Patent Application
20080165337
The subject matter disclosed is related to high power gas discharge laser systems for DUV light sources, e.g., used in integrated circuit photolithography, e.g., in highly line narrowed versions, e.g., for immersion lithography and other lithography uses requiring high power and/or requiring longer component life in lower power applications, and more particularly to a seed laser and amplification gain medium system with an improved power amplification stage providing higher gain and reduced ASE and coherency busting, e.g., for reduction in speckle. The method and apparatus can be applied to the reduction of and/or control of line-edge roughness (LER) and line-width roughness (LWR), collectively feature dimension roughness, of exposed photoresist for integrated circuit manufacturing to improve the ability to print device feature critical dimensions (CD), e.g., using deep-UV (DUV) lithography projection systems.
Deep ultraviolet light sources, such as those used for integrated circuit photolithography manufacturing processes have been almost exclusively the province of excimer gas discharge lasers, particularly KrF excimer lasers at around 248 nm and followed by ArF lasers at 198 nm having been brought into production since the early 90's, with molecular fluorine F2 lasers also having also been proposed at around 157 nm, but as yet not brought into production.
To achieve resolution reduction at a fixed wavelength and fixed NA (i.e., an 193 nm XLA 165 on a XT: 1400 with an NA of 0.93), one must optimize k1, where k1 represents process-dependent factors affecting resolution.
Based on Rayleigh's equation, for dry ArF tools today smaller resolution of state-of-the-art high-numerical-aperture ArF lithography can only be achieved with Resolution-Enhancement Techniques (RET's). RETs are a cost-effective way to maintain the aggressive evolution to smaller dimensions in IC manufacturing and are becoming integral to manufacturing lithography solutions.
These Process-related resolution enhancement efforts (lowering k1) have focused on reticle design, using methodologies such as phase shifting or pattern splitting on dual masks. While these techniques improve imaging, they also have significant drawbacks, including throughput loss. So when k1 is optimized for an application the only way to improve resolution further is to go back to the wavelength or NA.
Immersion lithography does just this for the 45 nm, the wavelength is constant at 193 nm so introducing water allows for NA's up to 1.35 and this relaxes the k1 requirement until processing at the 32 nm is required.
Since first introduction of excimer laser light sources in the DUV wavelength manufacturers of these light sources have been under constant pressure not only to reduce the wavelength, but also to increase the average power delivered to the wafer in the manufacturing process carried out by the steppers and scanners of the principle customers for such light sources, the stepper/scanner manufacturers, now including Canon and Nicon in Japan and ASML in the Netherlands. This requirement for smaller and smaller wavelength has come from the need of the integrated circuit manufacturer customers for the stepper/scanner makers to be able to print smaller and smaller critical dimensions on the integrated circuit wafers. The need for higher power has generally been driven by either the need for more throughput or higher dose for exposing certain photoresists on the wafer, or both.
This steady progression down the road formed by the so-called Moore's law about the progression of integrated circuit capabilities, and thus, basically the number of transistors per unit area and thus also basically smaller and smaller critical dimensions, has created various and serious problems for the light source manufacturers to address. Particularly the move to the 193 nm wavelength node of light sources has resulted in several challenges.
Resolution and critical dimension (CD) control in advanced lithography, at the 193 nm, requires a narrow spectral bandwidth, e.g., because all lens materials have some degree of chromatic aberration, necessitating a narrow bandwidth laser to reduce the wavelength variation in the light source, thereby diminishing the impact of chromatic aberration. Very narrow bandwidth can improve the ultimate resolution of the system, or, alternatively can give lens designers more focal latitude. However, there is an increase in the impact of speckle on printing more narrow critical dimensions on the wafer.
Partlo et al, Diffuser speckle model: application to multiple moving diffusers, ______, discusses aspects of speckle reduction. U.S. Pat. No. 5,233,460, entitled METHOD AND MEANS FOR REDUCING SPECKLE IN COHERENT LASER PULSES, issued to Partlo et al. on Aug. 3, 1993 discusses misaligned optical delay paths for coherence busting on the output of gas discharge laser systems such as excimer laser systems. Naulleau, Relevance of Mask-Roughness-Induced Printed Line-Edge Roughness in Recent and Future Extreme-Ultraviolet Lithography Tests, Applied Optics, Vol. 43, Issue 20, pp. 4025-4032 (2004) discusses the effects of speckle induced by roughness in the mask coupling to speckle in the arial image of the illuminating light. Lee, Effect of line edge roughness (LER) and line width roughness (LWR) on Sub-100 nm Device Performance, Advances in Resist Technology and Processing XXI, edited by John L. Sturtevant, Proceedings of SPIE Vol. 5376 (SPIE, Bellingham, Wash., 2004) discusses the impacts of photoresist materials and IC manufacturing process parameters on LER/LWR. Rydberg et al., Dynamic laser speckle as a detrimental phenomenon in optical projection lithography, J. Microlith., Microfab., Microsyst. Vol 53, No. 1-1-1-8 (July-September 2006), revision of a paper presented at the SPIE conference on Optical Microlithography XVIII, March 2005, San Jose, Calif., SPIE Proceedings Vol. 5754 refers to the effects of so-called dynamic speckle/coherence on LER/LWR. Yamaguchi et al., Impact of long-period line-edge roughness (LER) on accuracy in CD Measurement, Metrology, Inspection, and Process Control for Microlithography XIX, edited by Richard M. Silver, Proc. of SPIE Vol. 5752 (SPIE, Bellingham, Wash., 2005) discusses the impact of dimension feature size, e.g., length on LER. Leunissen et al., Full spectral analysis of line width roughness, Metrology, Inspection, and Process Control for Microlithography XIX, edited by Richard M. Silver, Proc. of SPIE Vol. 5752 (SPIE, Bellingham, Wash., 2005) discusses quantification of LWR. Patsis et al., Integrated simulation of line-edge roughness (LER) effects on sub-65 nm transistor operation: from lithography simulation, to LER metrology, to device operation, Emerging Lithographic Technologies X, edited by Michael J. Lercel, Proc. of SPIE Vol. 6151, 61513J, (2006) discusses LWR metrology and the impacts of LWR on device operating parameters. Pawloski et al., Characterization of line edge roughness in photoresist using an image fading technique, Unknown publication information, discusses the role of aerial image contrast and image-log-slope (ILS) on the resulting magnitude of line edge roughness (LER) in resist and whether the minimization of LER in current state-of-the-art, chemically amplified resist materials was limited by the quality of the projected aerial image or the materials and processing of the resist and the identification of the iso-fading condition, which in analogy to the iso-focal dose, results in a unique exposure dose for which the critical dimension (CD) of a resist feature does not change with increasing levels of fading. The paper suggests that there is a way to determine undesired iso-fading condition which applicants believe could be used in accordance with aspects of an embodiment of the disclosed subject matter to, e.g., control coherence busting to attempt to achieve a better or perhaps optimized state of LER on the wafer, and the paper suggests that while the aerial image plays a strong role on determining the magnitude of LER during resist printing, there also exists a fundamental limitation to LER from the resist materials that cannot be improved by further increase in the quality of the aerial image. Cobb et al., EUV photoresist performance results from the VNL and the EUV LLC, Emerging Lithographic Technologies VI, Roxann L. Engelstad, Editor Proceedings of SPIE Vol. 4688 (2002) © 2002 SPIE, discusses the role of optics and masks in LER using EUV light sources. Naulleau, The role of temporal coherence in imaging with extreme ultraviolet lithography optics, Optics Communications 219 (2003) 57-63, ©02003 Elsevier Science B.V., available at www.science direct.com, discusses the impact of temporal coherence in off-axis illumination using EUV broad band light sources. Goodman, Speckle Phenomena in Optics: Theory and Applications, may have information regarding speckle reduction and speckle in photolithography.
As used herein the term resonator and other related terms, e.g., cavity, oscillation, output coupler are used to refer, specifically to either a master oscillator or amplifier portion, a power oscillator, as lasing that occurs by oscillation within the cavity until sufficient pulse intensity exists for a useful pulse to emerge from the partially reflective output coupler as a laser output pulse. This depends on the optical properties of the laser cavity, e.g., the size of the cavity and the reflectivity of the output coupler and not simply on the number of reflections that direct the seed laser input through the gain medium a fixed number of times, e.g., a one pass, two pass, etc. power amplifier, or six or so times in the embodiment disclosed in Fork, et al. and not on the operation of some optical switch in the cavity. In some of the literature an oscillator in which the round trip through the amplification gain medium, e.g., around a loop in a bow-tie or racetrack loop, is not an integer number of wavelengths, may be referred to as an amplifier, e.g., a power amplifier, while also constituting an oscillator laser. The term power amplification stage and more specifically ring power amplification stage is intended herein to cover both of these versions of a power oscillator, i.e., whether the path through the gain medium is an integer multiple of the laser system nominal center wavelength or not and whether the literature, or some of it, would refer to such an “oscillator” as a power amplifier or not. The closed loop path or oscillation loop as used herein refers to the path through the amplification gain medium, e.g., an excimer or similar gas discharge laser amplification stage, around which the seed laser pulse light oscillates in the amplification stage.
Photolithographic processes, e.g., for integrated circuit manufacturing processes, for the formation of various integrated circuit structures and patterns, may be used, e.g., forming patterns by exposing a photoresist (PR) layer. Light is passed through a mask (or reticle), which is comprised of clear areas that transmit the incident light and dark regions that absorb the incident light. An optical projection arrangement forms an image of the mask pattern in the PR layer, e.g., at a defined plane of the wafer with a certain depth of focus. Areas of the PR layer sufficiently exposed by the light energy, can become either soluble or insoluble upon such exposure, so that desired portions of the PR layer can be removed, e.g., with a chemical developer solution applied to the surface of the PR layer. The removed areas uncover the underlying layer (e.g. a semiconductor layer, a metal or metal-containing layer, a dielectric layer, etc.), while the remaining PR protects the underlying layer. The open areas of the underlying layer can subsequently be treated, e.g., wet or dry etched, exposed to dopant for assimilation, subjected to ion implantation/etching, etc., while the remaining PR can, e.g., selectively block areas in the underlying layer where retardation or elimination of the effect of the treatment for the opened areas occurs. The remaining PR film can be stripped thereafter and more PR added and the next layer in the manufacturing process similarly treated.
The ever advancing requirement to increase the density of the patterned areas requires appropriately smaller scaling of the sizes of the imaged PR features and improvement in the resolution of the optical exposure systems used to image the PR layer. As critical dimensions continue to be reduced, the roughness of the edges of the patterned PR structures starts to become noticeable, being transferred into the underlying layer by the accompanying treatment, such as, etch, implantation or reaction processes, and adversely affecting the size and/or shape of features being formed or otherwise treated in the underlying layer. The effects of such feature dimension roughness/variation can be most important in features, such as transistor gates, where precision and uniformity of the edges of particularly the dimensional features defining the width of the gate region, e.g., separating the source and drain regions, and like features critical to performance of the device being made on the integrated circuit. The line-edge roughness (LER) or line-width roughness (LWR) of the PR patterns has particularly been noted in lithography utilizing Argon Fluoride (ArF) excimer laser sources, with exposure wavelengths near 193 nm, e.g., with the use of high numerical apertures and/or immersion lithography. Indeed reduction in LER and/or LWR is becoming one of the most important challenges in photolithography for IC fabrication. Therefore, it is highly desirable to reduce as much as possible the dimension roughness/variation of the PR patterns.
