WO2013185822A1

WO2013185822A1 - Maskless lithographic apparatus and method for generating an exposure pattern

Info

Publication number: WO2013185822A1
Application number: PCT/EP2012/061295
Authority: WO
Inventors: Axel Bodemann; Jochen Hetzler; Aksel GÖHNERMEIER
Original assignee: Carl Zeiss Smt Gmbh
Priority date: 2012-06-14
Filing date: 2012-06-14
Publication date: 2013-12-19

Abstract

The invention relates to a maskless lithographic apparatus (1), comprising: a light modulator (7) having a plurality of micromirrors arranged in a micromirror array for modulating an exposure beam (5) according to an exposure pattern, and an exposure optical system (8) for delivering the modulated exposure beam (5) onto a substrate (2). The apparatus (1) comprises a tilt error compensation unit (15) for compensating for tilt errors of the micromirrors of the micromirror array. The invention also relates to a corresponding method for generating an exposure pattern on a substrate (2), comprising: modulating an exposure beam (5) according to the exposure pattern using a plurality of micromirrors of a micromirror array, and delivering the modulated exposure beam (5) onto a substrate (2) in the form of a beam-spot array (40), wherein at least one of the modulating step and the delivering step comprises compensating for tilt errors of the micromirrors of the micromirror array.

Description

Maskless Lithoghraphic Apparatus and Method for Generating an

Exposure Pattern

Technical Field

The invention relates to microlithography, and more specifically to maskless lithography.

Background of the Invention

Lithography is typically the transfer of a pattern on a mask to a photosensitive material (photoresist) on a substrate, the patterned substrate being selectively etched for generating structures on the substrate which may be used as semiconductor devices. However, the expenditure for manufacturing masks for microlithography is considerable, especially for applications where a large mask size is required, as the manufacturing costs for the mask increase exponentially with the mask size. Manufacturing costs for lithography masks also become economically unreasonable when only a small volume production of

semiconductor devices is desired. As a result, maskless lithography techniques have recently become popular.

The article "High-resolution maskless lithography" by Kin Foong Chan et al., J. Microlithogr. Microfabrication, Microsyst. 2(4), 331 (2003), describes a maskless lithographic apparatus having a combination of low-numerical- aperture and high-numerical-aperture projection lens systems along with integrated micro-optics. A super video graphic array (SVGA) digital micromirror device (DMD) is used as a spatial and temporal light modulator. A mercury arc lamp filtered for the G-line (at 435.8 nm) or a custom H-line (at 405 nm) lens system in conjunction with a violet diode laser system may be used as a light source.

US 2010/0060874 A1 discloses a maskless lithographic apparatus having a light modulator modulating an exposure beam and an exposure optical system delivering the modulated exposure beam onto a substrate in the form of a beam spot array. A control unit switches off some rows in the beam spot array in order to make an exposure energy distribution uniform across the beam spot array. Moreover, a method for compensating an alignment error between a scan direction of the substrate and an arrangement direction of the light modulator is disclosed.

US 2011/0090479 A1 discloses an optical component for a maskless exposure apparatus which is capable of screening diffused light such that the image of a pixel of a DMD formed by a first image-forming lens in the maskless exposure apparatus has no influence on the image of a neighboring pixel and of totally reflecting the light after reflection or diffraction at the same time. The optical component may be devised as a micro-prism array or as a micro-mirror array such that a light incidence portion is formed in a wide manner and a light exit portion is formed in a narrow manner.

US 7,116,404 B2 discloses a lithographic apparatus having an illumination system that supplies an exposure beam, an array of individually controllable elements, a projection system and a substrate. The apparatus also includes a sensor system and a positioning system controllable to adjust a position and/or an orientation of at least one of the array of elements; a component of the projection system; and a component of the illumination system. The apparatus also includes a control system which controls the array of elements to pattern the beam, and which also receives the intensity signal and controls the positioning system according to the detected intensity distribution to adjust the projected radiation pattern. However, even by precisely aligning beam patterning components such as a Digital-Micromirror-Device and beam projection components of an exposure optical system of a maskless lithographic apparatus, there still remain errors which may degrade the quality of the projected exposure pattern and which may be detrimental to the correspondence between the light provided by the micromirrors of the micromirror array and the beam spots of the beam spot array.

Summary of the Invention

An embodiment of the present invention provides a maskless lithographic apparatus, comprising: a light modulator having a plurality of micromirrors arranged in a micromirror array for modulating an exposure beam according to an exposure pattern, an exposure optical system for delivering the modulated exposure beam onto a substrate (in the form of a beam spot pattern), and a tilt error compensation unit for compensating for tilt errors of the micromirrors of the micromirror array.

Controllable micromirrors of a micromirror array are typically supported by a hinge support and may be tilted between two discrete (fixed) tilt angles, corresponding to an "on" state and an "off state of a respective micromirror. The micromirror array is typically arranged with respect to the exposure optical system so that each micromirror either directs light of the exposure beam to the exposure optical system and thus onto the substrate ("on" state), or away from the exposure optical system, e.g. to an absorber ("off state). In general, each micromirror may be switched individually from the "on" state to the "off state and vice versa by a corresponding (digital) control signal provided by a suitable control unit. In this way, specific beam spots of a beam spot array formed on the substrate may be selectively switched on or off, in dependence of an exposure pattern to be produced on the substrate. The inventors have found that due to manufacturing imprecisions of the micromirror array, the tilt angle of the individual micromirrors in their "on" state does not necessarily coincide with the desired (nominal) value (typically being the same for all micromirrors), but deviates from the nominal value by an angle which typically varies from micromirror to micromirror, being referred to as tilt error in the following. The individual tilt error of each micromirror may lead to errors when generating the exposure pattern on the substrate. For instance, the tilt errors may lead to individual positional errors of beam spots of the beam spot array which is generated on the substrate by the exposure optical system, or to telecentricity errors of light bundles which are used for producing the beam spots on the substrate.

