 30 used to implement the projectors are Fig. 1 is only shown symbolically, as technical implementations of electrostatic imaging systems are well known from the prior art. The first projector stage images the plane of the apertures of the MD device into an intermediate image, which in turn is imaged onto the substrate surface by the second projector stage. Both stages employ a reduction imaging process by crossover c1, c2; thus, the intermediate image is reversed, but the final image generated on the substrate is upright (not inverted). The reduction factor is approximately 14x for both stages, resulting in a total reduction of 200x. A reduction of this magnitude is particularly useful for a lithography device to reduce miniaturization problems in the MD device. The electro-optical lenses are primarily constructed of electrostatic electrodes, but magnetic lenses can also be used.

[0039] As a means for introducing a small lateral displacement of the image, i.e. along a direction perpendicular to the optical axis cx, deflection means 16 are provided in one or both projector stages. Such deflection means can, for example, be designed as a multipole electrode system as described in US Pat. No. 6,768,125 B2. In addition, a magnetic coil can be used to generate a rotation of the pattern in the substrate plane, if required. The lateral deflections are significantly smaller than the lateral width of the designed beam pb itself. Although the lateral deflections required according to the invention are many times the width of an individual beam, this distance is still at least an order of magnitude less than the beam width, since the lateral extent of a partial beam is considerably smaller than that of the beam pb as a whole.

[0040] By controlling the pattern formed in the MD system 102, any desired pattern can be generated and transferred to the substrate. Conveniently, a stripe-scan exposure approach is used, with the substrate being moved beneath the incident beam, eliminating the need for a beam-scanning approach, in which the position of the beam is constantly changed and thus the beam is effectively scanned across the (more or less stationary) target surface, as in a single-focused beam system.

[0041] Fig. Figure 17 shows a longitudinal section detail of a possible implementation of the MD system 102 in a two-plate arrangement. The concept shown is known from the prior art, but is nevertheless suitable for use in the invention. A first plate 201 is an aperture plate with a set of apertures 21; only three of a large number of apertures are shown. The plate 201 defines, through its apertures, a corresponding number of sub-beams b1, b2, b3. The apertures 21 are imaged onto the target as already mentioned. The apertures 21 are systematically arranged to form staggered lines parallel to a direction corresponding to the relative movement of the images of the apertures across the target, as described in US Pat. No. 6,768,125 B2.In each row, the offset between successive apertures is preferably a multiple of the aperture width, while the rows run immediately adjacent to each other so that the aperture images completely cover the target during the scanning movement across the target.

[0042] The second plate 202 of the MD system 102 is called the blanking plate. It has a set of openings whose positions correspond to those of the apertures 21 in the aperture plate 201, but whose width is larger so that the partial beams can pass through them without striking the material of the blanking plate. The openings in the blanking plate 202 are equipped with electrodes 221 to impart a small but sufficient deflection to the respective partial beam. Each partial beam can be deflected individually as described in detail below. The blanking plate also includes the circuitry for electronically controlling and electrically supplying the electrodes. Further basic details of an MD device are covered in US Pat. No. 6,768,125 B2 and related prior art, while a preferred design of the blanking plate electronics according to the invention is discussed below.

[0043] Fig. Figure 18 illustrates the arrangement of the apertures in a top view of the MD device, as is typical in the prior art. The apertures are shown as cross-hatched squares, and their orientation is chosen so that the horizontal line coincides with the direction of (relative) substrate movement during exposure (cf. Fig. 19). Along this direction, the apertures are evenly spaced in rows, and between adjacent rows, the apertures are offset to form a staggered array, where the offset is an integer multiple of the aperture width w. In the example shown, the offset is 3w, and the aperture pattern repeats every three rows. Thus, the pattern repeats every 3×3w in the horizontal direction. Consequently, each aperture has an associated cell space of 9w. 2, or 3w × 3w, which is illustrated as squares surrounding the apertures. The prior art writing technique used with this aperture arrangement was described above with reference to the Fig. 19 and Fig. 20 discussed.

[0044] The Fig. The regular arrangement shown in Figure 18 may be interrupted by areas without apertures, as discussed in US 6 768 125 B2, for example, to provide space for electronics for internal processing of pattern data. Writing procedure

[0045] In contrast to well-known writing concepts (cf. Fig. 19), we propose a write approach that minimizes data transfer to and from the APS. We refer to this approach as the "trot" approach because the movement of the individual beams is reminiscent of the diagonal movement of a horse's legs during a trot.

