Field of the Invention

[001] The present invention relates generally to laser systems, More

specifically, d e invention relates to performing injects of reactive gas into the chambers of a gas discharge laser, such as a two chamber Master Oscillator- Power Amplifier exeimer laser.

Background of the invention

[002] One type. of gas discharge laser used in photolithography is known as an exeimer laser. An exeimer laser typically uses a combination of a noble gas, such as argon, krypton, or xenon, and a reactive halogen gas such as fluorine or chlorine. The exeimer laser derives, its name from the fact that under the.

appropriate conditions of electrical stimulation and high pressure, a pseudo- molecule called an exeimer (or in the case of noble gas halides, an exdplex) is created, which can only exist, in an energized state and can give rise to laser light in the ultraviolet range.

{003} Exeimer lasers are widely used in high-resolution photolithography machines, and are thus one of the critical technologies required for

microelectronic chip manufacturing. Current state-of-the-art lithography tools use deep ultraviolet i ^' DUV) light from the. KrF and ArF exeimer lasers with nominal wavelengths of 24.8 and ^' .193 nanometers respectively.

[004] While exeimer lasers may be built with a single chamber light source., the conflicting design demands for more power and reduced spectral bandwidth have meant a compromise in performance in such single chamber designs. One way of avoiding this design compromise and improving performance is by utilizing two chambers. This allows for separation of the functions of spectral bandwidth and pulse energy generation; each chamber is optimized for one of the two performance parameters.

[005] Such dual-gas-dis iarge-chamber excimer lasers are often called Master Oscillator-Power Amplifier, or "ΜΟΡΑ ' lasers, In addition to improving the spectral bandwidth and pulse energy, the efficiency of the dual chamber architecture can enable the consumable modules in MORA lasers to reach longer operational lifetimes than their counterpart modules in single chamber light sources,

[006] In each chamber, as the light source discharges energy across its electrodes to produce light, the halogen gas, fluorine in the case of ArF or rF lasers, is depleted. This causes a decrease i the laser efficiency which is seen,, for example, as an increase in discharge voltage required to create a given desired pulse energy. Since the discharge voltage has an upper limit determined by physical constraints of the hardware, steps must be. taken to: replenish the lost ^' fluorine so that the voltage remains below this, limit and the laser continues to f unefi on proper I y .

{007] One way to do this is with a full replenishment of the gas in the chambers, called a. refill, where all of the .gas is replaced while the laser is not firing to return the gas content in the chamber to the desired mix, concentration and pressure. However, refills are extremely disruptive as the laser must be shut off during the refill process, and thus the lithographic exposure o£ semiconductor wafers must also be paused in a controlled manner at the same time and then restarted when the laser is again operational to avoid improper processing of the waters. For this reason., t is typical to refill both chambers at once to save time., although this is not necessary .

[008] The need for a refill can depend on several, complex and often

unpredictable variables, including the light source firing pattern and energy, the age of the light source modules,, and others that will be familiar to those of skill in the art. For this reason, refills are typically done on a regular schedule,- which ensures that the light source operation will never suffer unanticipated

interruptioti due to the light source reaching its operaiional limit. Such a regular schedule generally yields very conservative upper limits on the time between refills, such that some users of the light source operatin at low pulse usages might be able to wait for a much longer period of time between refills than is provided b the simple schedule.

[009] Given the demands of increased throughput and light source availabilit ,- efforts have been made to minimize light source stoppage for refills. One way of doing this is by performing a partial replenishmen of the gas in the chambers, known as an. inject rather than a full refill, As long as the laser is able to continue to operate within certain parameters, it is not necessary to shut the laser down for the inject, and thus processing may continue during the inject process.

[0010] Another factor that decreases efficiency is the ageing of the laser; older exeimer lasers are in general less efficient than newer ones. However, it can be difficult to separate the effects of ageing from the effects of fluorine concentration that is not optimal, particularly when the control, of the fluorine is not precise.

[0011] A number of prior methods and systems have been described, for managing injects, including, for example, how to determine when an inject should occur and the amount of halogen gas to be provided by the inject, See, for example, U.S. Patents. Nos. 7,741,639 and 7,835,414, owned by the assignee of the present application. However, such prior art still requires that a refill be. done at some point to keep the laser within opera ting parameters.

