Gas performance control system for gas discharge lasers

[0002] This application is a Continuation-in-Part application which claims the benefit of priority to United States patent application no. 09/418,052, filed Oct. 14, 1999, now U.S. Pat. No. 6,243,406, which claims the benefit of priority to United States provisional patent application No. 60/123,928, filed Mar. 12, 1999.

[0003] 1. Field of the Invention

[0004] The present invention relates to a method and apparatus for stabilizing output beam parameters of a gas discharge laser. More particularly, the present invention relates to maintaining an optimal gas mixture composition by performing gas replenishment based on a monitored amplified spontaneous emission (ASE) signal which varies with the gas mixture status.

[0005] 2. Discussion of the Related Art

[0006] Pulsed gas discharge lasers such as excimer and molecular lasers emitting in the deep ultraviolet (DUV) or vacuum ultraviolet (VUV) have become very important for industrial applications such as photolithography. Such lasers generally include a discharge chamber containing two or more gases such as a halogen and one or two rare gases. KrF (248 nm), ArF (193 nm), XeF (350 nm), KrCl (222 nm), XeCl (308 nm), and F₂(157 nm) lasers are examples.

[0007] The efficiencies of excitation of the gas mixtures and various parameters of the output beams of these lasers vary sensitively with the compositions of their gas mixtures. An optimal gas mixture composition for a KrF laser has preferred gas mixture component ratios around 0.1% F₂/1% Kr/98.9% Ne. An F₂laser has a preferred gas component ratio around 0.1% F₂/99.9% Ne (or He, or combination thereof) (see U.S. Pat. No. 6,157,662, which is hereby incorporated by reference). Any deviation from these optimum gas compositions or those for other excimer or molecular lasers would typically result in instabilities or reductions from optimal of one or more output beam parameters such as beam energy, energy stability, temporal pulse width, temporal and spatial coherence, bandwidth, and long and short axial beam profiles and divergences.

[0008] Especially important in this regard is the concentration (or partial pressure) of the halogen containing molecules, e.g., F₂or HCl, in the gas mixture. The depletion of the rare gases, e.g., Kr and Ne for a KrF laser, is low in comparison to that for the F₂, although these rare gases are also replenished at longer intervals.

[0009] In industrial applications, it is advantageous to have an excimer or molecular laser capable of operating continuously for long periods of time, i.e., having minimal downtime. It is desired to have an excimer or molecular laser capable of running non-stop year round, or at least having a minimal number and duration of down time periods for scheduled maintenance, while maintaining constant output beam parameters. Uptimes of, e.g., greater than 98% require precise control and stabilization of output beam parameters, which in turn require precise control of the composition of the gas mixture.

[0010] Unfortunately, halogen depletion due to electrode erosion occurs in excimer and molecular lasers due to the aggressive nature of the fluorine or chlorine in the gas mixture. The halogen gas is highly reactive and its concentration in the gas mixture decreases as it reacts, leaving traces of contaminants. The halogen gas reacts with materials of the discharge chamber or tube. Moreover, the reactions take place and the gas mixture degrades whether the laser is operating (discharging) or not. The static gas lifetime is about one week for a typical KrF-laser. For a static KrF laser gas mixture, i.e., with no discharge running, increases in the concentrations of HF, O₂, CO₂and SiF₄have been observed (see Jurisch et al., above). During operation of a KrF-excimer laser, such contaminants as HF, CF₄, COF₂, SiF₄have been observed to increase in concentration rapidly (see G. M. Jursich et al., Gas Contaminant Effects in Discharge-Excited KrF Lasers,Applied Optics, Vol. 31, No. 12, pp. 1975-1981 (Apr. 20, 1992)).

[0011] One way to effectively reduce this gas degradation is by reducing or eliminating contamination sources within the laser discharge chamber. With this in mind, an all metal, ceramic laser tube has been disclosed (see D. Basting et al., Laserrohr für halogenhaltige Gasentladungslaser” G 295 20 280.1, Jan. 25, 1995/Apr. 18, 1996 (disclosing the Lambda Physik LP Novatube); and German Gebrauchsmuster DE 297 13 755.7 of Lambda Physik (disclosing an excimer or molecular laser having a window mount with a metal ceiling), each of which is hereby incorporated by reference). Gas purification systems, such as cryogenic gas filters (see U.S. Pat. No. 4,534,034) or electrostatic particle filters (see U.S. Pat. No. 5,586,134) are also being used to extend KrF laser gas lifetimes to 100 million shots before a new fill is advisable.

[0012] Notwithstanding using improved laser tubes and cryogenic purification systems, the most important techniques for stabilizing gas mixtures include gas replenishment procedures, whereby halogen injections (HI), rare gas injections, partial gas replacements (PGR) and total pressure adjustments are performed. The specific gas replenishment amounts and intervals are determined based on this evolution of the composition of the gas mixture, which is most prominently determined by depletion of the halogen species. U.S. provisional patent application no. 60/124,785 discloses the preferred gas replenishment procedure of the present invention and is hereby incorporated by reference. The '785 provisional application discloses to use several small gas actions such as micro-halogen injections (_HI) and mini-gas replacements (MGR_i) which are gas replacement procedures performed at multiple carefully selected intervals and amounts, in addition to total pressure adjustments, partial gas replacements (PGR) and new fills.

[0013] It is not easy, however, to directly measure the halogen concentration within the laser tube for making rapid online adjustments (see U.S. Patent No. 5,149,659 (disclosing monitoring chemical reactions in the gas mixture) and Japanese Patent No. JP 10341050 (disclosing a method wherein optical detection of a halogen specific emission is performed)). Therefore, advantageous methods applicable to industrial laser systems include using a known relationship between F₂or HCl concentration and a laser parameter. In such a method, precise values of the parameter would be directly measured, and the F₂or HCl concentration would be calculated from those values. In this way, the F₂concentration may be indirectly monitored.

