Patent Application
20150069260
The present application claims priority to European Application No. 13183963.1 filed Sep. 11, 2013, the disclosure of which is hereby incorporated by reference in its entirety.
The invention generally relates to an advanced programmable multi-beam pattern definition device for use in a charged-particle multi-beam processing (in particular nanopatterning or semiconductor lithography) or inspection apparatus. Furthermore, the invention generally relates to a charged-particle multi-beam processing or inspection apparatus (such as a nanopatterning or semiconductor lithography apparatus) including this multi-beam pattern definition device.
In particular, said pattern definition device comprises an aperture array field configured to be irradiated with a beam of electrically charged particles and allow passage of the beam through a plurality of apertures thus forming a corresponding number of beamlets, each of the beamlets traversing the aperture array field along a respective beamlet path through the pattern definition device and extending downstream of the pattern definition device to a respective nominal path for each beamlet, wherein the pattern definition device includes a deflection array device positioned across said aperture array field and comprises
A pattern definition device (PDD) of the mentioned kind is used to control the exposure of a target (such as a mask blank or wafer substrate) by means of a multitude of charged electron or ion beamlets in order to expose a desired pattern on the target. Within the multi-beam processing/inspection apparatus, the PDD is irradiated with a wide beam of charged-particles and generates a plurality of beamlets, typically several hundreds of thousands, in a two-dimensional array, such as a 512×512 array. The plurality of beamlets is then imaged onto the target.
Charged-particle multi-beam lithography and processing is of high interest for nanolithography and nanopatterning applications, such as for multi-beam mask writing and for maskless multi-beam direct write processes on silicon wafer substrates. With regard to the present application the terms ‘target’ and ‘substrate’ are used without difference in meaning.
In particular electron multi-beam writing is a promising concept for the future industrial fabrication of leading-edge photomasks as needed for 193 nm immersion lithography, of EUV-masks for extended ultra-violet lithography (EUVL), and of templates (1× masks) for nano-imprint lithography, in particular for sub-20 nm semiconductor technology nodes, with extendibility to sub-10 nm technology nodes. For the multi-beam mask writer (MBMW) the applicant has coined the acronym eMET (electron Mask Exposure Tool). Configurations of multi electron beam direct write (EBDW) processing on silicon wafers in multi-column PML2 (short for “Projection Mask-Less Lithography”), configurations are described in U.S. Pat. No. 7,214,951 and U.S. Pat. No. 8,183,543 of the applicant.
A charged-particle multi-beam apparatus with a multi-beam PDD of the above-mentioned kind is disclosed in the U.S. Pat. No. 6,768,125 of the applicant, which is hereby incorporated into the present disclosure as relevant prior art. That patent describes a charged-particle lithography and processing method and apparatus dubbed PML2 and further publications of the applicant describe eMET (short for “electron multi-beam Mask Exposure Tool”), both of which realize a multi-beam writing concept and use a PDD for structuring a charged-particle beam, which is extracted from a single source of electrically charged particles (electrons or ions). The PDD is, preferably, a programmable multi-aperture device realized as a so called aperture plate system (APS). The implementation based on a programmable APS allows a significant improvement of the achievable productivity in comparison with focused single spot beam systems as well as variable shaped beam (VSB) systems. The reasons for the improved productivity are, firstly, the parallelism of the process using a plurality of beams and, secondly, the increased current (of all beamlets writing in parallel) which can be imaged to a substrate at the same resolution. As compared with a single electron beam system, the current density at target of the electron multi-beam writing apparatus (in the order of A/cm2) is lower by approx. two orders of magnitude when compared to VSB systems, thus reducing instantaneous heating effects which are unavoidable when using single beam systems with high (>100 A/cm2) current density.
The different kinds of plates in the APS setup of the PDD realize a high integration density of apertures and deflectors. Within the PDD, accurate alignment between the two or more plates and adequate adjustment towards the direction of the incoming beam is required. This alignment can be accomplished in-situ by means of a configuration as disclosed in U.S. Pat. No. 8,546,767.