For delivering the modulated exposure beam onto the substrate, the exposure optical system may comprise one or more projection lenses. Moreover, typically, at least one of a microlens array, a (further) micromirror array, and a microprism array are arranged in the exposure optical system in order to form a beam-spot array on the substrate.

In the following, different techniques for compensating the tilt errors of the micromirrors will be described, which may be used alone or in combination for improving the precision of the exposure pattern to be produced on the substrate. It will be understood that the compensation of tilt errors of the micromirrors may be advantageously combined with other techniques for improving the precision of the exposure pattern on the substrate, e.g. by adjusting the position and/or orientation of the micromirror array (as a whole) and/or of specific optical elements of the exposure optical system.

In some embodiments, the light modulator is a Digital-Multimirror-Device (referred to as DMD in the following). DMDs are commercially available devices which may be provided with low expenditure. DMDs typically have a sufficient number of micromirrors for the present purposes, e.g. arranged in 1024 columns and 768 rows in accordance with the super video graphic array

(SVGA) standard. Moreover, the micromirrors of the DMD are typically arranged with a substantially equal pitch (e.g. of about 13.7 μιη) both in row and column directions. It will be understood that other electrically controllable light modulators having a sufficient number of micromirrors and allowing a high switching frequency between the "on" state and the "off state may be used as well in the present maskless lithographic apparatus.

In some embodiments, the tilt error compensation unit is an optical tilt compensation unit, typically in the form of an optical element for the

compensation of the tilt errors. For this purpose, the tilt error compensation unit is typically arranged in the beam path of the (possibly modulated) exposure beam which passes through the exposure optical system. The optical tilt error correction unit is typically transmissive for the light of the exposure beam which is directed thereon, as is the case with the exposure optical system. Depending on the type of optical compensation provided by the tilt error compensation unit, its position in the beam path may vary. In general, the optical tilt error compensation unit provides for an individual compensation of the tilt errors of each of the micromirrors. However, such an individual compensation does not necessarily require knowledge of the individual tilt errors of the micromirrors.

In some embodiments, the optical tilt error compensation unit comprises a plurality of light mixing elements, each for mixing light of the exposure beam modulated by one of the micromirrors, the light mixing elements having a constant diameter. Typically, the tilt error compensation unit provides a light mixing element for each of the micromirrors of the micromirror array, the light mixing elements also being arranged in a corresponding array, the spacing of the light mixing elements being chosen so that there is a one-to-one

correspondence between micromirrors and light mixing elements, so that light from one micromirror is provided to exactly one light mixing element. A light mixing element typically provides a homogenization of a light bundle which enters the light mixing element at an entrance surface by providing multiple reflections when the light bundle passes through the light mixing element to an exit surface thereof. In this way, the influence of the tilt error on the light distribution of an individual mirror may be compensated for, even though the exact individual amount of the tilt error of each micromirror is not known. As the light mixing elements have a constant diameter, a surface area of a light entrance face and a surface area of a light exit face are typically identical.

In one improvement, the light mixing element is one of a light mixing rod (optical integrator) and an optical fiber. The light mixing rods/optical fibers are typically arranged in an array, the array e.g. forming a plate-like structure, the

longitudinal axes of the light mixing rods/fibers extending perpendicular to the plate. A light mixing rod may be devised as a solid structure made of a material which has a high refractive index and thus provides multiple reflections due to total internal reflection. Alternatively, the rod may be a hollow structure having internal surfaces which are coated with a reflective material, also providing multiple reflections for the light which passes through the rod. An optical fiber (e.g. silica glass fiber) may also be used as a light mixing element for

homogenization of the light distribution being generated by a micromirror, having a cross section with a refractive index profile which allows to provide total internal reflection of a light bundle passing therethrough.

In some improvements, the (cylindrical) light mixing rod has a diameter of between 1 pm and 100 pm, more preferably of between 2 pm and 20 pm. The diameter of the light mixing rod is typically chosen in dependence of the diameter of a corresponding Airy disc (spot size) of a beam spot which is generated on the substrate by the exposure optical system. Typically, the diameter of the light mixing rod is chosen to be between one time and ten times of the diameter of the Airy disc (spot size) of a corresponding beam spot of the array (the spot size typically being identical for all beam spots in the array).

In one improvement, the light mixing rod has an aspect ratio of 25 or more, preferably of 50 or more, in particular of 100 or more. For the present purposes, a high aspect ratio of the light mixing rod has been proven to be of advantage. The aspect ratio of a cylindrical rod is defined as its length-to-diameter ratio, the length being measured in the longitudinal direction of the rod, the diameter being measured at an entrance surface / exit surface of the rod. One skilled in the art will appreciate that the aspect ratio of a rod having a non-circular cross- section, e.g. a rectangular cross section, may be defined in a similar way, the diameter being defined e.g. by a (diagonal) line which crosses the rectangular surface between opposite edges.

In some embodiments, the optical tilt error compensation unit comprises a plurality of light deflection elements, each for changing the direction of light of the exposure beam modulated by one of the micromirrors. With such an optical tilt error compensation unit, the tilt error of each micromirror may be

compensated for individually by changing the direction of light incident on the tilt error compensation unit by a defection angle which (in the ideal case) is identical to the tilt error in magnitude, but has the opposite sign. Unlike tilt error compensation by light mixing using optical integrators / light mixing elements, a tilt error compensation by deflection typically requires that an individual tilt error of a micromirror is known before manufacturing the optical tilt error

compensation unit, in order to produce the light deflection elements with the individual deflection angle which compensates for the tilt error. For determining the individual tilt errors, a DMD may be introduced into a maskless lithographic apparatus and errors, especially positional errors, in the aerial image of the exposure optical system may be determined for calculating the corresponding tilt errors. Of course, the tilt errors may also be measured by direct inspection of the DMD using appropriate measurement techniques, for instance, interferometric measurement techniques being referred to as "optical surface profiling" or "profilometers".