[0014] Previously described methods for laser gas characterization include measuring the spectral width of the laser emission (see U.S. Pat. No. 5,450,436 to Mizoguchi et al.), measuring the spatial beam profile of the laser emission (see U.S. Pat. No. 5,642,374 to Wakabayashi et al.) and measuring other characteristics of the output beam such as bandwidth, coherence, driving voltage or energy stability wherein a rough estimation of the status of the gas mixture may be made (see U.S. Pat. No. 5,440,578 to Sandstrom, U.S. Pat. No. 5,887,014 to Das and U.S. patent application Ser. No. 09/379,034, which is assigned to the same assignee, each patent and patent application being hereby incorporated by reference).

[0015] In the Vogler Application (09/379,034), a data set of an output parameter such as pulse energy and input parameter such as driving voltage is measured and compared to a stored “master” data set corresponding to an optimal gas composition such as is present in the discharge chamber after a new fill. From a comparison of the data values and/or the slopes of curves generated from the data sets, a present gas mixture status and appropriate gas replenishment procedures, if any, are determined and undertaken to reoptimize the gas mixture.

[0016] It is desired to improve upon all of the above-described techniques by performing gas control based on the monitored values of a parameter that varies with a known correspondence with the halogen content in the gas mixture. Moreover, the desired parameter would not be greatly influenced by other laser system and output beam parameters, such as resonator alignment, repetition rate, tuning accuracy, thermal conditions, aging of the laser tube and degradation of optics.

[0017] Another technique uses a mass spectrometer for precision analysis of the gas mixture composition (see U.S. Pat. No. 5,090,020 to Bedwell). However, a mass spec is an undesirably hefty and costly piece of equipment to incorporate into a continuously operating excimer or molecular laser system such as a KrF, ArF or F₂laser system which are typical light sources used in microlithographic stepper or scanner systems. Yet another technique measures fluorine concentration in a gas mixture via monitoring chemical reactions (see U.S. Pat. No. 5,149,659 to Hakuta et al.), but this method is not suitable for use with a rapid online correction procedure. It is desired to have a precise technique for monitoring gas mixture status which is easily adaptable with current excimer or molecular laser systems and provides rapid online information.

[0018] In typical gas discharge lasers such as excimer or molecular fluorine lasers, a constant laser pulse energy is maintained in short-term notwithstanding the degradation of the gas mixture by regulating the driving voltage applied to the discharge. As mentioned above, long term regulation is achieved by gas replenishment actions such as halogen injections (HI), total pressure adjustments and partial gas replacement (PGR). The smoothed long-term stabilization of the gas mixture composition uses a regulation loop where input laser system parameter data are processed by a computer (see U.S. patent application Ser. No. 09/167,653 to Vogler et al., above).

[0019] In these typical laser systems, an energy detector is used to monitor the energy of the output laser beam. The computer receives the pulse energy data from the energy detector as well as driving voltage information from the electrical pulse power module. This information is not selective enough since the energy monitor signal is influenced not only by the gas but also by resonator optics degradation or misalignment. The typical operation mode is the so-called energy-constant mode where the pulse energy is kept constant by adjusting the driving high voltage of the electrical pulse power module. In this way one gets constant values from the energy monitor. A change of the laser status which again can be caused by gas aging as well as by the status of the laser resonator leads to a change of the driving voltage. It is desired to operate the laser at an approximately constant driving voltage level. To achieve this an appropriate smoothed gas regulation procedure is necessary. In the example of an excimer laser usually the halogen gas component (F₂in KrF lasers) is depleted whereas the other gases (nobel gases Kr and Ne in KrF lasers) are usually not depleted. Therefore μHI's or other suitable smoothed gas actions are applied (see U.S patent application Ser. No. 09/447,882, which is assigned to the same assignee and is hereby incorporated by reference).

[0020] An excimer or molecular fluorine laser system is provided including a discharge chamber containing a gas mixture, multiple electrodes connected to a power supply circuit for energizing the gas mixture, a resonator for generating a laser beam, a processor, and means for monitoring an amplified spontaneous emission (ASE) signal of the laser. The processor receives a signal from the ASE monitoring means indicative of the ASE signal of the laser. Based on the signal, the processor determines whether to initiate a responsive action for adjusting a parameter of the laser system.

[0021] An excimer or molecular fluorine laser system is also provided including a discharge chamber containing a gas mixture, multiple electrodes connected to a power supply circuit for energizing the gas mixture, a resonator for generating a laser beam, a processor, and an amplified spontaneous emission (ASE) detector. The processor receives a signal from the ASE detector indicative of the ASE signal of the laser. Based on the signal, the processor determines whether to initiate a responsive action for adjusting a parameter of the laser system.

[0022] An excimer or molecular fluorine laser system is further provided including a discharge chamber containing a gas mixture, multiple electrodes connected to a power supply circuit for energizing the gas mixture, a resonator for generating a laser beam, a processor, and an amplified spontaneous emission (ASE) detector. The processor receives a signal from the ASE detector indicative of the ASE signal of the laser. The ASE detector detects a filtered signal, wherein a substantial portion of a stimulated emission signal of the laser beam is filtered from the laser beam prior to monitoring said ASE signal. The ASE detector includes a stimulated emission filter for substantially filtering the stimulated emission from a portion of the laser beam to permit the ASE signal to be resolved from the laser beam.

[0023] An excimer or molecular fluorine laser system is also provided including a discharge chamber containing a gas mixture, multiple electrodes connected to a power supply circuit for energizing the gas mixture, a resonator for generating a laser beam, a processor, and an amplified spontaneous emission (ASE) detector. The processor receives a signal from the ASE detector indicative of the ASE signal of the laser. The ASE detector detects a filtered signal, wherein a substantial portion of a stimulated emission signal of the laser beam is filtered from the laser beam prior to monitoring the ASE signal. The ASE detector includes a stimulated emission filter for substantially filtering the stimulated emission from a portion of the laser beam to permit the ASE signal to be resolved from the laser beam. The stimulated emission filter includes a central axis beam dump centrally positioned with the optical axis of the portion of the laser beam. The dimensions of the beam dump are set such that the stimulated emission is substantially blocked and mostly ASE radiation passes unblocked around the beam dump.

