TW541777B - Four kHz gas discharge laser system - Google Patents

Abstract

The present invention provides an excimer laser capable of producing a high quality pulsed laser beam at pulse rates of about 4,000 Hz at pulse energies of about 5mJ or greater. A preferred embodiment is an ArF excimer laser specifically designed as a light source for integrated circuit lithography. An improved wavemeter with special software monitors output beam parameters and controls a very fast PZT driven tuning mirror and the pulse power charging voltage to maintain wavelength and pulse energy within desired limits. In a preferred embodiment two fan motors drive a single tangential fan which provides sufficient gas flow to clear discharge debris from the discharge region during the approximately 0.25 milliseconds between pulses.

Description

541777 A7 ___B7_ _ 5. Description of the invention (1) The present invention is part of the following patent applications: Application No. 10 / 012,002, November 30, 2001; Application No. 10 / 000,991, November 14, 2001 ; Application No. 10 / 〇29,319, October 17, 2001; Application No. 09 / 943,343, August 29, 2001; Application No. 09 / 854,097, May 11, 2001; No. 09/837 , Application No. 150, April 18, 2001; Application No. 09 / 834,840, April 13, 2001; Application No. 09 / 794,782, February 27, 2001; Application No. 09 / 771,789, January 1, 2001 June 29; Application No. 09 / 768,753, January 23, 2001; Application No. 09 / 684,629, October 6, 2000; Application No. 09 / 597,812, June 19, 2000, and 09 / 473,852 application date is December 27, 1999. The present invention relates to gas discharge lasers, and more particularly to high repetition rate gas discharge lasers. BACKGROUND OF THE INVENTION Electric discharge gas lasers Electric discharge gas lasers are well known and have been used shortly after the invention of lasers in the 1960s. The high-voltage discharge excitation gas-phase gain medium between the two electrodes The resonance cavity of the dielectric medium allows the optical stimulus to be amplified and subsequently extracted from the resonance cavity in the form of a laser beam. Many of these electrical discharge gas laser systems operate in pulse mode. Excimer lasers and excimer laser beams have been known for certain types of electric gas discharge lasers since the 1950s. The best wishes for the use of excimer lasers for integrated circuit lithography are described in U.S. Patent No. 5,023, m, issued on June 1991, under the name "« Excimer Laser ". This patent case allows the applicant's employment to be of suitable paper size _ (Please read the precautions on the back before filling out this page) • Order ----- 541777 A7 B7 V. Description of the Invention 2 Case and Reference Ways are incorporated here. The 884 patent case is explained in the excimer "'is the seed repetition rate pulse laser. Figures 1 and 2 show the main 7 ° of the laser 10. (The first figure corresponds to the first figure of the 884 patent case. (Figure 2 corresponds to Figure 7 of the 884 patent case). The discharge 22 is between the two long (about 23 忖) electrodes 18 = 20 'and about 5/8 leaves. Similar to the previous description, the previous patent laser The repetition rate is typically in the range of 2 to pulses per frame. High repetition rate lasers are usually provided in a gas cycle. The laser is achieved by a long squirrel-cage fan 46 having about 23 fans. The fan structure is longer than the electrode, providing enough cycles, so at the pulse operation rate, the discharge between the electrodes: the interference hole body is cleared between the electrodes between two pulses. The main axis of the fan 46 46 Rudi 2 continued (for the 884 patent case Figure 9) is supported by two ball bearings 132. The gas used in the laser contains gas and has a very high reactivity. The fan rotor «fan spindle i 3 0 is sealed in the casing by a sealing element ⑶ Parts 丨 2 and U provide the same environment "inside, for example, the 8th and 9th lines are described on the 45th line; and the motor stator is attached to the outside of the sealing element, which means that the motor stator is not corroded by the fluorine-containing gas. However, the bearing 132 is affected by the cavity to the gas, and the bearing uses red "A" for lubrication. The bearing corrosion is contaminated by the bearing's lubrication. The module design quasi-knife laser is used for integrated electricity, ^ Λ, ^ When circuit breaking is performed, it is typical to work on the manufacturing line to work year-round. The cost of adding V and saving money is too high. Therefore, the large 4 excimer laser composition can be used in a few minutes. Replace. 6 Break-up assembly into a module and usually this paper size applies the Zhongguanjia Standard (CNS) A4 ^ (210 > ^ J7 (Please read the precautions on the back before filling this page) • Order · 541777 A7 —--- ---- B7________ 5. Description of the Invention (3) Line Narrowing The output beam width of the excimer laser used for lithography must be reduced to a micron component. This "line narrowing" is typically based on This is achieved by forming a line narrowing module (called a "line narrowing package" or "LNp") on the back side of the laser cavity. This type of line narrowing package is typically composed of dedicated optical components including chirps, mirrors, and gratings Composition. As the repetition rate increases, the borrow narrowing module maintains stable performance This poses a major challenge. This type of electrical discharge gas laser described in U.S. Patent No. 5,023,884 uses an electric pulse power system to generate a discharge between two electrodes as shown in Fig. 3. In this prior art system, a 'direct current power supply' The charger 22 is called a "charging battery" or a "C0" 42 capacitor bank to a predetermined and controlled voltage, and is called the "charging voltage" of each pulse. This charging voltage range is about 500 to 1000 volt range. After the charging capacitor has been charged to a predetermined voltage, the solid-state switch 46 is turned off, allowing electrical energy to be stored in the charging capacitor very quickly around a series of capacitor banks 52, 62 and 82 and inductors 48, 54 and 64 The magnetic compression circuit of the voltage transformer 56 generates a very high potential in the range of approximately 16,000 volts across the electrodes, which generates a discharge that can last for about 50 nanoseconds. Prior art systems on the market, between solid-state switch closure and discharge The time is about 5 microseconds; but it takes about 400 microseconds to charge the capacitor accurately to the predetermined voltage. This rate is lower than about 2,000 Hz. The pulse repetition rate is fast enough. However, the reader must understand that the accurate charging of the charging capacitor is extremely important, because the voltage level of the charging capacitor is controlled by the laser operator. This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm). ) 6 (Please read the notes on the back before filling out this page) Order this narrow grant 541777 V. Description of the invention (4 This system is the only control for discharging ^ e ^ ^ ^ except for the feasible control, and the discharge voltage决定 The main determinant of laser pulse energy. The heat exchanger used in integrated circuit lithography previously, human /, < monthless 丨 J technology excimer lasers typically require the laser gas system to be quenched, and the laser hole system Discharge heating and the heating of the rebel moon dagger, which was discussed in the previous article through the circulation fan rotation. Typically, this is done using a water-cooled heat exchanger with fins, as shown in Figure 1%. Doubling the laser repetition rate or more than doubling the heat generated by controlling the laser doubles, the main reason is that the power system required to circulate the laser gas increases with the cube of the required gas velocity. Control of Beam Quality When used as an integrated circuit lithography light source, the laser beam parameters (ie, pulse energy, wavelength, and bandwidth) are typically controlled within extremely tight specifications. For example, pulse-to-pulse pulse energy recovery control is required, and slightly slower linearized output beam wavelength feedback control is required. Doubling the pulse rate or more requires the control system to operate much faster. There is a need for a better laser design ' for pulsed gas discharge lasers to operate at a repetition rate in the range of about 4,000 pulses per second. SUMMARY OF THE INVENTION The present invention provides an excimer laser system capable of generating high-quality pulsed laser beams with a pulse rate of about 5 亳 joules or more at a pulse rate of about 4,000 Hz. The system can be configured to emit KrF or ArF lasers with small changes. operating. In the preferred embodiment, the system is a KrF excimer laser configured to operate at 4000 Hz and 6 亳 Joule pulses. It is specially designed for integrated circuit lithography. The paper size applies the Chinese National Standard (CNS) A4 specification. (210X297 mm) (Please read the notes on the back before filling out this page), OK | 541777

Description of the invention 2 4 8 Taihan factory (々, T, mesh) light source. Improved wavelength meter output beam parameters with special software monitor, and extremely fast PZT driven fine-tuning mirror, and (Please read the precautions on the back before filling this page} Control feedback configuration ^ Xia's pulse power The charging voltage 俾 maintains the wavelength and pulse energy within, and within. In a preferred embodiment, a two-fan motor drives a single all-line fan, recaptures the prostitute > "catching 1, sufficient airflow, and about 0.25 between pulses. Leap second time by discharge zone clear danger # φ A discharge chip. By changing the gas mixture, modifying the wire used for 193nm (namely) operation, narrowing the package output coupler and the wavelength port, and better slightly adjusting the electrode spacing The system can be configured to be used as an ArF laser without any substantial changes to the system. The two laser configuration systems and all components of the system are almost identical to the two configurations. Brief description of the drawings Figures 1, 2 and 2A show the prior art The structure of the laser system. Figure 3 shows the circuit diagram of the prior art pulse power system. Figure 4A shows a front view of a preferred embodiment of the present invention. Figure 4B shows a better tool. Section of the laser cavity of the embodiment. Figure 5 shows a circuit diagram of a preferred pulse power system. Figures 6 A and 6B show two preferred resonant power supplies. Figures 7, 7A and 8 show the cooling pulse power components. The technology. Figures 9A and 9B show views of saturable inductors. Figure 10 shows the power transformer. Figure 10 A shows the power transformer core. Figure 11 shows the cooling of the first saturable inductor technology. Figures 12, 12A and Figure 12B shows the cooling of the second saturable inductor technology. Figures 13 and 13A (l) ~ (8) show the chamber discharge in the preferred embodiment. The paper size applies the Chinese national standard A4 specification (210X297 mm) 541777 A7 B7 V. Description of the invention (view of area 6) Figure 14 shows the logout of the preferred embodiment. Figures 14A, MB, 14C and 14D are diagrams and line diagrams illustrating the calculation of wavelength and bandwidth. Pages ME, UF Figures 14G and 14H show a view of the wavelength and bandwidth monitoring etalon. Figure 15 is a block diagram showing the components used for wavelength and bandwidth calculations. Figure 16 is a block diagram showing the laser beam wavelength and Pulsed Laser System Structure. Figures 16A, 16B1, and 16B2 are schematic diagrams showing the LNP trimmer control technology. Structure 0 (Please read the precautions on the back before filling out this page), can | Brother 16C diagram shows the piezoelectric control effect of the trimmer. Figures 16D and 16E are schematic diagrams showing the principles of wavelength control deduction. Figures 17, 17A, 17B, and 17C show the technique of sweeping the surface of the grating. Figure 18 shows the two-motor blower control system. Figure 18A shows the structure of the fan blades. Figures 19B and 19B show the structure of a better nitrogen sweeping system. Figures 19C-19G show the seal beam path of the fighter structure. Figures 20, 20A, and 20B show the preferred shutter structure. Figures 21 and 21A show The preferred water-cooled ancient soil Λ has a hot plate heat exchanger. Figure 22 shows the special nitrogen sweeping technique used in the preferred embodiment. The paper size applies the Chinese National Standard (CNS) A4 specification (2) 0297 mm. 541777 V. Description of the invention (7 A7 B7 Brother 23 A-23T The drawings illustrate the pulse stretching technique. Figures 24A, B, and C illustrate another pulse stretching technique. Detailed Description of the Preferred Embodiment First Preferred Embodiment The following describes the first preferred embodiment of the present invention. It is a KrF excimer laser designed to generate a 7.5 Joule narrow band of approximately 248 nanometers (eg, 248.250 nanometers, 248,250 micrometers) at a pulse rate of up to about 4 kHz. This type of laser The radio specifications include a bandwidth specification range of less than 0.5 micron (FWHM bandwidth) and less than h4 micron (95% integrated bandwidth). The specification should also be 3 Σ wavelength feminine sex less than 0.05 micron and 30 pulse dose stability is less than 0.5 μjoule. Figure 4B is a front view of the preferred embodiment, the door has been removed, and a variety of laser components are identified. The laser is a modular unit, designed The entire module can be replaced very quickly and easily The laser downtime is kept to a minimum. Figure 4B shows the clockwise components of each component: • The status light 1K indicates the laser operating status, and • the control module 2K, based on the data from the handheld terminal device (Figure A Display) or the input control signal from the main controller of the lithography machine to control the laser function. • The shrink head module 3K is a part of the laser pulse power system and provides 彳 If㈣ the final stage 'charging is located in the laser cavity. The peak capacitor bank on the top of the room, • The stabilization module 4K is also called a wavelength meter, which controls the thunderbolt and provides a feedback signal that controls the wavelength and pulse energy. • The auto switch module 5K is equipped with a power meter. This paper Standards and Finances Standard_CNS A4 Specification (2 Huan 297 Gong 7) (Please read the precautions on the back before filling this page)

541777

Description of the invention (8 • MFT power supply 6K provides high voltage power to the metal fluoride trap (filter) located in the laser chamber module, two • left blower motor 7K, • laser chamber module 8K, • Interface module 9K provides interface circuit to match laser control cold lithography machine control device, and • Cooling supply module 10K, • Cooling water distribution module 11K, • Laser gas supply module 12K, • Ventilation The assembly 13K is used to ventilate the laser indoor gas to the outside atmosphere and includes a smoke detector. • The right blower motor 14K, • the line narrowing module 15K is also called the line narrowing package or lNP, • the right blower motor Controller 16K, • Left blower motor controller 17K, • Rectifier module 18K, which contains Co charging capacitor banks and circuits that trigger electrical pulses to provide early pulse compression and pulse voltage amplification. • For resonant charger module 19K In order to provide extremely fast resonant charging of Co charging capacitor banks, the 咼 voltage power supply module 20K is used to generate high voltage DC power from a three-phase AC power supply of standard mains , • AC / DC distribution module 21K • Helium sweep module (not shown in the figure) • Wavelength stabilization module (not shown in the figure) This paper size applies to China National Standard (CNS) A4 specification (210X297 mm) 11 (Please read the notes on the back before filling in this page j