The DAP has electronic circuitry associated with the apertures to allow “switching on ” and “switching off” of the individual beamlets in accordance with the pattern which is to be generated on the target. In particular, the DAP is provided with a deflection array composed of a plurality of deflection devices for deflecting selected beamlets (“switching off”). Each deflection device is located at an opening formed in the DAP and can generate an electrostatic field sufficient to deflect the respective beamlet traversing through the opening from its nominal path. The beamlets which are not deflected are, as already mentioned, imaged onto the target 14.
The DAP is manufactured by three main processing steps: 1) the fabrication of the CMOS circuitry; 2) an etch process for creating the apertures; and 3) a MEMS process for building the capacitive blankers. To date the fabrication process for a DAP is highly reliable, but nonetheless there is always a certain number of the beamlet deflectors in a DAP which are not working in the specified way. The complexity of this manufacturing process may cause a certain failure rate in the functioning of some blankers which can thus not be controlled, resulting in apertures that remain in an open state or apertures that are forever closed. Typically, the number of such beamlet deflector defects is small, below one per mille (i.e., in the order of a hundred for one DAP operated for more than 100-thousand programmable beams). Usually, the beamlet deflector defects are non-clustered, meaning that they are statistically distributed over the field of apertures. Thus, in general one can expect that there are no neighboring beamlet deflector defects, i.e., one can assume that between the locations of two such defects at least one operating beamlet deflector is present, and usually a few.
The beamlet deflector defects result in either “always-off” defects or “always-on” defects. An “always-off” defect is present where the beamlet is either blocked from passage through the PDD or deflected from its path so that it will not reach the target, at all times and in particular regardless of the pattern information. An “always-off” defect may be due to a closed aperture in the aperture array plate and/or DAP, or by the presence of a voltage at the deflection electrodes of the deflector device associated with the respective opening in the DAP, where this voltage cannot be turned off during operation of the DAP. Thus, the beamlet affected by this “always-off” defect is always filtered out and cannot be used to contribute to the exposure of a pattern on the target. Thus, during writing a pattern to a substrate, an “always-off” defect will not allow the passage of the beamlet, which would be needed for a correct exposure.
An “always-on” defect (also called “always-open”) occurs where a deflection device is defective, for instance due to a defective electrode or non-operative voltage-supply of the electrode, in such a way that the respective beamlet can always pass through the corresponding aperture, without being deflected at all or without being deflected to an amount sufficient for being completely filtered out before the target (for instance at the stopping plate 17).
It is possible to compensate for the effect of an “always-off” defect within an exposure algorithm which uses redundant writing, such as disclosed in U.S. Pat. No. 8,222,621 of the applicant, by taking into account the location of the beamlet affected by the flawed device and shifting the corresponding amount of exposure to one or more other beamlets within the same redundancy group. The function of the beamlet would thereby be substituted by one or more beamlets at functioning deflection devices. In contrast to such “always-off” defects, an “always-on” defect, besides being a nuisance, is generally more difficult though not impossible to deal with. For an “always-on” defect it is not possible to overcome the effect of unwanted exposure. Thus, it is desired to ensure that a PDD has no “always-on” beam defects.
Therefore, it is an aim of many embodiments of the present invention to provide PDD units without “always-on” beam defects, for instance by suitable ways to correct such PDD units which, upon manufacture thereof, exhibit apertures having this type of defect.
In prior art approaches for compensating a defect beam are described. For instance, JP 2006-019436 A describes compensating a beam failure of the “always-off” kind by transferring a neighboring beamlet to the position of the beam failure. In particular, arrays of charged-particle beam deflectors are inserted before or after an aperture array; these mufti-deflecting systems enable transferring selected beamlets to the paths of respective neighboring beamlets. This allows to shift a functional beam to the site which the defective beam would have in the aperture array plate.
This aim is achieved by a PDD as described in the beginning, in which one or more defective deflection devices are present which are permanently unable to deflect the beamlets traversing blanking openings with which the defective deflection devices are associated, thus leaving such beamlets as “non-deflected beamlets”, by means of an additional filtering device. The filtering device is positioned across the aperture array field and comprises
This solution provides a device capable of suppressing the unwanted beamlets, which correspond to “always-on” defects, by physical means. In particular, many embodiments of the invention regard modifying the PDD at the sites where an “always-on” defect is present in such way as to prevent the passage of the respective beamlets towards the target at all times of operation. In other words, all “always-on” defects are converted into “always-off” defects, which are much easier to compensate. It is remarked that the number of defective deflection devices is lower by at least one order of magnitude, but usually two or three orders of magnitude or more, than the complete number of deflection devices in the deflection array device.