It will be understood by those skilled in the art that disclosed in the present application is a method and application which may comprise reducing feature dimension roughness in an exposed photoresist layer exposed by a pulsed sub-400 nm light source by reducing dose variation by increasing the spatial frequency of speckle for a given speckle contrast. The method and apparatus may comprise reducing dose variation by increasing the spatial frequency of speckle for a given speckle contrast from pulse to pulse and utilizing a projection lens with a frequency cut off below substantially all of the speckle spatial frequencies. The method and apparatus may comprise producing integrated circuits by controlling laser light source speckle size and/or contrast variation from a known variation in image contrast to select a level of feature dimension roughness. The method and apparatus may comprise utilizing a photolithography tool comprising a light source input opening receiving sub-400 nm light pulses from a pulsed excimer or molecular fluorine gas discharge laser and a wafer photoresist exposure platform holding a wafer; detecting image contrast in feature dimension at the wafer and providing a feedback signal indicative of the image contrast; controlling laser light source speckle size and/or contrast variation based on the feedback signal to select a level of feature dimension roughness. The apparatus and method may comprise reducing feature dimension roughness in an exposed photoresist layer exposed by a pulsed sub-400 nm light source using a coherence busting mechanism comprising a coherence busting optical structure generating multiple sub-pulses delayed sequentially from a single input pulse, wherein each sub-pulse is delayed from the following sub-pulse by more than the coherence length of the light pulse and a beam pointing displacement mechanism displace the beam pointing between a main pulse and a delayed pulse. The apparatus and method may comprise reducing feature dimension roughness in an exposed photoresist layer exposed by a pulsed sub-400 nm light source by utilizing a laser light source comprising a seed laser and an amplification gain stage; utilizing a coherence busting mechanism intermediate the seed laser and the amplification stage comprising a beam splitter separating the seed laser output into a main beam and a beam entering an optical delay path having a delay length longer than the coherence length of a pulse in the seed laser output and a beam angular offset mechanism offsetting a delayed beam from the delay path and the main beam. Each sub-pulse may be delayed substantially less than its pulse length such that the delay creates substantially no increase in the pulse length or in the Tis of the pulse. The coherence busting mechanism may comprise a first coherence busting optical delay structure and a first angular offset mechanism, and a second coherence busting optical delay structure and a second angular offset mechanism. The optical delay time in the second coherence busting optical delay structure may be longer than the delay in the first optical delay structure by more than the coherence length of the input light pulse. The first angular displacement mechanism may be oriented to angularly displace the main pulse and at least one delayed pulse in one axis and the second angular displacement mechanism oriented to angularly displace the main pulse and at least one delayed pulse in a second axis. The first axis may be different from the second axis. The method and apparatus may comprise the coherence busting mechanism intermediate the laser system and the photolithography tool, within the photolithography tool, e.g., between the laser light source entry opening and the illuminator or between the illuminator and the wafer, or within the laser light source system, e.g., intermediate the laser and an OPuS or intermediate the OPus and the photolithography tool, or intermediate the laser system and the photolithography tool, e.g., within a beam delivery unit.
According to aspects of an embodiment of the subject matter disclosed a gain amplification medium suitable for use with, e.g., an excimer or molecular fluorine gas discharge seed oscillator laser in a multi-chamber (multi-portion) oscillator/amplifier configuration, this could be, e.g., a master oscillator power gain amplification configuration may take advantage of a improved seed laser coupling arrangements, fundamentally designed to insert seed laser, e.g., master oscillator seed output laser light pulse beam pulses, into an amplifying gain medium, generally with little loss and with protection against amplifier oscillation and/or ASE returning to the master oscillator while the master oscillator laser medium is excited, such could interfere with the proper operation of the master oscillator, e.g., in conjunction with the line narrowing module producing the appropriately narrowed seed oscillator output laser light pulse beam pulse bandwidth.
According to aspects of an embodiment of the subject matter disclosed, however, a preferred configuration may comprise, e.g., a ring cavity, e.g., a power oscillator or Power Ring Oscillator (“PRO”). Such a configuration has been determined by applicants to be a very effective solution to going to higher power laser operation in a line narrowed multi-portion (seed laser-amplifier) arrangement, particularly for gas discharge seed laser to identical gas discharge amplifier laser multi-portion laser systems. Such a laser system could be similar in operation to applicants' employers' XLA series lasers, though with a power ring amplifier stage. Improvement in CoC may be attained according to a aspects of an embodiment of the subject matter disclosed. Also, however, a power ring amplification stage may be useful for other applications, including with seed lasers of other than the same type of gas discharge laser, e.g., a solid state seed, e.g., matched to the lasing wavelength of an excimer or molecular fluorine amplifier, e.g., by frequency shifting and/or frequency multiplication. In such systems pulse trimming may be useful for ultimate control of laser system output laser light pulse beam pulse parameters, e.g., bandwidth, bandwidth stability, output pulse energy, output pulse energy stability and the like.
According to aspects of an embodiment of the subject matter disclosed a ring cavity could use, e.g., such a beamsplitter, e.g., with P-polarization instead of the normally used S-polarization, since applicants have found that the reflectivity of such an OPuS beamsplitter changes from 60% for S-polarization to about 25% for P-polarization.
Thus, according to aspects of an embodiment of the subject matter disclosed a ring cavity PO may be constructed, e.g., with a 24% output coupler, e.g., comprising an existing OPuS beamsplitter.
According to aspects of an embodiment of the subject matter disclosed applicants propose to re-arrange, e.g., an existing XLA product, e.g., with an excimer-based MO, from MOPA to MOPO (power oscillator), utilizing a power amplification stage, such as a ring power amplification stage, in accordance with aspects of an embodiment of the subject matter disclosed. Such a system can (1) improve energy stability, e.g., by operating the amplification stage at saturation, or effectively saturation, pulse to pulse, thereby more accurately insuring pulse to pulse energy stability, (2) achieve longer LNM life, e.g., by reducing the required MO output level to the μJ level, e.g., by increasing the amplification in the amplification stage (over applicants' assignees' traditional MOPA configuration) by approximately ten fold, and (3) and exploit the ability to operate the MO at less than 1 mJ in other ways beneficial to overall laser system operating life.
The advantages of the multi-chambered laser system according to aspects of an embodiment of the subject matter disclosed enable meeting the above discussed requirements for, e.g., higher power, better pulse energy stability, better bandwidth control and lower achievable bandwidth, higher repetition rates and lower cost of operation. For example, the lower MO output energy requirement can, e.g., allow for even better control of pulse parameters, e.g., bandwidth in the MO, with less energy loss in the MO, e.g., during line narrowing, and also lower the thermal impact, e.g., transients, and lower optical damage to the line narrowing optics, while maintaining or even increasing output power. Further according to aspects of an embodiment of the subject matter disclosed, e.g., through maintaining currently known MO pulse energy output levels, very high (e.g., 10×) increases in currently available laser system output light average power may be attained. This may be beneficial both for line narrowed systems and for broad band systems, e.g., XeCl multi-chamber laser systems used, e.g., for annealing amorphous silicon on substrates, e.g., in LTPS processes, e.g., for the manufacture of crystallized substrates for the production of, e.g., thin film transistors.
Turning now to
Relay optics 40, e.g., including a turning mirror 44 and a turning mirror 46, may serve to steer the seed oscillation laser 20 output laser light pulse beam 62 pulses exiting a line center (center wavelength) analysis module (“LAM”) 42 along a light path (optical axis) 60 to the input of the amplifier module lasing chamber 24. The LAM, in addition to center wavelength monitoring equipment (not shown) may include an MO energy monitor 48, which may be provided with a small portion of the laser output light pulse beam from the MO chamber 22 for metrology purposes, e.g., for nominal center wavelength and energy detection, by a beam splitter 50 inside the LAM 42. The turning mirror 44 may provide the master oscillator 22 output laser light pulse beam 62 pulses to the turning mirror 46 along an optical path beam path, which may reflect the beam 62 into the amplifier chamber 24 as a beam 64.
In the case of the system 20 of
Also in the path of the laser system 20 output laser light pulse beam 100 pulses may be, e.g., a beam expander 84, e.g., to decrease the energy density on downstream optics, including the OPuS 86 beam splitters 76, 78 and optical delay ling mirrors 90 and optics, e.g., in the scanner (not shown) utilizing the laser system 20 output laser light pulse beam 100 pulses. The laser system 20 may also include a shutter 96, including, e.g., a shutter beam splitter 98, e.g., taking off a portion of the laser system 20 output laser light pulse beam 100 pulses for energy measurement in an output energy monitor (not shown) in the shutter 96.
This existing XLA MOPA configuration, shown in
Turning to
According to aspects of an embodiment of the subject matter disclosed may include, e.g., placing the a beam expander 142 comprising, e.g., first and second beam expanding and dispersing prisms 146, 148 inside of an entrance window and beam expander housing 140 which may be affixed to the gain medium chamber 144 by a suitable means, e.g., welding or bolting with suitable sealing mechanisms. These optics 146, 148 may be placed inside the ring cavity formed between the orthogonal seed injection mechanism 160 input/output coupling partially reflecting mirror 162 and the beam returner 70, e.g., in order to reduce the energy density on the maximum reflector forming the beam returner 70, e.g., as illustrated in
Use of protective coatings on this optic 162 may be eliminated due, e.g., to the much lower energy output of the MO output laser light pulse beam pulses according to aspects of an embodiment of the subject matter disclosed. The beam expander optics 146, 148 and beam returner/reverser 70, due to their makeup including a fluorine containing crystal and their exposure to fluorine in the lasing medium gas in the chamber 144 and housings 140, 150 may be protected from optical damage. The AMPLIFICATION STAGE chamber window 168 similarly constructed, e.g., of a fluorine containing crystal, e.g., CaF2, need not have a protective coating on its face exposed to the highest energy density, facing the ring power oscillator oscillation cavity, due in part to its beam expansion in the beam expanded 14Z in the cavity and also to using around a 45° angle, e.g., a 47 degree orientation.
According to aspects of an embodiment of the subject matter disclosed a ring power oscillator cavity, e.g., as illustrated by way of example in
Further, all optics inside the chamber, e.g., including the snouts 140, 150 can be, e.g., further removed from the source of chamber dust. The configuration can also be made to fit, e.g., within a present XLA optics bay.
As explained elsewhere, the ring cavity, e.g., as shown in
In addition, such small MO energy would likely allow use of a low MO chamber pressure with a number of longer chamber life benefits.
Rather than having to contemplate ways to simply survive high raw power, e.g., in the 200 W range, applicants, according to aspects of an embodiment of the subject matter disclosed contemplate being able to instead focus on bettering energy stability, pointing stability, profile stability, and ASE stability of contemplated configuration while operating at full repetition rate, e.g., between 4 kHz and 6 kHz and even above.
Turning now to
In
Similarly a ring cavity may be set up according to aspects of an embodiment of the subject matter disclosed as illustrated in
There are a number of possible ways to couple the output laser light pulse beam pulses from the MO to the power amplification stage, e.g., as illustrated schematically and partly in block diagram format in
Illustrated schematically and partly in block diagram form in
As illustrated in
Illustrated in
Turning now to
According to aspects of an embodiment of the subject matter disclosed, the orthogonal seed injection mechanism may comprise an orthogonal injection seeding optic such as, e.g., optical element 350, illustrated in
According to aspects of an embodiment of the subject matter disclosed another version of a seed injection optic 360, illustrated in
According to aspects of an embodiment of the subject matter disclosed a variety of beam returners/reversers 370 may be utilized, e.g., as illustrated schematically in
It will also be understood by those skilled in the optics art that with three internal reflections, or with a three mirror arrangement, e.g., as is currently in use as a beam reverser on applicants' assignee's XLA model laser systems, the input beam 76 and output beam 78 will be effectively aligned and parallel and that relationship does not change, e.g., with rotation of the optic, e.g., optic 370, e.g., about an axis perpendicular to the plane of the page of
It will be understood by those skilled in the optics art that various combinations of the seed injection mechanism referred to in the present application and beam reversers/returners may be utilized to get the beams, e.g., on paths 76, 78 to cross, e.g., as illustrated in
Turning now to
As shown in
Similarly as illustrated schematically and partly in block diagram form, e.g., in
A ring cavity, e.g., with an output coupler seed laser coupling, e.g., an seed injection mechanism, while perhaps more complex a configuration, makes most efficient use of seed laser energy.