In one improvement, the optical tilt error compensation unit comprises a (three- dimensionally) structured surface having a plurality of wedge-shaped surface areas, each forming a light deflection element. The wedge-shaped surface areas are inclined to a propagation direction of the exposure beam by a deflection angle which is dependent on the individual tilt error. The slope of the wedge-shaped surface area is chosen so that the deflection angle has identical magnitude (absolute value) as the tilt error, but opposite sign, so that the tilt error may be compensated for. In the present example, the optical tilt error compensation unit is typically arranged close to the plane where the DMD is arranged, or close to a plane which is optically conjugate to the plane with the DMD.

In a further improvement, at least one of the light mixing rods and the wedge- shaped areas are devised as micro-structures being generated using

microlithography for patterning a surface of a substrate, and subsequent etching. For instance, hollow light mixing rods may be manufactured using microlithography for forming a two-dimensional pattern on a substrate such as a glass plate or a silicon wafer, and by selectively etching the substrate to form a plurality of holes passing therethrough. After generation of the holes, a suitable coating may be applied to the interior surfaces of the holes, using a material having a high refractive index, e.g. a metal, in order to enhance the reflectivity of the hollow light mixing rods. In a similar way, wedge-shaped tree-dimensional structures may be produced e.g. by direct-write / grey scale lithography, using an intensity-modulated laser to expose a relatively thick resist material on a substrate with a variable (position-dependent) dose, so that subsequent etching provides three-dimensional structures in the resist. In one embodiment, the tilt error compensation unit is adapted for correcting for positional errors of beam spots of a beam-spot array formed on the substrate by the modulated exposure beam, the positional errors being due to the tilt errors of the micromirrors. In this embodiment, the micromirror array is typically arranged in a pupil plane of the exposure optical system which is conjugate to the field plane where the substrate is arranged, so that tilt errors of the micromirrors result in positional errors on the substrate. The positional errors may be compensated by using an optical tilt error correction unit, or by using an appropriate tilt error compensation algorithm when generating the exposure pattern using the beam-spot array. One skilled in the art will appreciate that if the micromirror array is arranged in a field plane of the exposure optical system (instead of a pupil plane), tilt errors of the micromirrors will typically result in telecentricity errors on the substrate.

In a maskless lithographic apparatus, a scan direction of the substrate may be tilted at an alignment angle with respect to an arrangement direction in which the light modulator of the maskless lithographic apparatus is arranged. To be more precise, the direction of the rows of the beam-spot array is tilted at an alignment angle with respect to the scan direction of the substate. For correcting positional errors in such a system, positional errors in the scan direction typically have to be treated in a different way as positional errors in other directions, in particular in a direction perpendicular to the scan direction.

In some developments, the tilt error compensation unit is adapted to

compensate for the positional errors by at least one of: compensating for a positional error of at least one beam spot in a direction other than a scan direction of the substrate by modifying an assignment between the beam spots of the beam-spot array and specific pixel positions for generating the exposure pattern in the other direction, and compensating for a positional error of at least one beam spot in the scan direction by performing a temporal adjustment of the modulation of the exposure beam of a corresponding micromirror of the micromirror array used for producing the beam spot on the substrate.

Typically, each beam spot of the beam spot array is assigned to a predetermined pixel position in a direction perpendicular to the scan direction, the assignment of the beam spot to the pixel position being made based on the nominal position of the beam spot in the regular arrangement of rows and columns of the beam spot array. The pixel positions typically have an equal spacing in the direction perpendicular to the scan direction. However, if a positional error of a beam spot is larger than the spacing / pitch between adjacent rows of the beam spot array in the direction perpendicular to the scan direction, an assignment based on the nominal position of that beam spot in the regular arrangement of rows and columns is no longer appropriate. Thus, the assignment may be modified using the tilt error compensation unit in such a way that the beam spot whose position in the direction perpendicular to the scan direction is closest to a specific pixel position is assigned to that pixel position, changing the original assignment which is based on the nominal position of the beam spot in the beam spot array.

If the positional errors are such that no beam spot is located close to a specific pixel position, generating an exposure pattern at that specific pixel position is impossible. However, this situation may be avoided by choosing a specific value of the alignment angle (and a corresponding spacing between the rows and columns of the array) so that the pixel positions of two or more beam spots arranged in different rows of the beam spot array in the direction perpendicular to the scan direction are identical, providing a redundancy when

writing/generating the exposure pattern. Due to this redundancy, the case that no beam spot is arranged at a specific pixel position becomes highly

improbable. Positional errors of the beam spots in the scan direction may be compensated for by shifting the point of time when a respective micromirror is addressed (i.e. switched from the "on" state to the "off state or vice versa). Typically, a control unit of the apparatus addresses the micromirrors in a pre-determined way in order to produce the desired exposure pattern on the substrate (taking the scan speed of the substrate into account). Depending on the (positive or negative) deviation of the position of a respective beam spot in the scan direction, the switching between the "on" and "off states of the corresponding micromirror may either be performed earlier or later than at a nominal point of time which is used when the beam spot is at its nominal position in the beam spot array.

A further embodiment of the present invention provides a method for generating an exposure pattern on a substrate, comprising: modulating an exposure beam according to the exposure pattern using a plurality of micromirrors of a micromirror array, and delivering the modulated exposure beam onto a substrate in the form of a beam spot array, wherein at least one of the modulating step and the delivering step comprises: compensating for tilt errors of the micromirrors of the micromirror array.

As indicated above, tilt errors of the micromirrors (typically due to

manufacturing errors) may lead to errors when generating the exposure pattern on the substrate, i.e. the modulated exposure beam may not have the desired properties, leading to unwanted changes of the position of the beam spots on the substrate and/or to imaging errors, such as telecentricity errors, etc.