[0024] Additionally, an excimer or molecular fluorine laser system is provided including a discharge chamber containing a gas mixture, multiple electrodes connected to a power supply circuit for energizing the gas mixture, a resonator for generating a laser beam, a processor, and an amplified spontaneous emission (ASE) detector. The processor receives a signal from the ASE detector indicative of the ASE signal of the laser. The ASE detector detects a filtered signal, wherein a substantial portion of a stimulated emission signal of the laser beam is filtered from the laser beam prior to monitoring the ASE signal. The ASE detector includes a stimulated emission filter for substantially filtering the stimulated emission from a portion of the laser beam to permit the ASE signal to be resolved from the laser beam. The stimulated emission filter includes a polarization filter for filtering out a polarization component of the portion of the laser beam corresponding to a polarization of the stimulated emission.

[0025] An excimer or molecular fluorine laser system is further provided including a discharge chamber containing a gas mixture, multiple electrodes connected to a power supply circuit for energizing the gas mixture, a resonator for generating a laser beam, a processor, and an amplified spontaneous emission (ASE) detector. The processor receives a signal from the ASE detector indicative of the ASE signal of the laser. The ASE detector detects a filtered signal, wherein a substantial portion of a stimulated emission signal of the laser beam is filtered from the laser beam prior to monitoring the ASE signal. The ASE detector includes a stimulated emission filter for substantially filtering the stimulated emission from a portion of the laser beam to permit the ASE signal to be resolved from the laser beam. The stimulated emission filter includes a spectral filter for filtering out a spectral component of the portion of the laser beam corresponding substantially to a spectral distribution of the stimulated emission.

[0026] Also, an excimer or molecular fluorine laser system is provided including a discharge chamber containing a gas mixture, multiple electrodes connected to a power supply circuit for energizing the gas mixture, a resonator for generating a laser beam, a processor, and an amplified spontaneous emission (ASE) detector. The processor receives a signal from the ASE detector indicative of the ASE signal of the laser. The ASE detector detects a filtered signal, wherein a substantial portion of a stimulated emission signal of the laser beam is filtered from the laser beam prior to monitoring the ASE signal. The ASE detector includes a stimulated emission filter for substantially filtering the stimulated emission from a portion of the laser beam to permit the ASE signal to be resolved from the laser beam. The stimulated emission filter includes a temporal filter for temporally filtering out substantially all but the very leading edge of the laser pulse, wherein the leading edge substantially contains the ASE signal, and while substantially all but the leading edge is stimulated emission.

[0027] An excimer or molecular fluorine laser system is further provided including a discharge chamber containing a gas mixture, multiple electrodes connected to a power supply circuit for energizing the gas mixture, a resonator for generating a laser beam, a processor, and an amplified spontaneous emission (ASE) detector. The processor receives a signal from the ASE detector indicative of the ASE signal of the laser. Feedback from the resonator is blocked to reduce stimulated emission and facilitate detection of the ASE signal when the ASE detector detects the ASE signal corresponding to the signal received by the processor from the ASE detector indicative of the ASE signal of the laser.

[0028] FIG. 1 shows a schematic block diagram of a gas discharge laser system having a regulation loop including an ASE detector.

[0029] FIG. 2ashows a first embodiment of an ASE detector which uses polarizational ASE detection.

[0030] FIG. 2bshows a second embodiment of an ASE detector which uses angular ASE detection including a central axis beam dump.

[0031] FIG. 2cshows a fourth embodiment of an ASE detector which uses a shutter for blocking the resonator.

[0032] FIG. 2dshows a fifth embodiment of an ASE detector.

[0033] FIGS. 2eand 2fshow graphs of intensity versus time of a laser pulse illustrating detected signals by the ASE detector of the fifth embodiment.

[0034] FIG. 2gillustrates how the beam is transformed as it transmits optics of the ASE detector of the fifth embodiment.

[0035] FIG. 3 is a graph of ASE signal versus driving voltage measured under constant gas mixture conditions.

[0036] FIG. 4 is a graph of ASE signal versus number of halogen injections measured under constant driving voltage conditions.

[0037] FIG. 5 shows a first graph of driving voltage versus pulse count and a second overlayed graph of ASE amplitude versus pulse count measured for a laser system wherein a series of halogen injections were performed each to increase the halogen concentration by 2.2% of its value at various shot intervals from 1.5 to 4 million shots over a range of 60 million shots following a new fill.

[0038] FIG. 6ais a graph of photon density versus time of a laser pulse calculated for an excimer or molecular laser system.

[0039] FIG. 6bis a graph of photon density versus time of the ASE contribution to the laser pulse of FIG. 6a.

[0040] FIG. 7ais a graph of photon density versus time of a laser pulse calculated for the same laser system used in FIG. 6ahaving a lower halogen concentration and correspondingly higher driving voltage.

[0041] FIG. 7bis a graph of photon density versus time of the ASE contribution to the laser pulse of FIG. 7a.

[0042] A preferred embodiment provides a signal which is directly related to the status of the gas mixture of a gas discharge laser. An advantage is that the input signal of the processor controlled regulation loop is the amplified spontaneous emission signal (ASE) of the laser (for a general discussion of spontaneous emission, please see, e.g., V. F. Weisskopf and E. Wigner, Z Phys.63, 54 (1930); R. H. Dicke and J. P. Wittke, Introduction to Quantum Mechanics,Addison-Wesley (1960); and/or J. H. Eberly, American J Phys.,40, 1374 (1972)). It is here noted that it may be advantageous to be able to accurately measure the ASE of a gas discharge laser, such as an excimer or molecular fluorine laser, for other reasons than to determine a gas mixture status and/or to adjust a gas mixture, such as to ensure that the ASE signal remains below a certain tolerance threshold value for the laser system and optics, or to adjust an optical component of the laser system.