541777 A7 _________B7 _ 5. Description of the invention-(Please read the notes on the back before filling out this page) Figure 4A is a cross-sectional view of the laser cavity 10A of the first preferred embodiment of the present invention. The main components of the laser chamber are the housing structural elements pa and 14A, the cathode 18a and anode 20a upstream of the front dissipator tube 60, the spike capacitor bank 62 and the electrostatic trap unit 64 (all of which are similar to the first i shows the corresponding components of the prior art). The chamber includes a new anode support streamline structure 48, a new upper streamline structure 50, a gas rotating blade 52, a new 5-inch diameter tangential fan structure 46A, and four water-cooled heat exchanger units 58. Figure 13 shows a similar chamber configuration with a new current loop 548. Figure 16 is a block diagram showing the structure of an important laser system for controlling the output facility wavelength and pulse energy. Major improvements over prior art gas discharge lasers include: 1) Improved flow path of laser gas in the circulating chamber 2) Water-cooled pulsed power system 3) Ultra-fast wavelength meter with fast control deduction rule 4) Novel high load Periodic LNP with a combination of PTZ and stepper motor to drive the fine-tuning mirror 5) Large tangent fan with dual water-cooled brushless DC drive blower motor with special controller 6) Ultra-pure nitrogen sweep system for optical protection 7) Comes with Power meter's sealed switch 8) Improved heat exchanger configuration 9) Shedong seal system 10) LNP helium sweep system 11) Current loop with parallel air flow ribs This paper applies Chinese National Standard (CNS) A4 specifications (210X297 mm) -12-541777 A7 ^ --____ B7_ V. Description of the invention (10) 12) Inductive plate for optimal chamber impedance 13) Optimized electrode for performance life. The laser chamber heat exchanger is a preferred embodiment designed to operate at a 4,000 pulse repetition rate. Clearing the discharge between two pulses to affect the gas discharge area requires that the gas flow rate between the electrodes 188 and 208 be as high as about 67 m / s. In order to achieve this speed, the diameter of the tangential fan unit is set to 5 inches (the length of the blade structure is 26 inches), and the rotation speed is increased to 3,500 rpm. In order to achieve this kind of performance, this specific embodiment uses two motors to collectively transmit a driving power of about 4 kilowatts to the fan blade structure. At a pulse rate of 4000 Hz, the discharge will add about 12 kW of thermal energy to the laser gas. In order to remove the heat generated by the pen and the heat added by the fan, four separate water-cooled finned heat exchanger units 58A are provided. The motor and heat exchanger will be described in detail later. The preferred embodiment of the present invention uses FIG. 4A and FIG. 13 to outline the four-degree water-cooled heat exchanger 5 8 A with a heat radiation sheet. Each heat exchanger is slightly similar to the single heat exchanger shown in Figure 1 at 58 but with substantial improvements. The cross-sectional view of one of the heat exchanger components is shown in Figure 2i. In the heat exchanger, the J-side knife is not turned around at both ends. Figure 21A shows an enlarged end view of the heat exchanger with thermal expansion and contraction. The heat is replaced by a 7C component, which includes a heat sink structure 302, which is formed by a solid copper (CU 11 coffee) mechanism, and each pair contains a heat sink 303. Water flow to the HI ^ hole diameter of 33 忖. Plastic spoiler located in the axial channel This paper scale applies the Zhongguanjia standard (where) M specification (2Κ) χ297 public love) 13 (Please read the precautions on the back before filling in this page) Order:: ---- -*-541777 A7 V. Description of the invention (η device 306 can prevent water from flowing in the channel to form a laminar flow, and prevent the formation of a thermal boundary layer on the inside surface of the channel. The flexible flange unit 3 04 is the inner flange 3 04A 'Welding unit consisting of telescopic V 304B and outer flange 304C. The heat exchanger unit includes three c self-sealing 308 to seal the water flow from the laser heat exchanger. The telescopic band 304B allows the heat exchanger to expand relative to the chamber Reduction. Double-mouthed nut 400 connects the parent channel to the standard 5/16 inch positioning elbow fitting, which is in turn connected to the water source. The O-ring 402 provides a nut 400 and a heat sink structure 302. In a preferred embodiment of the spoiler, the spoiler is composed of four off-the-shelf elongated coaxial mixing elements, which are typically used to mix epoxy components and are obtained from 3M Company (static mixer, part number 06-D1229-00). Online Mixer Display Figures 30 and 30 of Figures 21 and 21. The on-line mixture forces water to flow along a rough spiral path, which reverses its clockwise direction at approximately every separation distance (0.3 inches). The spoiler substantially improves the performance of the heat exchanger. Apply Human tests have shown that the addition of a spoiler can reduce the required water flow by about a factor of 5 and maintain comparable gas temperature conditions. In a preferred embodiment, the airflow flows in and out of the discharge zone compared to the previous technology. The chamber is greatly improved. The vane structure 66 is designed to regulate the milk velocity of the region 68, which is regularized to the gas velocity of the region 68 downstream of the fan structure; 0 m / min. In the middle of the chamber, the top of the chamber is cut to give a similar water flow forming surface. Then the gas flow velocity in the discharge zone is accelerated to 67 meters / second. At about 4 inches downstream of the center of the discharge zone, the water flow section (CNS) Α4 ^ (210X297 ^) (Please read the precautions on the back before filling in this page} Order —… line-541777 A7 —-----B7_ 5. Description of the invention (η) is about 5/8 inches at 20 degrees Increased to about 4 inches, and then the gas was rotated by four rotating blades 52. This allows a large percentage of the voltage drop to be recovered through the discharge zone. The sacrum-shaped current loop 548 shown in Figure 13A (1) is formed by a 015-inch-thick nickel alloy (UNS N04400) by a laser mechanism and bent as shown in Figure 13A (3) to 13A (8), so the resistance of 23 ribs to gas flow is minimized. The cross section of each rib is 001 inch x 0.990 inch, and the longitudinal direction is parallel to the gas flow direction. Blower motor and large blower The first preferred embodiment of the present invention provides a large tangential fan driven by a dual motor for circulating laser gas. The preferred configuration shown in Fig. 18 provides a gas flow rate of 67 m / s between the electrodes, which is sufficient for a flow rate of 4,000 Hz to clear a space of approximately 1.7 cm between pulses. The fan blade structure of the fan is shown in cross section at 64A in Figure 4A. The perspective view is shown in Figure 18A. The blade structure is 5 inches in diameter and is made of solid aluminum alloy 6061-T6 rod-shaped stock. The sections of the blades are slightly offset by the adjacent sections. As shown in Figure 18A, the pressure disturbance in the discharge area caused by the blades is reduced. It is preferred that the leading edge of the individual blades be tapered to the blade edge, thereby reducing the reflection of the discharge sound waves from the leading edge of the blade during operation. The specific embodiment shown in FIG. 18 uses two 3-phase brushless DC motors, each having a magnetic rotor housed in a metal pressure cup. As described in US Patent No. 4,950,840, the pressure is separated by the stator portion of the motor and the laser. Gas environment. In this specific embodiment, the pressure cup is a thin-walled nickel alloy 400 with a thickness of 0.016 inches, which serves as a laser gas barrier. The two motors 53 and 2 drive the same -root axis and are programmed for reverse rotation. The second motor is a sensorless motor. The paper size is applicable to the Chinese National Standard (CNS) A4 specification (2) 〇 > < 297 mm. 15 (Please read the precautions on the back before filling this page). (13) (In other words, its operation does not include a position sensor). The right motor controller 534 that controls the right motor 53 is used as the main controller. It controls the bin motor control to 53 6 through analog and digital signals to indicate start / stop, current command, current feedback, etc. The communication with the field shooting controller 24A is done by entering the main controller 5342Rs_232 serially. Water-cooled pulsed power system Four kilohertz pulsed power system The operation of the laser system according to the present invention requires precise control of potentials in the range of approximately 12,000 volts to 30,000 volts in ⑽Hz (i.e. between = (Approximately 250 microseconds) is applied between the electrodes. As indicated in Section B of the background, in the first month J technology * pulse power system, the charging capacitor bank is charged to the exact predetermined & control voltage ', which is generated by turning off the solid state switch, so that the cymbals stored in the charging circuit are compressed by compression Amplifying the circuit to generate a predetermined potential across the electrodes @ off-off time to completion of the discharge is only a few microseconds (ie about 5 U heart). The charging interval of Ren Xianyi's technology requires far longer than 250U #. Reduce charging time by using a large power supply or several power supplies. For example, you are asked to use three prior art power supplies in parallel and operate at 4,000 Hz. In this preferred embodiment, as shown in FIG. 5, the applicant uses the same basic design as the solid-state downstream pulse power system part of FIG. 3, which is not read first as shown in FIG. 3. However, the applicant uses different technologies to charge Co. Resonant charging ... Shenyueren uses two types of resonant charging system to charge C very fast. . These systems are described with reference to Figures 6A and 6B. 541777 A7 ~~ ----- ^ _ 5. Description of the Invention (I4) The first resonance charger shows the circuit of the better resonance charging shown in Figure 6-8. In this example, a standard direct power supply of 208 VAC / 90 amp input and 800 VDC 50 amp output is used. The power supply is a DC power supply that can be adjusted from 600 volts to 8 volts. The power supply is directly attached to d, eliminating the need for voltage feedback to the power supply. When the power supply is activated, the power supply starts up and regulates a constant voltage in the C-1 capacitor. The system is independent of the voltage regulation of the c-1, so only the basic control loop is required for the power supply. Secondly, when the voltage of c-i drops below the voltage setpoint, the watchdog 'supply state will increase energy to the system. In this way, the power supply can replenish the energy transferred from C-1 to c0 during the entire time between the laser pulses (even the duration of the laser pulses). This further reduces the peak current demand of the power supply beyond the requirements of prior art pulsed power systems. The combination of requiring the power supply to have the most basic control loop and reducing the peak rated current of the power supply to the average power of the system can reduce the cost of the power supply by about 50/0 °. In addition, 'this better design allows manufacturing The manufacturers are flexible, because a fixed output voltage power supply with a fixed current is easily obtained from multiple source suppliers. Such power supplies are available from suppliers such as Elgar, Universal Voltronics, Kaiser, and EMI. Control Board This power supply continuously charges the 1033 pF capacitor 202 to the voltage required by the control board 204. The control board 204 also orders the opening and closing of the IGBT switch 206 to transfer energy from the capacitor 202 to the capacitor 42. The inductor 208 combines the capacitors 202 and 42 to set the transmission time to a constant value and limits the peak charging current. The paper size of this paper is subject to the Chinese national standard (A4 specification (210X297)) -17-(Please read the precautions on the back before filling this page). Order-: line 丨 541777 A7 ______B7 V. Description of the invention (15) The board 204 receives a voltage feedback 212, which is proportional to the voltage of the capacitor ", and receives a current feedback 214. The current feedback is proportional to the current flowing through the inductor 208. From this two feedback signals, If the switch 20 is turned on instantaneously at this time, the control board 204 can calculate the final voltage value of the capacitor 42 in real time. Therefore, the command voltage 210 is fed into the control board 204 and the stored energy can be stored in the capacitor 42 and the inductor 208. Make an accurate calculation and compare it with the required charging pressure command 2 10. From this calculation, the control board 204 will determine the correct time to turn on the IGBT switch 206 during the charging cycle. The system accuracy is after the IGBT switch 206 is turned on, The magnetic field energy stored in the inductor 20 will be transferred to the capacitor 42 through several constrained diode paths 215. The accuracy of the instantaneous energy calculation will determine the amount of fluctuations in the final voltage fluctuations of the capacitor 42. The extreme charging rate of this system may have too much jitter to meet the predetermined system adjustment requirements of ± 0.05%. In this way, additional circuits such as de-qing circuits or no circuits may be used, which will be described in detail later. Resonant charger The second resonant charger system is shown in Figure 6B. This circuit is similar to Figure 6A. The main circuit components are: II-three-phase power supply with constant DC current output 300. C-1 -A source capacitor 302, which is higher than the existing Cq capacitor 42 by a first power amplitude or more. Q1, Q2, and Q3-switches that control current flow to charge and maintain the regulated voltage of co.

Dl, D2, and D3- provide unidirectional current flow. This paper size applies to Chinese National Standard (CNS) A4 specifications (210X297 cm) 18 (Please read the precautions on the back before filling this page) -Order — ·: Line 丨 541777 A7 __ B7_ V. Description of the Invention (I6) R1 And R2- provides voltage feedback to the control circuit. R3 allows Cq voltage to discharge quickly when a small amount of overcharge occurs. The resonant inductor between the L1-C-1 capacitor 302 and the Co capacitor 42 limits the current and sets the charging transfer timing. Control board 304-based on circuit feedback parameter commands qi, 〇2 and ⑴ on and off. The operation example is as follows: The difference between the circuit in Figure 6B and the circuit in Figure 6A is the addition of switch Q2 and diode D3, which is called Dijing switch. This switch improves the regulation of the circuit by allowing the control unit to short-circuit outside the inductor. This "Dijing" prevents the extra energy stored in the charging inductor L1 current from being transferred to the capacitor c0. Before the laser pulse is needed, the C-1 voltage is charged to 600-800 volts and the switches Q1-Q3 are turned on. When there is a command from the laser, Q1 turns off. At this time, the current will flow from C · 1 to C0 through the charging inductor L1. As explained in the previous paragraph, the differentiator on the control board will evaluate the Cq voltage and the current flowing at L 丨 relative to the command voltage set point from the laser. Q1 turns on when the C1 voltage plus the equivalent energy stored in the inductor u is equal to the predetermined command voltage. The calculation formula is: V corpus [vC0s2 + ((Li * Ili,) / C ())] 0.5 Here: C voltage of V1 after Q1 is turned on and the current is zero.

Vc0s = voltage when Ql is on. lLls = Current flowing through L1 when Qi is on. After Q1 is turned on, the energy stored in L1 starts to be transferred to Cq via D2 until the CG voltage is approximately equal to the command voltage. At this point, Q2 turns off and the current stops flowing.

This paper is suitable for financial standards (cnsu ^ (2) OX297 ^) —ITT (Please read the notes on the back before filling this page)