[0012] Further, many of these prior art techniques do not have adequate control .of the fluorine concentration to be able to tell, whether a .given loss of efficiency is due to the fluorine level or to the ageing of the laser. As. a result, those using such prior art techniques generally assume that any loss of efficiency is due to a fluorine problem and attempt to adjust the fluorine to compensate, and thus do not compensate or the ageing of the laser in any way.

[0013] It would be desirable to have a method that allows a laser to operate within acceptable parameters for a longer period of time using only injects rather than refills to replenish the gas in the chambers, so that the laser need not be shut off during the refill procedures. It would also be desirable to be able to compensate for: die agein of the laser as part of such a method using only injects rather than refills.

Summary of the Invention

[0014] Systems and methods for operating a dual chamber gas discliarge laser, such as a MOPA excimer laser,, using only injects rather than refills while compensating for the ageing of the lase by increasing the pressure in one or both of the chambers are disclosed,

[0015] In one embodiment, a dual chamber gas discharge laser light source is described, comprising a master oscillator and an amplifier, each of the master oscillator and amplifier having a laser chamber containing a lasing medium gas comprising a halogen, and a gas replenishment system including a controller executing a replenishment scheme at regular intervals, the replenishment scheme comprising injecting into the selected laser chamber at each inject opportunity a quantity of a non-halogen containing gas and a quantity of the halogen containing gas estimated to result in a desired amount of halogen gas in the chamber after the inject opportunity; and after each N injects, where N is a predetermined number, adjusting the pressure in the chamber to compensate for any change in the efficiency of the lase -due to the length of operation of the laser, 0016] In another embodiment, a method, of replenishing gas in a dual chamber gas discharge laser light source having a master oscillator and an amplifier is described, each of the master oscillator and amplifier having a laser chamber containing a lasing medium gas com rising, a. halogen, the method comprising the steps of: selecting a plurality of inject opportunities occurring at regular intervals; injecting info the selected l ser chamber at each inject opportunity a quanti ty of a .non-halogen containing gas and a quantity of the halogen containing gas estimated to result in a desired amount of halogen gas in the chamber after the inject opportunity; and after each injects, where N is a predetermined number, adjusting the pressure in the chamber to compensate for any change in the efficiency of the laser due to the lengt of operation of the laser.

[0017] Still another embodiment discloses a non- transitory computer-readable medium having embodied thereon a program, the program being executable by a processor to perform a method of automatically replenishing the gas in a laser chamber of a dual chamber gas discharge laser light source having a master oscillator and an amplifier, each of the master oscillator and amplifier having a. laser chamber containing a lasing medium gas comprising a halogen, the method comprising the steps of: selecting a plurality of inject opportunities occurring at regular intervals; injecting into the selected lase chamber at each inject opportunity a quantity of a non-halogen containing gas and a quantity of the halogen containing gas estimated to result in a desired amount of halogen gas in the chamber after the inject opportunity; measuring an operating parameter of the selected laser chamber that is indicative of the efficiency of the laser during each injection of gas into the chamber; estimating the ratio of the change of the measured operating parameter to the number of shots fired b the laser;

measuring the change in pressure during each inject opportunity; determining the ratio of the change of the measured operating parameter to the change in pressure in the chamber; and after each N injects: estimating the change in the operating parameter over the N Injects; estimating the pressure in the chamber that is appropriate to reverse the change in the operating parameter over the injects; and adjusting the pressure in the chamber to the estimated pressure. Brief Description of. the Drawings

[0018] Figure 1 shows a simplified block diagram of a. gas replenishment system 100 for a dual chamber gas laser, such as a MOPA exdmer laser, according to one embodiment.

[0019] Figure 2 shows the relationship between the pressure in a laser chamber -of a dual chamber gas laser, each as a MOPA exdmer laser, and the reciprocal of efficiency at different points in the ife of the laser.

[0020] Figure 3A illustrates the typical effect of -continued shots and refills on the efficiency of a dual chamber gas laser, such as a MOPA exdmer laser.

[0021] Figure 3B illustrates a desired effect of maintaining the efficiency of a dual chamber gas laser, such as a MOPA exdmer laser, as shown in Figufe 2A but without the use of refills of the gas in the laser chambers.

[0022] Figure shows ^' the relationship between a series of injects and the resulting chamber pressure in a laser -chamber of a dual chamber gas laser,, such as a. MOPA exdmer laser, according to one embodiment.