[0043] FIG. 1 shows a schematic block diagram of a gas discharge laser system such as an excimer or molecular laser (e.g., F₂, KrF, ArF, XeCl, XeF, KrCl) in accord with a preferred embodiment of the present invention. The laser system includes a laser tube or discharge chamber 1 containing a gas mixture, a pair of main electrodes 11 connected to an electrical pulse power and discharge module 6 controlled by a processor 9, and a fan and cooling unit. The gas mixture typically comprises a halogen containing molecular species (e.g., F₂or HCI), an active rare gas (e.g., Kr, Ar or Xe) and a buffer gas (e.g., Ne or He), or in the case of the F₂laser, F₂and a buffer gas (e.g., Ne or He). The electrodes 11 are preferably configured of a specialized shape and material(s), depending on the laser being used. Preferred electrode configurations may be described at U.S. patent application Ser. No. 09/453,670, which is assigned to the same assignee as the present application, or alternatively at U.S. Patents No. 4,860,300, 5,347,532 or 5,729,565 or German Patent No. DE 44 01 892, which are hereby incorporated by references into the present application.

[0044] The gas mixture is filled into the discharge chamber during new fills. Halogen and rare gas injections, total pressure adjustments and gas replacement procedures are performed using a gas control box 7 including a vacuum pump, a valve network and one or more gas compartments. The gas control box receives gas via gas lines 8 connected to gas containers, tanks, canisters and/or bottles. The gas control box is controlled by the processor 9.

[0045] A resonator includes preferably a rear optics module 2 and a front optics module 3 on either side of the discharge chamber 1. The rear optics module 2 preferably includes a highly reflective optic which may be a mirror, or a mirror with a dispersive prism in front of it, or a dispersive reflection grating usually with one or more beam expanding prism(s) in front of it, and optionally one or more etalons may be disposed within the rear optics module 2. With a dispersive element, the natural bandwidth of the laser is reduced or a single line of a discrete spectrum is selected. The wavelength of the output beam 5 can be tuned by an adjustable orientation of the mirror or the grating, or may be pressure tuned, or tuned by rotating one or more prisms. The rear optics module 2 may be configured to permit a small amount of radiation to exit so it can be used with certain beam parameter monitoring activities not otherwise discussed below, such as wavelength calibration as is described in either of U.S. Pat. Nos. 6,160,832, 4,905,243 or 6,160,831, or U.S. patent application Ser. No. 60/202,564, which is assigned to the same assignee, each patent and patent application being hereby incorporated by reference. In particular, a line narrowing module would be used for a KrF or ArF laser preferably comprising multiple prisms, or a grating and one or more prisms, and optionally an etalon or other interferometric device, and line selection optics would be used for a F₂laser (e.g., besides the dispersive element mentioned above, an etalon, a birefringent plate and/or one or more prisms). Line-narrowing and/or selection may be used as described in any of U.S. Pat. No. 6,154,470 or U.S. patent application Ser. Nos. 09/317,695, 60/212,183, 09/657,396, 09/244,554, 09/602,184, 09/599,130, 09/629,256, 60/162,735, 60/173,993, 09/715,803, 60/166,967, 60/167,835, 601170,919, 60/212,257, and 60/243,462, which are assigned to the same assignee as the present application and are hereby incorporated by reference. Additional optical elements may be included with the rear optics module for beam steering.

[0046] The front optics module 3 preferably includes an outcoupler, and may include additional optics for beam steering, splitting or forming such as mirror(s), beam splitter(s), prism(s), etalon(s) and/or grating(s). An output coupling interferometer may be used such as is described at U.S. patent application Ser. No. 09/715,803, which is assigned to the same assignee and is hereby incorporated by reference. The output beam 5 of the laser system preferably exits the resonator via the outcoupler of the front optics module 3.

[0047] An energy monitor 4 is shown in FIG. 1 near the front optics module 3 for detecting a signal proportional to the pulse energy of the output beam 5. This energy information is sent to the processor 9 from the energy monitor 4, typically over a serial interface.

[0048] An ASE signal detector 10 is shown near the front optics module 3 for detecting a signal proportional to the ASE of the laser. This ASE signal information is also sent to the processor 9 from the ASE detector 10.

[0049] The ASE signal can be measured by any of a variety of ASE detection devices 10. For example, a photodiode detector configured to detect a peak intensity of the ASE signal or the ASE signal integrated over time. For very short wavelengths such as are generated by the F₂-Laser (157 nm), a specialized VUV sensitive detector is preferred. Preferred detectors are disclosed at U.S. patent application Ser. Nos. 60/122,145, 09/771,013 and 60/181,156, which are assigned to the same assignee and are hereby incorporated by reference into the present application. A detector such as is described in U.S. patent application Ser. No. 09/172,805, which is also assigned to the same assignee and is hereby incorporated by reference, may also be used.

[0050] It is desirable to detect the ASE signal substantially separate from the stimulated emission as it is otherwise difficult to resolve the ASE signal out of the entire laser pulse signal. This substantial separation of the ASE signal from the entire laser pulse signal is achieved in the preferred embodiment because it is recognized in the preferred embodiment that ASE and stimulated emission signals have some different properties. First, the ASE signal exhibits a greater divergence than the signal including the stimulated emission signal due to the collimating effects of the resonator and the nature of stimulated emission. Second, the ASE signal is non-polarized whereas the stimulated emission signal is usually polarized in one direction due to polarizing optical elements of the resonator (e.g., prisms, polarizing beamsplitters or windows oriented at Brewster's angle). Third, the ASE signal temporally arises before the stimulated emission signal or at the leading edge of the laser pulse. Fourth, the laser emission signal is largely dependent on photons appearing in the discharge chamber after reflecting from one or both resonator reflectors 2, 3 one or more times, whereas the ASE is independent of how many photons enter the discharge chamber after being reflected from a resonator reflector 2, 3. In addition, the ASE is not line-narrowed by passing through the rear optics module 2, and may tend to have a broader bandwidth than the stimulated emission, depending on the resonator configuration being used. These properties can each be used for discrimination that will be referred to as polarizational, angular, temporal and mechanical discrimination.