The solution according to many embodiments of the invention is fundamentally different from the approach of JP 2006-019436 A since the filtering device of many embodiments of the invention does not include a lens array. AH other beamlets are left unaffected so as to propagate along their respective paths, and in particular are not shifted so as to substitute a defective beamlet. Many embodiments of the invention provides a way to turn an “always on” defective beam into an “always off” defective beams, in contrast to JP 2006-019436 A where “always off” beams are not tolerated.
In one advantageous aspect of many embodiments of the invention, the (at least one) obstructing device is programmable to take two states, namely, an open state not affecting the respective non-deflected beamlet(s) and the obstructing state. In this case, the filtering device may suitably be realized as a plate-shaped device (“correction plate”) comprising an array of openings which correspond to the plurality of blanking openings of the deflection array device, and a plurality of obstructing devices, each obstructing device being associated with a respective opening and comprising at least one electrostatic electrode. Then, the deflection devices are selectively activatable to realize the obstructing state by diverting the beamlets traversing said respective openings from their nominal path, thus preventing the respective beamlets from traversing the pattern definition device along their respective nominal paths downstream of the pattern definition device for said minimum duration.
Whereas the switching times in the deflection array device need to be very fast in order to achieve productivity of the multi-beam writing or inspection apparatus, the switching times in the programmable obstructing device can be very slow. Thus, the obstructing device is configured to be switchable into the obstructing state for a certain minimum duration. This minimum duration may be greater than the product of the typical duration of activating the deflection devices in the deflection array device times the number of openings in a row of openings in the deflection array device; but often, this minimum duration is even longer and holds for the duration of one or more exposure processes of a target.
In another aspect of some embodiments of the invention the (at least one) obstructing device may be realized in the form of an opening provided with material obstructing the passage of the respective beamlet.
In one useful implementation of the PDD the deflection array device and the filtering device are respectively realized as plate-shaped devices comprising a respective array of openings; in particular, the filtering device may be realized as “correction plate” or “filtering array plate”. Furthermore, it is suitable when the filtering device is positioned across the aperture array field substantially parallel to the deflection array device.
In another useful development, the filtering device may also include the function of a (downstream) filed-boundary array plate. In this case, the filtering device may include—as the last component of the pattern definition device as seen along the direction of the beam—a plate-shaped device comprising an array of openings which correspond to the plurality of blanking openings of the deflection array device, and further including a downstream surface of smooth, preferably planar, shape except for said openings, to be oriented towards particle-optical components located downstream of the pattern definition device, such as a projection system for imaging the patterned beam onto the target.
Additionally an aperture array device may be provided comprising the plurality of apertures in the aperture array field; these apertures are configured to define the shape of the beamlets formed in the pattern definition device. The aperture array device is, preferably, positioned separate from the filtering device, i.e., in a small but finite distance.
In an advantageous alternative, the filtering device may be an aperture array device comprising the plurality of apertures in the aperture array field.
In the following, several embodiments of the present invention are described in more detail with reference to the drawings, which show:
The various embodiments of the invention discussed in the following are based on a development of the eMET-type charged-particle multi-beam exposure apparatus and its PDD system, having a large-reduction projecting system. In the following, first the technical background of the apparatus is discussed as far as relevant to many embodiments of the invention, then certain embodiments of the invention are discussed in detail. It should be appreciated that embodiments of the invention are not restricted to the following embodiments or the particular layout of PDD systems, which merely represent examples of possible implementations of the invention; rather, many embodiments of the invention are suitable for other types of processing systems that employ a charged-particle beam and a multi-beam patterning as well.
A schematic overview of an electron multi-beam mask exposure tool 100 (mask writer) employing an embodiment of the invention is shown in
The main components of the apparatus 100 are—in the order of the direction of the beam lb, pb which in this example runs vertically downward in FIG. 1—an illumination system 101, a PDD system 102, a projecting system 103, and a target station 104 with the target or substrate 14. The charged-particle optical systems 101, 103 are realized using either electrostatic lenses only (e.g. when using ion multi-beams) or a combination of electrostatic and magnetic lenses (e.g. when using electron multi-beams). The charged-particle optical parts 101,102,103 of the apparatus 100 are contained in a vacuum housing (not shown) held at high vacuum to ensure an unimpeded propagation of the beam lb, pb along the optical axis of the apparatus.