[0023] Figure 5 is a graph of a laser parameter having a linear relationship to the length of operation of the laser.

[0024 Figure 6 shows a single inject in a in a laser chamber of a dual chamber gas laser, such as a MOPA exdmer laser, and the .measurement of the sensitivity of -a parameter of laser operatio to a change in pressure., according to one embodiment,

[0025] Figure 7 Is a simplified flowchart showing one embodiment of the gas- replenishment method described herein. Detailed Description of the Invention

[0026] The present application describes a method and system for performing injects of halogen gas into the chambers of a two chamber gas discharge laser such as a MOP A extimer laser for the purpose of allowing operation of the laser within acceptable parameters without the necessity of performing refills. The method also allows for automatic compensation for the ageing of the laser by increasing the pressure in the chambers over time. While manual adjustments to pressure have been made previously, it is believed that such automatic and substantially continuous gas optimization without the need of refills has not previously been suggested or attempted in. the prior art.

[0027] It is expected that an inject process as described herein wilt result in an increased period of operation of the laser without needing to suspend operation during refills, as well as greater efficiency due to the compensation for ageing of the laser, A parameter indicative of the performance of the laser is measured, and the change in the parameter with respect to the length of operation of the laser is estimated to determine the ageing rate of the. laser. The change in the parameter with respect to the pressure in one of the chambers: is also measured to determine the change in efficiency due to pressure. At a selected time, the total change in the value of the parameter is estimated,, indicting the decrease in performance of the laser, and from this the amount of change in pressure that is needed to reverse the change in the value of the parameter is calculated. The pressure in the chamber is then changed, thus improvin the performance of the laser to correct for the amount of time that the laser has been in operation.

[0028] There: are a number of parameters indicative of the laser performance that may be selected, in many cases, laser efficiency may foe the easiest parameter to use. However,, other parameters may be selected if desired; for example, when considering injects to the power amplifier chamber, V may be the discharge voltage., while in the case of the master oscillator chamber V may be the delay lime ("dtMOPA") between an electrical discharge in the master oscillator chamber thai creates a laser sho and the subsequent electrical discharge in the power amplifier chamber that amplifies the shot

[0029] Still other parameters may be used,, such as bandwidth, which may be measured by the integral of a certain, percentage of the energy contained on either side of a center wavelength of a spectrum centered on the center wavelength. One bandwid th measure used in other contexts is the integral of 95% of the energy is common and is known as E95% or simply 195. Another parameter that may be used is the common voltage applied to both chambers, for example, at the peaking capacitor of a compression head of a solid state pulsed power system (SSPPM) for each chamber, and the energy output of one of the chambers. For the master oscillator chamber this is designated as E O. Other suitable parameters and /or combinations of parameters for estimating the amount of, or rate of consumption of, fluorine will be apparent to those of skill in the art,

[0030] As stated previously, in order to obtain the. full benefit of the method described ^' herein, it is desirable to have a method of controlling the fluorine level more accurately than has been done previously. Many prior art methods of fluorine control are actually biased by the long term, ageing effects, i.e., they confuse normal fluorine depletion with ageing, and thus cannot, decouple accurate fluorine control from the ageing effect. One such method of decoupling, fluorine control from ageing effects is described in U.S. patent application serial no. 11/094,313, commonly owned by the assignee, of -the present application. [0031] A simplified block diagram of a gas replenishment system 100 for a dual chamber gas laser., such as a MOPA exeimer laser, is shown in Figure 1. The MOPA exeimer laser has a master oscillator 102 containing a: laser chamber, ami a power amplifier 104 also containing a laser chamber, in operation, the master oscillator 102 produces a first laser beam 106 which is passed to the power amplifier 104. where it is amplified, to produce an amplified laser beam 10:8 which is output t a scanner machine (not -shown) for use in lithography,

{0032] Each laser chamber contains a mixture of gases; for example, in a given excimer laser each laser chamber might contain a halogen, e.g., fluorine, along with other gases such argon, neon, (commonly known as rare gases) and.

possibly others in different partial pressures that add up to a total pressure P. For simplicity, the halogen gas is hereafter described as fluorine, although the principles described herei may be applied to other halogen gases as well.