541777 A7 B7 V. Description of the invention (17) to co, led through D3. In addition to the "Dijing" circuit, Q3 and R3 from the bleeder circuit allow additional fine-tuning of the Co voltage. When the current flowing through the inductor L1 is stopped, the switch q 3 of the circuit 216 will be closed by the control board command, the CG voltage will be discharged to the predetermined control voltage and then Q 3 will be turned on. The time constants of the capacitor c G and the resistor R 3 are fast enough to discharge the capacitor C 0 to the command voltage so that they do not account for the perceptible amount of the total charging cycle. As a result, the resonance charger is equipped with a third-order adjustment control. The slightly coarse adjustment system is provided by the behavior of the moon and the switch Q 1 during the charging cycle. When the voltage of Co is close to the target value, the Dijing switch is turned off; when the voltage of CG is slightly higher than the target value, resonance charging is stopped. In a preferred embodiment, the switch Q1 and the Dickin switch are used to provide a tuned quasi-silicon degree better than +/- 0.1%. If additional adjustment is required, a third control device for voltage adjustment can be used. This is the 'Λ lower circuit of switch Q3 & R3 (Figure 6B shows at 216) to discharge CG to the accurate target value. C. Downstream Improvements As mentioned earlier, the pulse power system of the present invention utilizes the same basic design used in the prior art system shown in FIG. However, the basic design requires a number of major improvements, which caused a 3 factor increase in thermal load due to the large increase in repetition rate. This improvement will be detailed later. Detailed description of the rectifier and the compression head The main components of the rectifier 40 and the compression head 60 are shown in Fig. 3, which discusses the operation of the system in background section B. In this section, the manufacturing details of the rectifier and the compression head will be explained. Solid State Switch Solid State Switch 46 is provided by Powerex Corporation (Yangguwu, Penn.) This paper size is applicable to Chinese National Standard (CNS) A4 specification (210X297 mm) 20 (Please read the precautions on the back before filling in this Page). Order — 541777 A7 ______B7_ V. Description of the invention (ι〇 P / NCM 800 HA-34HIGBT switch. In a preferred embodiment, two such switches are connected in parallel. Inductors 48, 54 and 64 are optional. Saturated inductors, similar to the saturable inductors used in previous systems described in U.S. Patent Nos. 5,448,580 and 5,3 15,611, are hereby incorporated by reference. Figure 7 shows the betterness of the L0 inductor 48. Design. In this inductor, 4 electrical conductors from two IGBT switches 46B are formed by 16 ferrite toroidal coils 49 to form part 48A. The part 48A is an 8-inch long hollow working cylinder made of extremely high magnetic permeability material. , With 10 to 1 忖 and Od about 1 to 5 inches. Then the four conductors are wrapped twice around the insulating ring pie-shaped core to form a component 48B. Then the four conductors are connected to a board, which is in turn connected to C. Inductor row High voltage side of group 52. Saturable electricity A better schematic of the device 54 is shown in Figure 8. In this case, the inductor is a single-turn geometry, where the top and bottom covers 541 and 542 and the central mandrel 543 (all at high voltage) are assembled and passed. A single turn of the inductor's magnetic core. The casing 545 is connected to ground potential. The magnetic core is 0.0005 thick provided by Butler, Penn., Magnetics, or Adilanto, California. Inch tape wound 50_50〇 / () nickel-iron alloy. The heat sink 546 of the inductor housing can assist the internal heat transfer to forced air cooling. In addition, a ceramic disc (not shown) is installed under the bottom cover of the reactor Auxiliary heat transfer from the middle of the assembly to the base plate of the module. Figure 8 also shows one of the capacitors connected to the high-voltage capacitor bank 52 with high voltage, and the inductor unit connected to the 1:25 step-up pulse transformer 56 High voltage lead. The case 545 is connected to the ground of the unit 56. This paper size applies the Chinese National Standard (CNS) A4 specification (21〇 < 297mm) 21 (Please read the precautions on the back before (Fill in this page) -install 丨 、? —Line 丨 541 777 A7 B7 V. Description of the invention (19) The top view and sectional view of the saturable inductor 64 are shown in Figures 9A and 9B, respectively. In the inductor of this embodiment, the metal part 3 (n, 3, 2, 3 〇3 and 304 except the sum of the flux, as shown in Figure 9B, reduce the leakage flux of the inductor. The flux except the metal parts substantially reduces the area that the magnetic flux can penetrate, so it helps to saturate the inductor's saturable inductance. Reduced to the minimum. The current is based on the inductance of the winding around the core 3 0 7 as the assembly is wound by $ turns through the vertical inductor rod. The current entered at 3 $ goes down through the large-diameter conductor marked "1" in the center, to the six smaller conductors also marked "1" in the periphery, as shown in Figure 9A. Then the current flows down through the two conductors labeled 2 on the inside, then up to the 6 conductors labeled 2 on the outside, and then down to the flux except the metal on the inside, and then up to the 6 labeled 3 on the outside. The conductor then flows down to the two conductors marked 3 on the inside, and then flows up to the 6 conductors marked 4 on the outside, and then down to the conductor marked 4 on the inside. The flux except for metal components is maintained at a semi-full impulse voltage across the conductor, allowing the safe maintenance interval between the flux of the metal component and other turns of the metal rod to be reduced. The core 307 series is made of up to three coils 307A, B & c. The coils are provided by Butler, Pennsylvania, Meganet, Inc. or Adilanto, Calif., And 0.0005 inches in thickness. 80-20 % Nickel-iron alloy windings formed. Readers find nano-crystalline materials such as those obtained from VITROPEAN of vacuum SCHITELZE, Germany and nanometers of FINEMET from Hitachi Mekels, Japan A crystalline material can be used for the inductors 54 and 64. In prior art pulsed power systems, oil leaks from electrical components pose a potential problem. In this preferred embodiment, the constituent elements of oil insulation are limited to the paper size applicable to the Chinese National Standard (CNS) A4 specification (21〇297297) 22 (Please read the precautions on the back before filling this page)-, may : Line 丨 541777 A7 B7 V. Description of the invention (20 saturable inductors. In addition, the saturable inductor material shown in Figure 9B is covered in a basin-shaped oil-containing housing, and all of the sealed connections are located above the oil surface and are essentially Eliminate the possibility of oil leakage. For example, the bottom seal of the inductor 64 is shown in Figure 308. As the normal oil level is lower than the top lip of the housing 306, as long as the housing remains upright, the oil is almost impossible. Leaked outside the assembly. Capacitor banks 42, 52, 62, and 82 shown in Figure 5 of the capacitor (that is, c0, q, [... and ...) are all made of purchased off-the-shelf capacitors connected in parallel. Power Valley is 42 and 52 are film type capacitors, which can be obtained from suppliers such as Stawell, Northern California, Vishay Roederstein or Weimar, Germany. Applicants link capacitors and inductors. The best way is to class U.S. Patent No. 5,448,580 solders positive and negative terminals on special printed circuit boards with heavy nickel-coated copper leads. Capacitor banks 62 and 82 are typically obtained from suppliers such as Murata or tdk (both are (This company) is composed of a parallel array of high-voltage ceramic capacitors. For the preferred embodiment of the KrF laser, the capacitor bank 82 (ie Cp) is composed of 33 0.3 nF capacitor banks. Capacitance 9.9 nF; Cp ^ is composed of 24 0.40 nF capacitors, with a total capacitance of 9 · 6 nF; C i is a 5-7 μM capacitor bank and C0 is a 5.3 μM capacitor bank. Pulse Transformer Pulse Transformer 56 is also similar to the United States The pulse transformers described in Patent Nos. 5,448,580 and 5,3 13,481; however, the pulse transformer of this embodiment has only a single turn on the secondary winding, and 24 inductance units are equal to 1 of a single main turn with an equal step ratio of 1:24. / 24. The schematic diagram of the pulse transformer 56 is shown on page 1023 (please read the precautions on the back before filling this page), \ t — .line 丨 This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) ) 541777 A7 _____ B7 V. Description of the invention (21) Figure. Each of the 24 inductor units contains two flanged short stubs 56A (individual flat edges with threaded holes), which are bolted to the positive and negative terminals of the printed circuit board 56β, such as along the first As shown in the bottom edge of Figure 10 (the negative terminal is a high-voltage terminal with 24 primary windings). The insulator 56C separates the positive terminal of each short tube from the negative terminal of the adjacent short tube. There is a hollow cylinder length i 1 between the flanges of the short tube. / 16 inch, 0875 d, wall thickness is about 1/32 忖. The short tube was wrapped with a 1 忖 wide 0.77 mil thick Metglas 2605 S3 A and a 0 · mil mil thick Mylar film until the OD of the insulated Meigras package was 2.24 inches. A perspective view of a single bundled short tube forming a primary winding is shown in Figure 10A. The secondary winding of the transformer is a single OD stainless steel rod mounted on a PTFE (Teflon) tightly fitted to the inside of the insulation tube. The winding has four sections, as shown in Figure 10. The low-voltage terminal of the stainless steel secondary winding is shown at 56D in Figure 10, the primary HV lead to 56B of the printed circuit board 56B, and the high-voltage terminal is shown at 56F. As a result, the transformer forms an automatic transformer configuration, and the step-up ratio becomes 1:25 instead of 1:24. In this way, a pulse of about 1400 volts between the positive terminal and the negative terminal of the inductive unit will generate a pulse of about -35,000 volts at the terminal 5 of the secondary terminal. This single-resistance secondary winding design provides extremely low leakage inductance, allowing very fast output rise times. Laser chamber electrical components details

The Cp electric valley state 82 is composed of 33 0.3 nf capacitor banks arranged on the top of the chamber pressure vessel. The electrodes are about 28 inches long and spaced from about 0.5 to 10 inches. The preferred gap interval is 16.5 mm for KrF (ArF is preferably separated by 13.5 mm). The preferred electrode capacity is detailed later. In this embodiment, the top electrode is referred to as a cathode, and the bottom electrode is grounded, as shown in FIG. 5, and referred to as an anode. This paper size applies Chinese National Standards (CDs) from specifications (public love) 24 (Please read the precautions on the back before filling this page}

541777 V. Description of the invention (22 Component water cooling In order to match higher thermal load, the water provided by the internal cooling fan provided here is usually forced air cooling to operate in such a higher average power mode than 较 田. Σedge in Disadvantages of water cooling have traditionally been close to electrical components or have a high possibility. This specific concrete _ warp material-solid cooling pipeline due to module internal cooling, cooling usually dissipates most of the components deposited inside the module This potential problem can be substantially avoided. Since there are no joints or connections or junctions inside the module enclosure, and the cooling pipe is a continuous piece of solid metal (such as copper, stainless steel, etc.), the probability of leakage current inside the module It is greatly reduced. Therefore, the module connection of ^ = is on the outside of the metal sheet enclosure of the assembly, where it is matched with a quick-disconnect type connector. Detailed description of the rectifier 54 The example of the line group shaking the rectifier module, provided Water-cooled saturable inductor 54A, as shown in Figure 11A, is similar to inductor 54 in Figure 8, but the diffuser is replaced with a water-cooled jacket 54A1, as shown in Figure 丨 a. Cooling 54A2 is routed inside the module and is wrapped around the jacket 54A1. An IGBT switch and a series diode are set up on the aluminum base plate through the aluminum base plate. The three components constitute most of the power dissipation inside the branch group. Other items that can also dissipate heat ( Subtractor diodes, resistors, capacitors, etc.) are forced air cooling provided by two fans behind the module. Since the jacket 5 4 A1 is maintained at ground potential, the cooling tube is directly attached to the reactor housing and There is no voltage separation problem. This project is achieved by pressing and fitting the cooling pipe to the dovetail cut groove cut on the outside of the casing as shown in 54A3. The Chinese national standard (Q ^) A4 specification is applied to 25 paper sizes (2) 〇χ297mm) 541777 A7 __ B7 V. The description of the invention (23) and the use of thermally conductive compounds to assist in making good thermal contact between the cooling tube and the case. Cooling high-voltage components Although the IGBT switch is "floating" at high voltage, but The IGBT is mounted on an aluminum base, which is electrically isolated from the switch by a 1 / 16-inch rear aluminum oxide plate. The aluminum base plate, which operates as a thermal well and operates at ground potential, is much easier to cool because the cooling circuit has no High voltage isolation. The sketch of the water-cooled aluminum base plate is shown in Figure 7 A. In this example, the cooling tube is crimped to the aluminum base slot. Integrated connection with the base plate. During normal operation, the series diodes also float at high potential. In this case, the diode housing is typically used for design without high voltage isolation. In order to provide this necessary isolation, The diode "hockey rubber disc-shaped" package is sandwiched inside the heat well assembly and then mounted on top of a ceramic base and then on top of a water-cooled aluminum base plate. The ceramic bottom is just thick enough to provide the required electrical isolation. But it will not be too thick and cause more thermal resistance than needed. For this special design, the ceramic is 1/16 忖 thick oxide, but other foreign materials such as beryllium oxide can also be used to further reduce diode junction and cooling water. Thermal impedance. The second embodiment of the water-cooled rectifier uses the assembly of the base substrate attached to the 108-pole and -pole body as a single cooling plate assembly. The cooling plate can be made by brazing a single piece of nickel tube to Erming "top plate" and "bottom plate". As mentioned above, the IGBT and the pole body are designed to transfer heat to the cooling plate using the ceramic pottery discs underneath the assembly. In the preferred embodiment of the present invention, the cooling plate is used to cool the paper. The size of the paper is applicable to the Chinese National Standard (CNS) A4 specification (2) 〇χ297 公 楚 (Please read the precautions on the back before filling out this page).丨 541777 A7 _______B7_ 5. Description of the Invention (24) The method is also used to cool the IGBT and diode of the resonant charger. The heat pipe thermal rod is also used to transfer heat from the outer casing to the base substrate. Detailed description of the compression head The water-cooled compression head is similar to the electrical design of the previous art air-cooled version (using the same type of ceramic capacitors and using similar materials for reactor design). The main difference in this example is that the module needs to be performed at a higher repetition rate, so it is performed at a higher average power. In the case of a compression head module, most of the heat is dissipated into a modified saturable inductor 64A. The cooling of the assembly is not simple because the entire system is operated with short pulses of extreme rain voltage. As shown in Figures 12, 12a and 12B, the solution of this liquid is to isolate the casing from ground potential. This type of inductor is provided by wrapping a cooling tube around the two cylinders to form a ferrite core. The input and output cooling lines are coiled around the cylinder of the ferrite core formed by two cylinder sections and a ferrite block, as shown in Figures 12, 12A and 12B. The ferrite block is made of CN-20 material made by Ceramic Magnetics, Inc. of New Jersey. A single piece of copper tube (0.187 inches in diameter) was crimped and wound around the inductor 64A housing 64A1 and around the second winding form into a winding form. The end is left with sufficient length to extend through the fitting of the thin metal cover sheet, so there is no cooling pipe joint inside the base plate. Similar to the first-stage reactor housing for a water-cooled rectifier, the inductor 64A includes a dovetail notch, shown in 64A2. This housing is similar to the previous air-cooled version, with the exception of dovetail grooving. The copper cooling water pipe is compressed and fitted in the cut groove, and the heat pipe is well connected between the body and the cooling water pipe. Thermally conductive compounds are also added. This paper is sized for China National Standard (CNS) A4 (210 x 297 mm) 27 (Please read the precautions on the back before filling this page), τ.

•• Wire I 541777 Five invention descriptions (25 to minimize thermal impedance. The electrical design of the inductor 64A is slightly changed from the 64 shown in Figures 9 and 9). The inductor 64A only provides a surrounding core (64A3) The second circuit (instead of the five circuit) is composed of a four-tape coil (instead of a three-coil). As a result of the conduction path of this water-cooled pipe from the output potential to the ground potential, the bias current circuit is now slightly different. As mentioned above, the bias current is supplied from the rectifier's DC-DC transducer through the cable to the compression head. The current passes through the positive bias inductor lb2 and is connected to the Cp-1 voltage node. The current is then split, partly via the HV cable Wire (after grounding, return to DC converter through transformer) to the rectifier. The other parts pass through the compression head reactor L, P-1 (to the bias magnetic switch), and then "negative" bias the inductor through the cooling water pipe Lm, return to ground and DC-DC transducer. By balancing the resistance of each pin, the design project can ensure that there is sufficient bias current available for both the compression head reactor and the rectifier transformer. "Positive" bias current LB2 is very similar to the "negative" bias inductor Lb3. In this case, the same ferrite rod and block are used as the magnetic core. However, two return spacers with a thickness of 0.125 are used to form an air gap in the magnetic circuit, so that The magnetic core will not be saturated by the DC current. Instead of the cooling water pipe winding inductor, i8awg Teflon wire is wound around the molded part. Quick connection In this preferred embodiment, three of the pulsed power modules use blind matching electricity. The connection is made solely by sliding the module into the laser chamber to make all electrical connections to the laser system. These are AC distribution modules, groups, power supply modules and resonance charger modules. Each Module of the example