The illumination system 101 comprises, for instance, an electron or ion source 11, an extractor arrangement defining the location of the virtual source, a general blanker (not shown in
The beam lb then irradiates a blanking device which, together with the devices needed to keep its position (not shown), forms the PDD system 102, which is also shown in a schematic perspective detail view at the left-hand side of
Each switched-off beamlet is absorbed, for instance, on a stopping plate 17, which is situated at or near the second cross-over c2 of the charged particle projection optics; the other beamlets, which are switched-on, pass through a central opening of the plate 17 and thus are projected onto the target. The filtering of deflected beamlets can also (as e.g. shown in
In
The switched-on apertures are the only portions of the PDD transparent to the beam lb, which is thus formed into a patterned beam pb emerging from the apertures (i.e., in
A charged-particle optical projection system 103 then projects the pattern as represented by the patterned beam pb towards the substrate 14 (such as a 6″ mask blank with resist coating); since the beamlets which are switched off are absorbed at the stopping plate 17, only the switched-on beamlets will form an image of the switched-on apertures. The projection system 103 implements a demagnification of, for instance, 200:1, as realized by the applicant. The substrate 14 may be, for instance, in the case of an eMET-type system a 6-inch mask blank or a nanoimprint 1× mask or master template, covered with an electron sensitive resist layer, whereas for a PML2 system the substrate 14 may be a silicon wafer covered with a charged-particle sensitive resist layer. The substrate is held and positioned by a substrate stage (not shown) of the target station 104.
The projection system 103 is, for instance, composed of two consecutive charged-particle optical projector sections with a crossover c1 and c2, respectively. The charged-particle optical lenses 30, 31 (comprising, for instance, an electrostatic multi-electrode accelerating lens 30 and two magnetic lenses 31) used to realize the projectors are shown in
As a means to introduce a small lateral shift to the image, i.e. along a direction perpendicular to the optical axis cx, deflection means 16 are provided in one or both of the projector sections. Such deflection means can be realized as, for instance, a multipole electrode system, as discussed in the U.S. Pat. No. 6,768,125. Additionally, an axial current coil may be used to generate a an axial magnetic field providing rotation of the pattern in the substrate plane as needed. The lateral deflections are usually quite small in comparison to the lateral width of the beam array field at the substrate which is the width of the structured beam pb leaving the PDD divided by the reduction factor of the charged-particle projection-optics. Thus, the lateral deflections are at least one order of magnitudes below the width of the beam array field (it should be appreciated in this context that the lateral distance between beamlets is considerably smaller than the mentioned width of the beam array field).
Referring to
The PDD system 200 comprises a number of plates 201, 202 which are mounted in a stacked configuration, realizing a composite device whose components serve respective specific functions. Each of the plates is preferably realized with silicon microsystems technology in which the structures have been formed by micro-structuring techniques as outlined e.g. in U.S. Pat. No. 7,687,783 and U.S. Pat. No. 8,198,601.