In some variants, the tilt error compensating step comprises: compensating for the tilt errors of the micromirrors by at least one of mixing of light and deflecting of light of the exposure beam modulated by one of the micromirrors. In this way, an individual optical compensation of the tilt errors of the micromirrors may be provided. In some variants, the tilt error compensating step comprises: compensating for a positional error of at least one beam spot of the beam spot array on the substrate, the positional error of the beam spot being due to the tilt error of a corresponding micromirror. As indicated above, when the micromirror array is arranged in a pupil plane of the exposure optical system, tilt errors of the micromirrors will cause positional errors of the beam spots on the substrate.

In an improvement, the step of compensating for the positional error comprises at least one of: compensating for a positional error of at least one beam spot in a direction other than the scan direction by modifying an assignment between the beam spots of the beam spot array and pixel positions for generating the exposure pattern along the other direction, and compensating for a positional error of at least one beam spot in the scan direction by performing a temporal adjustment of the modulation of the exposure beam by a corresponding micromirror of the micromirror array, i.e. by a micromirror which produces the at least one beam spot (in its "on" state).

In this case, the compensation is typically performed by suitably modifying an algorithm for generating the desired exposure pattern by suitably addressing of the micromirror array for providing the desired temporal and spatial modulation of the exposure beam. Such an algorithm is typically implemented in a control unit for controlling the apparatus. One skilled in the art will appreciate that the algorithm, respectively, the control unit may implemented as a dedicated hardware, or as a hardware capable of executing software in association with appropriate software.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. All publications, patent applications, patents, and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control. In addition, the materials, methods, and examples are illustrative only and not intended to be limiting.

Other features and advantages of the invention will be apparent from the following detailed description, and from the claims.

Brief Description of the Drawings

Exemplary embodiments are shown in the diagrammatic drawing and are explained in the description below. The following are shown:

Fig 1 a a cross-sectional view of a conceptual diagram of a maskless lithographic apparatus having a Digital-Micromirror-Device (DMD) as a spatial and temporal light modulator,

Fig. 1 b a cross-sectional view of a conceptual diagram of a maskless lithography apparatus having an optical tilt error compensation unit in the form of a tilt compensation plate,

Fig 2a, b schematic diagrams of a positional error (Fig. 2a) and of a

telecentricity error (Fig. 2b) on a substrate due to a tilt error of a micromirror of the DMD,

Fig 3 an example of an optical tilt error compensation unit having a plurality of light mixing rods, each using multiple reflections for performing the light mixing;

Fig 4 light bundles passing one of the light mixing rods along a

longitudinal direction, resp., at an angle of 1° with respect to the longitudinal direction; Fig 5a-d a spatial and angular light distribution at an entrance surface of a light mixing rod (Fig, 5a) and at an exit surface thereof for three different lengths (Fig. 5b-d);

Fig. 6a-c three conceptual diagrams of a maskless lithography apparatus with a plate-like tilt error compensation unit having a micro- structured surface,

Fig 7 a plate-like tilt error correction unit having a micro-structured

surface comprising a plurality of wedge-shaped areas, each for changing a direction of light modulated by one of the micromirrors; and

Fig 8a,b a beam-spot array formed on a substrate by the modulated

exposure beam with a regular arrangement of beam spots (Fig. 7a) and with beam spots having positional errors due to tilt errors of the micromirrors (Fig. 7b).

Detailed Description of the Invention

Fig 1a shows a maskless lithographic apparatus 1 for generating an exposure pattern on a substrate 2 which is arranged on a stage 3 for moving the substrate 2. The apparatus 1 has a light source 4 providing an exposure beam 5 which is made uniform in an optical illumination system 6. The light source 4 may e.g. be devised as a semiconductor laser or an UV lamp. The optica! illumination system 6 may also include a wavelength filter for providing the exposure beam 5 at a desired wavelength.

The apparatus 1 further comprises a light modulator 7 for modulating the exposure beam 5 which has passed through the illumination system 6. The light modulator 7 modulates the exposure beam 5 in accordance with an exposure pattern to be provided on the substrate 2. The light modulator 7 is a spatial (and temporal) light modulator which comprises a plurality of micromirrors arranged in a micromirror array (not shown in Fig. 1 ). In the following, it is assumed that the light modulator 7 is devised as a Digital-Micromirror-Device (DMD), being a specific type of a Micro Electro Mechanical System (MEMS). The DMD 7 may include a substrate, memory cells formed on the substrate, and the plurality of micromirrors may be arranged in a matrix on the memory cells.

The exposure beam 5 which is modulated by the DMD 7 is delivered onto the substrate 2 via an exposure optical system 8 having an imaging optical system 9 and a microlens array 10 being arranged along a path in which the exposure beam 5 passes.

In the present example, the imaging optical system 9 is a magnifying projection lens (e.g. with a magnification of 1 : 3 or 1 : 4) that forms an image of the DMD 7 arranged in an object plane of the imaging optical system 9 on an image plane where the micro-lens array 10 is arranged. Each of the microlenses of the microlens array 10 focusses the modulated exposure beam 5, more specifically a portion of the modulated exposure beam 5 originating from a specific one of the micromirrors of the DMD 7 on a focus in a focal plane in which the substrate 2 is arranged. In this way, the modulated exposure beam 5 which passes the microlens array 10 generates a beam-spot array (not shown in Fig. 1) on the substrate 2 which is coated with a photosensitive material. One skilled in the art will appreciate that a further imaging optical system (not shown) which may be devised as a reduction lens (having a demagnification of e.g. about 5 : 1 ) may be used for generating a (demagnified) image of the plane of the microlens array 10 on the substrate 2, resulting in a beam-spot array with a reduced size.