[0033] Gas bottles 110 and 112 are connected to the master oscillator 102 and power amplifier 104 through valves 114 to allow for replenishment of the gas in. the laser chambers when desired. Gas bottle 110 typically might contain a mixture of gases includmg fluorine, argon, and neon, known as an "Ml mix," "tri-mix/' or often simply '" ^' fluorine/' while gas bottle 112 might contain a mixture of argon, neon and/or other gases, but no fluorine, known as an "M2 mix," "bi-mix," or "rare gas.'" ^' A controller 116, such as a processoror logic circuit, operates the valves 1.14 to transfer gases from bottles- 110 and 112 into the laser chambers of the master- oscillator 102 and power amplifier 104 based upon certain data as described furthe herein.

[0034] As is known in the art, two bottles of gas are needed, since the fluorine in gas bottle 110 is at a particular concentration that is typically higher than that desired for laser operation. In order to add the fluorine to the laser chamber of the master oscillator 102 or power amplifier .104 at a desired lower concentration,, the gas in bottle 110 must be diluted, and the non-halogen containing gas in bottle. 112 is used for this purpose.

[0035] Although not shown,, valves 114 typically include two valves for each laser chamber, an "injection" valve that allows gas to pass into and out of each chamber a a first rate, and a "chamber fill" valve that allows gas to pass into and out of each chamber at a. second, nd faster, rate. In addition, the laser chambers in the master oscillator 102 and power amplifier 104 contain blowers for mixing the gases that are in the chambers so that a homogenous mixture is maintained during operation. The blowers commonly also add heat to the gas.

[0036] As mentioned above, fluorine is consumed during operatio of the laser. The resulting decrease in. fluorine concentration typically causes a rise in the discharge voltage required to produce a laser pulse. In addition,, changes in. fluorine concentration also affect the delay time betwee the electrical discharges that cause production of the first laser beam 106 and the amplified laser beam 108; this delay time is often referred to as dtMOPA."

[0037] Thus, the fluorine concentration must be replenished to keep the laser operating within desired parameters _. ; Further, a satisfactory concentration of fluorine must be maintained while keeping the gas content in each laser chamber at a fixed pressure. Again, this is sometimes done by injects, i.e., partial replenishment of the gas i the chamber, .ratiher than, a full refill in which the chamber is purged and the gas completely replaced.

[0038] As with full refills, injects are typically done at fixed intervals, determined either by elapsed time between injects, or by the number of "shots," i.e., pulses of the laser, thai have been generated. In some prior art systems, injects are done in- each chamber after approximately every Ί million pulses by that chamber. For easier operation, the injects to the laser chambers are staggered, so that while each chamber receives an inject after about each million pulses,, the power amplifier 104 receives an Inject approximately 500,000 pulses after the master oscillator 102 receives an inject, and vice versa. Such timing of injects is described, for example, in U.S. Patent No. 7,835,414, owned by the assignee of the present application. f 0039] While a full refill simply replaces all of the gas in the laser chamber, an inject is intended mainly to replace the amount of fluorine that has been consumed since the last refill or inject. Since it is mostly the fluorine that is consumed during operation, it is known in the prior art thai infects to the laser chambers in both the master oscillator and power amplifier will include a fixed amount of the M2 mix, which contains no fluorine, and an amount of l mix containing enough fluorine to raise the concentration of fluorine in the chamber back to a desired level, thus replacing the fluorine which has been consumed.

[0040] However, as previously performed in the prior art, injects do not completely restore the operating condition of the laser to its- peak or desired efficiency. Rather, the efficiency of the laser decreases over time, even with injects; this is believed to be primarily due to the degradation of optical elements from exposure to the laser pulses, but may -also. ^' be due to other physics of laser operation.

(0041] Figure 2 shows an example of the relationshi between the pressure in a laser chamber and the reciprocal of efficiency at different points in the life of the laser. [0042] Because of the reduction of laser efficiency over time, a curve showing that relationship between efficiency and pressure,, such as those shown in Figure 2, typically moves up over time. Thus, a new laser might operate along the curve labeled 201, After some number of shots,- the laser might operate with less efficiency along the curve labeled 202, and after still more shots the laser might operate along the curve labeled 203,

[0043] As shown in Figure 2, in general the efficiency of the laser is higher (and thus the reciprocal of the efficiency is lower) at greater gas pressures, although this is only true up to the maximum chamber pressure at which the laser can operate, (There is also a minimum required pressure for the laser. Different lasers will have different maximum and minimum pressures.)