According to aspects of an embodiment of the subject matter disclosed, for the seed laser input/output coupling a range of maximally reflecting mirrors may be utilized, e.g., from about a square 45 degree Rmax to about a square 30 degree Rmax, e.g., as is used on applicants' assignee's ArF 193 nm LNMs. Reflectivity for P-polarization is only about 85% at 45 degrees. Instead of a perhaps more desirable 45 degree output coupler with more desirable P-polarization and S-polarization properties, due to time constraints, applicants have so far only examined the use of an OPuS beamsplitter, rotated to 45 degrees in the horizontal, which provides, e.g., 24% reflectivity for P-polarization and 60% reflectivity for S-polarization. A more desirable set of values may be 20% for p-polarization and a smaller number, e.g., around 10% for s-polarization. In the development unit according to aspects of an embodiment of the subject matter disclosed a chamber window was held at the standard 47 degree angle of incidence, rather than at Brewster's angle, which could also be employed.
Applicants have measured P-polarization round trip transmission and S-polarization round trip transmission, as follows:
with a ratio between S and P of 2.44:1.
According to aspects of an embodiment of the subject matter disclosed where attenuation of s-polarization may be needed, e.g., because of ASE, it may be achieved via Brewster reflections and insertion of partial reflectors in or into the power amplification stage cavity.
Also according to aspects of an embodiment of the subject matter disclosed, pulse duration may be controlled, e.g., to control ASE, and if so this may be done, e.g., using optics and electronics, e.g., useful in pulse trimming. Without controlling such timing and/or pulse trimming, according to aspects of an embodiment of the subject matter disclosed, e.g., the seed pulse duration may be longer than that desired, and, e.g., peak intensity may be lower for a given total energy.
MOPO energy vs. MO-AMPLIFICATION STAGE timing has been examined at different values of seed laser energy, ArF chamber gas mixture, percentage reflectivity of output coupler (cavity Q) and seed laser pulse duration, with the results as explained in relation to
ASE vs. MO-PO timing has been examined for different values of seed laser energy, ArF chamber gas mixture, percentage reflectivity of output coupler (cavity Q) and seed laser pulse duration with the results also explained in relation to
The relationship between forward energy and seed energy has also been examiner and the results of which are illustrated, e.g., in
ASE can be of great concern with MOPO designs. Improper timing may lead to increased ASE up to and including generations of only ASE when the MO and power amplification stages are so mis-timed that essentially only broad band(ASE) lasing occurs in the power amplification stage, which being an oscillator will lase when the discharge occurs between the electrodes in the power amplification stage. Unlike a power amplifier, such as in applicants assignee's XLA-XXX laser system where the seed beam passes through the amplification stage a fixed number of times depending on the optical arrangement, in systems according to aspects of an embodiment of the subject matter disclosed amplified spontaneous emission (ASE) lasing occurs whether the seed laser pulse is present for amplification or not. Back scatter from the amplifier cavity optics can form a parasitic laser cavity. Some amplifier cavity optics can form an unintended laser cavity between the amplifier and MO. Therefore careful control of timing is used, according to aspects of an embodiment of the subject matter disclosed, to keep ASE below limits that reduces or effectively eliminates the unwanted lasing. The back scattering or otherwise ASE at the MO may be measured by the ASE detector, which may comprise, e.g., a fluorescence detector.
ASE measurements have been made with medium and small seed input energy. For example for medium energy, e.g., seed energy of around 50 μJ, with a discharge voltage Vco of around 950V, and with a AMPLIFICATION STAGE gas fill of 38/380, fluorine partial pressure/total pressure, it has been shown that with a relative timing of between about −10 ns and +10 ns of optimum the ASE ratio is below about 3×10−5 With low seed energy, e.g., around 5 μJ, with the same voltage and fill the ASE ratio is kept below about 6×10−4 between about 10 ns to +10 ns of relative timing.
While a ring cavity can produce very low ASE with 50 uJ of seed energy, the present implementation according to aspects of an embodiment of the subject matter disclosed can reach the ASE upper limit specification limit at optimal relative timing of between about −10 ns and +10 ns of optimum with about 5uJ of seed energy. Also according to aspects of an embodiment of the subject matter disclosed the seed pulse can be, e.g., electro-optically trimmed to produce, e.g., a desired 10 ns pulse duration, even from an excimer seed laser with normally about a 30 ns pulse duration. According to aspects of an embodiment of the subject matter disclosed applicants expect reduced ASE with 5uJ of seed energy, e.g., due to higher peak intensity. Maintaining proper ASE performance may require selecting proper amplifier cavity optics that have appropriate selectivity to eliminate unwanted polarization (e.g., utilizing appropriate coatings/angles of incidence, etc.), which can result in better suppression of unwanted polarization, which can result in reduced ASE, e.g., from the S-polarization. Creating dispersion in the amplifier cavity, e.g., with beam expanding and dispersing prisms has also been determined by applicants, according to aspects of an embodiment of the subject matter disclosed, to be an effective method for further reducing the ASE ratio contributing to an effectively large enough margin against whatever ASE specification is selected.
According to aspects of an embodiment of the subject matter disclosed a method is proposed to reduce ASE in ring amplifiers, e.g., to take better advantage of other features of this architecture, e.g., low seed energy, high efficiency, energy stability etc. Applicants propose to introduce some broad band (at least much broader than the line narrowed seed radiation propagating in the opposite direction from that of the main radiation direction to increase ASE in this direction and reduce ASE in the main direction. That is to say broad band gain will be utilized in the opposite path around the ring to reduce the available gain for ASE in the main direction. This could be accomplished, e.g., with some scatter of the seed laser beam from the optics, e.g., by feeding florescence of the seed laser into the ring power amplification stage. The broad band emission can thereby, e.g., deplete gain available to the ASE and will be propagated oppositely to main radiation direction, reducing broad band emission in the main direction.
According to aspects of an embodiment of the subject matter disclosed it will be understood that solid state pulsed power systems such as the magnetically switched systems noted above in one or more of the referenced patents or patent applications and as sold with applicants' assignee's laser systems, having very tightly controlled timing of the firings of electric discharges between electrodes in the respective MO and amplification gain medium chambers, along with the properties of a ring power amplification stage in a MOPRO configuration (e.g., operating the ring power amplification stage at or very near total saturation), enables the delivery to a lithography tool or an LTPS tool, or the like, laser system output light pulse beam pulses having about twice the dose stability as is currently achievable, e.g., in applicants' assignee's XLA MOPA laser systems.
Referring to
For the arrangement shown in
Continuing with
The beam entering the beam mixer 1050 in
The beam mixer 1050 may be placed in between the seed beam laser portion and the amplifier laser portion of a MOPA or MOPRO (with, e.g., a ring power amplification stage), configured multi-chambered laser system, such as that shown in
It will be understood by those skilled in the art that as disclosed in the present application according to aspects of an embodiment of the subject matter disclosed, applicants have enabled the satisfaction of customer demands, both from scanner makers and semiconductor manufacturer end users, that have been placed on light source suppliers, e.g., ArF light sources, beyond even the traditionally expected power and bandwidth improvements. For example, further CoC Improvement is demanded because, e.g., ArF is now used in high volume production, e.g., on cost sensitive products, the industry expectation of equivalent reductions in cost of operation and thus cost of consumables for ArF as was historically demanded in KrF as that technology matured. In addition, energy stability improvements are met by the subject matter disclosed, e.g., critical dimension variation sensitivity to dose, which has become greater with the advent of low KI lithography techniques. The double exposure concept, e.g., also trades off between overlay and dose control. Optical maskless lithography will require single pulse exposure control, improved by aspects of embodiments of the subject matter disclosed.
According to aspects of an embodiment of the subject matter disclosed Cost Of Consumables Improvement, e.g., in XLA MO Chamber Life are enabled, e.g., because the parameters that lead to long chamber life did not produce sufficient MO energy for proper operation in the XLA MOPA configuration, e.g., a requirement of at least about a 1 mJ MO output energy. According to aspects of an embodiment of the subject matter disclosed significantly less than 1 mJ energy is required from the MO cavity, e.g., for lithography uses, thus allowing significant reductions in CoC for such low light sources. Therefore, e.g., with operating parameters more conducive to long electrode life, e.g., the system may produce, e.g., around 100 μJ from the MO cavity. This can, e.g., in ArF lasers, e.g., increase the MO chamber life time to at or about that of the PA chamber lifetime, an approximately a 10× increase in MO lifetime before replacement as a consumable while still attaining system output average power from the amplification stage for effective lithography with 60-100 watts or so output. MO cavity optics, LNM optics and the MO output coupler, according to aspects of an embodiment of the subject matter disclosed, experience lower 193 nm intensity, ensuring very much longer optics lifetimes.
With regard to energy stability improvements, the Cymer XLA light source led to a significant improvement in energy stability by exploiting the saturation effects in the PA of a MOPA configuration, e.g., with a two pass PA amplification. The slope of Eout vs. Ein for XLA is about ⅓. MO energy instabilities are reduced by a factor of 3× when passed through such a PA. However, even with the 3× improvement through the PA, the MOPA system energy stability is still greatly impacted by, e.g., the MO energy instability. MO and PA contributions are about equal. Other contributions such as, voltage regulation, timing jitter, and MO pointing jitter are relatively smaller contributors, but not insignificant. The PA energy stability performance falls somewhere between a typical broadband oscillator and a fully saturated amplifier.
According to aspects of an embodiment of the subject matter disclosed, a recirculating ring configuration, e.g., a power ring amplification stage, operates in a much stronger region of saturation. The slope of Eout vs. Ein for a seed laser/amplification gain medium system, e.g., with a ring power amplification stage has been measured by applicants' employer at 0.059. MO energy instabilities can be reduced by a factor of 17×, e.g., when passed through a recirculating ring oscillator, e.g., a power ring amplification stage. At the planned operating point of about 100 J of MO energy, the Eout vs. Ein slope has been observed at or simulated at 1/17.
With the recirculating ring configuration the amplification stage energy stability will exhibit the characteristics of a fully saturated amplifier. Applicants expect at a minimum to see about a 1.5-2× improvement in energy stability.
A larger improvement is expected in the MO σ2, e.g., due to reduced thermal transient effects at the approximately 10× lower required MO output pulse energy, a large improvement is expected in the amplification gain medium σ2, e.g., due to operating at or near saturation. Improvement is also expected for MO pointing σ2.
It will further be understood by those skilled in the art that according to aspects of an embodiment of the subject matter disclosed a power ring amplification stage. Optics may be utilized to create, e.g., a combined bow-tie and race-track oscillation path for four passes per oscillation path, or optics may be utilized to, e.g., create two or more overlapping bow-ties or race tracks. Bow tie arrangements may sometimes be referred to in the art as a cat's cradle. Characteristic of such amplifier media, e.g., regenerative or recirculating ring power amplification stage can include parallel planarity which could be a stable oscillator, e.g., in half planes or an unstable oscillator. The beam returner/beam reverser could utilize multiple mirrors or prisms or a combination thereof, positioned inside or outside the chamber or a combination thereof, e.g., dependant upon exposure to certain levels of energy density by one or more of the optical elements. Unwanted light, e.g., mostly ASE is discriminated against in a variety of ways, e.g., preferentially being created in a direction opposite from the regeneration path of the oscillations of the seed laser pulse beam, e.g., in the ring power amplification stage. Coated optics may be avoided not just in the amplifier gain medium cavity but also, e.g., for the output coupler or in the LNM of the MO, e.g., for the nominal center wavelength selection maximally reflecting mirror (Rmax) or on the grating. Expanding the beam within the amplifier stage cavity, e.g., corresponding to the vertical direction of a Brewster angle window, also can serve to protect optical elements in the ring power oscillator cavity as well as disperse the light to lessen ASE. The output coupler portion of the seed inject mechanism may, e.g., have a reflectivity of around 20% for the desired (in-band) frequencies (or polarization or both) and 100% for undesired light, e.g., ASE, as a further means of ASE reduction. The output coupler may have, e.g., both input and output side surfaces coated with a 100% reflectivity coating for the unwanted light wavelengths, e.g., for ASE which may also, e.g., clean up birefringence induce polarization. Beam expansion may also be able to be performed with multiple prisms, some one or more of which may be inside and/or outside of the chamber enclosure. That is, while one or more of such prisms may be inside the chamber enclosure and exposed to the fluorine containing laser gas mixture, at least one may also be outside the chamber.