The pattern generation process is coordinated by a control unit 13 which is adapted to control the DMD 7, more specifically, an on/off state of the micromirrors of the DMD 7, the stage 3 for moving the substrate 2, and possibly also the microlens array 10. In particular, the control unit 13 provides (digital) control signals to the DMD 7 for providing a desired spatial and temporal modulation of the exposure beam 5, as will be explained in greater detail with reference to Fig. 2a,b, showing two micromirrors 7a, 7b arranged in a row of a micromirror array 14 of the DMD 7, further rows not being shown for the sake of simplicity.

Each of the micromirrors 7a, 7b of the (two-dimensional) array 14 is supported by a hinge support (not shown) and may be tilted in a range of angles from - a to +a, e.g. from - 12° to + 12° by application of a (digital) signal from the control unit 13 to a memory cell of the DMD 7. In the present example, in a first state ("on" state, cf. micromirror 7a) which corresponds to a tilt angle of a micromirror of +12°, the micromirror directs light toward the exposure optical system 8. In a second state ("off' state, see micromirror 7b) which corresponds to a tilt angle of -12°, a respective micromirror reflects light towards a light absorber (not shown). Thus, by application of appropriate digital signals to the rows and columns of the DMD 7, a desired spatial (and temporal) modulation of the exposure beam 5 may be generated, resulting in a corresponding two- dimensional pattern of beam spots in a beam spot array produced on the substrate 2.

Yet, due to manufacturing errors of the DMD 7, the actual value of the tilt angle of each of the micromirrors 7a may deviate from the nominal value of +ct (e.g. + 12°), so that the direction of the reflective surface of the micromirror 7a in its "on" state is not parallel to the plane of the substrate 2, but deviates by an angle which is referred to as tilt error β in the following. The tilt error β

represents the difference between the nominal value of the tilt angle (which is identical for all micromirrors) and the actual value of the tilt angle which varies among the micromirrors of the DMD 7 due to manufacturing imprecisions, typically in a range from - 1 ° to + 1 °, so that the actual value of the tilt angle in the "on state" ranges from -a - 1 ° to +a + 1 °. As the DMD 7 is typically produced in a lithographic process, the micromirrors of DMDs which are produced in the same lithographic process (belonging to the same chip) typically have the same tilt errors i.e. for the DMDs belonging to the same chip, the tilt error β at a given position in the micromirror array is typically the same.

The errors in the image plane of the exposure optical system 8 (i.e. on the substrate 2) which result from a deviation (tilt error β) of 1° from the nominal tilt angle in the "on" state of the first micromirror 7a depend on the design of the exposure optical system 8, as can be gathered from Fig 2a,b: In Fig 2a, the DMD 7 is arranged in a pupil plane of the exposure optical system 8, the light distribution in the pupil plane being related by a Fourier transform to the image plane in which the substrate 2 is arranged. In the example of Fig. 2a, a tilt error β of the first micromirror 7a of the DMD 7 leads to a positional error δ of a beam spot which is generated on the substrate 2. In contrast thereto, as exemplified in Fig 2b, when the DMD 7 is arranged in a field plane of an exposure optical system which is conjugate to the image plane where the substrate 2 is arranged, the tilt error β of the first micromirror 7a will cause a deviation / tilt of the direction of the chief ray of the light bundle which impinges on the substrate 2 in the image plane, causing a telecentricity error in the image plane.

One way to compensate for the tilt errors of the micromirrors 7a, 7b which is due to the manufacturing tolerances of the DMD 7 is to make the light distribution of the modulated exposure beam 5 provided by each of the micromirrors 7a, 7b more uniform, resp. more homogeneous. This can be done e.g. by providing an optical tilt error compensation unit which is devised in the form of a plate 20 comprising a plurality of light mixing elements 21 in the form of light mixing rods, as shown in Fig 3. The light mixing rods 21 are arranged in an array, the array typically having the same number of rows and columns as the micromirror array 14. The tilt error compensation plate 20 is arranged in the beam path of the modulated exposure beam 5 at a position which facilitates focussing light from a respective micromirror 7a, 7b of the DMD 7 on an entrance surface of a corresponding light mixing rod 21. For instance, as shown in Fig. 1b, the tilt error compensation plate 20 may be arranged between a first microlens array 10 for focussing the light from the respective mircromirrors 7a, 7b and a second microlens array 10a for collimating the homogenized light distribution which exits the light mixing rod 21 and for focusing the

homogenized light of the exposure beam 5 on the substrate 2.

The light mixing rod 21 (also referred to as optical integrator) may be devised as a solid mixing rod, using the total reflection at the lateral interfaces to the surrounding medium (air) for reflecting the light which passes the rod 21 back and forth several times. Alternatively, the light mixing rod 21 may be devised as a hollow light mixing rod, which essentially consists of a hole in the plate-like tilt error correction unit / element 20, the lateral surfaces thereof being provided with highly reflecting coatings. Thus, multiple (Fresnel) reflections are produced inside the light mixing rod 21 , leading to a homogenization both of the angular and spatial distribution of the light passing through the rod 21. The light mixing rod 21 may be devised as a parallelepiped with a light entrance surface and a light exit surface having e.g. a rectangular geometry, or may have a cylindrical geometry with light entrance and exit surfaces having a circular cross-section. One skilled in the art will appreciate that instead of an optical integrator in the , form of a light mixing rod, other types of light mixing elements may be used as well for the present purposes, for instance optical fibers which provide multiple reflections due to total reflection at an interface between materials having different refractive indices, e.g. between a core and a cladding or an ambient medium, typically air.

In the present example, the plate-like tilt error compensation unit 20 has a plurality of hollow light mixing rods 21 (in a matrix arrangement) which are produced using microlithography for forming a two-dimensional pattern on a substrate such as a glass plate or a silicon wafer, and by selectively etching the substrate to form a plurality of holes passing through the plate, the distance (pitch) between adjacent holes being selected in dependence of the distance between adjacent micromirrors of the micromirror array. The holes forming the light mixing rods 21 are typically coated at the interior surfaces thereof using a material having a high refractive index, e.g. a metal, in order to enhance their reflectivity for the light passing through the holes.