A plurality of apertures is located in membranes mb formed by thinned regions of silicon wafers usually cut to a smaller quadratic or rectangular format, forming an aperture array field, designated by reference symbols af in
Referring to
Positioned in proximity to the AAP 201, preferably downstream thereof, a deflector array plate 202 (DAP; also referred to as blanking plate in view of its function in the context of the apparatus 100) is provided. The DAP 202 serves to deflect selected beamlets and thus to alter their beam path. The DAP has a plurality of so called blanking openings, which each correspond to a respective aperture of the aperture array plate 201. As already mentioned and shown in
Each blanking opening in the DAP is provided with a beamlet deflection device which allows to individually deflect charged particles transgressing through the opening, thus diverting the beamlet traversing the opening off its path. Each beamlet deflection device includes a set of beamlet deflection electrodes, usually a pair. Preferably, each set has electrodes of different types: a first type are ‘ground electrodes’, which are applied at the potential of the PD device, whereas another type, which is called here the ‘active electrodes’, are applied individual potentials in order to deflect the beamlet transgressing the corresponding blanking opening. The ground electrodes may be shared between adjacent beamlet deflection devices. They may be formed so as to have a substantial height over the height of the active electrodes. This is done in order to provide sufficient shielding of the blanking deflection device against cross-talking and other unwanted effects such as lens effects incurred by the deflection electrode geometry,
In the DAP illustrated in
For instance, beamlet b1 transgresses the subsequent larger opening of the pattern definition system 102 without being deflected, since the beamlet deflection device formed by the respective set of beamlet deflection electrodes is not energized, meaning here that no voltage is applied between the active electrode 221 and the associated ground electrode 220. This corresponds to the “switched-on” state of the aperture. Beamlet b1 passes the pattern definition system 102 unaffected and is focused by the particle-optical system through the crossovers and imaged onto the target with a reduction as induced by the charged-particle projection optics. For example, in systems implemented by the applicant, a reduction factor as large as 200:1 was realized. In contrast, as shown with beamlet b2, a “switched-off” state is realized by energizing the beamlet deflection device of this aperture, Le. applying a voltage to the active electrode 221′ with regard to the corresponding ground electrode. In this state, the beamlet deflection device formed by electrodes 220′,221′ generates a local electric field across the path of the corresponding beamlet b2 and thus deflects the beamlet b2 off its normal path p0 to a deflected direction. As a consequence the beamlet will, on its way through the charged-particle optical system, obey an altered path p1 and be absorbed at an absorbing means provided in the PDD system or the charged-particle optical system (e.g. at the stopping plate 17,
In the configuration shown in
It is to be noted out that the PDD system may further include an additional plate (not shown, but see the variant of
Furthermore, in
Such “always-on” defects can be detected by using in-situ beam diagnostics, or it might also be that the position of such beam defect is known from the characterization of the PDD unit prior to inserting it into the multi-beam writing or inspection tool. In particular, each DAP can be tested with a dedicated test bench in order to find openings having a deflector defect. A test bench suitable for this purpose is described in the article of S. Eder-Kapl et al. “Characterization of CMOS programmable multi-beam blanking arrays as used for programmable multi-beam projections lithography and resistless nanopatterning”, J. Micromech. Microeng, 21 (2011) 045038 (doi:10.1088/0960-1317/21/4/045038). This test provides a map of the locations of “always-on” and “always-off” defects. Thus, it is possible to determine the number and positions of DDDs for each DAP.
A first approach to correct “always-on” defects according to some embodiments of the invention is by using a correction plate, which is used as an in-situ device for eliminating “always-on” defects. In one preferable realization this correction plate is an additional plate-shaped device which is inserted to the plate stack of the PD device.
The correction plate 53 is a programmable device superficially similar to the DAP, having the ability to deflect or let pass the individual beamlets emanated by the DAP. In contrast to the DAP, the configuration of deflection will be kept during the entire processing of an exposure. The correction plate 53 is positioned within the PDD system, for instance at a short distance downstream the DAP 52, The correction plate 53 is usually configured with the same number and geometric positions of openings 70 as the DAP 52 with its openings 520.
In particular, the correction plate 53 is provided with a plurality of correction deflection device 71 (ODD), each of which is located at the site of one respective opening 70, such as to be able to deflect the beamlet traversing that opening 70 when activated. For instance, each of the CDDs 71 comprises at least two electrodes, for instance at least one active electrode 72 operating against a counter electrode 73. The counter electrodes 73 of the CDDs may be separate or connected with each other; they may be held at a common “ground potential”, or they may be applied a counter potential, for instance of same value but opposite sign as the potential of the active electrode 72, when the ODD is activated.