[0051] FIG. 2ashows a first apparatus for use with ASE signal detection in accord with a preferred embodiment which uses discrimination of p-polarized laser light. A prism 12 of the rear optics module 2 has a front surface 14 which may be oriented at the Brewster angle along the optical path of the resonating beam. As such, the front surface 14 transmits substantially all P-polarized light incident on it. Thus, substantially only s-polarized incident light is reflected from the front surface 14 of the prism 12. Referring to FIG. 2a,substantially all of the p-polarized light and a fraction of the s-polarized light incident at the front surface 14 of the prism 12 continue along the optical path of the resonator to the highly reflective optic 15. In this case, the stimulated emission signal is substantially p-polarized and as such it is not reflected from the front surface 12 to the detector 10. Instead of using the Brewster angle, a dielectric coating or coatings which reflect only s-polarized light may be used. Also, instead of using the front surface 12 of the prism 14, another optical component such as a polarizing beam splitter which can be an uncoated plate oriented near the Brewster angle or more effectively a plate with a special coating that reflects nearly all incident s-polarized light, or an optical Brewster window of the discharge chamber may be used.

[0052] FIG. 2bshows a second apparatus for use with ASE signal detection in accord with a preferred embodiment. The second apparatus permits a discrimination technique taking advantage of the fact that the ASE has a greater divergence than the stimulated emission. A beam splitter 16 is shown in FIG. 2breflecting out a portion of the main beam. As above, another optical component may be used for this purpose. In addition, a small portion of the main beam may be used which is permitted to exit the highly reflective optic 15. A central beam dump 18 is positioned substantially at the center of the separated beam portion. The beam dump 18 blocks the center portion of the separated beam which contains the substantial portion of the stimulated emission signal in the separated beam. Some of the ASE signal portion of the separated beam passes around the beam dump 18 owing to its greater divergence. The detector 10 is positioned to detect this portion of the ASE signal which passes around the beam dump 18. In this regard, the detector 10 may be as shown in FIG. 2b,or the detector 10 may be larger than the beam stop 18 wherein the beam stop 18 is in the middle of it such that the detector 10 may capture light on more than one side of the dump 18, or a lens may be positioned after the beam stop 18 to focus light from all sides of the beam stop onto the detector 10. The dimension of the beam dump 18 is selected to just block the substantial portion of the stimulated emission signal of the separated beam to allow a maximum amount of ASE signal to pass around the beam dump 18.

[0053] A third apparatus for use with ASE detection in accord with a preferred embodiment includes electronics for temporally discriminating all but the leading edge of a detected portion of the main beam. ASE photons are generated by the discharge and not by stimulated emission facilitated by the presence of other photons. Thus, the ASE photons are the first ones to appear from the discharge chamber after the discharge. As such, the leading edge which substantially contains the ASE signal portion of the detected portion is separated from the entire signal which may then be analyzed in accord with the present invention.

[0054] A fourth apparatus for use with ASE detection in accord with a preferred embodiment is illustrated at FIG. 2cand includes an excimer or molecular fluorine laser having a shutter 20 for blocking the resonator. The resonator may be blocked, e.g., using a shutter or other mechanism placed inside the resonator between the discharge chamber 1 and one of the resonator reflectors 2, 3 such as the outcoupling mirror 3 and/or a reflecting surface of the rear optics module 2. A large portion of the stimulated emission signal, i.e., that arising from photons reflecting back into the discharge chamber from the resonator reflectors 2 or 3 that are blocked by the shutter(s) 20, and otherwise appearing at the detector 10, is thus avoided, while practically none of the ASE signal is hampered. Only one of the resonator reflectors need be blocked to avoid stimulated emission signals at the detector 10, although both may be blocked. The shutter 20 can be closed quickly and an ASE measurement taken while, e.g., the stepper is moving to a different place on a wafer or to a different wafer, such that the ASE signal measurement may be performed online even though the resonator is briefly blocked in the fourth embodiment.

[0055] FIG. 2dschematically illustrates a fifth embodiment for measuring the ASE signal of the laser output beam 5. A front optics module 3 is shown having line-narrowing optics 21 therein including, e.g., a prism, an etalon and/or a grating. Having line-narrowing optics 21 in the front optics module 5 is advantageous because the ASE is in this way better filtered from the output beam than when all of the line-narrowing optics are located in the rear optics module 2 (see, e.g., U.S. patent application Ser. Nos. 09/738,849, 09/715,803 and 09/718,809, which are assigned to the same assignee as the present application and are hereby incorporated by reference).

[0056] A beam portion 22 is split off from the main beam 5 for diagnosis according to the fifth embodiment. A shutter 24 may be used for selectively permitting the beam to pass through. The diagnostic device of the fifth embodiment is preferably contained within an enclosure 26, such as may be evacuated or purged with an inert gas. The beam path to the enclosure and that for the output beam may each be contained within an enclosure (not shown), and preferably each is so contained for an ArF or F₂laser (see U.S. patent application Ser. Nos. 09/343,333, 09/594,892, 09/598,552, 09/131,580 and 09/771,013, which are assigned to the same assignee as the present application and are hereby incorporated by reference, and alternative configuration are set forth at U.S. Pat. Nos. 5,559,584, 5,221,823, 5,763,855, 5,811,753 and 4,616,908, all of which are hereby incorporated by reference. The beam 22 then preferably passes through an aperture or diaphragm 28 as it enters the enclosure 26.

[0057] The beam 22 after passing through the diaphragm 28 encounters a first beam splitter 30, wherein a portion is reflected and a portion is transmitted. The transmitted portion encounters a second beam splitter 32, wherein a portion is reflected and a portion is transmitted. The portion that is transmitted at the second beam splitter may be used for other diagnostic purposes or may be sent to a beam dump 34, as shown in FIG. 2d.