[0044j If it is desired to operate the laser at a particular target efficiency Ετ, while a new laser operating along curve 201 can achieve this at a pressure Pi, an aged laser that operates along curve 202 will require a higher pressure Pa to achieve the same efficiency. Similarly, an even older laser operating along curve 203 will require a still higher pressure Pa to achieve the target efficiency, again -assuming that Pa is within, the maximum operating pressure of the laser.

[0045] Thus, as shown in the example of Figure 2, it can also be seen that the decrease in efficiency can be compens ted for by increasing the pressure in the chamber, as long as the pressure remains less than the maximum chamber pressure. Thus, in one embodiment, when the efficiency of the laser drops below some minimum desired ef ficiency, the pressure ixi the chamber is increased until the efficiency rises to an acceptable level.

[0046] It is because of the maximum (and minimum) chamber pressure that accurate fluorine control is desirable for use with the method described herein. In the absence of good control of the fluorine, increasing the ressure in the chamber may result in the maxim m (or minimum) pressure being reached more quickly, thus reducing the time over which the lase can operate with only injects.

[0047] Figure 3A illustrates the typical effect of continued shots and refills on the efficiency of a laser. The horizontal axis represents the shots fired by die laser, and the vertical axis the reciprocal of the efficiency of the laser.

[0048] It is assumed here that upon startup the laser is optimized to operate at some, target efficiency Ετ. The efficiency of the laser typically drops over time as shots are fired, and thus the- reciprocal of efficiency increases as shown.

[0049] After some number N of injects, even with the replenishment of fluorine from the injects the efficiency of the laser has dropped to some minimum desired efficiency EM. .In the absence of some change from the known methods of performing injects, the efficiency of the laser will continue to decrease.

[0050] In the prior art, this problem is avoided by replacing all of the gas in the chamber by performing a refill in order to keep the laser operating within the desired range of efficiency. Thus, in Figure 2 A, after each injects a refill is performed, as shown by points 1, 2 3, , and 5 on the horizontal axis. At the conclusion of each refill, the gas may be optimized, including adjustment of the pressure, until the efficiency of the laser is restored to a desired efficiency, here shown as the targe efficiency Ετ, and the efficiency then again decreases as the laser continues to fire shots. An automated process of gas optimization after a refill is shown in U.S. Patent Application serial no. 13/174,640, filed June 30, 2011, and assigned to the assignee of the present application. [0051] As explained previously, operation of the laser. must stop during the refill process. Thus, it would be desirable to b able to restore the efficiency of the laser to its target efficiency Ετ (or some efficiency close to that) using only an inject of some type rather than a refill Figure 3 illustrates what a plot of the reciprocal of efficiency versus the number of shots fired might look. like in such a case. Now there are no refills,, but some othe process that occurs periodically, for example after each injects., and which would be able to at least

substantially restore the efficiency of the laser without the need for a refill (Note that M will typically be a different number than N as explained below.)

[0052] The increase in pressure in the chamber is obtained b reducin the size of the bleed at the end of an inject so that the pressure in the chamber after the inject is completed is at the higher level necessary to increase the efficiency of the laser as desired. After such an inject with a smaller bleed, injects are resumed in the normal fashion,, with normal bleeds that reduce the pressure in the chamber to the new higher level at the end of each, inject, until the efficiency of the laser has again dropped below the minimum desired efficiency.

[0053] This process may be seen in Figure 4, which shows a series of injects, and the resulting chamber pressure according to one embodiment, it is assumed that the injects begin after a refill of the lase chambers and optimization of the gas as described above. One of skill in the art will recognize that a typical inject actually includes three steps, as seen in more detail in Figure 6: first Ml mix is added to the chamber, followed by the. addition of M2 mix to ^' the chamber, both of which increase the pressure in the chamber, and finally- the mixed gas is bled from the chamber to reduce the increased pressure back down to the. desired pressure. [0054] As seen in Figure 4, after a refill is completed at point 1, the laser begins operating with a chamber pressure of Pi, and injects are performed at regular intervals as is known in the prior art. After some period of M injects, at point 2 the efficiency of the laser will fall to some minimum desired efficiency, as seen in. Figure 3B. The typical response to this in the prior art would be to purge the chamber of gas, refill it, and then optimize the gas as described above,