5. Description of the invention (26) or the female plug matches the corresponding plug installed on the back side of the laser chamber. In each case, two tapered pins with about 3 slave ends on the module guide the module to accurately locate the person and let the electrician match the field. Mesh-matched connectors such as AMP 194242-1 Model City / Yu Zi AMP A Division (Harrisburg, PA). In this specific embodiment, the connection is used for 乂 * 2G8 volts, AC volts, DC lions volts (out of the power supply and resonance charger) and several signal voltages. The blind matching link allows the module to be removed for service use and more within a few seconds or minutes. In this specific embodiment, because the output dust of the module is in Canada, it is deleted to the 3M00 volt range, so the blind matching link Not used for rectifier modules. Instead, use typical high-voltage connectors. Feeding to the cathode (inductance-enhancing technology) As indicated in Figure 5 and explained earlier, the peak capacitor bank Cp consists of 33 0.3 nF capacitors. Grounding at the bottom of the capacitor (ie, the head of the chamber) is connected to a metal plate called a corona barrier at the top of the capacitor. The corona barrier is connected to the cathode with 15 metal rods called drains. The metal rods pass through the one-piece main insulator and are bolted to the top of the cathode 514, as shown in Figure 13A. The one-piece insulator is described in U.S. Patent No. 6,208,674, which is incorporated herein by reference. In some cases, it is necessary to increase the inductance of the high voltage portion of the pulse power circuit. This increase may lead to an increase in laser power duration. The applicant decided that the better way to generate such a higher inductance is to provide short (for example, about 1 1/4 inch) vertical poles to connect the capacitors of the cp capacitor bank and the metal pipe of the sewer pipe to the corona plate. Prolonging the electrical path length between Cp and the cathode in this way substantially increases the inductance of this part of the circuit. The increase in inductance can be selected by selecting the square 俾 A7 in the vertical bar 541777. V. The description of the invention (27 Length adjustment. The discharge component elements 13 and 13 A⑴) show the details of the improved discharge configuration used in the preferred embodiment of the present invention. This This configuration includes what the applicant calls a fan-dielectric electrode electrode. In this design, the anode 54 includes a pure-end fan-shaped electrode 542, and a dielectric spacer 544 is placed on both sides of the anode (as shown). Improve the gas flow in the discharge area. The spacer is attached to the anode support rod ⑽, with screws at each end of the spacer beyond the discharge area. The screws allow the spacer and the support rod door to expand and slip. The anode is 26. .439Bf high. The bottom is ο · inches wide and the top is 0.141 hours wide. Attached to the flow-formed anode support plate 546, there are screws through the socket, allowing the electrode to do differential thermal expansion from its center position. The anode is mainly copper The alloy is preferably composed of c36_, C954GG, or C19400. The shape of the cathode surface is shown in Figure 13A. The preferred cathode material is CU000. Additional details on the configuration of the fan dielectric are shown in US Patent Month Number 0 No. 9 / 768,753, incorporated herein by reference. This configuration Qiao Dian * along the road 548 series is composed of bone-shaped parts, with 27 ribs evenly spaced along the direction of the electrode 度 = degree, and its cross section is shown in section 13A (1) Figure. As mentioned above, the motor front path is made of metal sheet. The ribs (the cross-section dimensions are about tx 0.09 inches) are twisted, so that the longitudinal dimension of each rib is in the direction of current flow. Another design of the dielectric spacer is shown in Figure 13A2 > ^ is more uniform. In this example, the spacer is more perfectly matched to the streamlined Yang Ren support cup to provide a better gas flow path. The applicant claims that r fast return Fan blade dielectric anode design. This paper is suitable for the country's specifications ⑵0Χ ^ Ϊ7

• ----- (Please read the precautions on the back before filling out this page). Yes.: Line 丨 541777 A7

--------- B7__ V. Description of the invention (28) Ultra-fast wavelength meter with fast control deduction rule controls pulse energy, wavelength and bandwidth used in prior art excimer mine of integrated circuit lithography The number of laser beams is closely regulated. It is typically required to measure the pulse energy, bandwidth, and center wavelength for each pulse as well as the feedback control of the pulse energy and frequency. In previous techniques, the feedback control of pulse energy is based on pulse by pulse. In other words, the pulses of each pulse can be measured quickly. The data obtained can be used to control the determinants to control the energy immediately after the pulse. For a 1,100 Hz system, this means that the measurement and control of the next pulse takes less than 1/1000 of a second. A 4x speed is needed for a 4000 Hz system. Techniques for controlling the center wavelength and measuring the wavelength and bandwidth are described in US Patent No. 5,025,455 "Beam Wavelength Adjustment System and Method" and US Patent No. 5,978,394 "Excimer Laser Wavelength and System". These patents are incorporated herein by reference. The wavelength and bandwidth are measured on a pulse-by-pulse basis for each pulse system, but due to the prior art techniques of controlling the center wavelength, which took several milliseconds, the feedback control of a typical wavelength took about 7 亳 seconds. Need for faster control. Preferred embodiments for rapid measurement and control of beam parameters A preferred embodiment of the present invention is an excimer laser system that can operate in the range of 4,000 Hz to 6,000, with extremely fast measurements Laser beam parameters' and extremely fast control of pulse energy and center wavelength. The measurement and control of beam parameters for such lasers are described below. The wavelength meter used in this embodiment is similar to that described in US Patent No. 5,978,394, and several descriptions are taken from this patent case. 31 This paper size applies the Chinese national standard (〇) S) A4 specification (210X297 mm) 541777 A7 I η · ιιΐΜ " --------- S7 V. Description of the invention (29) '' ^ Measuring beam parameters Figure 14 A layout diagram of a preferred wavelength meter unit 120, an absolute wavelength reference correction unit 190, and a wavelength meter processor 197 is shown. The optical equipment of these units measures pulse energy, wavelength and bandwidth. These and i are used in the return circuit to maintain the pulse energy and wavelength within the predetermined limits. The device refers to the atom based on instructions from the laser system control processor. The reference comes from my calibration. As shown in Fig. 14, the laser output beam crosses part of the mirror 17o, and the part of the mirror passes about 95.5% of the beam energy as the output beam 33, and reflects about 45% of the pulse energy, wavelength, and bandwidth measurement values. The pulse energy of about 4/0 reflected beam is reflected from the mirror 丨 7 丨 to the energy detector 1, the energy detector 172 contains an extremely fast photodiode 69, which can measure individual pulse energy occurring at a rate of 4,000 pulses per second . The pulse energy is about 7.5 millijoules. The output detected as 69 is fed to the computer controller. The computer controller uses the special force deduction rule to adjust the laser charging voltage. Based on the stored pulse energy data, it accurately controls the pulse energy of future pulses.俾 Limit the change in the energy of individual pulses and the integration of burst pulses. The linear photodiode array The details of the photosensitive surface of the linear photodiode array 180 are shown in Fig. 4A. The array is an integrated circuit chip that contains 1024 separate photodiode volumetric circuits and associated samples and sustaining readout circuits. The total length of the photodiode is 25 6 | pen meters (approximately 1 leaf) have a spacing of 25 microns. Each photodiode is 500 micrometers long. Photo-polar arrays are available, for example, from several sources. The preferred supplier is Bin .......... _ '-_____ This paper size is applicable to the national standard (Norwegian) A4 specification (210X297 male I) ": ~ 32I ~

Yiyi (please read the precautions on the back before filling this page) —Installation 丨 Thread 541777 A7 B7 V. Invention Description (30) '— Song Company. In a preferred embodiment, the inventor uses the model S3903-1024Q, which can be read at a rate of up to 4 × 10 pixels / second on the basis of FlFO, wherein pixel scanning can be performed at a rate of 4,000 Hz or more. The PDA is designed to operate at 2x106 pixels / second, but the applicant has found that it can be timed to operate faster, that is, up to 4x106 pixels / second. For pulse rates higher than 4,000 Hz, you can use the same PDA, but you can only take a portion (for example, 60%) of pixels per scan. Rough Wavelength Measurement The beam passing through the mirror 171 at about 4 ° / ° is reflected by the mirror 173 through the slit 177 to 174 to 175, the return mirror 174, and reflected to the infrared grating 176. The beam is collimated by a lens 178, which has a focal length of 458 4 mm. The light reflected by the grating 176 returns through the lens 178, is reflected again by the mirrors 174, 175, and 174 again, and is then reflected by the mirror 179 and focused on the left of the 1024-pixel linear light-diode array 180 in the area of pixels 600 to 9550. As shown in the upper part of Figure 14B (pixels 0-599 are reserved for precise wavelength measurement and bandwidth). The spatial position of the laser beam in the photo-polar array is a rough measurement of the output laser beam relative to the nominal wavelength. For example, as shown in FIG. 14B, light in a wavelength range of about 193,350 μm will be focused on the pixel 750 and its vicinity. Coarse Wavelength Calculation The coarse-wavelength optical device in the wavelength meter module 120 produces a rectangular image of about 0.25 mm x 3 mm on the left side of the photodiode ι80.丨 〇 or 丨 丨 An illuminated photodiode will generate a signal proportional to the intensity of the received illumination (shown in Figure 14c). The signal is read and digitized by the processor of the wavelength meter controller 197. Using this information and interpolation method, the deduction rule controller 197 calculates the image. The paper size applies the Chinese National Standard (CNs) A4 specification (210X297 public love). 33 C (Please read the precautions on the back before filling this page)

May I: Line 丨 541777 V. Description of the invention 31 Center position. This position (measured in pixels) is converted to a coarse wavelength using two calibration coefficients to obtain a linear relationship between position and wavelength. The calibration coefficient is determined with reference to the source of the atom's pan-length, which will be described later. For example, the relationship between image position and wavelength is the following deduction rule: λ-(2.3 micron / pixel) P + 191, 625 micron where P = rough image center position ^ In addition, if necessary The second power term such as "+ OP2" can be used to mention extra precision. N Precision wavelength measurement is shown in Figure 14. 'Approximately 95% of the laser beam passing through the mirror 173 is reflected and deviated. Brother 1 82 passes through the lens. 83 to the etalon assembly. The famous spear Ran "improved etalon" is described as a diffraction diffuser). The beam sent by the etalon is focused by the 458.4mm focal length lens on the etalon assembly. As shown in Figure 14, after the reflection leaves the second mirror, an interference sideband is generated at the center and right side of the linear light diode array 1 80. . ^ The spectrometer shall measure the wavelength and bandwidth substantially in real time. Due to the laser repetition rate page of 4,000 Hz to Hertz, it is necessary to use a quasi-stone but not computationally intensive deduction rule. # Achieve predetermined performance with economical and streamlined processing of electronic devices. Therefore, the arithmetic deduction rule is better to use integers rather than floating-point math, and mathematical operations are better to be effective (no square root, sine, logarithmic, etc.). Specific details of the preferred deduction rule used in this preferred embodiment will now be described. FIG. 14D is a curve with 5 spikes, which represents a typical etalon sideband signal measured by the linear photodiode array 180. The height of the central peak is drawn on this paper.Medium__ 准 (CNS) A ^ T ^ 97 公 楚)

(Please read the precautions on the back before filling out this page), can be line-| equipment · A7 —--- —- ___! Z_ V. Description of the invention (32) ^ * The clothes are lower than its daggers. As the wavelength of the light entering the etalon is different, the medium 2 large peak will rise and fall 'occasionally to zero. This form causes the central peak to be uncomfortable = used for measuring; ^ i. The other spikes will shift towards or shift in response to the change in wavelength = the center is two long, so the position of the spikes can be used to determine the wavelength, and the width can be measured. The second area of the individually marked data window is shown in the HD picture. Data-The location allows the sideband closest to the central spike to be used for analysis. However, when the wavelength is changed to 5, the sideband is too close to the central peak (resulting in distortion results in errors), the first peak is on the outside of the data window, but the second closest peak is on the inside of the data window. The software causes the processing of the control module 197. The device uses the second spike. Conversely, when the wavelength shift moves the current spike outside the data window, away from the center tip, the software will jump to the inner sideband inside the data window. The data window is also shown in Figure 14B. The pulse poles for repetition rates ranging from 4,000 Hz to 6,000 Hz are fast-tracked. The preferred embodiment uses the hardware of FIG. 15. The hardware includes model MpC 823 microprocessor supplied by Motorola, Phoenix, Arizona; Ep 6016QC240 programmable logic device supplied by San Jose, California, California; 402; execution and data memory bank 4. 4; The special extremely fast RAM406 is used to temporarily store the data of the photodiode array in the form of a table. The second 4X1024 pixel RAM memory bank 408 operates as a memory buffer; As described in U.S. Patent Nos. 5,025,446 and 5,978,394, the previous memory device required analysis of a large number of PDA data pixel intensity data representing the interference sideband generated by the photodiode array i 80 etalon i 84 to determine the center line wavelength and frequency search. This process is quite time-consuming even if a computer processor is used. The original paper size is applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm) 541777 A7 B7 V. Description of the invention (33) One reason is that it needs to analyze about 400 pixel intensity values Find and explain the etalon sidebands for each wavelength and bandwidth calculation. The preferred embodiment of the present invention can greatly speed up this process by setting up a processor, which is used to find the main sideband of the processor operation for computing wavelength information in parallel. The basic technology is to use a programmable logic device 402. When the pixel data is generated, a sideband data table is continuously generated from the PDA pixel data. The logic device 40 also recognizes which set of sideband data represents the sideband data of interest. Then, when it is necessary to calculate the central wavelength and bandwidth, the microprocessor simply picks up the data from the identified pixels, and the required values of the central wavelength and bandwidth are different. This processing can reduce the calculation time of the microprocessor by a factor of about 10. The specific steps for calculating the center wavelength and bandwidth are as follows: 1) PDA 180 is set to operate at 2.5 million hertz, PDA 180 is processed by processing 400 instructions at a scan rate of 4,000 Hz from pixels ^ to 6⑼, and Pixels 1 to 1028 are read at 100 Hz. 2) The analog pixel intensity data generated by PDA 1 8 0 is converted to digital 8-bits by analog / digital converter 410 (0 to 255) 'The digital data is temporarily represented as 8-bits of each pixel intensity of the photodiode array 1 80 The value is stored in the ram buffer 408. 3) Programmable logic device 402 searches for sidebands on an almost real-time basis, continuously analyzes the data sent to RAM buffer 408, stores all data in RAM memory 406, identifies each pulse sideband, generates a sideband table for each pulse, and stores the table In RAM 406, and identify a set of best two sidebands for further analysis of each pulse. The technology used by the logic device 402 is as follows: A) Programmable logic device 402 analyzes each paper size sent by the buffer 408. It is applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm) 36 (Please read the (Please note this page, please fill in this page again)-Installation-: Line-541777 A7 ---____ B7 V. Description of the invention (34) ^ 'Pixel value, determine whether it exceeds the intensity threshold, and track the minimum pixel intensity value. If the threshold is exceeded, it indicates that a sideband spike has arrived. The programmable editing device recognizes that the first pixel above the threshold is the "rising edge" pixel, and the "bottom" is the lowest pixel value stored in multiple pixels in front of the "rising edge" pixel. This pixel intensity value is identified as the "lowest value" of the sideband. B) The programmable logic device 402 then monitors the subsequent pixel intensity values and searches for sharp edges. Monitor by tracking the highest intensity value until the intensity drops below the threshold intensity. C) When a pixel below the threshold is found, the programmable logic device recognizes the pixel number of the falling edge and stores its maximum value. The programmable logic device then calculates the "width" of the sideband by subtracting the rising edge pixel number from the falling edge pixel number. D) The rising edge pixel number, maximum sideband strength, minimum sideband strength, and sideband search 4 values are stored in the circular table of the 406 sideband section of the RAM memory bank group. Data representing up to 15 sidebands can store data representing up to ^ sidebands for each pulse, but most pulses generate only 2 to 5 sidebands in the two data windows. E) The programmable logic device 402 also identifies the "best" sideband of each pulse for each pulse through programming. This is done by identifying the end bands completely from 0 to 199 ® and the first side bands completely from 400 to 599 windows. After one pulse, (1) the pixel data is collected, and (2) the total time required to form the pulse sideband circular table is about 200 microseconds. The main time-saving advantage of this technology is that the search sideband is performed while the sideband data is read, digitized, and stored. Once the second best sideband is identified for the special pulse, the microprocessor 400 obtains the original pixel data from the second sideband area from the RAM memory bank 406, and calculates the bandwidth and center wavelength from this data. The calculation is as follows: This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 Public Love ·)