Turning now to
It will be understood that one can select an operating point on the ASE curve at or around the minimum extremum and operate there, e.g., by dithering the control selection of dtMOPA to, e.g., determine the point on the operating curve 602 at which the system is operating. It can be seen that there is quite a bit of leeway to operate around the minimum extremum of the ASE curve 602 while maintaining output pulse energy on the relatively flat top portion of the energy curve to, e.g., maintain laser system output pulse energy and energy σ, and the related dose and dose a constant, within acceptable tolerances. In addition as shown, there can be a concurrent use of dtMOPO to select bandwidth from a range of bandwidths while not interfering with the E control just noted.
This can be accomplished regardless of the nature of the seed laser being used, i.e., a solid state seed or a gas discharge laser seed laser system. Where using a solid state seed laser, however, one of a variety of techniques may be available to select (control) the bandwidth of the seed laser, e.g., by controlling, e.g., the degree of solid state seed laser pumping. Such pump power control may, e.g., put the pumping power at above the lasing threshold in order to select a bandwidth. This selection of bandwidth may shift or change the pertinent values of the curve 604, but the laser system will still be amenable to the type of E and BW control noted above using dtMOPO to select both a BW and concurrently an operating point that maintains the output energy of the laser system pulses at a stable and more or less constant value in the flat top region of the illustrated energy curve 600. It is also possible to use a non-CW solid state seed laser and to adjust the output bandwidth. For example, selection of the output coupler reflectivity of the master oscillator cavity (cavity-Q) can adjust the output bandwidth of the seed laser system. Pulse trimming of the seed laser pulse may also be utilized to control the overall output bandwidth of the laser system.
It can be seen from
It is similarly known that the pulse duration of discharge pulses in a gas discharge seed laser, among other things, e.g., wavefront control may be used to select a nominal bandwidth out of the seed laser and thus also influence the slope and/or position of the BW curve 604 as illustrated by way of example in
Turning now to
The output beam 626 may then be turned by a maximally reflective mirror 634 (for the nominal center wavelength) to a seed injection mechanism 636. the seed injection mechanism may include, e.g., a partially reflective optical element 638 and a maximally reflective optical element 640, and may be two separate elements or a single optic as discussed elsewhere in the present application. As discussed elsewhere, the seed injection mechanism may inject the seed laser output pulse beam 626 into an amplification gain medium, such as a ring power amplification stage 650 along an injection path 652, whereby the pulse beam oscillates in a loop also comprising a return reverse path 654 and the partially reflective input/output coupler 638 until such time as enough amplification in the ring power amplification stage occurs by laser light oscillation in the cavity for the input/output coupler 638 to pass a laser system output light pulse beam 658 on to a tool using the output light. A beam splitter 654 can divert a small portion of the output beam 658 into a metrology unit 656 which may measure, e.g., output energy and bandwidth. A metrology unit 642 connected directly to the amplification gain medium laser 650 which can measure, e.g., ASE in the laser chamber 650.
A controller 660, which may comprise a processor 662, receives inputs from the various metrology units 632, 642 and 656, and others as appropriate, and utilize them as part of control algorithms referenced in one or more of the above noted patents and co-pending applications and also incorporate the control algorithm noted above regarding operating at or around the ASE curve minimum while maintaining energy constant and also selecting bandwidth within the limits imposed by a selected ASE limit. In addition, as is shown in one or more of the above referenced patents and co-pending applications the controller 660 may also control the timing of the creation of an output pulse in the seed laser and the creation of the output pulse in the amplification gain medium (dtMOPO for short) and also provide control signals to the line narrowing module, e.g., to control bandwidth, e.g., by wavefront manipulation or optical surface manipulation as discussed above and in one or more of the above referenced patents and co-pending patent applications.
Turning now to
According to aspects of an embodiment of the subject matter disclosed one may need to select an edge optic, that is an optic that may have to be used, and thus perhaps coated, all the way to its edge, which can be difficult. Such an optic could be required, e.g., between the output coupler, e.g., 162 shown in
As illustrated by way of example in
The seed injection mechanism 160 may comprise a partially reflective input/output coupler 162, and a maximally reflective (Rmax) mirror 164, illustrated by way of example and partly schematically in
With regard to the configuration of the beam expansion prisms 146a, 146b and 148 inside the ring power amplification stage cavity a similar arrangement may be provided to that of the beam expansion on the output of the power amplifier (“PA”) stage in applicants' assignee's XLA-XXX model laser systems, e.g., with a 4× expansion, e.g., provided by a 68.6° incident and 28.1° exit, e.g. on a single prism or on two prisms with the same incident and exit angles. This can serve to, e.g., balance and minimize the total Fresnel losses. Reflectivity coatings, e.g., anti-reflectivity coatings may be avoided on these surfaces since they will experience the highest energy densities in the system. According to aspects of an embodiment of the subject matter disclosed the beam expander/disperser 160 may be implemented with the first prism 146 split into to small prisms 146a, and 146b, which may be, e.g., 33 mm beam expander prisms, e.g., truncated, as shown by way of example in
The master oscillator seed beam 62 may enter the seed injection mechanism 160 through the beam splitter partially reflective optical element 162, acting as an input/output coupler, to the Rmax 164 as beam 62a, from which it is reflected as beam 74a to the first beam expander prism 146a, which serves to de-magnify the beam in the horizontal axis by about ½× (it remains about 10-11 mm in the vertical axis into the plane of the paper as shown in
The positioning and alignment of the prisms 146a, 146b and 148, especially 146a and 146b can be utilized to insure proper alignment of the output beam 100 from the ring power amplification stage into the laser output light optical train towards the shutter. The beam leaving the input/output coupler 162 may be fixed in size, e.g., in the horizontal direction, e.g., by a horizontal size selection aperture 130, forming a portion of the system aperture (in the horizontal axis) to about 10.5 mm. Another aperture, e.g., in the position roughly of the present PA WEB, e.g., in applicants' assignee's XLA-XXX laser system products, can size the beam in the vertical dimension. since the beam has about a 1 mRad divergence, the sizing may be slightly smaller in each dimension than the actual beam dimensions wanted at the shutter, e.g., by about 1 mm. According to aspects of an embodiment of the subject matter disclosed applicants propose that a system limiting aperture be positioned just after the main system output OPuS, e.g., a 4× OPus. A ring power amplification stage aperture may be located about 500 mm further inside the laser system. This distance is too great to avoid pointing changes turning into position changes at the specified measurement plane (present system aperture). Instead the limiting system aperture can be located just after the OPuS, and may have a 193 nm reflecting dielectric coating instead of a stainless steel plate commonly used. This design can allow for easier optical alignment, while at the same time reduce heating of this aperture.
According to aspects of an embodiment of the subject matter disclosed, applicants propose to implement a relatively stress-free chamber window arrangement similar to or the same as that discussed in the above referenced co-pending U.S. patent application, e.g., at least on the bean reverser side of the chamber, because of the use of, e.g., a PCCF coated window a this location.
According to aspects of an embodiment of the subject matter disclosed, applicants propose to, e.g., place ASE detection, e.g., backward propagation ASE detection, in either the LAM or in an MO wavefront engineering box (“WEB”), or in a so-called MOPuS, which can, e.g., include elements of the MOWEB from applicants' assignee's existing XLA-XXX model laser systems along with the mini-OPuSs discussed elsewhere in this application and in the co-pending application Attorney Docket No. 2005-0105-02 referenced herein, as well as, e.g., beam expansion, e.g., using one or more beam expansion prisms to expand the output beam of the MO in its short axis, e.g., to form generally a square cross-sectional beam. The current MO WEB and its beam turning function is represented schematically as the turning mirror, e.g., 44 shown in
According to aspects of an embodiment of the disclosed subject matter one or more mini-OPuS(s), which may be confocal, such that they are highly tolerant to misalignment and thus of potentially low aberration, e.g., for the off-axis rays needed in the proposed short OPuS(s), the so-called mini-OPuS, can have delay times of 4 ns and 5 ns respectively, where more than one is employed. These values were chosen so that both OPuSs exhibit low wavefront distortion with spherical optics in addition to appropriate delay paths for coherence busting. The low wavefront requirement may actually prevent significant speckle reduction from the mini-OPuS(s) unless an angular fan-out from the output of the mini-OPuS(s) is generated, e.g., by replacing a flat/flat compensating plate with a slightly wedged plate, so that the transmitted beam and the delayed beam in the mini-OPuS are slightly angularly offset from each other. The laser beam, e.g., from the master oscillator is partially coherent, which leads to speckle in the beam. Angularly offsetting the reflected beam(s) reentering the mini-OPuS output with the transmitted beam, along with the delay path separation of the main pulse into the main pulse and daughter pulses, can achieve very significant speckle reduction, e.g., at the wafer or at the annealing workpiece, arising from the reduction in the coherence of the laser light source pulse illuminating the workpiece (wafer or crystallization panel). This can be achieved, e.g., by intentionally misaligning the delay path mirrors, probably not possible with a confocal arrangement, but also with the addition of a slight wedge in the delay path prior to the beam splitter reflecting part of the delayed beam into the output with the transmitted beam and its parent pulse and preceding daughter pulses, if any. For example, a 1 milliradian wedge in the plate will produce an angular offset in the reflected daughter pulse beam of 0.86 milliradians.
The optical delay path(s) of the mini-OPuS(s) may have other beneficial results in terms of laser performance and efficiency. According to aspects of an embodiment of the disclosed subject matter, as illustrated schematically in
The length of the delay path 506 will delay the beam pulses so that there is a slight temporal shift between the part of the beam that is transmitted and the part that is reflected, e.g., more than the coherence length, but much less than the pulse length, e.g., about 1-5 ns. By selecting the appropriate path length, which determines the delay time, the addition of the two beams can be such that the energy in the pulse is spread into a slightly longer Tis, which in combination with later pulse stretching in the main OPuS(s) can improve laser performance, as well as providing other beneficial laser performance benefits.
Two mini-OPuSs may be needed to achieve the desired effect. The offset time between the pulses from the two mini-OPuss may be, e.g., one nanosecond. Based upon optical and mechanical considerations, the delays selected for the stretchers may be, e.g., a 3 ns delay path in the first mini-OPu s and a 4 ns delay path in the second. If the delay is shorter, the optical system, e.g., if it uses confocal or spherical mirrors can introduce unacceptable aberrations. If the delay is longer, it may be difficult to fit the system into the available space in the laser cabinet. The distance the beam must travel to achieve the 3 ns delay is 900 mm and to delay by 4 ns is 1200 mm. A confocal optical system 520, minimizing the sensitivity to misalignment, illustrated schematically in
The delay path time(s) in the mini-OPuS(s) for coherence busting and other purposes may be as short as about the temporal coherence length and as long as practical due to the noted optical and space considerations, such as misalignment and aberration tolerance. If there are two or more mini-OPuSs then the delay path in each must be different in length, e.g., by more than the coherence length and selected such that there is no significant coherence reaction (increase) due to the interaction of daughter pulses from the separate OPuS(s). For example the delay path times could be separated by at least a coherence length and by not more than some amount, e.g., four or five coherence lengths, depending on the optical arrangement.