[0046] Fig. Figure 2 illustrates the basic concept of the inventive "trotting" method with a first, simplified example. A target 40, e.g., a silicon wafer covered with a resist to be exposed, moves at a speed v, while a number of partial beams (only five beams, labeled A to E, are shown for clarity) are incident on the target. Each partial beam emits a pixel exposure period T1, which occurs between the top and bottom images in Fig. 2, the total dose for a pixel is transferred to the resist. In the figure, cross-hatched areas indicate pixels that are fully illuminated (according to the programmed pattern), while single-hatched areas indicate pixels that are currently being exposed and thus add up the required dose.

[0047] The target is moved relative to the MD device (which is fixed above the Fig. 2 is assumed to be constant), but the writing method requires that the location where each partial beam impinges on the target remains consistent with the location of a pixel despite this relative movement during each pixel exposure period, even when the target moves over a distance of several pixels. Consequently, the partial beams are deflected accordingly using the deflection system of the imaging optics (deflectors 16 in Fig. 1) distracted.

[0048] The width of a pixel exposed on the target is denoted by w, the distance between two adjacent partial beams is N'w, where N' is one more than the number N of partial beams writing in parallel on a line. The factor N+1 is a simple choice to ensure that N and N are relatively prime; any other number N' that is relatively prime with N could be used instead. Such a choice of N' and N ensures the exposure of adjacent pixels over a sweep of N' exposure periods. In the case of Fig. 2, N=5 and N'=N+1=6. Starting from an initial position at the beginning of an exposure period, shown in the first (top) sub-image at t=0:0, the times are successively labeled t=0:1, at which the movement of the target has progressed by one pixel width w, t=0:2 at two pixel widths 2w (not shown), etc. At t=0:5 (t=0:N), the movement has progressed to a distance of five (N) pixel widths 5w, which indicates the end of the current exposure. After that, the sub-beams are directed to another set of pixels, as in the bottom sub-image of the Fig. 2 to start the next pixel exposure period at t=1:0. (t=1:0 is a time shortly after t=0:N; in other words, t=0:N+δ=1:0, where δ denotes a small time interval used to switch the sub-beams.)

[0049] The switching on / off of each blanker in the MD device is controlled during each pixel exposure period according to the programmed pattern. For example, if a pixel irradiated by its corresponding sub-beam is to receive a dose of 25% of the maximum value, the blanker for that sub-beam is opened for 25% of the pixel exposure period.

[0050] It is worth mentioning that the inclination of the partial beams in Fig. 2 (and also in Fig. 19) is highly exaggerated; in a faithful representation of a realistic wafer process geometry, the tilt angles would be very small and seemingly inconspicuous to the human eye. Furthermore, it is also possible (by using, for example, a set of consecutive electrostatic multipoles) to shift the set of sub-beams laterally without affecting their alignment; in such a case, the tilt angle would not change at all. The deflection range can also be conveniently chosen to be symmetrical about the z-axis.

[0051] A typical value of the distance N'w between successive partial beams on the wafer is 180 µm.

[0052] The positioning of the partial beams on the respective pixels is controlled by suitable ion-optical devices in the imaging system, e.g. by means of an electrostatic multipole. Fig. Figure 3 (top panel) shows the function relationship of the spatial deflection x of the partial beams to time. Not all of the time can be used to open the blanker apertures because the partial beams and multipoles have a certain decay time, in Fig. 3, referred to as Ts, to stabilize after repositioning and settling. The decay time Ts is a small fraction of the pixel exposure period T1. The remaining part of the pixel exposure period T1, the usable time Tu=T1-Ts, is used for exposing the pixels. The time interval Tu is the pixel exposure time within which the appropriate dose must be delivered to the respective pixels. In Fig. 3 are also the times of the partial images of the Fig. 2. Accordingly, the time interval between consecutive frames (such as 0:0 and 0:1) is dt = Tu / N (here = Tu / 5), but the last two frames of the Fig. 2 separated by a time period Ts. The Fig. The write process described in section 2 does not include redundancy for the blanker array. This poses the potential risk that if a blanker cell fails, the corresponding pixel will always be illuminated or not illuminated at all (depending on the nature of the blanker cell's failure).

[0053] Redundancy can be added simply by doubling, tripling, etc., the number of blank cells illuminating a pixel; generally, a K-multiplication. A redundant tripling of the aperture (i.e., K=3) is exemplified in Fig. 4. Note that the spacing of these redundant apertures is an integer multiple of Nw. For example, it is possible to create K sets of (non-redundant) N' consecutive apertures, such that K blocks of redundant apertures write in a line, and the offset between these blocks is N'Nw. An alternative arrangement consists of K consecutive redundant apertures in each block, and N' suitably spaced blocks (e.g., with an offset of N'w between the last aperture of one block and the first aperture of the next block); the offset between blocks is then KNw.