(Please read the precautions on the back before filling this page) tr ..... 1 line: 541777 A7 B7 V. Description of the invention (35 The typical shape of the side band of the etalon is shown in Figure 14D. Based on the programmable avoidance clothing In the previous work of setting 402, the microprocessor 400 identified a sideband with a maximum value of about 180 pixels and a sideband with a maximum value of about 45 pixels. The pixel data surrounding the two maximum values was defined by the microprocessor 400 analysis. The shape and position of the sideband are as follows: A) The half-maximum is the maximum value of the sideband minus the minimum value of the sideband, the difference is divided by 2 and the result is added to the minimum value of the sideband. For the individual rising edges and individual lower yak margins of the two sidebands, g 'two pixels have the closest value above the half-maximum value and the closest value below the half-maximum value. Then for each case, the microprocessor extrapolates between the two pixel values to define the D1 and D2 endpoints, as shown in Figure 18B, with an accuracy of 1/32 pixels. From these equivalent values, the inner diameter D1 and the outer diameter D2 of the rounded band are determined. Precision wavelength calculation Precision wavelength calculation is performed using rough wavelength measurements and D 丨 and D2 measurements. The basic wavelength equation is: λ = (2 * n * d / m) cos (R / f) ⑴ where λ is the wavelength in micrometers, and η is the internal refractive index of the etalon, which is about 1000 , D is the etalon interval, about 1542 microns for KrF lasers, about 934 microns for ArF lasers, controlled at + / _ i microns, m is a power, and the integer value of the sideband peak wavelength is approximately 1244 〇, R is the sideband radius, 130 to 280 PDA pixels, one pixel = 25 microns, this paper standard home standard _ M specifications ⑵ 幻 38 (Please read the precautions on the back before filling this page) _ 装 _ ， 可 — -Line 丨 541777 m A7 B7 V. Description of the invention (36 f is the focal distance from the lens to the PDA plane. Expand the cos term 'abandoning too small and negligible higher power terms to obtain: λ = (2 * n * d / m) [l- (l / 2) (R / f) 2] (2) Rewrite the equation with diameter D = 2 * R to obtain: λ = (2 * n * d / m) [l- (l / 8) ( D / f) 2] (3) The main work of the wavelength meter is to calculate λ from 0. It is necessary to know that f, η, d and η and d are etalon characteristics, and combine ^ and d to form a single correction constant named ND. Consider f as another correction constant named FD, match D units in pixels to obtain the pure ratio The integer power jaw system according to wavelength selection, and its sidebands Use .m accurate enough to change the system a coarse wavelength side band decided on the.

The benefit of these equations is that all large numbers are positive. The WCM microcontroller can calculate this value while maintaining 32-bit accuracy. Inventors prefer to use square brackets as FRAC. FRAO [l- (l / 8) (D / FD) 2] 于 Internally, the FRAC system is represented by an unsigned 32-bit value, and its root system points to the left of the most significant bit. FRAC is often slightly less than 1 and therefore achieves maximum accuracy. For the D range of {560 ~ 260} pixels, pRAC is in the range of [buojgj to [1-25E-6]. When the calculated ND value is loaded, the wavelength meter calculates the internal unsigned 64-bit value. The name is 2ND = 2 * ND, and the internal wavelength unit is 亳 micron (fm) = 10A_15m = 0_001 micron. The wavelength Λ is expressed as fwl of the precise wavelength, and the unit is also a micron. Restate the equation with these variables: FWL = FRAC * 2ND / m (5) 39

This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 public love) 541777 V. Description of the invention (37) The root point displacement of FRAC is processed differently to obtain FWL (unit nanometer) ---------- (Dream first read the notes on the back before filling in this page) Bring in the equation and insert the known rough wavelength named CWL (also in nanometers). Remove m: m = closest integer (FRAC * 2ND / CWL) ( 6) The nearest whole number is equal to the old scheme plus or minus FSRs until the precise wavelength position closest to the rough wavelength is reached. The wavelength is obtained by canceling equation (4), then equation (6), and then equation (5). Calculate the WL separately from the inner and outer diameters. The average is the line center wavelength and the difference is the line width. Bandwidth calculation The difference between thunder and radio frequency is (λγλ!) / ^. Apply a fixed correction factor to consider the etalon spike characteristic width added to the true bandwidth of the laser. Mathematically, the reverse rotation deduction rule is to remove the characteristic width of the etalon from the measured width, but it is too computationally intensive, so subtracting the fixed correction factor △ λ e can obtain sufficient accuracy. Therefore, the bandwidth is Δ λ = [(〇2-〇!) / 2] -Δ λ g: line-Δ λ 6 is determined according to the standard of the etalon and the true radio frequency bandwidth. Typical uses for this purpose are in the range of 0.1-1 micron. Improved etalon This embodiment uses an improved etalon. Conventional etalon erection solutions Typically use elastomers to mount optical components on the surrounding structure, limiting the position of the optical components, but minimizing the force on the components. The chemical compound commonly used in this project is room temperature vulcanizable polylithium oxygen (RTV). However, the organic vapours emitted by such elastomers may be deposited on optical surfaces, resulting in poor performance. In order to extend the performance life of the etalon, it is desirable to set the etalon inside a sealed enclosure that does not contain an elastomer 541777 A7 B7 V. Description of the invention (38). (Please read the notes on the back before filling out this page.) The preferred embodiment includes an improved etalon assembly, shown in Figures 14 and 14E at 184. The fused silica etalon 79 shown in Figure 14G is composed of a top plate 80 and a bottom plate 82 with flanges 81, and the second plate is made of a special grade of fused silica. The etalon system is designed to generate a sideband with a free spectral range of 20,000 micrometers at 193.35 nm when surrounded by a gas having a refractive index of 1.0003 and an accuracy equal to or greater than 25. Three fused silica spacers 83 with ultra-low thermal expansion separate the plates, and the spacers are 934 microns ± 1 microns thick. Thus, techniques well known to those in the optical device manufacturing industry are used to hold the etalons together by optical contact. The reflection ratio of the inner surface of the etalon is about 88%, and the outer surface is anti-reflective coating. The transmission of the etalon is about 5000/0. -、? τ—: Line 丨 The etalon 79 uses gravity to maintain positioning on the aluminum casing 84. Three low-force springs 86 press the flanges toward the three pads (not shown), but the indicator 85 is located The radial position of the center of the flange 8 1 is 120 degrees below the bottom edge. There is only 0.004 inch to 87 clearance along the top edge of flange 81 to ensure that the etalon stays in place approximately. Such close tolerance matching can also ensure that if any vibration or impact is transferred to the etalon system during installation, the relative speed between the optical component and the contact point of the housing will be kept to a minimum. Other optical components of the etalon assembly 184 include a diffuser 88, a window 89, and a focusing lens 90 with a focal length of 458.4 mm. > The emitter 88 may be a standard prior art diffuser commonly used upstream of the etalon to generate a variety of angles of incidence required for proper operation of the etalon. The problem with the prior art technique is that about 90% of the light passing through the diffuser is not at a useful angle, and the result is not focused on the photodiode array. But the wasted light increases the heat of the optical system and contributes to the degradation of the optical surface.

541777 A7 B7 V. Description of the Invention (39) In a more specific embodiment, a diffractive lens array is used as the diffuser 88. With this type of diffuser, a pattern is produced in the diffractive lens array that completely scatters the light but is only within an angle of about 5 degrees. Results ^ 90% of the light falling on the etalon was shot at a useful angle, and a much larger proportion of the light hitting the etalon was finally detected by the photodiode array. As a result, the amount of light incident on the etalon is greatly reduced, thereby greatly extending the life of the optical component. Applicants estimate that the incident light can be reduced to less than 5% or 10% of the previous technology, and there is equivalent light to the photodiode array. Better Collimation Using a Diffuser Diffuser Figure 14H shows a preferred embodiment of a structure that further reduces the light intensity through the etalon. This specific embodiment is similar to the specific embodiments discussed above. The sample beam (about 15 mm x 3 mm) from the mirror 182 passes upward through the condensing lens 400 and is then re-collimated by the lens 402. The beam is now collimated, the dimensions are reduced to about 5 mm x 1 mm, through the etalon housing window 404, and then through the diffractive diffusing element 406, which in this case (for ArF laser) is Alabama Diffraction Diffusers from Mems Optics, Inc. The component part number is D023-193, which converts substantially all of the 193 nm light in any incident collimated beam of any cross-sectional configuration into 2 degrees in the first direction and 4 degrees in the vertical first and second directions. Beam. The lens 410 then "focuses" the unfolded light beam onto the rectangular pattern covering the photodiode array 180, as shown in FIG. The active area of the photodiode array is approximately 0.5 mm and approximately 25.6 mm long. The dot pattern formed by the lens 410 is approximately 15 mm x 30 mm. The diffractive diffusing element thoroughly mixes the spatial components of the light beam, but substantially maintains all the light beams within the limits of 2 degrees and 4 degrees. This paper size applies the Chinese National Standard (CNS) A4 specification (210X297). 42; ·- ----- U——r: (Please read the precautions on the back before filling this page), one t —: line — 541777 A7 B7 V. Description of the invention (40. Let the light passing through the etalon substantially reduce and can be used. Effective use. The reader understands that by further reducing the energy of the etalon beam, the dot pattern of the short dimension of the photodiode array can be realized. However, further reduction to less than 15 mm will make optical alignment and communication more difficult. Therefore, the designer should consider the size of the dot pattern to be compromised. Another design is used in the KrF.radiation system operating at about 248_327 nm, providing a similar design to adjust the wavelength. In this specific embodiment, the focal length of the lens 400 is about 50mm (lens is part number OILQP001 of MeUes GHOT). The collimator lens 40 has a focal length of _2mm (Eνι Laser part number PLCC_1〇 (M〇3_uv) The diffractive diffusing element is Optical Co., Ltd. part number D023-248. In this specific embodiment and in the specific embodiment of ArF, the distance between the two lenses is appropriately positioned by the spacer 416. The applicant estimates that the laser operating at 2000 Hz passes the laser of the etalon The beam energy is about 10 watts, which is not enough to cause significant thermal problems with the etalon. In other preferred embodiments, the laser beam will come to the focal point between the lens 400 and 400. A suitable lens in this example must be used The well-known choice of optical technology. The feedback control of pulse energy and wavelength is based on the previous measurement of the pulse energy of each pulse, and controls the pulse energy of subsequent pulses to maintain a predetermined pulse energy and also maintain a predetermined total integrated dose of a specific number of pulses. In U.S. Patent No. 6,005,879 "Pulsed Energy Control of Excimer Lasers", incorporated herein by reference. Laser wavelengths can be used in feedback configurations using wavelength measurements and techniques known in prior art such as US patents Controlled by Known Technology No. 5,978,394, "The standard of the paper is turned over the standard _ standard (格 机 格 (2 songs 297)) ------—— (Please read the notes on the back first 嗔(Fill in this page), Yiding-line-five, invention description (41 knife laser wavelength system ", also incorporated here by way of bow 丨 W. The applicant recently developed the wavelength fine-tuning technology, which uses the I electric driver to provide The fine-tuning mirror moves extremely fast. Several technologies are described in US Patent Application No. 608,543, "Thunder Radio Frequency Band", filed on June 30, 2000, and incorporated herein by reference. Sections 16A and 16 The relationship is hidden by this purpose, showing the main elements of the technology. Piezoelectric stacks are used for extremely fast mirror adjustments. The adjustments obtained by the lever arm of the stepper motor of the prior art are much slower. The piezo stack adjusts the arm fulcrum position. π is a novel design of a novel piezoelectric actuator driven by a combination of PZT-stepping motor. Figure 16 is a block diagram showing the characteristics of a laser system. This system is very important for controlling the output laser beam wavelength and pulse energy. The display line narrowing module 15K includes a two-laser laser beam expander, a fine-tuning mirror μ, and a grating. The wavelength meter 104 monitors the output beam wavelength and provides a feedback signal to the LNp processor 106. The LNP processor 106 controls the position of the fine-tuning mirror 14 by operating the stepper motor and ρζτ stack, which will be described in detail later. The operating wavelength can be selected remotely by the laser controller 102. The pulse energy is also measured at the wavelength meter 104, which provides a signal for the controller 102 to control the pulse energy in the feedback configuration, as described above. Fig. 16A is a block diagram showing a PTZ stack 80, a stepper motor 82, a mirror 4 and a mirror holder 86B. 16B 1 is a schematic view showing the structural details of a preferred embodiment of the present invention. The stepper motor is used to generate a large change in the position of the mirror 14 through the 26.5: 1 lever arm 84. In this case, the diamond pad 81 at the end of the piezoelectric driving device 80 is provided with a spherical tool for contacting the fulcrum of the lever arm 84. The top of the lever arm 84 and the mirror 44 r are installed ...: (Please read the precautions on the back before filling in this page) Order —: Line — This paper size applies to China National Standard (CNS) A4 (210X297 mm) (42 ~ & quot ~~ 541777 The contact between the 86 of the A7 seat 86 is provided with a cylindrical double-pointed nail on the lever arm, and four spherical ball bearings (only shown both) are mounted on the lens holder, as shown by M. Piezo Drive 80 Mounted on 1 ^ 1 > frame with piezo mount 808, and stepper motor woman I on the frame with stepper motor mount 82A. Mirror 14 uses three aluminum balls (16B1 The figure shows only one) is mounted on a lens mount 86 with a three-point mount. Two elastic bullets apply a compressive force to fix the mirror back to the sphere.

