EP1374650A1

EP1374650A1 - Method and device for producing extreme ultraviolet radiation and soft x-radiation

Info

Publication number: EP1374650A1
Application number: EP02729817A
Authority: EP
Inventors: Jürgen Klein; Willi Neff; Stefan Seiwert; Klaus Bergmann; Joseph Pankert; Michael Löken
Original assignee: Fraunhofer Gesellschaft zur Forderung der Angewandten Forschung eV; Philips Intellectual Property and Standards GmbH; Koninklijke Philips Electronics NV
Current assignee: Fraunhofer Gesellschaft zur Forderung der Angewandten Forschung eV; Philips Intellectual Property and Standards GmbH; Koninklijke Philips NV
Priority date: 2001-04-06
Filing date: 2002-03-23
Publication date: 2004-01-02
Also published as: DE10291549D2; EP1248499B1; CN1531840A; DE50115489D1; EP1248499A1; TWI284916B; ATE469533T1; JP2004530269A; WO2002082872A1; JP4330344B2; DE10139677A1; US7126143B2; US20040183037A1; CN1311716C

Abstract

A discharge area (10) of pre-determined gas pressure and two electrodes (11,12) each have an opening set up on the same symmetrical axis (13) and form a plasma (17) in an area of their openings during an increase in voltage as they reach a preset ignition voltage. This is a source of radiation (17') to be generated, triggers plasma ignition and feeds stored energy into the plasma via the electrodes on igniting the plasma. An Independent claim is also included for a device for generating extremely ultraviolet radiation and weak X-rays.

Description

Method and device for generating extremely ultraviolet radiation and soft X-rays

The invention relates to a method for generating extremely ultraviolet radiation with the features of the preamble of claim 1.

A method with the aforementioned method steps is known from DE-A-197 53 696. The method is carried out with a device which has an electrode system forming the discharge space. This electrode system generates extremely ultraviolet radiation and soft X-rays, which are used in particular for EUV lithography. The electrode system consists of two electrodes, namely a cathode and an anode, which are each formed with an opening. The opening is essentially a hole and both openings lie on a common axis of symmetry. The cathode is designed as a hollow cathode, so it has a cavity. This is used to form the electric field in a predetermined way. In particular, the arrangement of the electrodes is such that the field lines in the area of the boreholes are sufficiently stretched to meet the breakdown condition above a certain voltage. The The discharge space is filled with gas and the gas pressure is in the range of 1 Pa to 100 Pa at least in the area of the electrode system. The geometry of the electrodes and the gas pressure are chosen so that the desired ignition of the plasma takes place on the left branch of the Paschen curve and, as a result, there is no dielectric breakdown between the electrodes outside the openings. As a result of the ignition, a current-carrying plasma channel is formed in an axially symmetrical shape, namely in the region of the openings of the electrodes. With the help of the energy store, a current is sent via this channel. The resulting Lorentz force constricts the plasma. As a result of this constriction effect and through ohmic heating, very high temperatures occur in the plasma and radiation of a very short wavelength is generated. The known device can generate EUV light in the wavelength range of 10-20 nm.

It is essential for the method that a switching element between the electrode system and the energy store can basically be dispensed with. A low-inductive and effective coupling of the electrically stored energy into the electrode system can therefore be achieved. Pulse energies of a few joules are sufficient to trigger current pulses in the range from several kiloamperes to a few 10 kiloamperes. The energy coupling is triggered in the controlled or self-breakthrough discharge in coordination with a predetermined ignition voltage. The ignition voltage is influenced, for example, by the gas composition, the temperature, a pre-ionization, the electrical field distribution and other variables. It can be set according to the Paschen curve by means of the gas pressure of the discharge vessel. Up to this ignition voltage, the energy store must also be charged, so that as much energy as possible can be fed into the plasma in the event of ignition.

The invention is based on the object of a method with the method steps mentioned at the beginning to be improved in such a way that the radiation yield, in particular the yield of EUV light per pulse, is improved, as is the pulse-to-pulse stability of a large number of successive discharges which occur in the process for generating the EUV Light can be exploited.

The above object is achieved by the features of the characterizing part of claim 1.