[0054] In a realistic implementation of the invention, the non-redundant and redundant writing are carried out according to the above, based on the Fig. 2 and Fig. 4 explanations. A simple example is shown in Fig. 5, with N=3, N'=4, and K=2. The sub-beams are designated A1, A2, B1, B2, C1, C2, ..., namely a combination of a capital letter A, B, ... (for the sub-beams of different redundancy groups) with a number 1 or 2, which designates the individual redundant sub-beams. Thus, sub-beams A1 and A2 are redundant and offset from each other by Nw=3w, while the offset between the sub-beams of consecutive redundancy groups, e.g., between an A sub-beam and a B sub-beam, is a value (nN+N')w, which is relatively prime to Nw.

[0055] Referring again to Fig. 3, the time Tu and the particle flux in the beam are selected such that the maximum dose for a pixel, divided by a redundancy factor K', can pass through an aperture within the duration Tu. The redundancy factor K' used here is usually equal to K, but can be selected smaller if not all redundancy columns are activated. In contrast to the prior art method according to Fig. 20, in which the loading of the information is carried out individually for each pixel and the deflection x is reset for each pixel, according to the invention the usable time Tu covers several pixels within each sawtooth period of the deflection x. Furthermore, reloading of the pixel information can occur several times within a period Tu according to the respective gray level of the pixel (see below "Data coding"), as in Fig. 3 for the pixel deflection signals p A to pE shown. Architecture of the blanking plate

[0056] The following discloses an electronics architecture for a blanking plate 202 that enables the implementation of a write operation according to the above write method principle. An overview block diagram of such a blanking plate 60 is shown in Fig. 6. The blanking plate may physically consist of a single wafer chip field containing logic circuits and blanking apertures, or it may be a stack or a composite architecture of two or more components. For example, a chip field containing the logic circuits may be bonded to a chip field containing the blanking electrodes. The maskable openings are arranged in an aperture field 61 in 2R rows of C apertures each. Thus, the blanker cells may be viewed as C cell columns running perpendicular to the rows (perpendicular in Fig. 6; the direction of the rows is horizontal). The number C of blanker elements (apertures) per row is preferably equal to N*N' or an integer multiple of this number. The blanker cells are addressed column-wise using two memory arrays 62N, 62S of the APS, which are, for example, static RAM blocks with doubled ports that allow for high bit density (N and S stand for north and south, respectively).

[0057] The memories can be accessed in read and write mode from the external interface 63 of the blanking plate 60. A control logic 64 interprets commands and data passed to it via the external interface 63 and accordingly addresses 65 the memory arrays 62N, 62S and provides the control signals 66 (clock signal, enable, etc.) for the blanker cells in the aperture array 61. The northern RAM data provides the data and control signals to the upper R blanker rows, and the southern RAM is connected to the lower R blanker rows.

[0058] In the embodiment shown here, the memory arrays 62N, 62S consist of a number of memory blocks 67 in the form of individual memory units. Each of these blocks is implemented as a dual-port RAM (DPRAM). Such DPRAMs are standard library elements from libraries of ASIC suppliers. It is obvious that in other embodiments, the memory of the blanking plate 60 can be organized differently, e.g., in two memory banks (corresponding to the arrays 62N, 62S) or just one memory, or a number of memory blocks 67 can be implemented by corresponding memory units, depending on the memory size of the memory units.

[0059] Referring to Fig. 7, the RAM of each memory block 67 is composed of static memory elements, allowing for a very high memory density within the ASIC. The entire contents of each RAM block are addressable by two ports, designated Port A and Port B, each accessible by its own access means, including an address bus, a data bus, and additional control signals as explained below. The two ports allow simultaneous access to information from two different locations. It should be noted that the data width and address width do not necessarily have to be the same on both sides; the memory can, for example, be organized as 8 bits on Port A and 32 bits on Port B.

[0060] Fig. Figure 8 shows the layout of a single half-column HC1 corresponding to the upper half (north half) of a blanker column according to a preferred embodiment of the invention. The blanker cells are designated Blanker0, ..., Blanker(R-1) and symbolically represented as square areas with a small square (the aperture). The lower half of the column is symmetrical (mirrored about the central horizontal axis of the aperture array 61), and further columns are simply added to the right and left of the shown column. The blanker information is stored in the DPRAM 82 through port A and then transmitted to the individual blanker cells through port B as control signal data via the control line 84. Each blanker cell is equipped with a 1-bit memory (e.g., a flip-flop) to store the current data supplied by the DPRAM.

