Illumination system with a plurality of light sources

[0001] The present application is a continuation of U.S. patent application Ser. No. 09/627,559 filed on Jul. 27, 2000, now U.S. Pat No. 6,570,168, and a continuation in part application of U.S. patent application Ser. No. 10/201,652 filed on Jul. 22, 2002.

[0002] The Ser. No. 10/201,652 application is (a) a continuation-in part of U.S. patent application Ser. No. 10/150,650, and (b) a continuation-in part of U.S. patent application Ser. No. 09/679,718. The Ser. No. 10/150,650 application was filed May 17, 2002, and is a continuation-in-part of the Ser. No. 09/679,718 application. The Ser. No. 09/679,718 application was filed Sep. 29, 2000, issued as U.S. Pat. No. 6,438,199, and is a continuation-in-part of U.S. patent application Ser. No. 09/305,017. The Ser. No. 09/305,017 application was filed May 4, 1999, and issued as U.S. Pat. No. 6,198,793. The present application is also claiming priority of (a) International Application No. PCT/EP00/07258, filed Jul. 28, 2000, (b) German Patent Application No. 299 02 108.4, filed Feb. 8, 1999, (c) German Patent Application No. 199 03 807.4, filed Feb. 2, 1999, and (d) German Patent Application No. 198 19 898.1, filed on May 5, 1998, and (e) German Patent Application No. 199 35 404.9, filed on Jul. 30, 1999.

[0003] 1. Field of the Invention

[0004] The invention concerns an illumination system for wavelengths ≦193 nm, i.e., VUV and EUV-lithography with a plurality of light sources, for example, as well as a mirror or lens device for producing secondary light sources, comprising several mirrors or lenses, divided into raster elements.

[0005] 2. Description of the Prior Art

[0006] In order to allow even further reduction in the structural width of electronic components, especially to the submicron range, it is necessary to reduce the wavelength of the light used in microlithography.

[0007] For wavelengths smaller than 193 nm, lithography with weak x-rays or so-called EUV-lithography is discussed.

[0008] A suitable illumination system for EUV-lithography should homogeneously or uniformly illuminate, with as few reflections as possible, a pregiven field for EUV-lithography, especially the annular field of an objective lens, under lithography requirements. Furthermore the pupil of the objective lens should be illuminated up to a particular degree of filling σ, independently of the field, and the exit pupil of the illumination system should be situated in the entrance pupil of the objective lens.

[0009] Regarding the basic layout of EUV-illumination systems, we refer to the applicant's pending applications EP 99 106348.8, submitted on Mar. 2, 1999, entitled “Illumination system, especially for EUV-lithography”, U.S. Ser. No. 09/305,017, submitted on May 4, 1999 entitled “Illumination system particularly for EUV-lithography”, and PCT/EP 99/02999, submitted on May 4, 1999, entitled “Illumination system, especially for EUV-lithography”, whose disclosure contents are incorporated in their entirety in the present application.

[0010] The following are discussed herein as light sources for EUV-illumination systems:

[0011] laser plasma sources

[0012] pinch plasma sources

[0013] synchrotron radiation sources

[0014] Since light is emitted from these light sources, they are also examples of primary light sources.

[0015] In the case of laser plasma sources, an intensive laser beam is focused onto a target (solid, gas jet, droplet). The target is heated so strongly by the excitation that a plasma is formed. This emits EUV-radiation.

[0016] Typical laser plasma sources have a spherical beam, i.e., a radiation angle of 4 π, as well as a diameter of 50 μm to 200 μm.

[0017] In pinch plasma sources, the plasma is produced by means of electrical excitation.

[0018] Pinch plasma sources can be described as volume radiators (D=1.00 mm), which emit in 4 π, whereby the beam characteristic is dictated by the source geometry.

[0019] In the case of synchrotron radiation sources, one can distinguish three types of sources at present:

• • [0020] bending magnets
  • [0021] wigglers
  • [0022] undulators

[0023] In bending magnet sources, the electrons are deflected by a bending magnet and emit photon radiation.