[0058] The beam portion reflected at the first beam splitter 30, which may be 5% of the intensity of the beam that is incident on the first beam splitter 30, is shown encountering a half-wave plate 36. At this point, i.e., prior to encountering the half-wave plate 36, most of the intensity of the beam is due to stimulated emission and is sigma polarized, as shown. The ASE is not polarized, and thus roughly half of the ASE intensity is of each polarization both before and after passing the half-wave plate. The principal polarization direction of the beam reflected at the first beam splitter 30 is transformed at the half-wave plate 36 from sigma polarization to pi polarization, as shown.

[0059] The beam then encounters a beam expander for expanding the beam preferably by around eight times. A first prism 38 of the beam expander preferably has an entrance face oriented at Brewster's angle to the incident beam, such that only pi polarized light is reflected to a beam dump 39 and substantially all sigma polarized light is transmitted which includes substantially only ASE light. The beam expander shown in FIG. 2dincludes a second beam expander prism 40.

[0060] The expanded beam then encounters an optical etalon 42 for selecting a single narrow portion of the linewidth of the beam. An optional second etalon 44 may be used for further line-narrowing. The main line-narrowed portion of the beam reflected at the beam splitter 30 is filtered out at the etalon 42 or etalons 42, 44. A far higher proportion of the beam transmitted by the etalon 42 or etalons 42, 44 is ASE radiation than of the beam reflected at the beam splitter 30. Instead of using the etalon 42 or etalons 42, 44, another spectral filter may be used such as a grating, a prism or another interferometric device.

[0061] After passing through the etalon 42 or etalons 42, 44, the beam is focused by a lens 46 and through a diffuser 48. The beam is then detected at a photodiode 50. As mentioned, the proportion of the ASE in the beam that is detected at the photodiode 50 is far higher than the proportion of the ASE in the main beam 5. In a preferred embodiment, the ratio of the ASE to the stimulated emission detected at the detector 50 is around 1:1. An alternative embodiment would use a spectral filter such as an etalon, grating and/or prism alone, without also incorporating a polarization filter, and another alternative embodiment would use a polarization filter without using a spectral filter, although the use of both is preferred for resolving the ASE signal from the narrow band emission of the laser.

[0062] The beam that transmits the beam splitter 30 and is reflected at beam splitter 32 is not subjected to polarizational and spectral filtering within the enclosure 26. The beam reflected at beam splitter 32 encounters a lens 52, a diffuser 54, an attenuator 56, another diffuser 58 and a second photodetector 60. Preferably, the ratio of the ASE to the stimulated emission detected at the detector 60 is less than 1:50.

[0063] The detected signals at the detectors 50 and 60 is then input into an oscilloscope 62. The attenuator 56 preferably reduces the intensity of the signal detected at the second photodetector 60 to about the intensity detected at the first photodetector. When the traces are compared next to each other and each having an intensity somewhere preferably late in the temporal profile of the pulse normalized to the other. The trace from the first detector 50 will have a greater intensity at the beginning of the pulse, due to its having a greater contribution of ASE than that detected at the second detector 60. By observing how much greater that intensity is in comparing the two traces, the amount of ASE in the beam can be determined. Preferably, a computer is used to facilitate the calculations.

[0064] FIGS. 2eand 2fshow examples of measurements of the ASE in accord with the fifth embodiment of this invention. Referring to FIG. 2e,the trace 62 measured at the first detector 50 is shown clearly having a greater intensity at the front edge of the trace than the trace 64 measured at the second detector 60, and the area A is dependent on the amount of ASE that is in the main beam 5. Referring to FIG. 2f,the trace 66 measured at the first detector 50 is shown clearly having a greater intensity at the front edge of the trace than the trace 68 measured at the second detector 60, and the area B is dependent on the amount of ASE that is in the main beam 5.

[0065] It appears that the proportion of ASE in the beam measured in FIG. 2fis greater than the proportion of ASE in the beam measured in FIG. 2e.In fact, the proportion of ASE determined to be in the beam measured in FIG. 2fwas around 0.3%, while that determined to be in the beam measured in FIG. 2ewas around 0. 13%, or about half of that of FIG. 2f.

[0066] The point 70 where the traces 66 and 68 may be observed to have equal intensities in FIG. 2fwas the selected point of intensity matching of the two traces 66 and 68. Note that the point 70 is later in the trace 66, 68, due the presence of the ASE predominantly earlier in the traces 66, 68. Note also that plot 1 of FIG. 2eand plot 2 of FIG. 2fshow traces of the main beams of the laser systems used in the measurements of FIGS. 2eand 2f,respectively. The fifth embodiment advantageously utilizes Applicants' knowledge of polarization and narrowband performance of the laser emission and provides an apparatus and technique for measuring in real time the ASE level with tolerable certainty.

[0067] FIG. 2gillustrates a reduction of the ASE to stimulated emission intensity ratio of the beam reflected from the beam splitter 30 of FIG. 2das the beam transmits the elements 36-48 toward the detector 50. The beam reflected by the beam splitter 30 may have an intensity ratio of ASE to narrow band laser emission of 1:50 or higher, wherein the beam contains a 98% or greater contribution of the narrow band emission of the laser system and a 2% or less ASE contribution as indicated at block B1. As the beam transmits elements 36-40, polarization separation occurs as indicated by block B2. The beam transmitted through the prism 40 then may have an intensity ratio of ASE to narrow band laser emission of 1:6, as indicated at block B3. As the beam transmits element 42 or elements 42 and 44, spectral separation occurs as indicated at block B4. The beam transmitted through the element 42 or elements 42 and 44 then may have an intensity ratio of ASE to narrow band laser emission of 1:1, as indicated at block B5.