[0061] In the interpretation of the Fig. 8, the width bw of the control line data is 2 bits (bw=2), but it can also have any other value. It is convenient to choose the number R so that it is an integer multiple of bw. Higher values ​​of bw allow a shorter time to reload an entire column. The individual blanker cells are enabled by row enable signals ('enable row') er0, er2, ... er(R-2) to store the information supplied by the DPRAM at the correct time, with each row enable signal activating a group of bw cells. Only one row enable signal is active at a time, so that the data on control line 84 is loaded into the correct blanker cell group; then the next blanker cell group is loaded, and so on. For example, for R=128 and bw=8, a column can be reloaded in 128 / 8 = 16 clock cycles. For a 70 MHz design, this means a time of 230 ns.

[0062] Preferably, sufficient memory is provided to enable interleaved reloading of the blank cells: The DPRAM memory is divided into two address ranges, A0 and A1, so that one address range, A0, can be reloaded via port A, while the data stored in the other address range, A1, can be loaded into the blank cells. In the next pixel exposure period, the banks are swapped; A1 is reloaded via port A, while A0 is loaded into the blank cells.

[0063] For efficient writing algorithms and for testing purposes, it is advantageous to provide a static data setting on control line 84 to enable simultaneous switching of all blanker cells to masking mode (switched-off apertures). This is achieved by means of a "masking block" 83, which can be activated by a masking signal mask-out. When the mask-out signal is set, the data of all blanker cells in the shown half-column are set to 1, whereas the masking block 83 is transparent to the data coming from port B in normal operating mode. Furthermore, it is possible to optionally provide a global blanker enable signal, which allows simultaneous switching of all apertures of the aperture array, e.g., as a signal that activates all mask-out signals and row enable signals of all (half) columns. The present design also allows the implementation of "grayscale," i.e.Exposure doses between no exposure (minimum dose 0%) and full exposure (maximum dose per pixel, 100%; see also US 2005 / 0242303 A1). The gray levels are realized as discrete steps from 0 to g-1, where g is an integer representing the 100% level. The number g of displayable gray levels can be realized in the present setup as follows.

[0064] Fig. Figure 9 shows a timing diagram of the control sequence with 8 gray levels (g=8, three bits), with time along the horizontal axis. The time Tu is conceptually divided into g time slots, and according to the desired gray level, the pixels are activated during some of these time slots. For a gray level of 5 out of a total of 8, for example, there are eight time slots in each Tu, five of which are set to on. When the dose imparted to a pixel in the time Tu is digitized into g gray levels, the blanker cells are recharged g times within Tu. The blanker cell groups are labeled sequentially by the row enable designation, and the gray level components are labeled g0 to g7. The entire charging cycle of the blanker cells occurs within the usable cycle time Tu. The data is loaded into the blanker cells at g consecutive times (corresponding to time slots) and as shown in Fig. 9 shown at the beginning of the respective time slots; at the same time, the next set of data for the following exposure period is loaded. It should be noted that, as shown in Fig. 9, the charging times between the different blank cell groups are staggered. In the final part of Tu, each blank cell group is switched off after the last grayscale time slot g7 has expired; this is conveniently done using the mask-out signals according to Fig. 8 controlled masking block 83. The off state lasts until the first grayscale time slot g0 of the next exposure period is loaded and activated. The timing diagram shown provides optimal dose preparation by the APS.

[0065] In Fig. Below the representation of the time-dependent deflection x, exemplary data streams for five different apertures during the time period Tu are illustrated in Figure 3. These data streams can be generated directly on the APS from the grayscale data. This process is explained in more detail in the following section. Data encoding

[0066] In realistic applications, the patterns to be generated contain grayscale, with g grayscale levels ranging from 0 to 100% of a pixel's full exposure. Since g grayscale levels require g bits to be stored in the SRAM, this results in overhead in the data. To reduce the data rate via the external interface, it is possible to send only the grayscale G (G=ld(g) bits) to the APS and expand the data from one grayscale level (a G-bit number) to the g aperture bits in the APS. This can be done before the data is stored in the DPRAM, or when reading the data from the DPRAM and passing it to the blanker cells.