[0024] Wiggler sources comprise a so-called wiggler for deflection of the electron or an electron beam, and this wiggler comprises a multiple number of alternating polarized pairs of magnets arranged in rows. If an electron passes through a wiggler, it is subjected to a periodic, vertical magnetic field and the electron oscillates in the horizontal plane. Wigglers are also characterized by the fact that no coherency effects occur. The synchrotron radiation produced by a wiggler is similar to a bending magnet and radiates in a horizontal solid angle. In contrast to the bending magnet, it has a flux that is intensified by the number of poles of the wiggler.

[0025] There is no clear dividing line between wiggler sources and undulator sources.

[0026] In case of undulator sources, the electrons in the undulator are subjected to a magnetic field of shorter period and smaller magnetic field of the deflection poles than in the case of a wiggler, so that interference effects occur in the synchrotron radiation. The synchrotron radiation has a discontinuous spectrum based on the interference effects and emits both horizontally as well as vertically in a small solid-angle element; i.e., the radiation is highly directional.

[0027] It is critical for an EUV-illumination system to provide a sufficiently high Lagrange optical invariant or etendu. The Lagrange optical invariant of a system is defined as the product of the illuminated surface times the square of the aperture.

[0028] If the numerical aperture in the plane of the wafer is in the range NA_wafer=0.1-0.25, then in the case of 4:1 systems, a numerical aperture in the reticle plane of NA_reticle=0.025-0.0625 is needed. If the illumination system is supposed to illuminate this aperture homogeneously and independent from the field up to a filling degree of σ=0.6, for example, the EUV-source must have the following 2-dim Lagrange optical invariant or etendu: (LC).
LC_ill.=σ²LC_Obj=0.149 mm²−0.928 mm²

[0029] The Lagrange optical invariant LC, is generally defined as follows for the lithography system described herein:

[0030] LC_ill.=σ²x·y·NA²=σ² A·NA², wherein A is the illuminated area. The area comprises 110 mm×6 mm, for example, in the reticle plane.

[0031] The Etendu of a laser plasma source is defined as the product of the illuminated surface of an imaginary unit sphere around the source and the square of the Numerical Aperture at which each field point of the imaginary unit source sees the spherical source.
LC=A·NA²
A^LPQ=2π[cos(θ₁)−cos(θ₂)]×(R_sphere)², with R_sphere=1 mm
NA≈r^LPQ/R_sphere=0.100
where θ₁ is the minimum beam angle with respect to the optical axis and θ₂ is the maximum beam angle with respect to the optical axis
LC_LPQ=2π[cos(θ₁)−cos(θ₂)]·r²_LPQ

[0032] With the typical source parameters:

• [0033] 1. r_LPQ=0.1 mm, θ₁=0°, θ₂=90° yields: LC_LPQ=0.063 mm². This corresponds to 27% of the required value of the Lagrange optical invariant LC_ill of, for example, 0.236 mm².
• [0034] 2. r_LPQ=0.025 mm, θ₁0°, θ₂=90° yields: LC_LPQ=0.0039 mm². This corresponds to 1.7% of the required value of the Lagrange optical invariant of, for example, LC_ill=0.236 mm².

[0035] The Lagrange optical invariant LC_Pinch of a pinch plasma source with a diameter of 1 mm, Ω=0.3 sr, for example, is:

 LC_Pinch=A·NA²=(π·1 mm²/4)·0.3053²=0.073 mm².

[0036] Thus, the pinch plasma source provides 31% of the required value of the Lagrange optical invariant of, for example, LC_ill=0.236 mm².