In
In the embodiment shown in
It is important to note that there are decisive differences between a DAP 52 and a correction plate 53 of an embodiment of the invention: While the DAP 52 has to include a fast CMOS circuitry in order to account for the fast-changing pattern information, the circuitry of the correction plate 53 can be slow since the mode of operation in the correction plate will stay constant over at least the duration of an entire exposure treatment of the target 14. It is sufficient that the circuitry of the correction plate 53 can be held in a programmed state for a minimum duration which corresponds to the duration of the exposure of a wafer, which is a multiple (by orders of magnitude) of the duration of the exposure of a row of pixels on the target. In terms of the elements in the PDD, this duration is a multiple of product of the duration of activating a deflection device 521 in the DAP 52 times the number of apertures in a row of apertures in the aperture filed af. For instance, an implementation of the correction plate circuitry may include simply wired components or may be realized as simple (flip-flop) slow CMOS circuitry, The latter is discussed more in detail in the following, referring to
Referring to
a shows two registers 77, 77′ for programming respective blankers 76, representing the first two of 64 registers forming a queue sequence S1. A sequence of 8 queues S1, S2, . . . S8 connected serially forms a row R01 comprising 512 flip flop registers FF1 to FF512, as visible in
Prior to operating the PD device, a bit pattern is loaded into the shift register array RR of the correction plate and will then remain there either statically or until a new pattern is loaded (or power is lost). Each bit in one of the shift registers 77 has two states (low and high). The low state of the register sets the blanker 76 of the respective deflection device to “open”, by not applying a voltage and thus allowing the beamlet to pass undeflected, whereas the high state causes the blanker 76 to apply a voltage to the electrodes, so as to effect deflecting the respective beamlet. The bit pattern loaded into the array is by default made up of low bits, allowing all beamlets to pass. Only those bits that correspond to defective, permanently open deflection devices in the DAP, are set high.
As illustrated in
Under vacuum conditions, the temperature of the correction plate 53 will be controlled by radiative cooling only. In order to minimize the current being drawn from the power source and to reduce heat production from switching processes during loading of the pattern, the eight groups G1 . . . G8 will be loaded with the pattern data in succession, one at a time, by means of separate data Data 1 in . . . Data8in and clock signals Clock1 in . . . Clock8in. The clock cycle time is chosen suitably so that the propagation delay of the cascaded clock repeaters is accounted for, since the clock signal arriving at the last flip flop takes 8 propagation delays of repeaters A1 to A8.
The row-style layout of the circuit array RR shown in
In another approach according to many embodiments of the invention a filtering device according to many embodiments of the invention is realized as a passive plate in the APS, where “passive” denotes that the plate does not have a circuitry to deflect beamlets. In the passive filtering plate an array of openings/apertures is provided, but it is modified such that openings are missing, or have been dosed, at those sites which correspond to the locations of the “always-on” defects.
To obtain such a modified array, for instance, one possibility is to use a repair system based, e.g., on a focused on beam (FIB) system, in which a gas injection supported deposition is used to dose the openings at the sites of an “always-on” defect. This approach uses the fact that with FB-induced deposition there is not only a deposition onto the surface, but always a side growth as well. This can be exploited to bridge openings, thus covering them, starting the FIB-induced deposition at one or more edges of the opening and then moving on the beam across the area of the opening, until the complete area of the opening is covered by deposited material. Instead of FIB-induced deposition, laser-beam or electron-beam induced deposition methods may be used as well.
Thus, referring to
The filtering array plate 83 is then combined with the (partially defective) DAP 82 while the original aperture array plate 81 is kept. The material in the site 84 will absorb or otherwise obstruct the passage of the respective beamlet b0 and thus prevents its passage through the PDD system 8. Thus, a PDD 8 is produced which avoids the “always-on” defects for the beamlets, and the obstruction site 84 serves as obstructing device according to several embodiments of the invention,
In an analogous manner as with
It is to be noted that in all embodiments, the first plate of the POD system can also serve as upstream FAP, by virtue of its planar surface which is oriented towards the illumination system 101. In order to enhance this function it may suitably include an electrically conducting cover layer (not shown).
It is further to be noted that also a programmable correction plate (cf.
In another variant shown in
Alternatively, a filtering aperture array plate may be produced having an aperture array with pre-defined missing apertures at the sites of “always-on” defects, by means of a fabrication process of state of the art where the plate pattern is written with an electron beam. This kind of fabrication process is straightforward. Therefore, this alternative of producing a modified aperture array plate is expected to be very effective as well.
| Number | Date | Country | Kind |
|---|---|---|---|
| 13183963.1 | Sep 2013 | EP | regional |