FIG. 16B2 is a preferred embodiment slightly different from FIG. 16B1. This embodiment includes a telescopic joint 87 to isolate the piezoelectric driving device from the internal environment of the LNp. Such isolation prevents damage to the piezoelectric element caused by ultraviolet rays, and prevents pollution that may be caused by the exhaust gas of the piezoelectric material. Refer also to Figure 9D and the accompanying text for another LNP design that provides a much improved lnp seal. Early prism plate A preferred embodiment of the present invention includes a single cymbal plate on which three beams for three beams are provided. Each of the three cymbals is positioned on the plate 7 during manufacturing, and the cymbals are inserted between the three movable positioning pins, and the positioning holes of the precision boring are inserted through the bottom of the plate. Then each piece (using 3M 2216 two-part epoxy resin) was glued to the three small points (0.1 inches in diameter) of the board, the glue was hardened and the latch was removed. This technique ensures accurate positioning of plutonium relative to each other and simplifies the manufacturing of LNP. During the gluing step, each tadpole used a side-by-side test fishing line to clamp the iron to fix the distance from the board by 0.004 inches, so that the resulting glue points produced three extremely short feet with extremely slightly elastic feet. It can minimize the pressure caused by temperature changes. The position of the diaphragm is manually adjusted to the direction of the laser beam passing vertically through the chamber, as shown in Fig. 16A. This paper size applies to China National Standard (CNS) A4 (210X297mm) -45 _ (Please read the precautions on the back before filling this page)

541777 A7 V. Description of the invention (43) ^-Test result-Figure 16C shows the actual test data obtained by the laser of the specific embodiment of Figure 16B2. The line graph is the deviation of the target wavelength from the average viewing window. The deviation was reduced from about 0.05 micrometers to about 0.05 micrometers. This specific embodiment compares the aforementioned stepper motor drive system with a major system acceleration, but the speed is not fast enough to do pulse-to-pulse adjustment. Early mirror positioning methods required about 7 milliseconds to move the mirror 14, resulting in a pulse of 2 Hz that made it impossible to adjust the pulse wavelength. In the early technology, the lever arm was pivoted with the axis of the zone as the axis, and the movement of the mirror was less than that of the stepper. The total travel distance of the stepper is 1/2 inch (127 mm) and the first step, so the distance of each step is about 2 microns. With a reduction of 1-26, the first-order moving mirror is about nanometers, and the laser wavelength is typically changed by about 0.1 micron. In the fast-acting technology shown in Figure 12A, a piezoelectric stack 80 is added to the pivot position of the lever arm. Preferred Piezoelectric Reactor = Model ρ_84〇 · ^ for Des Vodborn, Physical Instrument Company. This C stack produces a linear adjustment of about 3.0 microns, with a driving voltage change of volts. This range is equal to the earth step of the stepper motor. The piezoelectric stack corresponds to the control signal in less than 丨 microseconds, and the system responds to the update signal at a frequency of 4000 Hz. In a preferred embodiment, the pulse rate is controlled at 4000 Hz. The pulse rate is not based on the previous pulse, but is based on the pulse before the previous pulse to allow sufficient time for wavelength calculation. However, this specific embodiment provides a 7 factor improvement over the previous art design with a 7 millisecond delay. This provides faster feedback control. The better feedback control deduction rule is described in Section 16D®. In this deduction, the average wavelength is calculated for each pulse control wavelength, the last 4 pulses, and the last 2 pulses. If any average value deviates from the target, the Chinese paper standard (CNS) A4 specification (210X297) is applicable. (Please read the precautions on the back before filling this page) Order | Line-541777 V. Description of the invention (44 skins If the long-term is 0.02 micron, no adjustment is needed. If the deviation is greater than 0 · ⑽ UU meters ^ ㈣ Piezo Stack® 8G adjusts the mirror assembly to provide a wavelength calibration average. Which one is used and how much has passed since the last adjustment Time to decide. When the piezoelectric stack approaches its range of postal warfare (or in order to provide a more usable feu 'use 45 to 55% instead of the buckle to the 7.% range), advance the piezoelectric motor by two steps The stacking is maintained within its control range. It takes about 7 ί: seconds to complete the step because of the stepper motor. The deduction rule can make several piezoelectric adjustments during one step of the stepper motor. Pre-fine adjustment and active fine-tuning The example can be used for purposes other than chirp correction. In some cases, integrated circuit lithography operators want to change the wavelength at a predetermined reference. In other words, the 'target wavelength λτ may not be a fixed wavelength, but may be based on material requirements. Change, follow a predetermined pattern change, or use early past wavelengths or its number of daggers, and continuously or periodically update the learning deduction rule to change the adaptability feedforward. Another: Send a preview of the preferred embodiment of the present invention including feedforward deduction. Laws. These algorithms can be coded by the laser operator based on a known burst operation pattern. Outside, this rule can be adapted, so the laser control device is called the test cluster; Display, and then modify the control parameters to provide the mirror 14 to adjust the pre / month skin length to avoid or minimize displacement. The adaptive feedforward technology involves software that uses data from one or more previous burst pulses to specify the repetition rate. Establishing the Zeber mode and using the PZTT stack to transform the Zeber effect. To properly design the Zeber transformation, two pieces of information are needed: (1) the impulse response of ρζτ · and (2) the Heber shape. For each repetition rate " That is, the paper size of this paper is applicable to the national standard (CNS) of A4 specification (〇〇297297) 541777 V. Description of the invention (45) The impulse response of the pzt stack is reversed to the milk stack (with Suitable series of pulses' which, when applied) will lock against Li. This operation can be performed offline through the sound m through research behavior. The data sequence can be used for two purposes: T and the repetition rate as indexes. This table is picked at the end of the operation period, and the waveform data is used for adaptive feedforward transformation. Possibly, when the mother-to-child repetition rate is changed, when using a small number of bursts, the material-like mode and the PZT impulse response mode can be obtained almost instantaneously based on the progressive calculation between the mode and the data. wrong

Errors can be updated every N bursts (eg, adaptation). A better deduction rule is described in Figure 16E. The beginning of the burst pulse can be controlled using the deduction rule shown in Figure 16E. The letter k indicates the number of bursts. The burst is divided into two areas, one for k and one for area. The k-zone is used for the pulse number less than [^^] (defined long enough to cover the mouth). The proportional constant Pk, the integral constant W, and the integral and SLCEk of the line center error are used for each pulse number. The Pζτ voltage corresponding to the pulse number of the k-zone in the next burst is determined by this constant and sum. After the k-th pulse, the traditional proportional-integral routine controls the PTZ voltage. The next burst pulse voltage is the current voltage plus P1CE + P2LCE. A flow chart illustrating the main steps of this deduction rule is shown in Figure 6E.

High-load-cycle LNPs are known to narrow the package; however, previous techniques taught that the sweep flow is maintained directly on the grating surface. The sweep flow is typically provided through a port located behind the grating surface. However, the applicant found that at a very high repetition rate, a layer of hot gas (nitrogen) developed on the surface of the grating caused wavelength distortion. This distortion of 48 paper standards is applicable to the Chinese National Standard (CNS) A4 specification (210X297 public love) 541777 A7 B7 V. Invention description (46 (Please read the precautions on the back before filling this page) to V. You can borrow The active wavelength control correction discussed earlier. Another method is to sweep the surface of the grating, as shown in Figure 17. In Figure 17, a 10-inch long 3 / 8-inch sweeper is directly bolted to the top of the 61 mm hole, 1 / 4-inch Interval) to provide a sweep stream. As described in the next section, the sweep gas may be nitrogen from a pure nitrogen supply. In the specific embodiment of the car, LNp is swept with helium, which will be described later. Other techniques are shown in Figures 17A, πβ and 17C. Specific techniques for controlling wavelength and bandwidth are described in the following patent applications, each of which is incorporated herein by reference: U.S. Application No. 09 / 794,782, dated February 27, 2001, entitled "Piezoelectric The driver does laser wavelength manufacturing system ", U.S. Application No. 10 / 〇27,210, the application date is December 21, 2001, and the name is" Plasma Driver for Laser Wavelength Control "; and Application No. 10 / 〇 No. 36,925, please refer to December 21, 2001, the name "Laser Light I " General Project of Lithography Process". Knowing the system: The first embodiment of the present invention includes an ultra-pure nitrogen sweeping system, which provides greatly improved performance and substantially extends the life of constituent components. Fig. 19 is a block diagram showing the main structure of the first preferred embodiment of the present invention. In this specific embodiment of the system, the five types of excimer laser constituent elements purged by nitrogen are LNP2P, a high-voltage constituent element 4P installed in the laser chamber, a high-voltage constituent element 4p, and an upstream pulse power constituent element. high-voltage cable 8P, output coupler 12p, and wavelength meter 141 >. Each component 2P, 4P, 8P, 12P, and 14P is housed in a sealed container or chamber, each of which has only two ports, a nitrogen inlet and a nitrogen outlet. The source of lice milk 16P is typically a large nitrogen tank of the integrated circuit manufacturing plant (a typical paper size is suitable for financial and economic standards (CNS) M specifications cricket tearing 541777 A7 ------------ B7. Note (Type 47) is maintained at liquid nitrogen temperature), but can also be a relatively small vial of nitrogen. Nitrogen source The gas is sent from the gas source 16P, enters the nitrogen sweep module 17P, and passes through the nitrogen filter 18P to the distribution panel 20P. It contains a flow control valve to control the flow of nitrogen to the component being purged. As for each component, the sweep flow guide return module 17P to the flow monitor unit 22p ′ is used to monitor the flow passage returned by each sweep unit, and when the monitored flow is lower than a predetermined value, an alarm is activated (not shown in the figure). Figure 19A is a line diagram showing specific constituent elements of the preferred embodiment, including a number of additional nitrogen structures that are not particularly relevant to the sweep structure of the present invention. Nitrogen filter An important structure of the present invention includes a nitrogen filter. In the past, in the fabrication of integrated circuit lithography for excimer laser manufacturing, it was considered that a nitrogen purging filter was not necessary 'because the gas specifications of commercially available nitrogen were almost good enough, and the specifications that the gas met were clean enough. However, the applicant found that occasionally the source gas was out of specification 'or that the nitrogen line leading to the sweep system contained contamination. The pipeline may also be contaminated during maintenance or operating procedures. The applicant judged that the addition of the filter to further reduce the probability of damage caused by pollution obtained an excellent guarantee. A preferred nitrogen filter is a Model 500K inert gas purifier available from Aeronex, San Diego, California. This filter can remove ΗΑ, ο], co, CO2, effluent and non-methane hydrocarbon gas to a level below ppb. This filter can remove 99.9999999% of all particles of 0.003 microns or more. This paper size applies Chinese National Standard (CNS) A4 specifications (210X297 mm) 50 (Please read the precautions on the back before filling out this page) • Binding-Order ----: Line-541777 A7 B7 V. Description of the invention ( 48 flow monitor single tl 22 The flow monitor is set for each of the 5 sweeping components. These components are commercially available units with a structure that issues an alarm when the flow is too low. Stainless steel (316SST) composition with electro-polished interior. Some types of plastic tubes composed of PFA 400 or ultra-high-purity Teflon can be used. Part or all of the sweep gas can be recycled as shown in Figure 19B. This type ^ In this case, add a blower or water-cooled heat exchanger to the sweeping module. For example, the sweeping flow from the optical component can be circulated, the sweeping flow from the electrical component can be exhausted, or part of the combined flow can be exhausted. LNP helium sweep In a preferred embodiment, the LNP is swept with helium, and the remaining components of the beam path are swept with nitrogen. Helium has a lower refractive index than nitrogen, so the use of helium minimizes the thermal effects of LNP. But helium is about 1,000 times more expensive than nitrogen. Improved seals have found significant advantages in providing an extremely "clean" beam path. Laser optics degrade rapidly in the presence of high-energy ultraviolet radiation combined with multiple forms of pollution, including oxygen pollution. The preferred technique for enclosing the beam path is described in US Patent Application No.-I_. The application date is November 14, 2001. The name is "Gas Discharge Laser with Improved Beam Path." Applicable to China National Standard (CNS) A4 specification (210X297 mm) J ------—— (Please read the precautions on the back before filling this page), = tt line-541777 A7 —— _ —_B7 V. Description of the invention (49) Enter here. Figures 19C, D, E and F are taken from the application. Figure 19C is a schematic diagram showing similar expansion and contraction between the various components of the gas discharge system of the main oscillator described above. Figure 19D shows a modified example, including the telescopic joint configuration of the LNp stepper motor and the interface between the sealed motor and the LNp enclosure. Figure 19 £ shows the thermal decoupling orifice of the LNP, which minimizes the heating and also surrounds the LNP inlet , So Sweep with relatively cheap helium gas. The gas is sent by the LNP as shown in Fig. 19C and Fig. 95 through the moon-to-chamber unit. Figs. 1, 2, 3, 4 and 5 of 19F show the easy-to-seal telescopic joint seal to provide The laser module is sealed, but allows the module to be quickly and easily decoupled. Allows the module to be replaced quickly. Figure 19G shows a special sweep configuration to sweep the high-intensity part of the wavelength meter. This special sweep is described in the next section. System advantages The system described here represents a significant improvement in the KrF laser, the special ArF laser, and the & laser length J quasi-knife field radiation performance. It can basically eliminate the problem of contamination, resulting in a substantial extension of the component component life and a substantially improved beam quality. . In addition, since the air leakage is substantially eliminated except for the air outlet, the flow rate can be controlled to a predetermined value, which has the effect of reducing the nitrogen demand by about 50%. Sealed shutter unit with power meter The first preferred embodiment includes a sealed shutter unit of $ 00. There is a built-in power meter as shown in Figures 20, 20A and detailed drawings. With this major improvement, the shutter has two functions, the first being a shutter that blocks the laser beam; and the second full beam power meter to monitor the beam power when measurement is needed. Figure 20 is a top view showing the main components of the shutter unit. The components include the device 502, the beam dumper 504 and the power meter 506. With this paper size Standard Finance Standard (CNS) A4 ^ (210) ^ 57 52 (Please read the notes on the back before filling this page)

541777 A7

Order

541777 A7

5. Description of the invention (51) 95% of the energy of B7 is used as the output beam. Approximately 4% of the reflected beam is reflected from the mirror 171 to the set detector 172, where the pulse energy is measured. (Please read the notes on the back side of the reflected beam through the mirror 171 before filling in the rest of this page shown at 61A and entering other monitors of the wavelength meter). At 4,000 Hz, 5% output can represent a large amount of ultraviolet light, so special care must be taken to ensure that the gas in this part of the beam path is extremely clean and pure. To achieve this, the wavelength meter was modified to seal the area between the upstream end of mirror 170, the upstream end of mirror 171, and the front end of the window of detector 172 and the rest of the wavelength meter. Special sweep streams flowing in and out of this area are shown at 62A. The rest of the wavelength meter is swept by the second sweeping stream shown in Figure 64. The sweep stream 62A is limited to the wavelength juice by the seals of the mirrors 170, 171 and the detection transom 172. The sweep stream is sent from this region along the laser output beam path and is returned to the output coupler module 68A through the stretched region 6A to sweep the module. Then it flows through the telescopic joint unit 70A, flows into the window block 72A, flows out from the outlet of the window block and the outlet of the telescopic joint unit 7GA, and then returns to the nitrogen selection and removal module 17 through the pipeline, as shown in Fig. 19 and Fig. 74. The downstream end of the window 170 is swept using a sweeping flow from the shutter module 5K. As shown in Figure 19, the sweep stream is from module 17; or in some cases, the window 76A is removed and the shutter module outputs an open connection to the beam line being cleaned. In this case, at the exit of 78a The sweep line can be directed to a customer sweep return system or vented to the atmosphere. Constant optical parameters As lithography lasers age, beam quality characteristics change. usually. Quality tends to deteriorate slowly. Even if the beam quality no longer meets 2 divisions after use, it is usually necessary to replace the main component components (such as the f-beam chamber, coffee and / or wavelength meter). For example, during the life of the ambiguity laser, the quality of the laser beam is substantially 54. The paper size is applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm) 541777. 5. Description of the invention (52) Micro-materials modified within the specification range. For the use of sigh leaves for optical quality laser beams: integrated circuit lithography manufacturers can be problematic. "Outperformed" by Yi Shutian, the quality results will lead to undesired changes in the quality of integrated circuits. The way of H-determination is to provide a kind of lightning. The laser material can keep the quality of the laser beam substantially stable during the use of the laser. This project can be achieved using the technology described in US Patent Application No.-which is filed at the same time as this case. Here, a piezoelectric actuator fine-tuning mirror is used to provide wavelength stability values and bandwidth values that correspond to expected nominal values rather than lightning Shoot the best value. Pulse energy can be controlled using the aforementioned feedback control device to maintain a stable energy value at the expected normal value instead of the best possible value. Fluorine concentration and laser gas pressure can also be passed, section: Generate the expected laser beam quality value, instead of the most stable value of pulse energy and wavelength, and the narrowest possible bandwidth value. Extending the pulse duration The optical components of the integrated circuit lithography equipment, especially the lens components, are extremely expensive. These constituent elements are known to deteriorate by using short-wavelength ultraviolet light. It is known that for a given pulse energy, the degradation increases as the pulse length decreases. Therefore, a longer pulse length is equivalent to a longer lens life. So 'longer pulse lengths are a significant improvement over this excimer laser. The ArF laser at 193 nm has a far worse serialization rate than the 248 nm KrF laser. Designed for the most effective laser performance, typical pulse times for KrF and ArF narrowband lasers with pulses in the range of about 5 millijoules to about j 〇 millijoules are about J 5 nanoseconds to 25 nanoseconds. Figure 23A shows the characteristics of a typical excimer laser pulse. This special feature is voltage or Cp, laser-free and lateral fluorescence with laser. 55-Order this paper size applies Chinese National Standard (CNS) A4 specification (210X297 mm) 541777 5. Description of the invention ( 53 / =? Of laser pulses. The applicant conducted experiments with several techniques, using field pulse durations such as capacity expansion, but these techniques caused the overall performance to be much lower and worse. The pulse J y configuration of the "Current Overshoot Maximizer" successfully extended the pulse duration. The current overshoot maximizer pulse power and current overshoot maximizer ("c〇m") circuit is shown in the figure. One clue is that the current overshooting strategy is feasible, as shown in Figure 23A. It is obviously easy: the laser intensity in this picture has two exact bulges. The first bulge is caused by the initial positive current cycle, and the second smaller bulge is due to the first negative. The current cycle is caused by the discharge. The difference in intensity is because the energy stored in the CP just before the negative current cycle is much less than the energy stored in Cp just before the first positive current cycle. If the laser amount at the negative current cycle can be increased , Then knot The resulting laser pulse will be about twice as wide. For reference, the results of the ⑯ laser displayed on the baseline are shown at 23BBI. At this time, it is not like KrF, and there is no laser at all in the negative current period of ArF. , ^