According to aspects of an embodiment of the subject matter disclosed applicants propose to employ a coherence-busting optical structure that, e.g., generates multiple sub-pulses delayed sequentially from a single input pulse, wherein also each sub-pulse is delayed from the following sub-pulse by more than the coherence length of the light, and in addition with the pointing of each sub-pulse intentionally chirped by an amount less than the divergence of the input pulse. In addition applicants propose to utilize a pair of coherence-busting optical delay structures, where the optical delay time difference between the pair of optical delay structures is more than the coherence length of the input light. Each of the two optical delay structures may also generate sub-pulses with controlled chirped pointing as noted in regard to the aspects of the previously described coherence busting optical delay structure.
According to aspects of an embodiment of the disclosed subject matter two imaging mini-OPuSs, which may be confocal, such that they are highly tolerant to misalignment and thus of potentially low aberration, e.g., for the off-axis rays needed in the proposed short OPuSs, the so-called mini-OPuSs, and can have delay times of 4 ns and 5 ns respectively. These values were chosen so that both OPuSs exhibit low wavefront distortion with spherical optics. The low wavefront requirement may prevent significant speckle reduction from the mini-OPuSs unless an angular fan-out from the mini-OPuSs is generated, e.g., by replacing a flat/flat compensating plate with the slightly wedged plate.
It will be understood by those skilled in the art that according to aspects of an embodiment of the disclosed subject matter, adequate coherence busting may be achieved sufficiently to significantly reduce the effects of speckle on the treatment of a workpiece being exposed to illumination from the laser system, such as in integrated circuit photolithography photoresist exposure (including the impact on line edge roughness and line width roughness) or laser heating, e.g., for laser annealing of amorphous silicon on a glass substrate for low temperature recrystallization processes. This may be accomplished by, e.g., passing the laser beam, either from a single chamber laser system or from the output of a multi-chamber laser system or from the seed laser in such a multi-chamber laser system before amplification in another chamber of the multi-chamber laser system, through an optical arrangement that splits the output beam into pulses and daughter pulses and recombines the pulses and daughter pulses into a single beam with the pulses and daughter pulses angularly displaced from each other by a slight amount, e.g., between, e.g., about 50μ Rad and 500μ Rad and with each of the daughter pulses having been delayed from the main pulse(s), e.g., by at least the temporal coherence length and preferably more than the temporal coherence length.
This may be done in an optical beam delay path having a beam splitter to transmit a main beam and inject a portion of the beam into a delay path and then recombining the main beam with the delayed beam. In the recombination, the two beams, main and delayed, may be very slightly angularly offset from each other (pointed differently) in the far field, referred to herein as imparting a pointing chirp. The delay path may be selected to be longer than the temporal coherence length of the pulses.
The angular displacement may be accomplished using a wedge in the optical delay path prior to the delayed beam returning to the beam splitter which wedge imparts a slightly different pointing to the delayed beam (a pointing chirp). The amount of pointing chirp, as noted above may be, e.g., between about 50 and 500μ Rad.
The optical delay paths may comprise two delay paths in series, each with a respective beam splitter. In such an event each delay path can be different in length such that there is not created a coherence effect between the main and daughter pulses from the respective delay paths For example, if the delay in the first delay path is 1 ns the delay in the second delay path could be about 3 ns and if the delay in the first delay path is 3 ns the delay in the second could be about 4 ns.
The wedges in the two separate delay paths may be arranged generally orthogonally to each other with respect to the beam profile, such that the wedge in the first delay path can serve to reduce coherence (speckle) in one axis and the wedge in the other delay path can reduce coherence (speckle) in the other axis, generally orthogonal to the first. thus, the impact on speckle, e.g., contribution to line edge roughness (“LER”) and/or line width roughness (“LWR”), e.g., at the wafer in exposure of photoresist in an integrated circuit manufacturing process can be reduced along feature dimensions in two different axes on the wafer.
According to aspects of an embodiment of the subject matter disclosed, with, e.g., a 6 mRad cross of the bowtie in a bowtie ring power amplification stage, the magnification prisms inside the ring cavity may be slightly different for the in-going and outgoing beams, and could be arranged so that the beam grows slightly as it travels around the ring or shrinks slightly as it travels around the ring. Alternatively, and preferably according to aspects of an embodiment of the subject matter disclosed, a result of breaking the larger beam expansion prism into two separate pieces, e.g., enabled by larger spacing between out-going and in-coming beams, e.g., about 5-6 mm, as illustrated by way of example in
According to aspects of an embodiment of the subject matter disclosed applicants propose to place the Rmax, e.g., 164 and the OC, e.g., 162 portions of the version of the seed injection mechanism containing an Rmax 164 and an OC 162, e.g., along with the positioning of the system horizontal axis beam output aperture on that same stage. This enables, e.g., prior alignment of each as an entire unit and removes the need for field alignment of the individual components. This can allow, e.g., for the position of the Rmax/OC assembly, e.g., 160, shown in
According to aspects of an embodiment of the subject matter disclosed applicants propose to position a mechanical shutter to block the MO output from entering the ring, when appropriate, similar to such as are utilized on applicants' assignee's OPuSs, e.g., to block them during alignment and diagnosis. The exact location could be, e.g., just above the last folding mirror prior to the ring power amplification stage, where the mini-OPuSes are protected during unseeded ring power amplification stage alignment and operation.
Turning now to
The coherency buster 160a may have an input beam 162a incident on a beam splitter 164a, e.g., a partially reflective mirror 164a for the pertinent wavelength. Part of the beam 162a that is reflected into the delay path comprised of a plurality of mirrors, e.g., confocal mirrors 166a, is negatively imaged, e.g., twice, and on the final leg of the delay path pass through a pulse trimmer 170a. It will be understood that such optical coherence busters may have more than four mirrors, e.g., six mirrors, but are illustrated schematically with only four for convenience and clarity. A portion of the light exiting the pulse trimmer 170a is reflected into the output beam 172a and a portion reenters the delay path. The delay path may be much shorter than the seven to ten meters or so of, e.g., a 4× OPus, such that the second and third passes through the delay path do not substantially overlap the pulses entering and leaving the coherency buster 160a, but rather do not even substantially stretch the pulses. There may be, however, overlapping in the high frequency components of the pulses, which serves in coherence busting. The pulse trimmer 160a may be used, e.g., to shorten the ultimate output pulse 172a, e.g., by cutting off a portion of the pulse circulating in the coherency buster delay path using the pulse trimmer 170a, or much or all or substantially all of the second and subsequent passes through the delay path. The pulse trimmer 170a may be, e.g., a Pockels cell or other suitable fast acting light transmission switch, e.g., a light beam modulator/deflector, e.g., an electro-optic or acousto-optic device, e.g., a crystal that changes refractive index when excited by a field, e.g., an electric field, an acoustic field or a magnetic field.
At the output of the seed laser 372a is illustrated a representation of the seed laser output laser light pulse beam pulse coherency 374a containing a single dot indicative of relatively high coherency. The output of the seed laser 372a may be passed through one or more coherency busters, e.g., 376a, 378a, e.g., as shown by example in
The preferred embodiment uses a first delay something more than Ins due to increased alignment problems with the shorter delay and increased aberrations in the pulse as stretched in a shorter delay path. Each of the delay paths is, however longer than the coherence length of the pulse and the second delay path is longer than the first, to achieve coherence busting, e.g., due to high frequency mini-peak interleaving effects discussed herein.
The mini-OPuS pulse stretchers may be selected and arranged to, e.g., fold the beam on itself or fan it out in first one axis, e.g., in a first mini-OPus 376a, resulting in the coherency representation 378a and then in another orthogonally related axis, e.g., in a second mini-OPuS 380a, resulting, e.g., in the coherency representation 390a. A pulse trimmer/pulse steerer 392a, e.g., and electro-optical (“E-O”) element 392a may sweep (paint) the seed beam into the input/output coupler 400a of the amplifier portion 394a resulting in the blurring in one axis as shown in the pulse coherency representation out of the power oscillator 410a (and also the coherence representation 410 into the amplification gain stage 394a). The “regular” or “standard” OPuS, e.g., a 4× Tis OPuS (roughly tem meters of delay path), which may contain, e.g., 2 delay paths 412a, 420a initiated by a first beam splitter 414a and a second beam splitter 422a, similarly may be arranged to fold the beam on itself in first one axis and then a second resulting, e.g., in the pulse coherency representations of, respectively, 414a and 424a. The final coherency representation 424a shows schematically that the coherency of the seed beam has been greatly reduced, i.e., the beam has been smeared in its passage from the seed laser 372a to the amplifier gain medium 394a and as amplified in the amplifier gain medium 394a and subsequently further having its coherency busted in the 4× regular OPuS 412a, 420a.
It will be understood by those skilled in the art that depending on the initial coherency of the pulse, e.g., out of the seed laser, e.g., almost completely coherent in the case of solid state seed lasers to very little coherency, but still coherency that is desired to be even further reduced, e.g., with an excimer seed laser the type, number and arrangement of coherency busting elements may vary. For example, it may only be necessary to do active coherency busting, e.g., with one form or another of pulse steering/painting, for solid state seed lasers, and this may in some cases for some applications prove to need only a ramp or only AC pulse deflection, i.e., in one axis or the other, or may prove to need both DC and AC pulse painting (Hybrid painting) along with OPuS effect coherency busting both between the MO and amplifier gain medium, e.g., PO or PA or other amplification gain medium stage, e.g., a ring power amplification stage, and also may need to employ the effect of the regular OPuS pulse stretcher(s) on the output of the amplifier gain medium. With an excimer gas discharge laser MO, with relatively much lower coherency than from a solid state seed laser, only passive coherency busting, e.g., between the MO and gain amplifier medium may be needed, e.g., with one or both of the mini-OPuSs 376, 380 or other passive optical elements as noted above between the MO and amplifier gain medium.
One may still need, however, to do beam steering also, e.g., with an active beam steering mechanism for even more smearing of the pulse (more divergence), that may be less essential and need a smaller sweeping angle. Such a seed laser mini-OPuS is believed to need approximately only a 1 foot total path delay each and can also be conveniently built onto the seed laser optical table as is currently the practice for relay optics in applicants' assignee's XLA series laser systems.
According to aspects of an embodiment of the disclosed subject matter it is contemplated to apply a time changing voltage on a timescale similar to the seed pulse duration, e.g., by applying a DC voltage level until triggered, at which point the high voltage may be shorted to ground, e.g., via a stack of fast MOSFETS, e.g., illustrated schematically in
One can normalize these contrast values to the maximum value in order to evaluate the percentage reduction in contrast, e.g., brought about by the dynamic pointing shift. At the optimum relative timing point the speckle contrast was found to be reduced to about 40% of its peak. Using the 1/√{square root over (N)} assumption for equivalent number of independent pulses the data can be used to derive the number of pulses required to achieve this level of speckle contrast reduction. At the optimum relative timing, and with 3 kV applied to the E-O cell, the contrast reduction was found to be equivalent to 6 pulses. Even higher voltage levels (and thus even larger pointing shift during a single pulse) could improve this result. Applicants performed similar measurements with the seed laser pulse entering the MOPO amplification stage cavity, but no discharges between the AMPLIFICATION STAGE electrodes and noted that reflections from the OC and the Rmax in the XeF cavity, from the OPuS effect, beam spreading alone, indicated that the maximum speckle contrast was reduced by the amount predicted by the OPuS effect (N=1.56 with a 20% OC, giving 1/√{square root over (n)}=0.80. Thus 70% contrast becomes 56%). The effect of smearing, even though the initial speckle contrast is lower, appears not to change when adding the secondary reflections from the full XeF cavity. The equivalent pulse for speckle reduction is still about 6.