[0037] The Lagrange optical invariant or Etendu for the undulator source can be estimated by a simplified model assuming a homogeneous two-dimensional radiator with diameter $\begin{array}{c}\varnothing =1.0\mathrm{mm}\mathrm{and}\mathrm{aperture}{\mathrm{NA}}_{\mathrm{Und}}=0.001\mathrm{with}\\ {\mathrm{LC}}_{\mathrm{Und}}=A·{\mathrm{NA}}^2\\ \begin{array}{c}{A}_{\mathrm{Und}}=\pi ·{\left(\varnothing /2\right)}^2\\ =0.785{\mathrm{mm}}^2\end{array}\\ {\mathrm{NA}}_{\mathrm{und}}=0.001\\ \mathrm{as}\\ {\mathrm{LC}}_{\mathrm{Und}}=A·{\mathrm{NA}}^2=0.00000079{\mathrm{mm}}^2=7.9e-07{\mathrm{mm}}^2\end{array}.$

[0038] As can be seen from this rough calculation the Etendu or Lagrange optical invariant of the undulator source is much too small in comparison to the required value of the Lagrange optical invariant.

[0039] To increase the Lagrange optical invariant, an illumination system comprising a synchrotron radiation source known from U.S. Pat. No. 5,512,759, comprises a condenser system with a plurality of collecting mirrors, which collect the radiation emitted by the synchrotron radiation source and form it to an annular light beam that corresponds to the annular field being illuminated. By this, the annular field is illuminated very uniformly. The synchrotron radiation source has a beam divergence >100 mrad in the plane of radiation.

[0040] U.S. Pat. No. 5,439,781 shows an illumination system with a synchrotron radiation source, in which the Lagrange optical invariant, is adjusted by means of a scattering plate in the entrance pupil of the objective lens, wherein the scattering plate can comprise a plurality of pyramidal structures. Also, in U.S. Pat. No. 5,439,781, the synchrotron radiation source has a beam divergence >100 mrad. The synchrotron radiation according to U.S. Pat. No. 5,439,781 is also focused, for example, by means of a collector mirror.

[0041] The disclosure contents of both of the above-mentioned documents

• • [0042] U.S. Pat. No. 5,512,759
  • [0043] U.S. Pat. No. 5,439,781
    are incorporated into the disclosure contents of the present application by reference.

[0044] The object of the invention is to supply an illumination system of easy construction having the required Etendu in the object—or reticle—plane.

[0045] The object of the invention is solved in that several light sources are coupled in order to illuminate the exit pupil of the illumination system up to a predetermined degree of filling.

[0046] The coupling of several light sources also results in an increase of intensity. A coupling of several light sources is possible as long as the total Lagrange optical invariant of all coupled sources is less than the Lagrange optical invariant of the illumination system LC_ill.

[0047] There are three basic possibilities for coupling:

• [0048] 1. Addition method: Identical or similar illumination systems are distributed about an axis of the system. The exit pupil of the illumination system is illuminated by the circular pupils of the system parts, which must not overlap. The partial pupils are located on the side face of a pyramid-shaped input mirror, which superimposes the light bundles on the object or reticle.
• [0049] 2. Mixing method: In this case, each system part illuminates the entire exit pupil of the illumination system, but with regions free of light between the secondary light sources. The individual grids of the secondary light sources are staggered by superimposing, so as to uniformly fill the pupil. The coupling mirror consists of a raster element plate, whose raster elements or facets have the shape of pyramids. Each side face of an individual raster element pyramid is illuminated by a secondary light source.
• [0050] 3. Segment method: Similar to the addition method. Unlike the addition method, a segment of any desired shape is illuminated by appropriate beam deflection, instead of a circular segment of the exit pupil.

[0051] Preferred embodiments of the invention making use of at least one of the above-mentioned methods are the subject of the subsidiary claims.

[0066] FIG. 1 shows the layout of a system in which the light sources are coupled together according to the addition method. The light sources 1.1, 1.2 in the present case have a small source diameter, in this case laser plasma sources are further investigated.

[0067] Regarding the basic layout of EUV-illumination systems, we refer to the applicant's pending applications EP 99 1 06348.8, submitted on Mar. 2, 1999, entitled “Illumination system, especially for EUV-lithography”, U.S. Ser. No. 09/305,017, submitted on May 4, 1999, entitled “Illumination system particularly for EUV-lithography”, now U.S. Pat. No. 6,198,793 B1, and PCT/EP 99/02999, submitted on May 4, 1999, entitled “Illumination system, especially for EUV-lithography”, whose disclosure contents are incorporated in their entirety in the present application.