[0067] If the data is first expanded and then stored in DPRAM, each gray level must be converted into a vector of bits. Consecutive vectors (i.e., vectors of consecutive apertures) are arranged in a matrix. After a certain number has been collected, e.g., the same number as the gray levels, to obtain a g ' g matrix, the matrix should be transposed (swapping rows and columns) and written to DPRAM.

[0068] On the other hand, expansion after storage in the DPRAMs (“late extraction”) saves memory space, and the controller does not have to change the data format of the data supplied via the external interface.

[0069] Fig. Figure 10 shows a basic logic diagram that implements a late extraction with redundant columns. Two corresponding half-columns are shown in Fig. 10 and are designated as HC1.0 and HC2.0 respectively. (For clarity, only two columns are shown, which are shown in the Fig. 4 explains are redundant, in Fig. 10; the second column HC2.0 represents all other redundant columns. The remaining columns, especially columns for writing further non-redundant pixels, are suppressed.) An additional timing signal 840 is used, which is common to all columns and is updated after each grayscale time slot. The DPRAM logic 871 for the first column HC1.0 derives the full data with a width G bw for the set of redundant half-columns from the received data. An extractor 841 extends the grayscale information and extracts the single-bit data (bw bit) for the first half-column HC1.0 based on the timing signal, as explained in more detail below. The full data of width G bw is passed to the DPRAM of the next redundant column HC2.0. The DPRAM logic 872 of the half-column HC2.0 can therefore use the copied data in the next exposure period; only the address data is supplied by the control logic.This procedure is repeated analogously for further (not shown) redundant half-columns as for column HC2.0.

[0070] It should be noted that the columns HC1.0, HC2.0, because they are designed for redundant writing, necessarily work with identical grayscale data; the expansion of the grayscale information into individual on / off information takes place in the extractors 841, 842, respectively. In the Fig. In the structure shown in Figure 10, the DPRAM of column 1.0 also provides buffer functions for the data input from the control logic.

[0071] Various algorithms can be used for data expansion, such as pseudorandom number generation algorithms, which distribute the on-time to the redundant apertures evenly across Tu in a quasi-stochastic manner.

[0072] The time information 840 is used by the extractors 841, 842 to generate an exposure bit according to the desired gray level and the actual time in order to obtain the bit sequence suitable for the gray level. At each time step corresponding to a gray level time slot, the extractor generates bw bits using or simulating an extraction function f for each of the bw rows, where the function f takes the time information 840 (G bits) and the gray level assigned to the respective row (G bits) and delivers a bit signal (Boolean function); symbolically: f = f(time, gray level). The result of this function is then passed on as a control signal to the corresponding aperture. In a simple embodiment, the time information 840 is generated by a linear counter (not shown) that is incremented at each time step, and the extraction function f is a comparison f = compare(time <Graustufe).

[0073] In a preferred embodiment, the timing signal 840 is the output of a G-bit random number generator (e.g., a linear feedback shift register LFSR) that generates the numbers 0 to g in random order for calculating each bit g of the exposure control signal stream, but each number once over a set of g grayscale time slots. Then, the function f can again be a comparison function, f = compare(time <Graustufe), die direkt das an die Aperturen weitergegebene Steuersignal betrifft. Die Verwendung eines solchen Zufallsgenerators dekorreliert den Bitstrom für eine Apertur über die Zeit.

[0074] Another parameter can be included in the comparison to decorrelate the data streams between the individual pixels. For example, in addition to its grayscale information, each pixel also receives an individual number that serves as a "ranking key" to determine the actual bit sequence generated for that gray level. The ranking key remains unchanged throughout the exposure cycle. The exposure key is used as an offset value that is added or subtracted (modulo g) to the value provided by the random number generator before the grayscale comparison is performed; i.e., symbolically: f = compare( [(time-ranking key) (mod g)] < grayscale). Redundant arrangement of the blanking plate

[0075] One possible (and probably the most likely) type of failure in an APS system as shown here is a blanking cell failure due to mechanical or electrical damage such that the associated partial beam cannot be blanked ("always-open failure"). By providing a stack of two blanking aperture plates positioned one above the other and running electronically in parallel, the probability of an always-open failure can be significantly reduced. This redundant configuration with two blanking plate devices 202a, 202b operating in parallel is shown in Fig. 11. The two outer ones in Fig. The apertures shown in Figure 11 are intact, whereas the two inner apertures have a defect in the first and second blank plates, respectively. Due to the redundant stacking of blank plates, proper operation of both apertures is ensured. Arrangement of the apertures

[0076] Fig. Figure 12 shows an example of an arrangement of apertures according to the invention. This drawing should be compared to the arrangement of the Fig. 18 according to the state of the art. In the supervision of the Fig. 12 the orientation is the same as in Fig. 18, and the orientation directions are consistent with those of the Fig. 6 and Fig. 8; the apertures are shown as cross-hatched squares. As in Fig. 18, each aperture has an assigned cell space of 3w × 3w. In contrast to the prior art, an additional space is inserted between each block of three apertures to realize an offset of N'w (here: 10w) that is coprime with the base offset in a row (here: 3'3w).