Applicants performed similar measurements with AMPLIFICATION STAGE cavity electrodes discharging and thus implicating the effects of the amplification within the AMPLIFICATION STAGE cavity, which indicated as shown in
According to aspects of an embodiment of the disclosed subject matter applicants propose to use a 6 mirror coherency busting mechanism (for convenience herein optical pulse delay paths are indicated schematically as having four mirrors per delay path) which has been developed by applicants' assignee for additional path delay inside either or both of the 1st or 2nd pulse stretchers in the OPuS used with applicants' assignee's XLA model multi-chamber laser systems. Such a delay path can, e.g., produce −1 imaging for each sub-pulse. This is illustrated schematically and in cartoon fashion, e.g., in
According to aspects of an embodiment of the disclosed subject matter it may be necessary to combine two separate laser beams at various points within a system according to aspects of an embodiment of the disclosed subject matter. If only half of the entrance to a 6 mirror pulse stretcher is illuminated, the sub-pulses flip between top and bottom as shown, e.g., in
Use of a solid state laser source for lithography has been proposed in the past and not pursued for two reasons. Solid state lasers are not considered capable of the high average power required for lithography and a solid state laser produces single mode output which is highly (perfectly) coherent. According to aspects of an embodiment of the disclosed subject matter applicants propose to address the low average power problem with, e.g., a hybrid solid state seed/excimer amplifier combination. The high coherence properties of the solid state seed can be addressed in a number of ways according to aspects of embodiments of the disclosed subject matter, e.g., by creating sub-pulses, e.g., that are separated in time longer than the coherence length, or by, e.g., changing the seed laser pointing, e.g., over very short time scales, e.g., within a single laser pulse, or a combination of both. Coherency busting has been found by applicants to be of benefit in dual chamber gas discharge (e.g. excimer) seed/gas discharge (e.g., excimer) amplifier portion lasers as well.
De-phasing of a speckle pattern can be seen from a diffuser 670a to occur with a λ/2 d where d is the illumination distance for a slot aperture and diameter for a circular aperture, e.g., as illustrated schematically and in cartoon fashion in
The effective number of equivalent independent laser pulses can be seen to be equal to the Tis magnification of the each pulse stretcher. Each OPuS pulse stretcher of the kind noted above may have a multiplication of around ˜2.4×. With, e.g., three stages of pulse stretching, the number of independent sub-pulses will be (2.4)3=13.8. Since speckle contrast scales with the number of independent sub-pulses, N, as 1/√{square root over (N)}, pulse stretchers can provide an output speckle contrast of 1√13.8=26.9% with an input speckle contrast of 100%. Since this may still be too high a speckle contrast, according to aspects of an embodiment of the disclosed subject matter a mechanism(s) may be provided to reduce the speckle contrast into or out of the pulse stretcher(s). The same can be said for the so-called mini-OPuSs discussed elsewhere.
Pulse trimming has been demonstrated, e.g., with the utilization of electro-optics, e.g., at 193 nm. Rather than polarization rotation, used in some other forms of pulse trimming, electro-optics can be used for beam steering, e.g., steering a seed laser light pulse beam within a single pulse in the beam. Utilization of such, e.g., at the output of the seed laser, can result in, e.g., according to aspects of an embodiment of the disclosed subject matter, the electro-optic material(s) only needing to be subject to a low average power seed laser beam. By, e.g., randomly and/or continuously changing the beam steering, e.g., within a single laser pulse, the angular acceptance of the power amplification stage can be “painted” or filled in for each laser pulse. As a result, a main pulse can have a divergence set, e.g., by the MO/power amplification stage optical configuration and not, e.g., by the seed laser characteristics. A greatly reduced coherence for the laser system output laser light pulse can be the result.
According to aspects of an embodiment of the disclosed subject matter an injection controlled amplifier laser system, e.g., with a plane cavity and flat rear mirror, may have suitable energy stability, e.g., for seed pulse inject energies in the range of 0.0085 to 0.99 mJ. This energy of the beam may be, e.g., incident on the rear mirror of, e.g., a power amplification stage, which may form the input coupler from the seed laser. This reflector may have, e.g., about a 90% reflection and about 8% transmission. Therefore, the seeding energy entering the amplification stage cavity itself may be, e.g., about an order of magnitude smaller than what is incident onto the back reflector. With a ring cavity, especially with a partially reflecting seed injection mechanism according to aspects of an embodiment of the disclosed subject matter, discussed elsewhere herein, e.g., the input seed energy may be much less wasted, e.g., admitting around 80% of the seed laser light. An Rmax and OC can be in an F2 containing environment, and thus more robust, though, e.g., if polarization coupling is used, coupling efficiency may still be less than optimum for certain applications. A suitable architecture, e.g., in a MOPA configuration may be a 2-channel (“tic-toc”) solid state seed laser, e.g., a 3rd harmonic Nd:YLF MO or Nd:YAG system (tuned, e.g., to 351 nm) along with a pair of two 3-pass XeF PA modules. Such a system in a MOPO, e.g., a master oscillator/power amplification stage (such as a ring power oscillator amplification stage) configuration is also considered as an effective alternative. Such a two channel MOPO approach may be similar to the MOPA configuration, i.e., with two seeded power oscillators. Various coupling techniques could be used, e.g., MO coupling using a polarization technique or a seed inject mechanism. Efficiency v. Emo for differing PO/PA configurations has been found to be better for a MOPO or a three pass MOPA, though four pass MOPAs were not tested. Exemplary pulse width (FWHM) has been found to be for an MOPO about 17.3 ns, for a MOPA, single pass, about 13.9 ns and for a MOPA triple pass about 12.7 ns.
Applicants have examined speckle patters for decorrelation with angular shift, e.g., in a MOPO output beam, e.g., with a Nd—YLF seed laser and a XeF power oscillator (e.g., a flat-flat polarization coupled arrangement). With the relative timing between the XeF discharge and the seed laser pulse adjusted and angular and spatial adjustment also made for maximum suppression of the weak line (353) produced by the XeF gain.
The maximum intensity of the seed pulse has been observed to occur during the initial, very low level, fluorescence of the amplification stage. This very low level fluorescence (and thus gain) is believed to be enhanced by this seed light, as observed in MOPO output. Adjustment of the timing of the seed earlier than or later than, e.g., about 20 or so ns before the amplification stage firing can, e.g., lead to an increase in weak line output, an indication of, e.g., when the “primordial” photons are generated in the AMPLIFICATION STAGE.
Turning now to
According to aspects of an embodiment of the disclosed subject matter, the pointing/divergence sensitivity of a pulse stretcher, e.g., a 4 mirror 6 mirror pulse stretcher, e.g., a regular OPuS such as a 4× Tis OPuS, or a so-called mini-OPuS, or the delay path discussed in more detail in regard to
One may assume that each ray is incoherent from all others, e.g., where the path length between the OC and the Rmax is maintained to be longer than the temporal coherence length. Each ray may also be assumed, e.g., to be angled slightly different from all others since, e.g., perfect alignment is believed to be extremely difficult, especially in the vertical direction. Applicants believe that about 37 μrad of angle difference in the vertical direction is needed to create uncorrelated speckle. Summing the normalized energy weighting to give the equivalent number of independent pulses and taking the square root to give the reduction in standard deviation, the sum from the above is 1.56. The square root is 1.25 and thus the standard deviation when using both OC and Rmax reflections is predicted to be 0.551/1.25=0.440, which comports well with a value that applicants have measured, i.e., 0.427.
Static fan out, otherwise referred to herein as a hall of mirrors effect, believed to be essentially unavoidable with manual alignment, produces a single pulse speckle contrast with amplification in an amplification gain medium that is 2.50× smaller than the seed laser alone. This reduction is the equivalent of 6.3 uncorrelated sub-pulses. Some of this contrast reduction is due to the weak line content from the XeF power oscillator used for testing the effects of the oscillation amplification stage, but most is believed to be due to the static fan out effect. Likely, many of the sub-pulses created by the OPuS-like static fan out characteristics of the OC-Rmax (OC-rear cavity mirror) reflections are all amplified to nearly equal intensities and thus create more equivalent independent pulses than shown in the above table.
Tilt angle required to produce uncorrelated speckle patterns may be significant. The first big jump in equivalent pulses, from 1.0 to 1.55, is believed by applicants to be mostly due to the poor pulse-to-pulse repeatability of the speckle patterns when running as a MOPO. Even without changing the mirror tilt at all, two pulses are correlated no better than 30-35%. With seed only, this pulse-to-pulse correlation has been found to be about 85-90%. The long slow rise in equivalent pulse number does not even reach a value of 2.0 until about 400 μrad of mirror tilt as illustrated, e.g., in
Through experimentation relating to coherence applicants' employer has learned that, e.g., sub-pulses produced by a pulse stretcher are incoherent and lead to a different fringe pattern if their angles are slightly shifted. The pin hole fringe pattern shifts maximum to minimum when input angle is λ/2 d.
A plot of pointing shift (inferred by applicants from speckle shift measurements) v. E-O cell applied voltage is shown in
Turning now to
Applicants' experimental measurements have determined that with no ramp and no AC voltage the 2D speckle contrast overall is 76.8% and varies from the horizontal to the vertical axis. With painting using the ramp alone the speckle contrast overall was 29.4%, again varying in the two axes. Painting with the AC alone gave a speckle contrast overall of 59.9%, again varying in the two axes. With the ramp and AC voltages applied the spectral contrast was 28.1% overall and varying in both axes. This was using a less optimized circuit than the one of
Applicants believe that a more optimized circuit, shown by way of example in
Turning to
It will be understood by those skilled in the art that disclosed herein is a method and apparatus which Applicants have simulated through calculations speckle reduction as relates to the location of coherence lengths within a single gas discharge (e.g., ArF or KrF excimer) laser system output pulse after such a pulse has passed through the two OPuS pulse stretchers sold on laser systems manufactured by applicants' assignee Cymer, Inc., used for pulse stretching to increase the total integrated spectrum (Tis) to reduce the impact of peak intensity in the laser output pulse on the optics in the tool using the output light from the laser system, e.g., a lithography tool scanner illuminator. There are two OPuS in series, with the first having a delay path sufficient to stretch the Tis of the output pulse from about 18.6 ns to about 47.8 ns and the second to stretch the pulse further to about 83.5 ns, e.g., measured at E955 (the width of the spectrum within which is contained 95% of the energy of the pulse.
Starting with the unstretched pulse, applicants divided the pulse into portions equal to the approximate coherence length, assuming a FWHM bandwidth of 0.10 pm and a Gaussian shape for the coherence length function. The impact of the pulse stretching on the coherence length portions of the pulse after passing through the first OPuS was to show that a first intensity hump in the spectrum of the stretched pulse was made up of the coherence length portions of the main pulse, a second intensity hump was mad up of coherence length portions of the main pulse overlapped with coherence length portions of a first daughter pulse. A third hump in the intensity spectrum is the result of overlapping of the first and second daughter pulses. Looking at the individual coherence length portions of the two humps applicants observed that the multiple versions (including daughters) of the coherence length portions remained sufficiently separated to not interfere with each other.
After passage through the second OPuS the simulated spectra, again only looking at the content of the first three humps in the stretched pulse, in the simulation (under the second hump were contributions from the original undelayed pulse, as before, the first delayed pulse from the first OPuS, as before and the first delayed pulse from the second OPuS), applicants observed that in this second pulse the multiple versions of the coherence length portions were very close together. This is caused by the fact that the first OPuS has a delay of ˜18 ns and the second has a delay of ˜22 ns. Thus only 4 ns separates the versions of the coherence length portions, which is still not close enough for interference.
Under the third hump applicants observed contributions from the first delayed pulse from the first OPuS, the second delayed pulse from first OPuS, the first delayed pulse from the second OPuS, and the second delayed pulse from second OPuS. applicants observed that the separation between some related coherence portions is larger than for others in the third hump in the intensity spectrum of the pulse stretched by two OPuSs. This increase in separation is due to the fact that two round trips through each OPuS equal ˜36 ns=18*2 and ˜44 ns=22*2. Thus the separation between coherence lengths grows with each round trip.