[0068] Each system part 10.1, 10.2 is essentially identical in construction and comprises a light source 1.1, 1.2, a collector mirror 2.1, 2.2, and a field raster element plate 4.1, 4.2.

[0069] The light of each source is collected by means of the collector mirror assigned to a particular source and transformed into a parallel or convergent light bundle. The field raster elements of the particular field raster element plate 4.1, 4.2 decompose the light bundle and create secondary light sources 6 in the diaphragm plane of the illumination system. These secondary light sources are imaged by the field lens (not shown) or field mirror in the exit pupil of the illumination system, which is the entrance pupil of the objective lens (not shown). The field raster elements of the field raster element plate are arranged on the plate and oriented so that the images of the field raster elements are superimposed in the reticle plane 9.

[0070] The systems are brought together where the field raster element plates are located. The field raster element plates are located on a pyramid, the number of the sides of the pyramid corresponds to the number of coupled partial systems. The angle of inclination of the pyramid sides is chosen such that the illuminated fields of the partial systems in the reticle plane 9 are superimposed.

[0071] The partial systems parts 10.1, 10.2 are arranged such that their partial pupils fill the diaphragm plane of the illumination system optimally.

[0072] In the embodiment shown in the drawings, the partial systems are oriented such that they possess a common system axis. The angular spacing of the partial system is then 360°/number of systems.

[0073] For four partial systems, FIG. 2 shows the illumination of the pyramid, on each of the four lateral surfaces 20.1, 20.2. 20.3, 20.4 of the pyramid one field raster element plate of a partial system in the area of the illuminated surface 22.1, 22.2, 22.3, 22.4 is arranged.

[0074] The field raster elements are arranged and oriented such that the images of the field raster elements overlap in the reticle plane 9. The angle of inclination of the pyramid surfaces 20.1, 20.2. 20.3, 20.4 is chosen such that the illuminated fields of the partial system superimpose in the reticle plane.

[0075] The illumination in the diaphragm plane is provided by four circular partial pupils 30.1, 30.2. 30.3, 30.4, as shown in FIG. 3, which in turn are divided into individual secondary light sources 6, corresponding to the number of illuminated field raster elements of the field raster element plates.

[0076] In FIG. 3, the aperture of the total system is NA_Obj=0.025 and the aperture of the system parts is NA_Teilsystem=0.0104.

[0077] Depending on the number of coupled partial systems, one can imagine the arrangement and symmetries of the partial pupils 30.1, 30.2. 30.3, 30.4, 30.5, 30.6 as shown in FIGS. 4A through 4D with coupling of 3, 4, 5 and 6 sources.

[0078] The maximum diaphragm diameters of the partial systems are derived from the total aperture NA_Obj of the objective lens in the diaphragm plane and the number of partial systems or subsystems. ${\mathrm{NA}}_{\mathrm{Teilsystem}}=\frac{{\mathrm{NA}}_{\mathrm{Obj}}}{1+\frac{1}{\sin \left(\frac{\pi }{\mathrm{Anzahl}}\right)}}$
Whereby:

[0079] Teilsystem=partial system; Anzahl=number of partial systems

[0080] When the pupil of each subsystem is filled, the pupil can be illuminated to η% of the maximum. $\eta =\mathrm{Anzahl}·\frac{1}{{\left(1+\frac{1}{\sin \left(\frac{\pi }{\mathrm{Anzahl}}\right)}\right)}^2}$
Whereby:

[0081] Anzahl=number of partial systems

[0082] The following table gives NA_{system part} and the filling factor η for NA_Obj=0.025:

[0083]

Number of system parts	NAsystem parts	Filling factor ηmax
2	0.0125	0.500
3	0.0116	0.646
4	0.0104	0.686
5	0.0093	0.685
6	0.0083	0.667
7	0.0076	0.641
8	0.0069	0.613
9	0.0064	0.585
10	0.0059	0.557