[0055] However, as stated above, in one embodiment of the present method, instead of a refill, when gas has been added to the chamber in the next inject M+l, the amount of gas hied from the chamber is reduced so that the pressure in the chamber does not drop back to Pi as in the prior injects. As a result, at the end of this inject, the pressure in the lase chamber is now at a pressure P¼ which is higher than the previous pressure Pi. Injects are then performed .normally to replenish the fluorine in the chamber and keep the chamber pressure at Hie increased pressure p2, rather than the prior pressure Pi, as the laser continues to operate.

[0056] After another M injects, the efficiency of the laser will have fallen again. As was done after the first M injects, at the end of the next inject. Inject 2M+1, the ■amount of gas bled from the chamber will again be reduced from the prior injects so th t the pressure does not drop back to P¾, again resulting in an increase of pressure in the chamber to a new pressure Ps which is higher than I\ and thus again leading t an increase in the efficiency of the laser.

[0057] Thus, over the first M injects the pressure in the chambers will be Pi as shown in Figure 2, and over this period the efficiency of the laser will drop from i s starting position of the target efficiency Έτ toward a minimum efficiency EM as shown in Figure 3B, At the point labeled 2 on Figure 3B, inject M+l raises the pressure in the chambers from Pi to Pi as shown 011 Figure 2, and restores the efficiency to Ετ as shown at point 1 in Figure 3B, Similarly, over the next M injects the pressure in the chambers will be Pi (Figure 2), and again the efficiency will dro -as shots are fired (Figure 3B). At point 3. on Figure 3B, inject 2M+1 will again raise the pressure in the chambers, now from P2 to Pa, and again retur the efficiency to the target efficiency Ετ.

[0058] The values of and are somewhat arbitrary and will depend on the particular laser and the choices made by a laser operator or engineer. In some lasers, refills may ^' be done approximately every 2 billion pulses, and injects approximately every million pulses, so th t there may foe about 2,000 injects in between refills (i.e., N = 2000), In some embodiments of the- resent method, it is believed to be preferable to adjust the pressure as described more often than every 2,000 injects, i order to keep the efficiency of the laser closer to its optima! value than typically occurs between refills:, It is believed, that adjusting the pressure in the range of every 50 to 200 injects (i.e., 50 < M < 200.) may result in smaller variations in the efficiency of the laser.

[0059] It may thus be seen that in principle it is possible to compensate for the aging of the laser by increasing the pressure in the chambers in this- fashion. The question is thus how to determine how much to increase the pressure in order to obtain the proper compensation for the age of the laser. 0060] In one embodiment, it is first determined how fast the efficiency of the laser deteriorates, and thus the relationship between the age of the laser and the change in the efficiency curves shown on Figure 3, i.e., how many shots it takes for the laser's efficiency curve to move:, for example, from curve 201 to curve 202, and from curve 202 to curve 203. This is considered the ^' ''aging rate" of the laser. [0061] Next, to determine the ageing rate, first a parameter V is selected and measured. As a ove; while efficiency may be the easiest parameter to consider, other parameters may be used as well. Again, in one embodiment when considering injects to the power amplifier chamber, V may be the discharge, voltage, while in the case of the master oseillator chamber V it may be the delay time dtMOPA. In other embodiments, V mav be some other measurement such as bandwidth, for example E95, the common voltage applied to both chambers, for example^ at the peaking capacitor of a compression head of a solid state pulsed power system (SSPPM) for each chamber, or the energy output of one of the chambers such as E O for the master oscillator chamber. Other suitable parameters and/or combinations of parameter for estimating the amount of, or rate of consumption of, fluorine will be apparent to those of skill in the art,

[0062] The value of V in general will be related to the amount of fluorine as well as the age of the laser. However, if the amount of fluorine can be accurately controlled and remains constant, then any change in the value of V (ΔΥ or dV) will only be due to the effect of ageing., and thus good fluorine control is desirable to obtain the: maximum benefit of the described method. It is assumed hereafter that such control is present so that the amount of fluorine does remain approximately constant, so that the change, of V is solely due to the ageing, i.e„ the firing of shots, of the laser,

[0.063] ^' With approximately constant fluorine, measuring the change in V o ver a number of shots of the laser (At or dt) results in a value dV/dt, the estimation of which will be apparent to one of skill in the art. One example of accurate fluorine control and the estimation of dV/dt is shown in U.S. patent application ser, no. 13/25.1,181, filed on September 30, 2011. and assigned to the assignee of the present application, which is incorporated herein by reference. Other methods of fluorine control and the estimation of dV/dt are present in the prior art.