cp value and Lp value determine the phase of electric pumping during each part of the waveform ♦ Important. In this case, the pumping level needs to be increased during the negative current cycle. Two ways can be used to achieve the "first way" is to lower the bid, so a larger 1 negative Vcp is applied when discharging; the second "reduces the Cp, so it is higher for the specified storage energy" and can overcome the head of the higher peak current. Faster inductance problem. Red = 6.hH drops to 4 · 1ηΗ, and the negative period 豕 raises the sending level above ^ 's limit value' as shown in Figure 23C. It can be seen from the figure that the radiation energy is caused by the negative current period. Decrease CP value to further increase laser transmission during the negative current cycle. I This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) 541777 A7 -----— B7 — V. Description of the invention (54 ) Level, as shown in Figure 23D. The laser output during the negative current cycle shown in the figures is still much lower than. Intensity level. One of the reasons is that the matching of the equivalent value and the value change improves the discharge impedance. When LpACp decreases, a larger amount of cp storage can be transferred to discharge during periods when the discharge has a high impedance. The discharge impedance over time was measured and the results are shown in Figure 23 £. The discharge impedance takes about a nanosecond to collapse from the million ohm value to the i ohm region. If the “and formed discharge circuit is fast enough, all the energy stored in Cp will be dissipated in a high-impedance discharge, and there will be no voltage overshoot at Cp. The reduction of Cp and Lp is required for the negative current cycle to achieve laser, The decrease in current also increases the coarse Cp energy component, which is absorbed by the discharge during the first positive current cycle. As the absorbed energy increases during the positive current cycle, less energy remains in Cp during negative voltage overshoot, and therefore less The energy can be used to drive the negative current cycle. The schematic diagram of COM is shown in Figure 23f. This circuit is similar to the double pulser used in the 2X program in the early stage, but the external spike capacitor is replaced by a so-called inductor. The setting between Lcom and Cp can be set. The size of the v * s product of a saturated inductor can be switched at the same time as the discharge is triggered. With this configuration, the current will flow from the peak capacitor to the discharge and COM inductors. The sequence of events is as follows * 1) Cp is a pulsed power mode The charging pulse of the group (Figure 23G). 2) The COM saturable inductor switches at the same time as the burst discharge (Figure 23H). 3) The current starts to accumulate, flowing Cp to discharge and Lcom (Figure 231). 4) The Cp voltage passes through zero, and the current passing through the discharge begins to subside, but the Lp paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) 57 (Please read the precautions on the back before filling this page) 丨Assembly 丨 • Order—: Line 丨 541777 A7 B7 V. Description of the Invention (55 Inductance and Lcom inductance force the current to continuously flow out of Cp, forcing Vcp to be negative. Because the Lcom inductance is greater than Lp, its current flow is much calmer than the discharge current flow Slower (Figure 23J). 5) The negative voltage of Cp (thus the negative voltage of Lp) reverses the discharge of the current flow, but the current flowing through Lcom remains in the same direction because the flow

Lcom's current has a larger value than Lp and therefore has a longer time constant (Figure 23K). 6) The current flowing through Lcom no longer flows out of Cp. Instead, it flows through discharge and presents a much lower impedance than Cp. The reverse discharge current is driven by the negative voltages of Pa and Kp flowing through Lcom. The half-cycle amplitude of the negative current flowing through the discharge in this way is increased by the presence of the COM inductor (Figure 23L). The optimal operation of COM depends on minimizing the total inductance Lcom of the coM inductor. Two factors can cause the value of Lcomm to increase. The first factor is that the Lcom saturable switch needs to have a sufficient v * s product. As the product requirement of ¥ * 8 increases, the ferrite core with a larger volume is required, so the saturation inductance of the switch also increases. The preliminary COM experiment uses a 100 nanosecond transmission time pulse power system. When charging Cp, the core volume is required to be 8 toric cores, each of which is 1.0 inch inner diameter X 2.5 inch outer diameter X 0.5 inch thickness. Later experiments used a 75 nanosecond transmission time pulse power system developed for 2X programs. Using this transmission time, only 6 cores are needed, and the saturation inductance of the com switch is reduced by 25%. The second factor of Lcom's inductance is the longitudinal inductance of the spike capacitor bus. All schematic diagrams shown so far represent Cp and Lp as a single block circuit element. These circuit elements are actually scattered. For example, this paper size applies Chinese National Standards (CNs) A4 specification (210X297). (Please read the notes on the back before filling this page.) The peak capacitance consists of 28 capacitors with individual values of 0.59 nF. A more realistic schematic representation of the laser cavity is shown in Figure 3. The peak capacitance is now decomposed into Cpl, CP2, ... Cpn. The head inductance is also broken down into individual elements. Except for Cp and Lp, the constituent elements are bus inductance elements LM to. The components of the bus bar inductance represent the longitudinal inductance between the spike capacitors. Such circuit components are usually not important because they do not participate in the Cp-Lp-Zd circuit and their inductance is smaller than the Lp_l inductor on the inside of the last stage of the pulse power module. When using COM, COM is not a decentralized component, but instead feeds the peak capacitor busbar at a single point, so this kind of busbar inductor does become important. Therefore, before reaching the discharge, the current flowing from Lcom passes through the unequal amount of bus inductors. Design a novel spike capacitor interconnect board that eliminates as much as possible the spike valley as the magnetic flux between the top and bottom of the bus. The effect of this magnetic flux rejection plate can be seen in the energy frequency captured on the spike capacitor busbar after the main discharge. Add the lower flux rejection board, and get a 14% reduction in inductor inductance per length. A similar decrease was obtained by increasing the upper flux exclusion plate. Minimizing the bus inductance can greatly improve the efficiency of the COM system. Because the value of Lc0m is small, Lc0m forms a reverse current that is higher than the surface discharge peak. The initial COM measurement was performed using a total pressure of 36 kPa and an ArF gas filled with partial pressure fluorine. This gas mixture indicates that the pressure of the baseline mixture is greatly reduced from the 44o total pressure. Over the past 13 months, multiple developments in pulse power and chamber design (most of which have been listed above) have allowed for a gradual reduction in gas pressure while still maintaining a satisfactory 193 nm optical gain. I hope that the reduction of operating pressure comes from the experiments that the lower laser gas pressure was found. Is this paper applicable? Ia standard-59-(Please read the precautions on the back before filling this page)

541777 A7 _________B7___ One. V. Description of the invention (57) The force 'high repetition rate performance (1000 Hz and above) is relatively stable. At 360 kPa pressure level, the efficiency of the laser system is reduced to 20% to 30% when using coM, and the characteristic double pulse width can only be obtained for Vco values of 900 Volts or above. After several experiments, it was found that using a lower total pressure of 30 kPa, the reduction in COM efficiency is similar to using a KrF laser, and a low Vco value of 800 volts can achieve successful operation. Vcp and laser waveforms of various charging voltages without COM are shown in Figure 23M (1-4). Vcp and laser waveforms including COM are shown in Figure 23N (1-4). The change in efficiency when using COM is shown in Figure 230. As the total gas pressure decreases, the difference between the COM-free configuration and the COM configuration becomes smaller. Figure 23P shows measured 193nm pulse duration with and without COM. COV on force π [can increase the laser pulse duration from about 8 nanoseconds FWHM to 31 nanoseconds FWHM. Unfortunately, the FWHM value of 8 nanoseconds is not the baseline value of the 2X program. For comparison, the first! The graph shows the 16.8 nanosecond FWHM pulse duration of the 2X program baseline performance. The results in Section 4 show that COM can effectively double the duration of the 193 nm optical pulse. Other performance criteria are equally important, such as line narrowing spectral performance, energy stability, and electrode erosion. This section contains performance measurements for COM-narrowed laser systems. The KrF configuration is used because there are still many uncontrolled variables in the height-narrowing ArF system. The narrowing efficiency of the line with COM in place is essentially the same as the unused COM, as shown in Figure 23Q. At lower operating voltages, the output can show only small defects relative to the operating voltage, while at higher operating voltages there is no difference at all. FWHM spectral performance measurements are shown in Figure 23R. Make this paper size applicable to Chinese National Standard (CNS) A4 specifications (2) 0X297 mm 60 (Please read the precautions on the back before filling this page)

541777 A7 _B7_ V. Description of the invention (58) The COM result is used to obtain a slightly lower FWHM spectrum and a more consistent spectrum over the entire operating range. The most likely reason for using the consistent spectrum of COM is that using COM in the operating voltage range does not change the pulse shape, but the pulse shape changes dramatically when COM is not used. 95% spectral performance is shown in Figure 23S. As the operating voltage decreases, the 95% spectral integration of the two configurations decreases. This effect is well known and is believed to be due to the reduction in cavity gain at lower operating voltages. With the decrease of cavity gain, the weak laser "wing" of the spectrum is more impacted than the main body of the spectrum, because it is closer to the laser threshold. It is also well known that longer pulse times result in a 95% reduction in spectral integration. For normal cases where COM is not used, the two conditions of suppressed spectrum wing and suppressed pulse duration conflict with each other, because the pulse duration decreases with lower operating voltage, but tends to reverse and increase the suppression of the wing 95% spectral integration. Since COM can maintain a constant pulse shape over almost the entire operating voltage range, the suppressed spectrum can also achieve a complete effect. Extremely improved 95% spectral integration for lower operating voltages. For example, at the output level of 10 millijoules, a 95% reduction in spectral integration can be achieved by 45% using COM. For broadband and line narrowing operations, energy stability is another important performance parameter. There are several ways to specify the laser energy stability, and there are several modes of operation that can measure the laser energy stability. It has been found in the past that all of these methods are proportional to a simple measurement of the energy standard deviation σ to some extent when operating in constant voltage mode. Figure 23T shows the measured value of the energy change in the constant voltage mode of the line narrowing KrF operation. The low repetition rate performance of COM is slightly worse than the performance without COM, but there is no substantial difference. As there is currently no Chinese paper standard (CNS) A4 size (210X297 mm) for this paper size 61 (Please read the precautions on the back before filling this page)