[0068] It is recognized herein that the ASE signal measured by the detector 10, e.g., by any of the first though fifth embodiments described above, or through another technique understood to those skilled the art, is strongly correlated to the status of the gas mixture. The ASE signal is a direct measure of population inversion density of the gas mixture. The population inversion density is proportional to the small signal gain go. To a good approximation, the ASE signal intensity varies depending on the small signal gain go according to the following equation:

[0069] ASE signal α exp[g₀●L]  (1)

[0070] In equation (1), L is the length of the gain medium. This is particularly clear from observations using an arrangement including an inverted gain medium without a feedback mechanism such as a resonator. Such an arrangement can be realized according to the description above of the fourth apparatus using a laser system wherein the beam path is blocked somewhere within the resonator. For example, a shutter may be introduced between the discharge chamber and one of the resonator reflectors such as the outcoupling mirror or the highly reflective mirror. In such an arrangement without a feedback mechanism, where laser oscillation is blocked, the ASE signal generally depends on the number of active gas molecules (i.e., for a KrF laser the number of F₂and Kr molecules, for an ArF laser the number of F₂and Ar molecules, and for a F₂laser the number of F₂molecules), and the electrical power brought into the discharge which is a function of the driving voltage. Of course, the specific dependence is more complex and based on the discharge quality and rate equations.

[0071] FIG. 3 shows a measured ASE signal for a laser with a blocked resonator versus the driving discharge voltage at constant gas mixture conditions. In the experiment shown at FIG. 3, the driving voltage began at 1.5 kV and was incremented in steps of 0.2 kV from 1.5 kV to 2.1 kV. The ASE signal increased with each increment of driving voltage from about 500 mV at 1.5 kV to about 1275 mV at 2.1 kV.

[0072] FIG. 4 shows a measured ASE signal for a laser with a blocked resonator versus the halogen concentration at constant driving voltage. In the experiment shown at FIG. 4, a KrF laser with a blocked resonator was used. The halogen concentration of the F₂/Kr/Ne gas mixture was initially 0.1% (1% Kr, 98.9% Ne) and the measured ASE signal was about 750 mV. Then, the halogen concentration was increased using microhalogen injections (μHIs)such as are described in United States provisional patent application No. 60/124,785, which is hereby incorporated by reference into the present application. Each μHI resulted in an approximately 1.5% higher halogen concentration in the gas mixture. The ASE signal was shown in FIG. 4 to increase from around 750 mV to around 1100 mV after the fifth halogen injection.

[0073] FIG. 4 demonstrates the sensitivity of the ASE signal to small changes in the halogen concentration which usually occurs in excimer and molecular lasers such as KrF, ArF and F₂lasers, as mentioned above. This sensitivity of the ASE signal to small changes in the halogen concentration also holds for a laser where the gain medium is within a resonator which is not blocked. Moreover, even in a laser operating in an energy-constant mode where the driving voltage is not constant but adjusted within a certain band to maintain the energy at a constant value, the ASE signal is strongly correlated to the gas mixture status. Therefore, continuous online control of the status of the gas mixture is possible without blocking the resonator and without interrupting industrial laser processing.

[0074] FIG. 5 shows two plots. Plot A is ASE signal amplitude versus pulse count. Plot B is driving voltage versus pulse count. The laser used for generating the data of FIG. 5 was operating at a repetition rate of 2 kHz in constant energy mode. Halogen injections were performed at each vertical line in the graph. As can be observed, the pulse count intervals between halogen injections were varied between about 1.5 and 4 million pulses. Although in a preferred embodiment of the present invention, a gas mixture regulation loop based on the measured ASE signal is used, such a regulation loop was not used when the data used in FIG. 5 were measured. Each halogen injection resulted in a 2.2% increase in the partial pressure of the halogen containing molecular species in the gas mixture.

[0075] As is clearly observed from a comparison of plot A and plot B in FIG. 5, the relative change in ASE signal is much larger than the change in driving voltage as a result of the gas actions. This shows a clear advantage in sensitivity of the present invention over conventional systems which monitor changes in driving voltage to track the gas mixture status. Both a sharp rise in ASE signal due to the halogen injections and a decrease in ASE signal between halogen injections are clearly resolved in plot A. In contrast, it is difficult to resolve any changes in driving voltage resulting from the halogen injections in plot B. The first three halogen injections have no effect on the ASE signal. This could be due to residual buffer gases remaining in the gas supply lines following the last new fill prior to the injections of FIG. 5 being performed. After around 20 million pulses, the rate of the halogen injections was increased from around once every 4 million pulses to about once every 1.5 million pulses. Though the starting ASE level was exceeded at the end of the experiment by about 50%, one skilled in the art would understand that the ASE signal level could be kept constant by appropriate adjustment of the HI rate. In the present invention, such adjustments would be based on using the measured ASE signal as an input to a gas regulation loop.

[0076] A brief explanation is now provided for the enhanced sensitivity of the ASE signal to changes in gas mixture composition, most notably halogen concentration. The small-signal gain go is proportional to the deposited power density (U·I), where U is the steady state voltage and I is the discharge current. The steady state voltage U is a characteristic quantity of a gas discharge. For a fixed total pressure of the gas in an excimer laser such as a KrF or ArF laser, U depends largely on the ratio of the concentrations of the halogen and rare gases, i.e., c[halogen]/c[rare gas], where “c” represents “concentration”. The rare gas concentration in the discharge chamber decreases very slowly over the lifetime of the laser gas, while the halogen concentration decreases faster due to chemical reactions. The time dependence of the discharge current I(t), or in other words, the power deposition (U×I(t), where U is nearly time independent during the steady discharge state) into the plasma depends on the characteristics of the external discharge circuit (i.e., the charging voltage U_c, the capacitance C_pof the peaking capacitors connected in parallel to the electrodes, and the inductance L) and on the discharge resistance R which is U/I(t). The halogen concentration influences the whole time behavior of the small-signal gain and consequently the time behavior of the laser output. This qualitative analysis is modeled on a base of rate equations describing the time evolution of inversion density, laser photon density and ASE photon density.