[0064] In the simplest example, the value of dV/d twill he a constant, so that the value of V over time is simply a straight line having some slope, as shown in Figure 5. As shown here, after injects, the value of V changes from Vi at time TJ to Vi at time T2. (injects typically occur at fixed intervals based on the number of shots of tlie laser.) To compensate for this change in V due to the agein of the laser, it is desired to return the value of V to Vi.

|0065] Ne the change in pressure required to return the value of V to Vi is determined. The sensitivity of the parameter V to pressure, i.e., dV/dP, is first determined. In one embodiment, this is done by comparing: the value of the signal V to the pressure in the chamber at certain points during the inject process as shown in Figure 6. Figure 6 shows a single inject, which includes the injection of Ml mix and the. subsequent injection of M2 mix, which raises the pressure in the chamber to Px, and then a bleed of the chambe to reduce t e pressure to the desired pressure, here designated PY,

{0066] A value of V designated Vx is measured a the time when the pressure is Fx, i.e.. after the Ml and M2 mix. have been injected, and again when the pressure is PY (designated VY), i.e., after gas has been bled from the chamber to reduce the pressure to the desired pressure. The difference in the signal V is taken by- calculating Vx - y, as is the difference in -pressure. Px - P*. The value of d V dP is then given by; dV V _x - V _Y

dP ^ p ZTpZ [0067] Note that while Figure 6 shows the value of ^' V decreasing with increasing pressure, the opposite may be true, depending upon which particular parameter is selected as V.

[0068] As above, after M injects, the change in V is known or estimated (i.e., dV/'dt times t, where t is again the number of shots fired), as shown in Figure 5. If dV/dP is approximately constant during the period of the M injects, in the simplest approximation, the appropriate change in pressure that will compensate for tile change in V is easily determined by; dV

V ₂ — y _t =

dP where Vi and V¾ are the values of V as measured, at time Ti and T2 as above, and Pi is the pressure at time Ti, and P2 is the desired pressure at time ¾.

[0069] However, it is likely that dV/dP is not constant during the period, and thus in some embodiments, a more detailed determination of the value of dV/dP after injects is desirable, in one embodiment this may be accomplished by calculating dV/dP after each inject as shown in Figure 6, and. filtering the resulting values of dV/dP through a filter, for example, a low pass filter.

[0070] Thus, every M injects, another estimate of the current value of d.V/dP is obtained from the values of dV/dP over the most recent M injects to get the new desired pressure, in general, as seen in Figure 3, the described method will generally result in increasing the pressure in the chamber to compensate for the ageing of the. laser. However, in some rare circumstances, it may be appropriate to reduce the pressure i the chamber. In such a case, the bleed of the next inject is simply extended until the pressure drops, to the desired level. [0071] In some cases, the value of dV/dP may- vary very quickly, and the use of a low pass filter will tend to smooth out the variations and eliminate noise. It will be. appreciated that such a low pass filter may also result in an. estimate of dV/dP at a given moment that differs from the value of dV/dP calculated tor the last inject. If the bandwidth of the filter is very small, most of the noise in the value of dV/dP will he eliminated,, but if there is a large change in the value of dV/dP, the lag time for the output of the filter to catch up to the actual value of dV/dP is increased. One of skill in the art will appreciate how to select a filter to obtain a desired balance of noise reduction and convergence of the output value.

[0072] In some embodiments, it may be desirable to change the pressure by slightly less than what is calculated,, i.e., to remove slightly less than the entire change in V, in order to allow for noise in the -measurements.

[0073] The method described herein can be used with both chambers of a dual chamber laser. However, it will typically be desirable to use a different parameter V for each chamber, and the resulting pressures for efficient operation may be different for each chamber. For example, for the power amplifier laser chamber it may be appropriate to use the discharge voltage as the parameter V, while for the master oscillator chamber it may be preferable to use dtMOPA, the laser bandwidth, or E95. Alternatively, the injects to the PA chamber may remain fixed as described in co-pending application Serial No. 13/098,259. owned by the assignee of the current application, and the method described herein used only for the MO chamber.