[0077] Fig. Figure 13 shows a variant of the arrangement that incorporates a redundancy of two (K=2). Each block of apertures is now doubled. The enlarged section of the Fig. Figure 14 shows the unit cell of this arrangement. As can be seen, the offset of the apertures within the redundancy group of apertures is an integer multiple of the base offset (here, 9w), while the offset to the first aperture of the next block includes an additional offset; here, the offset is 1w, so that N'w=(1+9)w. Thus, the apertures in this cell are offset to provide redundancy, and the APS can be constructed by concatenating multiple versions of this unit cell. Mark reading

[0078] The proposed architecture can also be used for reading markings 41, 42 on the target 40. Such markings are made of a material that reflects a high percentage of the beam current or generates secondary radiation upon irradiation with the beam, so that these reflections can be measured by a detector 44 positioned to the side of the beam. Preferred types of secondary radiation can be, for example, secondary electron or Auger electron radiation. In particular, Auger radiation has the advantage of providing characteristic electron radiation at a specific energy with a very narrow linewidth, making it possible to tune the detector 44 accordingly. For example, with reference to Fig. 15 and Fig. 16, a two-stage mark reading can be provided to enable coarse and fine synchronization of the positioning between the table and beam positions. In the first stage ( Fig. 15), to find the coarse position on the resist, a marker coding with a high autocorrelation peak can be used, such as a bar code or lines arranged at marker intervals according to prime numbers. The MD device is then set up to illuminate the marker pattern 41 on the resist, and if the programmed pattern matches the pattern on the target, a correlation peak of the secondary current is observable. Optimal performance is achieved when all markers have a width w, namely equal to the sub-beam width. The correlation peak of the coarse marker pattern 41 is used to find the location of a fine marker 42, which is spaced from the coarse markers 41 to allow real-time processing of the detector data with an accuracy of ±w.

[0079] In the second stage ( Fig. 1) The fine marker 42 can be used to increase the resolution of the marker-finding process. Because it is known that the position of the fine marker lies in a 2-wavelength field, the MD device and the deflectors are set up to move through this field with a step size d according to the required resolution, while the relative motion of the target is preferably maintained so that it moves with the same uniform motion and speed as during the pixel exposure process described above. A first sub-beam 1 is positioned at the beginning of the 2-wavelength interval (x=0), a second sub-beam 2 at x=d, and so on until the end of the interval is reached. Thus, the sub-beams overlap and result in an intensity distribution that gradually increases to a maximum and then decreases again, as in Fig. 16. When the largest reflection current is measured, the best possible congruence between the partial beams and the marking is found. This process can occur independently of any movement of the resist.

[0080] The detector 44 used for position synchronization should have a sufficiently high temporal resolution corresponding to the required spatial resolution (namely w or d) divided by the speed v of the relative target movement. Therefore, the sampling rate of the detector 44 is expediently at least less than 1 / Tu. To increase the signal integration time, when the beam passes the area where the markings are located, the time during which the beam is held relative to the substrate can be chosen to be greater than the time required to expose one pixel (e.g., 10 Tu), without changing or reducing the target's scanning speed. In other words, the distance traveled during position synchronization can be a multiple of the distance covered by the relative movement of the target during a typical pixel exposure period Tu in a subsequent (or previous) exposure process.This approach may appear to result in some pixels in the area immediately following the marking region in the scanning direction not being fully exposed; however, this is actually avoided because the marks are sufficiently spatially separated from the surrounding lithographic pattern (or microstructure pattern) to be created.

[0081] Fig. Figure 15a illustrates the use of a variant of the markings, namely, hidden markings 411. The markings 411 may have been produced during previous processing steps and covered with a cover material produced in an additional process step. The structuring of the markings 411 may also have been achieved in situ by suitable methods, such as deep ion irradiation, wherein the ions penetrate a cover layer and significantly influence only a hidden layer (which would then be the markings). A suitable implementation of the markings are metallic layers, preferably in a multilayer structure, which lie beneath but close to the target surface. Fig. 15a shows a first stage as above based on the Fig. 15, but it is obvious that the hidden markers 411 are also used for positioning with fine resolution as in Fig. 16 can be used.