Light which enters the light mixing rod 21 at an angle relative to a longitudinal direction of the rod 21 , e.g. at an angle β of 1° thereto (cf. Fig 4), is

homogenized due to the multiple reflections inside the light mixing rod 21. It will be understood that the degree of homogenization of the light which enters the light mixing rod 21 depends on the parameters of the light mixing rod 21 , especially on its length L along the longitudinal direction, as well as on the size of its cross-section. In the following, it is assumed that the light mixing rod 21 has a circular cross section with a diameter D.

Fig 5a shows an example of a spatial light distribution (top) and an angular light distribution (bottom) at the entrance surface of the light mixing rod 21. Fig 5b to Fig 5d show the corresponding light distributions at the exit surface of the light mixing rod 21 for three different lengths L of 0.1 mm, 0.5 mm, and 1.0 mm, respectively. As can be gathered from Figs. 5b-d, the uniformity of the light distribution at the light exit surface of the light mixing rod 21 increases with increasing length of the light mixing rod 21 due to the increased number of reflections provided therein.

The diameter D of the cross section of the light mixing rod 21 is typically chosen in a range from 1 micron (μιτι) to 100 microns, in particular from 2 microns to 20 microns. The diameter D of the light mixing rod 21 may be selected based on the diameter of an Airy disc of a corresponding beam spot on the substrate 2, the Airy disc corresponding to a center region of a diffraction pattern produced by a uniformly illuminated circular aperture, the diameter of the Airy disc being given by D(Airy) = 0.5 x λ / NA (1 ) wherein λ designates the wavelength of the exposure beam 5 produced by the light source 4 (being for instance about λ = 400 nm), and NA designates the numerical aperture when focussing a light bundle on the entrance surface of the light mixing rod 21. The numerical aperture NA is defined as NA = n x sin (φ), wherein n is the refractive index of the ambient medium (air having n = 1.0) and φ designates the half-angle of the light bundle focused on the entrance surface of the light mixing rod 21 (cf. Fig 3). The factor of one half in equation (1) is due to the fact that the diameter of the Airy disc is defined as a "full width half maximum", FWHM, value.

Typically, the diameter D of the light mixing rod 21 is chosen to be: D(rod) = X * D(Airy), with X being chosen in a range of values from one to ten (1 < X < 10). The reason for choosing this range is as follows: If the diameter D(rod) of the light mixing rod 21 is smaller than the diameter D(Airy) of the Airy disc, the light mixing rod 21 will cut off parts of the incident light bundle, leading to a reduced light intensity at the exit surface of the light mixing rod 21. Also the diameter D(rod) of the light mixing rod 21 should typically not be larger than ten times the diameter D(Airy) of the Airy disc, as the etendue of the light bundle leaving the light mixing rod 21 increases with increasing rod diameter D(rod). However, the etendue should be made as small as possible in order that the diameter of a beam spot on the substrate 2 which is generated by a respective microlens of the microlens array 10 approximately corresponds to the diameter D(Airy) of the Airy disc. Moreover, for a given diameter D(rod) of the light mixing rod 21 , a length-to- diameter ratio (aspect ratio) A which allows to produce a desired number of reflections should be chosen, the length L of the light mixing rod 21 depending on the number of reflections (referred to as #R in the following) by the following equation:

L = #R * D(rod) / NA = #R * X * 0.5 * λ / NA² , (2) the aspect ratio A = L / D being given by #R / NA. It will be understood that eq. (2) holds only for a hollow light mixing rod 21 and that for a solid light mixing rod the refractive index n of the material constituting the rod has to be taken into account.

Appropriate exemplary values of the aspect ratio A and length L of (hollow) light mixing rods are summarized in the following table (table 1):

Number of Reflections 5 5 5

Lambda [nm] 0.4 0.4 0.4

NA 0.05 0.1 0.2

X = D(rod) / D(Airy) 3 3 3

Rod diameter [microns] 12 6 3

Rod length [microns] 1200 300 75

Aspect ratio A 100 50 25 table 1

As can be gathered from table 1 , the aspect ratio A of the light mixing rod 21 should typically be at least 25 or more, preferably 50 or more, in particular 100 or more. Another possibility to compensate for tilt errors of the micromirrors 7a, 7b of the DMD 7 by using a different plate-like tilt error compensation unit 30 will be explained in the following with reference to Fig 6a-c and Fig. 7. The tilt error compensation unit 30 comprises a plurality of light deflection elements 31 , each for changing a direction of the exposure beam 5 modulated by a specific one of the micromirrors 7a to 7e. In the example of Fig 7, the plurality of light deflection elements 31 are devised as wedge-shaped areas of the plate 30, wherein each wedge-shaped area 31 is inclined to a plane parallel to the substrate (and perpendicular to a nominal direction of the beam path) by a deflection angle γ. The deflection angle γ is chosen, individually, i.e. varies among the wedge-shaped areas 31 and is chosen for compensating a corresponding tilt error β of a corresponding one of the micromirrors 7a-e, being shown in their "on" state in Fig. 7. In particular, when no tilt error occurs, as is the case with the second micromirror 7b represented in Fig. 7, no tilt error compensation is required, and a deflection angle γ = 0° is chosen for the corresponding sub-area 31 of the surface 30a of the tilt error compensation plate 30.