[0077] FIGS. 6a-6band FIGS. 7a-7bshow the calculated laser pulses and the ASE pulses for two different halogen concentrations of a laser operating in constant energy mode (and thus operating at two different discharge voltages U₀and U₁), using the parameters set forth in Table 1. In constant energy mode, the same number of laser photons are emitted in each pulse.

[0078] FIGS. 6a-6bshow the photon density of the laser pulse and the ASE signal versus time, respectively, in the higher halogen concentration case. FIGS. 7a-7b,show the photon density of the laser pulse and the ASE signal versus time, respectively, in the lower halogen concentration case. In the lower halogen concentration case, the same number of laser photons are emitted as in the higher halogen concentration case (i.e., in constant energy mode), but in a longer pulse. That is, the area under the curves in FIGS. 6aand 7ais equal, even though there is a greater contribution from photons arriving at later times in the curve of FIG. 7athan in FIG. 6a.Specifically, the first peak in FIG. 7a,of two identifiable peaks observed in each of FIGS. 6aand 7a,is shorter and the second peak is taller as compared with those of FIG. 6a.Since we know the ASE signal is almost totally represented at the leading edge of the pulse (see discussion of the third embodiment or the references cited above), we expect the ASE signal contribution to be smaller in FIGS. 7aand 7bthan in FIGS. 6aand 6b.

[0079] In fact, the ASE signal is much smaller in the lower halogen concentration case of FIG. 7bthan in the higher halogen concentration case of FIG. 6b.This is easily observed from the fact that the ASE peak is simply smaller in FIG. 7bthan it is in FIG. 6b.Thus lower halogen concentrations produce longer laser pulses and smaller ASE signals. The parameters used in the calculations of FIGS. 6a-6band 7a-7bare listed in Table 1.

[0080] Experiments were performed measuring ASE signal versus halogen concentration, and driving voltage versus halogen concentration. In the experiments, the composition of the gas mixture was initially F₂:Kr:Ne=0.075%:1%:98.925%, and the halogen concentration was decreased incrementally in steps of 1.25%/step. The halogen depletion was realized by replacing portions of the gas mixture with the same amount of Ne/Kr (99% Ne, 1% Kr) resulting in the 1.25% dilution of only F₂per step. The ASE signal was separated for detection using the apparatus shown in FIG. 2adescribed above. The ASE signal is observed in plot A to decrease by about 40 mV per step, while the driving voltage is increased by about 0.01 kV per step to maintain constant output energy of the laser pulses.

[0081] The ASE signal was again observed to be very sensitive to slight changes in the halogen concentration. The signal becomes significantly smaller when the halogen is removed, even though the driving voltage is increased to keep the laser pulse energy constant.

[0082] In accord with the present invention, the sensitivity of the ASE signal to slight changes in the status of the gas mixture, especially the halogen content, is used advantageously as an input signal for a gas performance control system. The following applications of ASE controlled laser regulation may be used in accord with the present invention. First, ASE control may be used with smoothed long-term stabilization of the laser gas by micro halogen injections (μHIs) as described in the Ser. No. 09/447,882 application incorporated by reference above. A typical approach might be to maintain the ASE signal S at a constant level or within a certain narrow range, e.g., S-ΔS<S<S+ΔS, using μHIs or other smoothed gas actions wherein one or more gas components are replenished by injection or mini gas replacement MGR or partial gas replacement PGR.

[0083] Second, a second parameter such as the driving voltage (HV) could be used as an additional input for a HI regulation loop. While the ASE signal is directly correlated with the gas mixture status, in our example particularly with halogen concentration, the HV is influenced by other parameters like resonator alignment or degradation of resonator optics. In this way, the correspondence between the ASE signal and the HV signal indicates problems arising within the resonator. For example, if the ASE signal is held constant by maintaining a constant halogen concentration, but HV is nonetheless increased to achieve constant laser pulse energy, an indication of either resonator optics degradation or resonator misalignment is realized. In this scenario, when the ASE signal and HV information indicate such a problem with the resonator, an error signal or warning message could be triggered by laser control software. Examples of some of the system specifications that could be indirectly monitored by directly monitoring the second parameter, e.g., driving voltage, in addition to monitoring the ASE signal to monitor gas mixture status are alignment of an optical component, degradation of an optical component, synchronization of two or more optical components such as two reflective elements, one or both being a resonator reflector, or two or more prisms, or one or more prism(s) and either a grating or HR mirror, etc., contamination within the resonator such as ozone or oxygen especially along the beam path or dust especially on the optics or on a tube window, or an insufficiently clean beam path or purge strength especially for the F₂laser, but also potentially for the ArF-laser, since the F₂-laser uses a beam path free of photoabsorbing species such as water, oxygen and/or hydrocarbons, and is preferably operated having a purging inert gas flowing within a beam path enclosure (see U.S. patent application Ser. No. 09/343,333, which is assigned to the same assignee and is hereby incorporated by reference into the present application).

[0084] Third, the ASE signal S may be used to control a new fill procedure. Halogen concentration can vary from gas supply bottle to gas supply bottle and can also be influenced by long gas lines. The ASE signal S can be used for fine tuning of the total gas replenishment performed in a new fill procedure. A typical procedure of such a “self-calibrating” new fill might be to fill the laser with laser active gases, e.g., halogen and rare gases in an excimer laser or F₂in an F₂molecular laser, to a concentration that is somewhat less than the expected concentrations fixed in the laser settings. The result would be a lower ASE signal than is expected from a laser after a new fill with the expected concentrations, e.g., S<S_expected. There would follow smoothed gas injections such as μHIs until S=S_expectedis achieved.

[0085] The specific embodiments described in the specification, drawings, summary of the invention and abstract of the disclosure are not intended to limit the scope of any of the claims, but are only meant to provide illustrative examples of the invention to which the claims are drawn. The scope of the present invention is understood to be encompassed by the language of the claims, and structural and functional equivalents thereof.

[0086] In addition, in the method claims that follow, the steps have been ordered in selected typographical sequences. However, the sequences have been selected and so ordered for typographical convenience and are not intended to imply any particular order for performing the steps.