[0082] In summary, the invention allows for the realization of designs using fully synchronous design processes with only a single clock signal. Although the apertures do not switch simultaneously (which is an advantage for the power consumption characteristics of the chip), the time during which an individual aperture is open or closed is extremely precise, as it depends only on the accuracy of the clock (which is typically in the range of a few ppm). Further advantages of the present blanking plate structure include: - Reduced number of flip-flops, which is significantly reduced compared to previous PML2 concepts. - Number of flip-flop switching reduced by a factor of several thousand. - Drastic reduction in energy consumption (in the same operating mode). - Flip-flop switching distributed over a longer period of time. - Switching of the apertures spread over a longer period of time. - The conception is completely synchronous and uses only one clock. - Write procedure allows corrections of the optical system within the data stream.

Claims

[1] A method for irradiating a target with a beam of energetic, electrically charged particles, comprising the steps of: - providing a pattern definition means (102) having a plurality of apertures, - illuminating the pattern definition means by means of the beam which passes through the pattern definition means through the apertures and thereby forms a shaped beam (pb) consisting of a corresponding plurality of partial beams, - forming the shaped beam into an image on a location of the target (14, 40), the image comprising images of at least a portion of the plurality of apertures, the images of the apertures having a first width (w), - Moving the target relative to the pattern definition means along a path within an image plane in which the images of the apertures are formed, where during pixel exposure times (Tu) the location of the image is moved along with the target, characterized by , that the pixel exposure time has a duration within which the relative movement of the target covers a distance, the ‘forward distance’, which is greater than the first width (w), measured on the target, but less than or equal to the distance between adjacent partial beams along the direction of the relative movement, measured on the target, and after the pixel exposure time, the location of the image relative to the target is changed, this change in location substantially compensating for the movement of the location of the image during the pixel exposure time (Tu) relative to the pattern defining means (102), the relative movement of the target during successive pixel exposure times (Tu) occurring over a distance which exceeds the distance between adjacent partial beams along the direction of the relative movement. [2] A method according to claim 1, wherein the forward path is an integer multiple (Nw) of the first width (w). [3] A method according to claim 1 or 2, wherein the forward distance is smaller by a first width than the distance between adjacent sub-beams measured along the direction of relative movement. [4] Method according to one of claims 1 to 3, in which the partial beams which lie in the same row along the direction of relative movement of the target form groups of at least one partial beam, wherein adjacent partial beams in each group of more than one partial beam are spaced apart by a distance which, measured on the target, is an integer multiple of the forward distance, and wherein the number of these groups is selected such that the product of the first width and the number of groups is greater than or equal to the forward distance. [5] Method according to one of claims 1 to 4, wherein the relative movement of the target is a continuous linear movement within time intervals containing a plurality of pixel exposure times. [6] A method according to any one of claims 1 to 5, wherein outside the pixel exposure times the shaped beam is masked out in a manner in which it is absorbed by an absorbing means before reaching the target. [7] A method according to any one of claims 1 to 6, wherein the relative movement of the target during a period of time between the start times of successive pixel exposure times constitutes a second width (Nw) which is an integer multiple of the first width. [8] Method according to claim 7, in which the distances between two partial beams lying in the same line along the direction of relative movement of the target are integer multiples (N'w) of the first width, and at least some of these integer multiples (N') are relatively prime to the value of the integer multiple (N) underlying the second width. [9] A method according to claim 8, wherein said integer multiples differ by one. [10] A method according to any one of claims 1 to 9, wherein the target moves relative to the pattern defining means along a direction substantially perpendicular to the direction of propagation of the shaped beam. [11] Method according to one of claims 1 to 10, in which the pattern definition means (102) has a plurality of electrostatic partial beam deflectors (221), each of which is assigned to an aperture of the pattern definition means and is each controlled by an individual control signal (84) and is arranged, when activated by the associated control signal, to deflect the partial beam passing through the respective associated aperture by an amount sufficient to deflect the partial beam from its desired path, and the apertures of the pattern definition means are controlled in a time-dependent manner according to a pattern to be formed on the target by means of the plurality of control signals. [12] Method according to claim 11, in which, due to a grouping of the apertures into disjoint groups (HC1), the control signals (84) of the partial beam deflectors (221) of the apertures of each group are supplied in a time-graded manner via a common control line. [13] A method according to claim 12, wherein the apertures of each group in the pattern defining means are arranged along an orientation