One skilled in the art will appreciate that for manufacturing the tilt error compensation unit 30 shown in Fig. 7, the individual tilt error β of each micromirror 7a-d of the micromirror array 14 has to be determined beforehand. Such a determination may be done by an appropriate measurement, e.g. by introducing the DMD 7 into an apparatus 1 of the kind shown in Fig. 1 and observing individual position errors of the beam spots of an aerial image in the image plane of the exposure optical system 8. By determining the magnitude (and angle) of the positional error of a respective beam spot, a tilt error β of a corresponding micromirror 7a-e may be calculated and individual deflection angle γ for a corresponding wedge-shaped area 31 of the tilt error

compensation plate 30 may be determined. The tilt error compensation plate 30 may be structured using microlithography, more specifically, grey scale lithography, in order to generate the three- dimensional, wedge-shaped areas 31 at the surface 30a thereof. Such three- dimensional structures may be generated e.g. by a direct-write lithography system, using an intensity-modulated laser to produce three-dimensional microstructures on a photoresist layer arranged on a transparent substrate, e.g. a glass plate.

The tilt error compensation plate 30 may be arranged at a position in the beam path of the exposure beam 5 which is preferably close to a plane which is optically conjugate to the plane where the DMD 7 is arranged, for instance close to the micromirror array 10, as shown in Fig. 6a. As shown in Fig. 6b, the tilt error compensation plate 30 may also be arranged at or close to the plane where the DMD 7 is arranged. In this case, the exposure beam 5 from the illumination system 6 may pass the plate 30 a first time, and the modulated exposure beam 5 deflected from the DMD 7 may pass the plate 30 a second time. Of course, the fact that the light passes the plate 30 two times has to be taken into account for determining the deflection angle γ required to

compensate for the tilt error of a specific one of the micromirrors 7a-7e.

As shown in Fig. 6c, the tilt error compensation plate 30 may also be arranged in the beam path of the unmodulated exposure beam 5 before the DMD 7. In this case, it is advantageous to arrange the error compensation plate 30 in a plane parallel to the plane where the DMD 7 is arranged. A further projection lens 12 may be used for imaging the plane with the error compensation plate 30 to the plane with the DMD 7, making both planes optically conjugate. in any case, as the tilt error is compensated for individually for each of the micromirrors 7a-e, a specific tilt error compensation plate 30 will be required for each DMD 7 (or for a group of DMDs produced in the same microlithography process). Alternatively or in addition to a tilt error compensation based on optical compensation structures or elements, a compensation of positional errors in the image plane of the exposure optical system 8 which are due to the tilt errors of the micromirrors 7a-e of the micromirror array 14 may be performed. For this purpose, the control unit 13 of Fig. 1a (of Fig. 1b or of Fig. 6a-c) may comprise a tilt error compensation unit 15 which is adapted to correct for positional errors δ of the beam spots P1 to P5 of a beam spot array 40 which is formed on the substrate 2 by the modulated exposure beam 5, see Fig 8a,b. In the present example, it is assumed that a scan direction (y direction) of the substrate 2, i.e. a direction in which the substrate 2 is moved by the stage 3, is titled at a so- called alignment angle Θ with respect to the beam spot array 40, more precisely with the rows of the beam spot array 40 (and thus also with respect to the micromirror array 14 of the DMD 7).

Fig 8a shows a beam spot array 40 on the substrate 2 in which each of the beam spots P1 to P5 is located at its nominal position in the array 40. Such a situation may occur, e.g., when using one of the optical tilt error compensation units 20, 30 described above. Fig. 8b shows a beam-spot array 40 in which some of the beam spots 41 have a position which deviates from their nominal (regular) position in the array by an (individual) positional error δ, being a vector (having both a magnitude and a direction). In order to compensate for the (individual) positional errors δ of the beam spots P1 to P5, the vector

components of the positional error δ in the scan direction y and in a direction perpendicular thereto (x direction) may be compensated for in a different way, as explained further below.

As can be gathered from Fig 8a, b, a projection of the beam spot array 40 in a direction x perpendicular to the scan direction y produces a scanning line which is represented at the bottom of Fig 8a, b. For generating a desired pattern / structure on the substrate 2, each of the beam spots P1 to P5 is assigned to a specific (pixel) position along the scanning line in the x direction. For producing a desired exposure pattern, a specific position in the x direction of the substrate 2 is addressed by the control unit 13 by switching a specific one of the micromirrors 7a-e and thus a corresponding beam spot P1 to P5 from its "on" state to its "off state or vice versa.

The beam spots P1 to P5 of each row are typically assigned to the pixel positions in the x direction in a natural way, i.e. the first beam spot P1 in a row corresponds / is mapped to the first position in the x direction, the second beam spot P2 is attributed to the second position in the x direction, and so on.

However, although such an assignment of beam spots P1 to P5 to pixel positions in the x direction is desirable when no positional errors are present and the beam spots P1 to P5 are in their nominal position in the array 40 (cf. Fig. 8a), such an attribution may not be desirable as soon as positional errors δ occur which are large enough so that the order of the projections of the beam spots P1 to P5 on the scan line along the x direction does not correspond to the nominal order (cf. Fig. 8b).

In this case, the control unit 13, more precisely, the tilt error compensation unit 15, may change the attribution, resp., the mapping of the beam spots P1 to P5 to the pixel positions along the x direction in order to re-establish the correct attribution of beam spots P1 to P5 to the pixel positions in the x direction. For instance, in the example shown in Fig. 8b, the attribution of the second to fourth beam spots P2 to P4 to the pixel positions in the x direction is changed in accordance with the respective positional errors δ.

It will be understood that a case may occur in which the (systematic) positional errors δ of the respective beam spots P1 to P5 are such that none of the beam spots P1 to P5 has a projection along the x direction which comes close to a specific pixel position, so that no exposure pattern may be generated at that specific pixel position. However, in contrast to the simple example shown in Fig 8a,b, the DMD 7 and thus the beam spot array 40 typically has a considerably larger number of rows and columns, and the alignment angle Θ may be chosen such that a nominal position of more than one beam spot (e.g. five beam spots) may correspond to a given position in the x direction (as explained e.g. in the article by Kin Foog Chan et al., cited above), such the occurrence of a case in which no pattern may be generated at specific pixel positions in the x direction becomes highly unlikely.