可 I

Line I 541777 A7 B7 V. Reason for the determination of (59) method of the invention, the energy stability performance for repetition rate higher than 1 500 Hz is degraded due to the use of COM. Once at a higher repetition rate, the probability of poor energy stability performance increases, and after the main discharge, the energy remaining in the COM inductor causes bursting electron flow activity. Because the COM inductor stores energy through the flow of current, when the discharge goes out and becomes open, the voltage across the COM inductor will ignite, trying to maintain any current flow that still exists. Sufficient voltage ignition across the COM inductor and thus across the electrode will again trigger a burst voltage, but this time it results in a non-uniform and poorly pre-dissociated gas medium, which results in an arc or a burst of electron flow. These bursting electron currents have extremely low energy compared to the main discharge, the former being 1 millijoules and the latter being 2-5 joules. Although these energies are small, the energy density of the burst electron flow is much higher than that of the main discharge because it is concentrated in the polar cell. The high energy density of such a community forms acoustic energy, which is in the form of an overvoltage wave emitted from an arc region. At low repetition rates, this acoustic energy may be dissipated before the next laser pulse, and thus has little impact on laser operation. However, at higher repetition rates, a large amount of audio energy remains, resulting in pressure fluctuations that interfere with discharge quality. In addition, the higher repetition rate approaches the first resonance of the chamber volume, and an increase in the frequency of sound from the burst electron flow may pump these resonances to form an acoustic standing wave inside the chamber. Other evidence for increasing the burst electron flux density when using COM is based on the comparison of fluorine consumption. With COM, the time between injections is shortened, suggesting that the symbol usage is increased by 33%, so the electrode dissolution rate is also increased. Cymer found that the electrode erosion system is strongly related to the size, number, and frequency of the bursting electron currents that appear between the electrodes. The use of COM seems to improve the burst electron flow activity 62 (Please read the precautions on the back before filling this page) This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) 541777 A7 V. Description of the invention (60 'Therefore it is expected to adversely affect the life of the chamber. Another technique for pulse multipliers to stretch the duration of a pulse is to divide each pulse into two or more parts, delay the first part and then join the parts. One of them This technology is described in U.S. Patent No. 6, 〇67,311, which was assigned to the applicant's employer in a manner that is incorporated here. The other technology is described in U.S. Patent Application No. 10 / 006,913. , Application date: January 29, 2001, incorporated herein by reference. Figures 24A, 24B, and 24C from this application illustrate one technique. Figures 24A and 24C show the results before and after pulse stretching. May be unchanged The scope of the invention makes various modifications to the invention. Those skilled in the art will recognize many other possible variations. The system can be designed for high pulse repetition rates beyond 4,000 Hz, such as about 3,000 Hz to 600. Any repetition rate in the range of 0 Hz or 8000 Hz. Repetition rates in excess of 8000 Hz may require the use of different radon fan systems. The aforementioned laser system previously explained that the system specifically for KrF can be modified by changing the gas mixture and modifying 193 nanometers. The LNP and wavelength meter are used as lasers. The electrode spacing is preferably as small as 毛 5 hair to 13.5 ，. For example, use a motor driver to adjust the bending mechanism shown in Figure 22A. The curvature of the grating can be improved. Frequent active feedback control. Or by using a piezoelectric device to control the curvature of the grating, it can provide much faster bandwidth control. The other heat exchangers are the configurations shown at b. Obviously easy to modify. For example, all four units can be combined into a unit. Using a large heat sink in the heat exchanger to mitigate the rapid change in gas temperature caused by the burst mode laser operation constitutes its significant advantage. Readers must understand that at very high pulse rates, the pulse energy is the same as the standard paper size (please read the precautions on the back before filling this page). 541777 A7 B7 V. Description of the invention (The control of the 61 switch does not need to be fast enough to use the immediately preceding pulse to control the pulse energy of a specific pulse. For example, a control technology can be provided in which the pulse energy of a specific pulse is measured Can be used to control subsequent second or third pulses. There are many changes and modifications in the deduction rule for converting wavelength etalon and grating data into wavelength values. For example, the applicant found that the focus error of the etalon optical system caused extremely small The error result may cause the measured line width to be larger than the interline width. As the diameter of the etalon under measurement becomes larger, the error increases slightly. This can be achieved by scanning the laser and a range of wavelengths and monitoring The measured step change of the sideband as it leaves the window is corrected. The correction factor can be determined based on the position of the sidebands measured by the casement. Thus, the foregoing disclosure is not intended to be restrictive. The scope of the present invention is determined by the scope of the attached patents and their legal equivalents. (Please read the notes on the back before filling this page),?! 1 line-64 This paper size is applicable to the Chinese national standard A4 specification (210X297 mm) 541777 A7 B7 V. Description of the invention (62) Component label comparison 1. .Conductor 12A, 14A ... Shell structure 1K ... Status light 12P ... Output coupler 2-4 ... Conductor 12, 14 ... Shell structure element 2K ... Control module 13K .. .Ventilation assembly 2P ... LNP 14K ... Blower motor 3K ... Compression head module 14 P ... Wavelength meter 4 K ... Stable core group 14 ... Steering mirror 4 P ... Voltage components 14A ... Impulse 5K ... Automatic shutter module 15K ... Wire narrowing module 6K ... MFT power supply 16K ... Blower motor controller 6P ... Laser cavity Chamber 16P ... Nitrogen source 7K ... Blower motor 17K ... Blower motor controller 7 ... Plate 17P ... Nitrogen sweep module 7A ... Prism plate 18K ... Rectifier module 8K .. .Laser chamber module 18A ... cathode 8P ... high voltage cable 18P ... filter 9K ... interface module 18 '2 0 ... electrode 10K ... cooling supply module 19K ... Resonant charger module 10A ... Laser chamber 20K ... High voltage power supply 10P ... Impulse power component module 10 ... laser 20A ... anode 11K ... cooling water distribution module 20P ... distribution board 12K ... laser gas supply module 21K ... AC / DC distribution module (please Please read the notes on the back before filling this page) This paper size is applicable to Chinese National Standard (CNS) A4 specification (210X297 mm) 65 541777 A7 B7 56D ... 56V for low voltage terminal ... 56F for main HV lead ... Terminals 58, 58A ... Heat exchanger 60 .... Front dissociator tube, compression head 61 ... Sweep tube 61A ... Reflected beam passing section 62 .... Capacitor bank 62A ... Sweep stream 64 .. .Inductor, electrostatic capture unit 64A ... Sweep current 64A ... Fan leaf structure, saturable inductor 64A1 ... Housing 64A2 ... Slotted 6 4 A 3 ... Core 66 ... Blade structure 68 ... Zone 68A ... Output coupler module 69 ..... light diode, detector 70A ... retractable joint unit 72A ... window block 74A ... export 76A .... window 78A .. .Clean the outlet of the pipeline (please read the precautions on the back before filling this page) V. Description of the invention (63) 22P ... Flow monitor unit 22. DC power supply 24A ... Laser controller 33 .. Output beam 4 0 ... rectifier 42 .. capacitor bank 46 .. fan, solid-state switch 4 6 A ... fanlai structure 46B ... IGBT switch 48 .. fanjuu 'inductor' anode support flow forming structure 48A. .. 48B ... 49. Ferrite toroidal coil 50 .. Upflow forming structure 52 ... Capacitor bank, gas rotating vane 54 ... Inductor 54A1 ... Jacket 54A2 ... Cooling Pipeline 54A3 ... Groove 56 ..... Transformer 56A ... Short tube 56B ... Printed circuit board 56C ... Insulator This paper size applies to China National Standard (CNS) A4 (210X297 mm) 66 79. ·· Etalon 80 ... Top plate, PZT stacking, pressed 173-5 ..... Mirror 170 ... Partial mirror electric drive 171 .. Mirror 80A. Piezo base 81 .. Flange, diamond pad 172. ·· Energy detector 176 ... Infrared grating 82. · · Base plate, capacitor bank, 177. · · Slit stepper module 82A · · · Stepper motor mount 178 ... · Lens 17 9 ... Mirror 83 · ·· Fusion silicon oxide spacer 180… Linear light diode array 84 .. · Case, lever arm 182 ·· Mirror 85 ··· Indicator, lens holder 183 ... Lens 8 6 ··. Block 184 ... Calibrator assembly 87 ... Top edge, telescopic joint 190 ... Reference correction unit 88 ... Diffuser 197 ... Wavelength processor, interpolated 89 ... Window deduction controller 90 ... Focus Lens 200 ... Standard DC power supply 95 ...... Chamber window unit 202 ... Capacitor 102 ..... Laser controller 204 ... Control board 10 4 ... Wavelength meter 206 ... IGBT switch 106. .. LNP processor 208 ... inductor 120 ... wavelength meter unit 210 ... command voltage 130 ... spindle 212 ... voltage feedback 132 ... ball bearing 214 ... current feedback 136 ... · Sealing element 215… unconstrained diode path 140… stator 300… .three-phase power supply 541777 A7 ------ B7 V. Description of the invention (64) This paper size applies to Chinese National Standard (CNS) A4 Specifications (210X297mm) (Please read the notes on the back before filling this page)

67 541777 A7 B7 V. Description of the invention (65) 301-304 ... Metal lens 302 ... heat sink structure, source of electricity 416 ... spacer container 500 ... sealed shutter unit 303 ... heat sink 502 ... shutter 304 ... flange unit, control board 504 ... beam dumper 304A ... inner flange 5 0 6 ... power meter 304B ... telescopic joint 510 ... laser output path 304C ... outer flange 512. ..Path 305 ... current into 514 ... cathode, coil 3 0 6 ... spoiler, on-line mixing 516 ... beam stop unit, housing 518 ... flexible arm 3 0 7 ... Shizixin 520 ... permanent magnet, coil 307A-C ... coil 522 ... magnet 308 ... c-shaped seal 524 ... detector unit 400 ... double port nut, micro-processing 526 ... flexible Arm, focusing lens 530, 532 ... motor 402 ... 0 ring, programmable logic 534 ... motor controller device, collimating lens 536 ... bin motor controller 404 ... executing and data memory Body 540 ... Anode row set, etalon housing window 541 ... Cathode 406 ... RAM, RAM memory, 541-542 ... Cover diffusion element 5 4 3 ... Mandrel 408 ... .RAM memory bank, 544 ... Media spacer RAM buffer 546 ... heat sink, anode support rod 410 ... analog / digital converter, 548 ... current loop (please read the precautions on the back before filling this page) This paper size applies to China National Standard (CNS) A4 Specification (210X297 mm) 68

Claims (1)

541777 々, patent application scope 1. An extreme pulse rate gas discharge laser system capable of operating with a pulse repetition rate exceeding 3 micropulses per second, the laser body: A) —laser chamber containing a laser body gas There are two slender electrodes that define the discharge area, and the airflow path has a gradually widened cross section downstream of the electrode. It allows a large percentage of the static voltage drop that occurs in the discharge area to be recovered. B) A tangential fan is used to generate discharge gas. The sufficient gas velocity of the laser gas, when the repetition rate in the range of 4,000 pulses per second or more is "made 4, after each pulse, but before the next pulse, substantially all the discharge is cleared by the discharger. Generated ions, c) a heat exchanger system that can remove at least 16 kW of thermal energy from the laser gas, D) a pulsed power system configured to provide electrical pulses to the electrodes, the electrical pulses being sufficient to The rate of 4, 000 pulses with a precisely controlled pulse can generate laser pulses in the range of about 5 millijoules, and E)-a laser beam measurement and control system that can measure the pulse energy , Energy wavelength and energy pulse bandwidth, or substantially every pulse accompanied by feedback control pulse energy and wavelength. 2. If the laser in the patent application scope item 丨 the discharge laser system is a KrF excimer laser system; And the gas contains krypton, fluorine, and neon. 3 · As in the laser in the scope of patent application No. 2, the chamber also contains a blade structure upstream of the discharge zone, which is used to regulate the gas velocity upstream of the discharge zone. Applicable to Chinese National Standards (CNs) A4 specification (210 X 297 mm) 69 / 、 The fan contains the main shaft driven by Eryi • If the patent application scope item 1 the lightning brush DC motor. 丨, 5.If the patent is applied for Laser in scope item 4_ where the motor is water-cooled. One spoonful of laser V in the scope of patent application for item 4, which ^ each of the motor contains a certain stator and each of the motors contains a magnetic force. Open the inside of the pressure cup of the rotor and the laser gas. Ϋ 7. If the laser in item 4 of the scope of the patent application, the tangential fan includes a blade structure cut by the aluminum stock. ^ I 8 · If the scope of the patent application The thunder of item 7 The diameter of the outer side of the fan blade structure is about 5. 9. As described in the patent application scope item 4 ^ where the motor is sensorless Tanda and further includes a main motor controller that controls one more motor and controls another motor The guest motor controller. 讥 If the thunder g of the scope of the patent application, the heat exchange system with fins is water-cooled.; | Air: η · The heat exchanger system as the thunder of the scope of the patent application. Contains at least four separate water-cooled heat exchangers. 12. As described in item 10 of the scope of patent application, the heat exchanger system includes at least one heat exchanger with a tubular water flow, and at least one spoiler is provided in the heat exchanger system. path.丨. As described in item u of the patent application, the four heat exchangers each include a tubular water flow channel and a spoiler. Η.If the patent application scope is the oldest thunder ~ among which the pulse work package M1777 The scope of patent application for water-cooled electrical components. 15 = Shenca_14㈣㈣Cold-type component At least-the component is a component with a high voltage higher than 12 release *. 16. If the 15th item of the scope of the application for patent is ^ Gan's field body ^, where the two voltages are electrically isolated with ground potential, and the cooling water flows through the inductor. 7 ·: Lei in the scope of patent application No. 1 in which the pulsed power system includes a resonance charging system, which charges the charger to an accurate control voltage. Bu: 丨 (Please read the precautions on the back before filling this page) 18 · If the thunder of item 17 of the patent application scope, the resonant charging system has a De-Qing circuit. : ≫ 19. The resonant charging system according to item 17 of the scope of patent application contains a drain circuit. ί '? ::; 2〇 · If the thunder ^ in the scope of the patent application, the resonance charging system includes Dijing circuit and discharge circuit. 21. The lightning power of item i in the scope of patent application, wherein the pulse power includes a charge system consisting of at least three power supply units arranged in parallel. The beam measurement and old system includes an etalon unit, a photodiode array, a programmable logic device, and an optical device for focusing the light emitted from the etalon unit to the photodiode array, wherein the programmable logic The device analyzes the data from the photodiode array by programming sea and determines the position of the etalon in the photodiode array. Package package package package package package Order — 541777 A8 B8 C8 D8 patent application scope 23. If the patent application scope of the thunderbolt is 22, the measurement and control system also includes a microprocessor, which is programmed by the program to locate the sideband by the programmable logic device Data calculated wave width. 24. For example, in the scope of patent application No. 22, the programmable logic device is planned by a deductive rule program to calculate the wavelength and bandwidth based on the measurement of the sidebands. L ^ 2 5. As described in item 24 of the scope of patent application _ where the programmable logic device is configured to calculate the wavelength and bandwidth at a speed comparable to 1 / 4,00j > flavor per second. _ 、 26. If the scope of the patent application is # 22 的 雷 #, wherein the etalon unit contains a -diffractive diffusing element. "Age 27. For example, the laser of the scope of the patent application; ^ contention, ^ steps include a line narrowing unit, which contains at least part of the fine-tuning mirror driven by the PZT driver. 28. Such as the scope of the patent application of the laser 27 The fine-tuning mirror part is driven by a stepper motor. 29 · If the thunder of item 27 of the scope of patent application ^ ,: further includes a front trimming device. 1 :. 30 · If the thunder of scope 27 of patent application_ It further includes an active fine-tuning device which includes a learning deduction rule 7: 31. If the step of Lei in the 27th patent application scope includes an adaptive feedforward deduction rule. 32. In the 27th of the patent application scope, Lei #, where The line narrowing unit includes a grating to define a grating surface and a device for forcibly sweeping the gas adjacent to the surface of the grating. This paper size applies the Chinese national standard (_) M specification (21〇297297 mm) 72 (please first Read the precautions on the back and fill in this page). You can apply for patent scope 33. If you apply for patent gas. 34. If you apply for thunder silt in the special scope of item 32: the sweep gas is nitrogen. Shoot Y, where the sweeping gas is helium 35. If the patent application scope item! Thunder <, further contains _ species atmosphere removal system which contains a nitrogen sound nitrogen cleaning system ?, • two of the oldest patent application scope Lei, further includes a nitrogen sweep display unit including a flow monitor. The mine cup instrument contains a sweep vent for the ventilation of the sweep gas by the laser. 37. If the scope of the patent application is No.!, The step contains a shutter Shan Fan, which includes an electrically operated shutter power meter, the shutter and power meter can be positioned on the output beam path with a command signal. 38. If the thunder step of the 27th patent application scope includes a beam sealing system, provide A first beam seal interposed between the first window of the chamber and the line narrowing unit, and a second beam seal interposed between the second window of the chamber and the output facet unit, each of the beam seals includes a metal stretch fiber. 39. If the thunder of the scope of the patent application is 38, the first and second beam seals are each configured to be easily replaced by the laser cavity. If the thunder of the scope of the patent application is the middle scope, Elastomer-free Provide vibration isolation from the chamber, provide isolation of the beam train from the atmospheric gas, and allow the laser chamber to be replaced or replaced by the laser unit without replacement of the laser chamber Λ: Withdrawal-41. Such as applying for a patent The% of thunder in the range i. Among them, the measurement and control system is used to split one of the output pulses of the small director from the laser. Zhongqing Patent Fanyuan beam splitter, which is used to indicate a small proportion to the pulse energy. The first beam splitter of the measuring device and an isolation device are used to isolate the volume of the human beam splitter, the secondary beam splitter, and the volume of the pulse energy detection transom and the rest of the measurement and control system.俾 Boundary of an isolation zone 〇42. For example, the mine hall in item 41 of the patent scope includes a step of hip-sweeping competition for sweeping the isolation zone with a sweeping gas. 43. If the laser body No. 42 in the scope of patent application, the laser further includes an output coupler unit and an output window unit, the cleaning device is configured to ventilate from the isolation area and also to clean the output_ 马 合 器Unit and window unit. f Λ The scope of application for patent No. # 之 之 雷 # Wherein the chamber also contains a motor circuit, the circuit has a ribbed profile: rectangular cross section, and its longitudinal direction is dependent on the laser gas flow direction. 45. As in the application of the scope of patent application item i, the chamber contains an anode and an electrical spacer located on the anode_fresh top, which is improved in the laser current between the two electrodes. !! 1 package 46. If you apply for the range of w thunder | Where the pulse power system. The corona-containing plate is electrically connected to the tip of the capacitor bank by a metal rod. The length of the metal rod can be selected to provide a higher sense.