direction transverse to the direction of relative movement of the target. [14] A method according to claim 13, wherein the apertures of each group extend over one half the width of the area occupied by the apertures of the pattern defining means measured transversely to the direction of relative movement of the target, and the control signals are fed to the partial beam deflectors (221) in this area from apertures from two opposite sides. [15] Method according to one of claims 11 to 14, in which data for activating the control signals are fed into data storage means (67) of the pattern definition means and stored there, and these data are read out individually for correspondingly activating the control signals. [16] A method according to claim 15 using dual-port data storage means, wherein during each pixel exposure time, data for controlling the apertures are read out through a first port of the data storage means to the control signals, and data relating to the subsequent pixel exposure time are stored via a second port of the data storage means. [17] A method according to claim 15 or 16, wherein the apertures of the pattern definition means are grouped into groups of mutually redundant apertures, wherein in each group the control signals associated with the apertures of the group are generated from a common signal (871) which is copied in turn to the control signals. [18] A method according to claim 17, wherein the common signal is copied to the control signals with respective time delays corresponding to the spatial offset of the respective aperture along the direction of relative movement of the target. [19] A method according to claim 17 or 18, wherein the common signal includes information of a gray value to be converted by means of the apertures belonging to the group, and the gray value is developed into a sequence of inactive and active control signals, the gray value defining the relative number of inactive control signals with respect to the total number of control signals in the group. [20] A method according to claim 19, wherein the development of the gray value occurs after the data has been read from the storage means. [21] A method according to claim 19 or 20, wherein developing the gray value into a sequence of control signals includes at least one of the following methods: Expansion into an evenly spaced sequence, expansion into a pseudorandom number sequence, expansion using linear feedback shift registers, expansion using different shifts for the individual control signals; or a combination of these. [22] A method according to any one of claims 11 to 21, wherein the apertures are arranged in at least two parallel devices offset from each other by a distance along the beam direction, the apertures of the parallel devices being operated in a parallel manner. [23] A method for positioning a target using a beam of energetic, electrically charged particles, comprising the steps of: - providing a pattern definition means (102) having a plurality of apertures, - illuminating the pattern definition means by means of the beam which passes through the pattern definition means through the apertures and thereby forms a shaped beam (pb) consisting of a corresponding plurality of partial beams, - forming the shaped beam into an image on a location of the target (14, 40), the image comprising images of at least a portion of the plurality of apertures, the images of the apertures having a first width (w), and at least a portion of the beam impinging on markers (41, 42) provided on the target, which then generate secondary radiation, - moving the target relative to the pattern definition means along a path within an image plane in which the images of the apertures are formed, wherein during a pixel exposure time (Tu), the relative movement of the target covers a distance (“forward distance”) which is greater than the first width (w), measured on the target, but less than or equal to the distance between adjacent partial beams along the direction of the relative movement, measured on the target, wherein the relative movement of the target during successive pixel exposure times (Tu) occurs over a distance which exceeds the distance between adjacent partial beams along the direction of the relative movement, - measuring the secondary radiation in a detection system with a sufficiently high sampling rate, and - Adjusting the target position in such a way that the measured secondary radiation reaches an extreme value (maximum or minimum). [24] The method of claim 23, wherein the substrate guide moves the target at a uniform speed equal to the speed at which the target moves during a subsequent exposure process. [25] A method according to claim 23 or 24, wherein the markings comprise structures which are part of a multi-layer structure and which have been formed on the target in at least one preceding process step. [26] A method according to any one of claims 23 to 25, wherein the markers comprise components for generating secondary radiation containing Auger electron radiation, and the detection system uses an energy-selective detector arranged to detect said Auger radiation. [27] A method according to any one of claims 23 to 26, wherein the markers comprise components with a characteristic backscatter yield. [28] A method according to any one of claims 23 to 27, wherein the forward distance is a multiple of a distance travelled in the relative movement of the target during a typical pixel exposure time during a subsequent exposure process.

Citation Information

Patent Citations

  • Maskless particle beam system for exposing pattern on a substrate

    GB2389454A

  • Advanced pattern definition for particle-beam exposure

    US20050242303A1

  • Maskless particle-beam system for exposing a pattern on a substrate

    US6768125B2

  • Particle-beam exposure apparatus with overall-modulation of a patterned beam

    WO2007112465A1

  • Apparatus and method for imaging a multiple particle beam onto a substrate

    DE102004055149A1