[0074] Where a different parameter V is used for each chamber, the described calculations become somewhat more complicated due to the coupling effects between the chambers. While these effects are typically small, they are easil taken into account by the use of matrix inversion techniques that will be familiar to those of skill in the art

[0075] Figure ? is a simplified flowchart showing one embodiment of the gas replenishment method described herein, As above, it is assumed that the laser chambers have bee filled, and the gas optimized. At step 701, it. is. determined what inject intervals have been selected. As above, these may have, been selected either b elapsed time o number of shots fired.

[0076] At step 702, the initial pressure PI is determined, and dV/dt is estimated. Alternatively, dV/dt may be determined by measurement as explained herein as the injects occur.

[0077]; For each of the first M-l injects, gas is injected into the chamber, and dV/dP for the inject is calculated at step 703,

[0078] At step 704, for the Mth inject, gas is injected, dV/dP for the inject is calculated, and the change in pressure ΔΡ appropriate to compensate for the change in. V over the N injects is calculated,

|0079] At step 705, during inject M+l, after the gases have been injected, the chamber is bled until the new desired pressure F2, i.e.. Pi + ΔΡ, is reached, Inject M+l is. also treated as the first inject for the next period of M injects, and the process repeats as long as the laser is operated, and the pressure stays within the minimum and maximum pressure at which the laser will operate.

[0080] in some embodiments, upon, the commencement of laser operation after a refill or a significant pause in operation, for example over approximately an. hour, it may be advantageous to wait for small number of injects before performing the method of Figure 7, in order to let any thermal ^' transients that ma arise die out and arrive at a steady state of operation. Thus, for example, it may be determined, that for the first few injects, o the order of 5 to 10, dV/dt is not calculated, and that for a few more injects, on the order of 10 to 20, dV/dt is calculated but will not be used if the value appears to vary excessively. This may allow for a more accurate calculation of dV/ t, and thus for more accurate control of the laser efficiency with, the described method.

[OOSi] The disclosed system and method has been explained above with, reference to several embodiments. Other embodiments will be apparent to those skilled in the art in light of this disclosure. Certain aspects of the described method and apparatus may readily be Implemented using configurations or steps other than those described in the embodiments above, or in conjunction with elements other than or in addition to those described above.

[0082] For example, It will be understood by those skilled in the: art that while the preferred embodiment is a master oscillator-power amplifier multi-chambered excimer or molecular fluorine gas discharge laser system ("MOPA") the system may also be configured to have other oscillator/amplifier conilgiu'af ons, such as a master oscillator-power oscillator (" ΌΡΟ"), a power oscillator-power amplifier (TOP A") or a power oscillator-power oscillator ("POPO")

configuration, or the like. It will also be understood that in each of such configurations the output of the first oscillator stage is amplified in some fashion in the second stage, regardless of whether the second stage is a power amplifier or a power oscilla or. 0083} Similarly, unless otherwise indicated specifically to the contrary reference to a master oscillator stage or chamber ("MO") in the Specification or the appended claims, and/or a. power amplifier stage. or chamber ("PA") in the Specification or appended claims, shall be considered to be broad enough to cover any oscillator first stage or chamber feedin an output into any amplifier second stage or chamber for amplification, and the term oscillator chamber or oscillator stage is broad enough to cover any such, oscillator stage and the term amplifier chamber or stage is broad enough to cover any such amplifier stage. Further, while the above description uses a two stage or chamber laser as an example, the system and method disclosed might also be applied to a single chamber laser or any multi-chambe laser.

[0084] It should also be appreciated that the described method and apparatus can be implemented in numerous ways, including as a process, an apparatus, or a system. The methods described herein may be implemented by program instructions for instructing a processor to perform such methods, and such instructions recorded on a computer readable storage medium such as a hard disk drive, flopp disk, optical disc such as a compact disc (CD) or digital ersatile disc (DVD), flash memory, etc. The methods may also be incorporated into hard -wired logic if desired, it should be noted that the order of the steps of the methods described herein may be altered and still be within the scope of the disclosure.

[0035] These and other variations upon the embodiments are intended to be covered b the present disclosure, which is limited only by the appended claims.