[0084] Hence, the maximum attainable filling factor with the addition method using four subsystems and NA_Obj=0.025 is achieved with η_max≈0.69. As a boundary condition, the overall Etendu of the coupled sources may not exceed the system Etendu LC_ill=η_max·LC_Obj; thus, we must always have:
ΣLC_i≦LC_ill
all sources

[0085] FIG. 5 shows a second form of embodiment of the invention, in which the light sources 50.1, 50.2 are pinch plasma sources, for example. The pinch plasma sources 50.1, 50.2 are also denoted as primary light sources. The source diameter of the pinch plasma sources is not negligible.

[0086] A partial illumination system with pinch plasma source comprises the light source 50.1, 50.2, a collector mirror 52.1, 52.2, which collects the light and illuminates the field raster element plate 54.1, 54.2. The field raster elements of the plate produce secondary light sources. At the location of the secondary light sources, the pupil raster elements are arranged on a pupil raster element plate. The field raster elements of the field raster element plate are used to shape the field and the pupil raster element of the pupil raster element plate correctly image the field raster element in the reticle plane. Preferably, each field raster element is assigned to a pupil raster element. The light is guided by reflection from the field raster elements of the field raster element plates to the pupil raster element of the pupil raster element plate 56.1, 56.2 and from there to the reticle, or object 58.

[0087] The systems are brought together at the location of the pupil raster element plates. The pupil raster element plates are located on a pyramid. The number of sides of the pyramid corresponds to the number of coupled subsystems. The angle of inclination of the pyramid sides is chosen such that the illuminated fields of the partial systems or subsystems are brought together in the reticle plane.

[0088] If the subsystems have a common system axis, then the angular spacing of the system parts is 360°/number of systems and the pupil raster element plates of the subsystem are preferably arranged on the lateral surfaces of a pyramid, as shown in FIG. 2.

[0089] The advantage of the addition method of coupling is that identical or similar illumination systems can be coupled together. The raster element plates of the subsystems are separate and can thus be fabricated separately.

[0090] In the addition method, it should be noted that intensity differences of the individual sources are directly passed on to the illumination of the pupils, and thus the intensity of the partial pupils is dictated by the source power.

[0091] The intensity distribution in the diaphragm plane becomes independent of the intensities of the individual sources if one mixes the secondary light sources in the pupil plane. This technique is also hereafter designated as the mixing method.

[0092] Whereas in the addition method the beam bundles of each source only penetrate after passing through the diaphragm plane, in the mixing method the beam bundles penetrate in front of the diaphragm plane and are mixed in the diaphragm plane. The maximum aperture for each subsystem is adapted to the desired angle of filling of the objective aperture. As in the addition method, systems of identical construction can be coupled together for the individual sources. They are uniformly arranged about a common system axis. The systems are coupled together in the plane of the secondary light sources.

[0093] FIG. 6 shows an illumination system based on the mixing method for coupling of several light sources.

[0094] The light sources once again are laser plasma sources. The same components as in FIG. 5 are designated with the same reference numbers. In contrast to FIG. 5, for example, there is a single pupil raster element plate 100, which includes a plurality of pyramids. The pupil raster element plate 100 is arranged at the location of the secondary light sources, which are produced by the field raster elements. A secondary light source is located on each flank, or lateral side, of the plurality of pyramids.

[0095] The schematic representation of FIG. 7 shows a typical arrangement of the field raster elements 110 on the field raster element plate. Each field raster element plate produces a grid of secondary light sources in the diaphragm plane. The distribution of the secondary light sources in the diaphragm plane corresponds to the arrangement of the field raster elements.

[0096] By shifting the subsystems, as depicted in FIG. 8, the grids of secondary light sources can be brought to be located next to each other, corresponding to the number of subsystems.

[0097] If four sources are coupled together, the arrangement of secondary light sources 6 shown in the schematic representation of FIG. 8 is obtained. For the correct superimposing of the four subsystems, each set of secondary light sources is located on a mirrored pyramid. The flanks of the pyramid are inclined such that the images of the field raster elements are superimposed in the reticle plane. The schematic representation of FIG. 9 shows a segment of the pupil raster element plate. One clearly recognizes the individual pupil raster elements 104 that are formed by the flanks of an equilateral pyramid 106.