In order to compensate for positional errors δ of the beam spots P1 to P5 in the scan direction (y direction), the tilt error compensation unit 15 is adapted to make a temporal adjustment of the modulation of the exposure beam 5, i.e. the time instant for providing a control signal for switching of a respective

micromirror 7a-e between its "on" state and its "off state is shifted in time so that the switching occurs at the nominal position of the respective beam spot P1 to P5 in the array 40 and consequently, the positional error δ in the scan direction y is compensated for. In this way, the desired exposure pattern may be produced without deviations occurring in the scan direction y. The tilt error compensation by such a temporal adjustment is only limited by the maximum switching frequency of the micromirrors 7a-e between the "on" state and the "off state, and by the scan speed in the scan direction y.

While example embodiments have been particularly shown and described, it will be understood by those of ordinary skill in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the present invention, as defined by the following claims and their equivalents.

Claims

1. A maskless lithographic apparatus (1), comprising:

a light modulator (7) having a plurality of micromirrors (7a-e) arranged in a micromirror array (14) for modulating an exposure beam (5) according to an exposure pattern, and

an exposure optical system (8) for delivering the modulated exposure beam (5) onto a substrate (2),

characterized by

a tilt error compensation unit (15, 20, 30) for compensating for tilt errors (β) of the micromirrors (7a-e) of the micromirror array (14).

2. The maskless lithographic apparatus of claim 1 , wherein the light modulator is a Digital-Micromirror-Device (7).

3. The maskless lithographic apparatus of claim 1 or 2, wherein the tilt error compensation unit is an optical tilt error compensation unit (20, 30).

4. The maskless lithographic apparatus of claim 3, wherein the optical tilt error compensation unit (20) comprises a plurality of light mixing elements (21 ), each for mixing light of the exposure beam (5) modulated by one of the micromirrors (7a-e), the light mixing elements (21 ) having constant diameter.

5. The maskless lithographic apparatus of claim 4, wherein the light mixing element is one of a light mixing rod (21 ) and an optical fiber.

6. The maskless lithographic apparatus of claim 5, wherein the light mixing rod (21 ) has a diameter (D) of between 1 μιη and 100 μητι.

7. The maskless lithographic apparatus of claim 5 or 6, wherein the light mixing rod (21) has an aspect ratio (L / D) of 25 or more.

8. The maskless lithographic apparatus of claim 3, wherein the optical tilt error compensation unit (30) comprises a plurality of light deflection elements (31 ), each for changing a direction of light of the exposure beam (5) modulated by one of the micromirrors (7a-e).

9. The maskless lithographic apparatus of claim 8, wherein the optical tilt error compensation unit (30) comprises a structured surface (30a) having wedge- shaped areas (31) forming the light deflection elements.

10. The maskless lithographic apparatus of any one claims 5 to 7 or 9, wherein at least one of the light mixing rods (21 ) and the wedge-shaped areas (31) of the tilt error compensation unit (20, 30) are micro-structures generated in a microlithography process.

11.The maskless lithographic apparatus of anyone of the preceding claims, wherein the tilt error compensation unit (15) is adapted for correcting positional errors (δ) of beam spots (P1 to P5) of a beam-spot array (40) formed on the substrate (2) by the modulated exposure beam (5), the positional errors (δ) being due to the tilt errors (β) of the micromirrors (7a-e).

12. The maskless lithographic apparatus of claim 11 , wherein the tilt error

compensation unit (15) is adapted to compensate for the positional errors (δ) by at least one of:

compensating for a positional error (δ) of at least one beam spot (P1 to P5) in a direction (x) other than a scan direction (y) of the substrate (2) by modifying an assignment between the beam spots (P1 to P5) of the beam- spot array (40) and pixel positions for generating the exposure pattern in the other direction (x), and compensating for a positional error (δ) of at least one beam spot (P1 to P5) in the scan direction (y) by performing a temporal adjustment of the modulation of the exposure beam (5) by a corresponding micromirror (7a-e) of the micromirror array (14) which corresponds to the at least one beam spot (P1 to P5).

13. A method for generating an exposure pattern on a substrate (2), comprising: modulating an exposure beam (5) according to the exposure pattern using a plurality of micromirrors (7a-e) of a micromirror array (14), and

delivering the modulated exposure beam (5) onto a substrate (2) in the form of a beam-spot array (40),

characterized in that

at least one of the modulating step and the delivering step comprises:

compensating for tilt errors (β) of the micromirrors (7a-e) of the micromirror array (14).

14. The method of claim 13, wherein the tilt error compensating step comprises: compensating for the tilt errors (β) of the micromirrors (7a-e) by at least one of mixing of light and deflecting of light of the exposure beam (5) modulated by one of the micromirrors (7a-e).

15. The method of any one of claims 13 or 4, wherein the tilt error

compensating step comprises:

compensating for a positional error (δ) of at least one beam spot (P1 to P5) of the beam-spot array (40) on the substrate (2), the positional error (δ) of the beam spot (P1 to P5) being due to the tilt error (β) of a corresponding micromirror (7a-e).

16. The method of claim 15, wherein the step of compensating the positional error (δ) comprises at least one of:

compensating for a positional error (δ) of at least one beam spot (P1 to P5) in a direction (x) other than the scan direction (y) by modifying an

assignment between the beam spots (P1 to P5) of the beam-spot array (40) and pixel positions for generating the exposure pattern in the other direction (x), and

compensating for a positional error (δ) of at least one beam spot (P1 to P5) in the scan direction (y) by performing a temporal adjustment of the modulation of the exposure beam (5) by a corresponding micromirror (7a-e) of the micromirror array (14).