Applicants concluded that for a mini-OPuS as described in this application a single mini-OPuS with delay equal to one coherence length will create a train of pulses that dies out after about 4 coherence length values. Thus, applicants determined that for a single mini-OPuS to be effective, the two main OPuSs should not bring any daughter coherence lengths to within 4 coherence lengths of each other. but, applicants have observed in the simulation that the main OPuSs do just that, though only marginally so. The separation between coherence lengths for the third and greater humps is sufficient. Applicants believe that the impact of a single mini-OPuS between MO and amplification gain medium will be nearly the full expected coherence busting effect. A second mini-OPuS between MO and PA may not adequately interact with the two main OPuSs. The empty spaces, not filled with related coherence length portions of the spectra pulse humps get more scarce when one combines a single min-OPuS and two regular OPuSs, and the second may be too much. according to aspects of an embodiment of the present invention applicants propose the coordinated change of the regular OPuS delay lengths when the mini-OPuS(s) are installed, including whether they are part of the laser system or installed down stream of the regular main OPuSs, e.g., in the lithography tool itself. Applicants believe that such mini-OPuS(s) can fill in the valleys of the pulse duration somewhat, leading to an increase in Tis, e.g., allowing a reduction in the delay lengths of one of the two main OPuSes for better overall coherence length separation.
According to aspects of an embodiment of the disclosed subject matter, applicants propose to provide a method and apparatus for the reduction of and/or control of line-edge roughness (LER) and line-width roughness (LWR), collectively, feature dimension roughness, of exposed photoresist for integrated circuit manufacturing to improve the ability to print device feature critical dimensions (CD), e.g., using deep-UV (DUV) (sub-400 nm wavelengths) lithography projection systems noted elsewhere in the present application. A method and apparatus is proposed for reducing and/or controlling the line edge roughness and line width roughness during exposure of deep-UV lithographic process, in which, e.g., the speckle contrast and/or speckle size of the laser source light can be adjusted to minimize and/or control the magnitude and spatial frequency of the line edge roughness and/or line width roughness of critical dimensions produced using deep-UV projection lithography. The properties of the light illuminating the mask and exposing the PR can be controlled to accomplish the desired results.
According to aspects of an embodiment of the disclosed subject matter a method and apparatus is proposed to reduce/minimize dose variation along a pattern length, i.e., edge, i.e., a feature dimension A change in dose (intensity) variation (e.g., on the order of 0.05 mJ/cm2 to 2 mJ/cm2 can result in lateral movement of the PR pattern edge (e.g., on the order of 1 to 10 nm). Therefore, in order to reduce the LER/LWR it is desirable to reduce the dose variation along the image pattern edge. This can also impact the variation in line width, LWR by making both edges of the feature having the line width less rough. Reducing/minimizing speckle or spatial coherence of the laser beam, by any of the techniques noted in the present application and related applications as referenced above can be used to appropriately reduce speckle effects on the roughness along feature dimensions, e.g., in general of aimed at particular axis of the integrated circuit chip. Speckle, induced at least in part by coherence of the illumination pulse (or plurality of pulses), when projected in a photolithography system contributes to variation in intensity (or dose) along the image pattern edge. Reducing the coherence/speckle will reduce LER/LWR. Reducing the contrast (intensity modulation) of the speckle given fixed speckle size and a reduction in speckle contrast of each laser pulse (or plurality of pulses) contributes directly to the reduction in the cumulative dose (intensity) variation. The amount of reduction in dose variation can also depend on the number of pulses and correlation of the speckle fingerprint from pulse to pulse. By reducing the speckle size (increasing the spatial frequency of the speckle), e.g., with a generally fixed speckle contrast, can also be effective. Since the projection lens (objective), e.g., contained with a photolithography tool (e.g., a scanner) illuminator, has a fixed frequency cut-off, determined by the resolution of the lens (given by the numerical aperture, illumination condition and exposure wavelength), frequencies of speckle higher than the cut-off can be eliminated (i.e., not transmitted) by the optical system. Therefore if the speckle size is reduced (spatial frequency of the speckle is increased), the total contribution of the speckle to the projected dose (intensity) variation can be reduced, and/or even selected at some desired amount of roughness, which can in some case, along with such other CD definition techniques in photolithography, e.g., optical proximity correction techniques, may be employed to improve the overall ability to more precisely define CD shapes, uniformities, etc. Reduction of both speckle contrast and speckle size can be done, as noted elsewhere in the present application within the laser system, the lithography tool or intermediate the two, e.g., in a beam delivery unit, and can also be done with moving diffusers, vibrating lens, etc., as is know in the art for coherence reduction. Reduction in the speckle correlation, e.g., between pulses, i.e., the variation of the speckle fingerprint from pulse to pulse, which if caused to increase or decrease can cause the cumulative speckle contrast in an exposure comprised of a plurality of pulses to be reduced as governed by statistical averaging. If it is desirable to maintain the LER/LWR as a fixed quantity, e.g., over the image field, multiple fields or wafer to wafer exposure, a method is proposed whereby laser speckle contrast and size can be controlled (increased/decreased, when and as desired) based on the known variation in image contrast, e.g., as sensed by an image contrast sensor, e.g., to maintain a fixed level of LER/LWR, e.g., also in different axes.
Turning now to
The system may also include a sensor, e.g., an image contrast sensor 1372, such as a ______, made by ______, which can be arranged to detect the impact of speckle on the patterning of an integrated circuit as a whole, or portions thereof, or in selected axes, e.g., one generally parallel to main feature dimension extending in one axis of the integrated circuit and one in another axis, e.g., generally orthogonal to the first axis. the sensor output may be provided to a controller 1374, which may be part of the control system for the laser or the scanner or an overall control system for both the laser and scanner, and may provide a control signal, based on the feedback from the sensor 1372 to the coherence busting mechanism. The control signal may alter the operation of the coherence busting mechanism, e.g., by modifying the actuation signals for, e.g., a beam sweeping mechanism(s) in either or both axes, of modifying the position compensation plates 520, 532 in the mini-OPuS as discussed above, e.g., in regard to
Turning now to
It will be understood by those skilled in the art that there is disclosed herein an apparatus and a method which may comprise reducing feature dimension roughness in an exposed photoresist layer exposed by a pulsed sub-400 nm light source by reducing dose variation by increasing the spatial frequency of speckle for a given speckle contrast. The method and apparatus may comprise reducing dose variation by increasing the spatial frequency of speckle for a given speckle contrast from pulse to pulse and utilizing a projection lens with a frequency cut off below substantially all of the speckle spatial frequencies. The method and apparatus may comprise producing integrated circuits by controlling laser light source speckle size and/or contrast variation from a known variation in image contrast to select a level of feature dimension roughness. The method and apparatus may comprise utilizing a photolithography tool comprising a light source input opening receiving sub-400 nm light pulses from a pulsed excimer or molecular fluorine gas discharge laser and a wafer photoresist exposure platform holding a wafer; detecting image contrast in feature dimension at the wafer and providing a feedback signal indicative of the image contrast; controlling laser light source speckle size and/or contrast variation based on the feedback signal to select a level of feature dimension roughness. The apparatus and method may comprise reducing feature dimension roughness in an exposed photoresist layer exposed by a pulsed sub-400 nm light source using a coherence busting mechanism comprising a coherence busting optical structure generating multiple sub-pulses delayed sequentially from a single input pulse, wherein each sub-pulse is delayed from the following sub-pulse by more than the coherence length of the light pulse and a beam pointing displacement mechanism displace the beam pointing between a main pulse and a delayed pulses, such as, e.g., a mini-OPuS. The apparatus and method may comprise reducing feature dimension roughness in an exposed photoresist layer exposed by a pulsed sub-400 nm light source by utilizing a laser light source comprising a seed laser and an amplification gain stage; utilizing a coherence busting mechanism intermediate the seed laser and the amplification stage comprising a beam splitter separating the seed laser output into a main beam and a beam entering an optical delay path having a delay length longer than the coherence length of a pulse in the seed laser output and a beam angular offset mechanism offsetting a delayed beam from the delay path and the main beam. Each sub-pulse may be delayed substantially less than its pulse length such that the delay creates substantially no increase in the pulse length or in the Tis of the pulse. The coherence busting mechanism may comprise a first coherence busting optical delay structure and a first angular offset mechanism, and a second coherence busting optical delay structure and a second angular offset mechanism. The optical delay time in the second coherence busting optical delay structure may be longer than the delay in the first optical delay structure by more than the coherence length of the input light pulse. The first angular displacement mechanism may be oriented to angularly displace the main pulse and at least one delayed pulse in one axis and the second angular displacement mechanism oriented to angularly displace the main pulse and at least one delayed pulse in a second axis. The first axis may be different from the second axis. The method and apparatus may comprise the coherence busting mechanism intermediate the laser system and the photolithography tool, within the photolithography tool, e.g., between the laser light source entry opening and the illuminator or between the illuminator and the wafer, or within the laser light source system, e.g., intermediate the laser and an OPuS or intermediate the OPus and the photolithography tool, or intermediate the laser system and the photolithography tool, e.g., within a beam delivery unit.
While the particular aspects of embodiment(s) of the LASER SYSTEM described and illustrated in this patent application in the detail required to satisfy 35 U.S.C. §112 is fully capable of attaining any above-described purposes for, problems to be solved by or any other reasons for or objects of the aspects of an embodiment(s) above described, it is to be understood by those skilled in the art that it is the presently described aspects of the described embodiment(s) of the subject matter disclosed are merely exemplary, illustrative and representative of the subject matter which is broadly contemplated by the subject matter disclosed. The scope of the presently described and claimed aspects of embodiments fully encompasses other embodiments which may now be or may become obvious to those skilled in the art based on the teachings of the Specification. The scope of the present LASER SYSTEM is solely and completely limited by only the appended claims and nothing beyond the recitations of the appended claims. Reference to an element in such claims in the singular is not intended to mean nor shall it mean in interpreting such claim element “one and only one” unless explicitly so stated, but rather “one or more”. All structural and functional equivalents to any of the elements of the above-described aspects of an embodiment(s) that are known or later come to be known to those of ordinary skill in the art are expressly incorporated herein by reference and are intended to be encompassed by the present claims. Any term used in the specification and/or in the claims and expressly given a meaning in the Specification and/or claims in the present application shall have that meaning, regardless of any dictionary or other commonly used meaning for such a term. It is not intended or necessary for a device or method discussed in the Specification as any aspect of an embodiment to address each and every problem sought to be solved by the aspects of embodiments disclosed in this application, for it to be encompassed by the present claims. No element, component, or method step in the present disclosure is intended to be dedicated to the public regardless of whether the element, component, or method step is explicitly recited in the claims. No claim element in the appended claims is to be construed under the provisions of 35 U.S.C. §112, sixth paragraph, unless the element is expressly recited using the phrase “means for” or, in the case of a method claim, the element is recited as a “step” instead of an “act”.