[0098] If the Etendu (LC) of the individual sources is small, the pupil raster elements can be designed as plane mirrors, i.e., the flanks of the equilateral pyramids 106 are planar.

[0099] When the source diameter is not negligible, such as with pinch plasma sources, the pupil raster elements 104 must image the field raster elements in the object plane, for example, the reticle plane. In this case, a concave mirror surface 108, as shown in FIG. 10, must be worked into the pyramid flanks.

[0100] The schematic representation of FIG. 10 shows a system in which several pinch plasma sources are coupled with a pupil raster element plate comprising pupil raster elements with concave surfaces. The same components as in FIG. 6 are given the same reference numbers.

[0101] The examples shown in FIGS. 5 through 10 are designed for four coupled sources. However, the same method can be used for three, five, six or more sources. The grids should then be shifted such that the secondary light sources are located on the side faces of pyramids. The degree of filling of the pupil is limited similar to the addition method.

[0102] The advantages of the mixing method are that the individual sources are mixed in the pupil plane. Fluctuations in source intensity are not shown in the pupil as inhomogeneous pupil illumination. Furthermore, the system pupil can be filled more uniformly with secondary light sources.

[0103] As a third method of coupling several light sources together, the segment method shall be described.

[0104] The segment method works similar to the addition method. The coupled illumination systems are uniformly distributed about a common system axis. Each system has a corresponding segment to fill the diaphragm plane. Instead of filling this segment with a circle as in the addition method, one can uniformly fill up the segment by orienting the field raster elements on the field raster element plate. FIG. 11 shows the illumination of one of four segments 200 of the system pupil 202, when four sources are coupled together. In segment 200 secondary light sources 6 corresponding to the number of illuminated field raster elements are formed.

[0105] In order for the individual light bundles to be correctly superimposed in the reticle plane, pupil raster elements must be arranged at the location of the secondary light sources, which deflect the light bundles so that the images of the field raster elements are superimposed in the reticle plane. Depending on the size of the source, the pupil raster elements have planar surfaces for point like sources or concave surfaces for extended sources. Accordingly, field and pupil raster elements are tilted individually and without symmetry.

[0106] The advantage of the segment method is the optimal filling of the diaphragm plane with secondary light sources 6 by a pairwise tilting of field and pupil raster elements.

[0107] Although no optical components have been depicted in the preceding examples of embodiments of the illumination systems after the lenses or mirrors with raster elements, it is obvious to the person skilled in the art that field lenses or field mirrors must be provided after the lenses or mirrors with raster elements in order to shape the annular field in the reticle plane and to image the diaphragm plane into the exit pupil of the illumination system, for example. This is shown in FIG. 12. The illumination system of the second embodiment, shown in FIG. 5 was adapted by introducing a field lens 300 between the pupil raster element plates 56.1 and 56.2. The field lens 300 represents an optical unit, which can also comprise two or more mirrors. The field lens 300 images the plurality of secondary light sources formed on the pupil raster element plates 56.1 and 56.2 into the exit pupil 310. In this regard, concerning the basic layout of EUV illumination systems, refer to the applicant's pending applications EP 99 1 06348.8, submitted on Mar. 2, 1999, entitled “Illumination system, especially for EUV-lithography”, U.S. Ser. No. 09/305,017, submitted on May 4, 1999 entitled “Illumination system particularly for EUV-lithography”, and PCT/EP 99/02999, submitted on May 4, 1999, entitled “Illumination system, especially for EUV-lithography”, whose disclosure contents are incorporated in their entirety in the present application.

[0108] An EUV-projection exposure system is shown in FIG. 13. The illumination system is already shown in FIG. 12. The reticle 58 is imaged by the projection objective lens 320 onto the wafer 330. The EUV-projection exposure system can be realized as a stepper or scanning system.