It will be understood also be those skilled in the art that, in fulfillment of the patent statutes of the United States, applicant(s) has disclosed at least one enabling and working embodiment of each invention recited in any respective claim appended to the Specification in the present application and perhaps in some cases only one. For purposes of cutting down on patent application length and drafting time and making the present patent application more readable to the inventor(s) and others, applicant(s) has used from time to time or throughout the present application definitive verbs (e.g., “is”, “are”, “does”, “has”, “includes” or the like) and/or other definitive verbs (e.g., “produces,” “causes” “samples,” “reads,” “signals” or the like) and/or gerunds (e.g., “producing,” “using,” “taking,” “keeping,” “making,” “determining,” “measuring,” “calculating” or the like), in defining an aspect/feature/element of, an action of or functionality of, and/or describing any other definition of an aspect/feature/element of an embodiment of the subject matter being disclosed. Wherever any such definitive word or phrase or the like is used to describe an aspect/feature/element of any of the one or more embodiments disclosed herein, i.e., any feature, element, system, sub-system, component, sub-component, process or algorithm step, particular material, or the like, it should be read, for purposes of interpreting the scope of the subject matter of what applicant(s) has invented, and claimed, to be preceded by one or more, or all, of the following limiting phrases, “by way of example,” “for example,” “as an example,” “illustratively only,” “by way of illustration only,” etc., and/or to include any one or more, or all, of the phrases “may be,” “can be”, “might be,” “could be” and the like. All such features, elements, steps, materials and the like should be considered to be described only as a possible aspect of the one or more disclosed embodiments and not as the sole possible implementation of any one or more aspects/features/elements of any embodiments and/or the sole possible embodiment of the subject matter of what is claimed, even if, in fulfillment of the requirements of the patent statutes, applicant(s) has disclosed only a single enabling example of any such aspect/feature/element of an embodiment or of any embodiment of the subject matter of what is claimed. Unless expressly and specifically so stated in the present application or the prosecution of this application, that applicant(s) believes that a particular aspect/feature/element of any disclosed embodiment or any particular disclosed embodiment of the subject matter of what is claimed, amounts to the one an only way to implement the subject matter of what is claimed or any aspect/feature/element recited in any such claim, applicant(s) does not intend that any description of any disclosed aspect/feature/element of any disclosed embodiment of the subject matter of what is claimed in the present patent application or the entire embodiment shall be interpreted to be such one and only way to implement the subject matter of what is claimed or any aspect/feature/element thereof, and to thus limit any claim which is broad enough to cover any such disclosed implementation along with other possible implementations of the subject matter of what is claimed, to such disclosed aspect/feature/element of such disclosed embodiment or such disclosed embodiment. Applicant(s) specifically, expressly and unequivocally intends that any claim that has depending from it a dependent claim with any further detail of any aspect/feature/element, step, or the like of the subject matter of what is claimed recited in the parent claim or claims from which it directly or indirectly depends, shall be interpreted to mean that the recitation in the parent claim(s) was broad enough to cover the further detail in the dependent claim along with other implementations and that the further detail was not the only way to implement the aspect/feature/element claimed in any such parent claim(s), and thus be limited to the further detail of any such aspect/feature/element recited in any such dependent claim to in any way limit the scope of the broader aspect/feature/element of any such parent claim, including by incorporating the further detail of the dependent claim into the parent claim.
It will be understood by those skilled in the art that the aspects of embodiments of the subject matter disclosed above are intended to be preferred embodiments only and not to limit the disclosure of the subject matter disclosed(s) in any way and particularly not to a specific preferred embodiment alone. Many changes and modification can be made to the disclosed aspects of embodiments of the disclosed subject matter disclosed(s) that will be understood and appreciated by those skilled in the art. The appended claims are intended in scope and meaning to cover not only the disclosed aspects of embodiments of the subject matter disclosed(s) but also such equivalents and other modifications and changes that would be apparent to those skilled in the art. In additions to changes and modifications to the disclosed and claimed aspects of embodiments of the subject matter disclosed(s) noted above others could be implemented.
The present application is a Continuation-in-Part of U.S. patent application Ser. No. 11/787,180, filed on Apr. 13, 3007, entitled LASER SYSTEM, Attorney Docket No. 2006-0083-04, which was a Continuation-in-Part of U.S. patent application Ser. No. 11/584,792, filed on Oct. 20, 2006, entitled LASER SYSTEM, Attorney Docket No. 2006-0083-01, which claimed priority to U.S. Provisional Application Ser. No. 60/732,688, filed on Nov. 1, 2005, entitled 200 W GAS DISCHARGE EXCIMER OR MOLECULAR FLUORINE MULTICHAMBER LASER, Attorney Docket No. 2005-0094-01, and to Ser. No. 60/814,293 filed on Jun. 16, 2006, entitled 200 WATT DUV GAS DISCHARGE LASER SYSTEM Attorney Docket No. 2005-0103-01, and to Ser. No. 60/814,424, filed on Jun. 16, 2006, entitled LONG LIVED MO IN MOPO CONFIGURED LASER SYSTEM, Attorney Docket No. 2006-0012-01, and was a Continuation-in-Part of U.S. patent application Ser. Nos. 11/521,904, filed on the Sep. 14, 2006, entitled LASER SYSTEM, Attorney Docket No. 2005-0103-02; and 11/522,052, filed on Sep. 14, 2006, entitled LASER SYSTEM, Attorney Docket No. 2005-0104-01; and 11/521,833, filed on Sep. 14, 2006, entitled LASER SYSTEM, Attorney Docket No. 2005-0105-01; and 11/521,860, filed on Sep. 14, 2006, entitled LASER SYSTEM, Attorney Docket No. 2006-0007-01; and 11/521,834, filed on Sep. 14, 2006, entitled LASER SYSTEM, Attorney Docket No. 2006-0012-02; and 11/521,906, filed on Sep. 14, 2006, entitled LASER SYSTEM, Attorney Docket No. 2006-0013-01; and 11/521,858, filed on Sep. 14, 2006, entitled LASER SYSTEM, Attorney Docket No. 2006-0018-01; and 11/521,835, filed on Sep. 14, 2006, entitled LASER SYSTEM, Attorney Docket No. 2006-0020-01; and 11/521,905, entitled LASER SYSTEM, filed Sep. 14, 2006, Attorney Docket No. 2006-0071-01, the disclosures of each of which are hereby incorporated by reference. The present application is related to U.S. patent application Ser. No. 11/447,380, entitled DEVICE AND METHOD TO STABILIZE BEAM SHAPE AND SYMMETRY FOR HIGH ENERGY PULSED LASER APPLICATIONS, filed on Jun. 5, 2006, Attorney Docket No. 2006-0039-01, and is related to U.S. patent application Ser. No. 10/881,533, entitled METHOD AND APPARATUS FOR GAS DISCHARGE LASER OUTPUT LIGHT COHERENCY REDUCTION, filed on Jun. 29, 2004, and published on Dec. 29, 2005, Pub. No. 20050286599, the disclosures of which are hereby incorporated by reference. The present application is also related to U.S. Pat. Nos. 6,549,551, issued on Apr. 15, 2003, to Ness et al, entitled INJECTION SEEDED LASER WITH PRECISE TIMING CONTROL; and 6,567,450, issued on May 20, 2003, to Myers et al, entitled VERY NARROW BAND TWO CHAMBER HIGH REP RATE GAS DISCHARGE LASER SYSTEM; and 6,625,191, entitled VERY NARROW BAND TWO CHAMBER HIGH REP RATE GAS DISCHARGE LASER SYSTEM, issued on Sep. 23, 2003 to Knowles et al; and 6,865,210, issued on Mar. 8, 2005, to Ershov et al, entitled TIMING CONTROL FOR TWO CHAMBERED GAS DISCHARGE LASER SYSTEM; and 6,690,704, entitled CONTROL SYSTEM FOR TWO CHAMBER GAS DISCHARGE LASER SYSTEM, issued on Feb. 10, 2004 to Fallon et al; and 6,561,263, issued on May 6, 2003, to Morton et al., entitled DISCHARGE LASER HAVING ELECTRODES WITH SPUTTER CAVITIES AND DISCHARGE PEAKS; and U.S. Pat. No. 6,928,093, entitled LONG DELAY AND HIGH TIS PULSE STRETCHER, issued to Webb et al. on Aug. 9, 2005; the present application is also related to co-pending U.S. patent application Ser. No. 10/781,251, filed on Feb. 18, 2004, entitled VERY HIGH ENERGY, HIGH STABILITY GAS DISCHARGE LASER SURFACE TREATMENT SYSTEM, Attorney Docket No. 2003-0105-02; and Ser. No. 10/884,547, filed on Jul. 1, 2004, entitled LASER THIN FILM POLY-SILICON ANNEALING SYSTEM, Attorney Docket No. 2004-0062-01, published on Jun. 30, 2005, Pub. No. US-2005-0141580; and Ser. No. 11/173,988, entitled ACTIVE BANDWIDTH CONTROL FOR A LASER, filed on Jun. 30, 2005, Attorney Docket No. 2005-0001-01; and to Ser. No. 11/169,203, entitled HIGH PULSE REPETITION RATE GAS DISCHARGE LASER, filed on Jun. 27, 2005, Attorney Docket No. 2004-0094-01; and to Ser. No. 11/095,293, entitled GAS DISCHARGE LASER OUTPUT LIGHT BEAM PARAMETER CONTROL, filed on Mar. 31, 2005, Attorney Docket No. 2004-0119-01; and Ser. No. 11/095,976, entitled 6 KHZ AND ABOVE GAS DISCHARGE LASER SYSTEM, filed on Mar. 31, 2005; and Ser. No. 11/201,877, filed on Aug. 11, 2005, Attorney Docket No. 2004-0063-01, entitled LASER THIN FILM POLY-SILICON ANNEALING OPTICAL SYSTEM, Published on Dec. 8, 2005, Pub. No. US-2005-0269300; and Ser. No. 11/254,282, entitled METHOD AND APPARATUS FOR GAS DISCHARGE LASER BANDWIDTH AND CENTER WAVELENGTH CONTROL, Attorney Docket No. 2005-0076-01; and Ser. No. 11/346,519, filed on Feb. 1, 2006, entitled, VERY NARROW BAND, TWO CHAMBER, HIGH REP RATE GAS DISCHARGE LASER SYSTEM, Attorney Docket No. 2001-0090-16; and Ser. No. 11/323,604, filed on Dec. 29, 2005, entitled MULTI-CHAMBER GAS DISCHARGE LASER BANDWIDTH CONTROL THROUGH DISCHARGE TIMING; and Ser. No. 11/363,116, entitled VERY HIGH REPETITION RATE NARROW BAND GAS DISCHARGE LASER SYSTEM, filed on Feb. 27, 2006, Attorney Docket No. 2003-0129-04; and Ser. No. 10/881,533, entitled METHOD AND APPARATUS FOR GAS DISCHARGE LASER OUTPUT LIGHT COHERENCY REDUCTION, filed on Jun. 30, 2004, Attorney Docket No. 2003-0120-01; and Ser. No. 10/847,799, entitled LASER OUTPUT LIGHT PULSE STRETCHER, filed on May 18, 2004, Attorney Docket No. 2003-0121; and U.S. patent application Ser. No. 11/394,512, entitled CONFOCAL PULSE STRETCHER, filed on Mar. 31, 2006, Attorney Docket No. 2004-0144-01; the disclosures of each of which are incorporated herein by reference.
| Number | Date | Country | |
|---|---|---|---|
| 60814293 | Jun 2006 | US | |
| 60814424 | Jun 2006 | US | |
| 60732688 | Nov 2005 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 11787180 | Apr 2007 | US |
| Child | 11981450 | US | |
| Parent | 11584792 | Oct 2006 | US |
| Child | 11787180 | US | |
| Parent | 11521904 | Sep 2006 | US |
| Child | 11584792 | US | |
| Parent | 11522052 | Sep 2006 | US |
| Child | 11521904 | US | |
| Parent | 11521833 | Sep 2006 | US |
| Child | 11522052 | US | |
| Parent | 11521860 | Sep 2006 | US |
| Child | 11521833 | US | |
| Parent | 11521834 | Sep 2006 | US |
| Child | 11521860 | US | |
| Parent | 11521906 | Sep 2006 | US |
| Child | 11521834 | US | |
| Parent | 11521858 | Sep 2006 | US |
| Child | 11521906 | US | |
| Parent | 11521835 | Sep 2006 | US |
| Child | 11521858 | US | |
| Parent | 11521905 | Sep 2006 | US |
| Child | 11521835 | US |