The operation of the process with ignition delay leads to an extension of the construction of the conductive plasma. This achieves an improvement in the cylinder symmetry of the low-resistance starting plasma required for the discharge, that is to say that plasma which builds up in the region of the openings of the electrodes after the ignition voltage has been reached. The ignition delay consequently leads to an improvement in the EUV yield / pulse and pulse-to-pulse stability. With a method in the range of pulse operation from 50 Hz to 500 Hz, an increase in the EUV yield of approximately 10 percent was observed when selecting an ignition delay of approximately 1 ms.

To influence the ignition delay, the procedure is such that the ignition delay is reduced by increasing the gas pressure or increased by reducing the gas pressure. Such changes in the gas pressure are particularly easy to achieve if the gas flows through the area of the electrode system, for example in order to influence the repetition frequency, that is to say to be able to carry out the method with higher pulse frequencies.

In order to influence the ignition delay, the method can be carried out in such a way that the ignition takes place by triggering a trigger pulse which is applied to a trigger electrode which influences an ignition area of the plasma. With triggering, the distribution of charge carriers is likes to be influenced in the ignition area of the plasma and thus also the time at which the ignition then takes place effectively.

It is expedient to proceed in such a way that the triggering for reaching a predetermined ignition delay takes place in combination with an application of a pressure interval of the gas pressure. In this case, both the pressure and the triggering time are set, since the discharge can only be operated stable or even at all within a certain pressure interval even in triggered operation.

In the context of triggering described above, the method can be carried out in such a way that triggering is applied with a predetermined trigger delay. The ignition delay is increased accordingly.

The coupling of stored energy into a discharge operated in self-breakthrough occurs automatically with the breakthrough, that is to say with the ignition of the plasma, care being taken to ensure that the energy store is charged before ignition takes place, taking into account the Pbetrieblsbetriebs. It is therefore necessary to have information about the voltage rise and reaching a predetermined ignition voltage. As a result, the method can be carried out in such a way that the voltage rise and / or the reaching of a predetermined ignition voltage is / are measured and that the gas pressure and / or the triggering is influenced taking into account the measurement result. If the influence is exerted as part of a continuous control, the gas pressure or a trigger delay is used as the manipulated variable. The desired ignition delay can thus be achieved or monitored using measurement technology.

It can also be done in such a way that the ignition timing is measured. This will make it possible to record the time between the time of reaching the Ignition voltage and the effective ignition time passes, which corresponds to the ignition delay.

To carry out the measurement of the ignition timing, the procedure can be such that the ignition timing is measured by measuring a voltage differential of the electrode voltage and / or by measuring a current differential of the electrode current. At the beginning of the ignition, the voltage applied to the electrodes changes abruptly, as does the current flowing in the discharge. The voltage collapses and the current swells, both can be reliably detected.

The ignition delay can be regulated in that the time between reaching the predetermined ignition voltage and the ignition point is measured and in that the gas pressure is adjusted according to the predetermined ignition delay by means of the measurement result. The time between reaching the predetermined ignition voltage and the ignition point is measured, for example, in an analog manner using an integrator or digitally using a counter. The time is fed to a controller as a measured variable, which accordingly adjusts the gas pressure to stabilize the ignition delay. It can be averaged over a series of discharge processes, that is over a predetermined number of pulses.

A special method is characterized in that the voltage applied to the electrodes is measured from the beginning of the voltage rise over a predetermined period of time, which includes a presumed ignition timing, an ignition voltage integrator being preferably used for the measurement. The time period therefore exceeds the time required for the charging process or the voltage rise at the electrodes. As a result, information about the ignition voltage and the ignition delay can be determined in the same signal. The ignition voltage integrator enables a multitude of information from the same measurement signal. Furthermore, the method can be modified in such a way that a measurement of the voltage applied to the electrodes includes storage of the ignition voltage value reached until the start of the subsequent voltage rise. The storage takes place, for example, with a sample-and-hold circuit.

Appropriately, the procedure can be such that the charge state of a capacitor bank connected directly to the electrodes as an energy store is continuously monitored during a voltage rise, and that after the predetermined ignition voltage has been reached, triggering is carried out, if necessary, with the predetermined trigger delay. Information about the state of charge of the capacitor bank can be obtained and evaluated using suitable electronics. They form the basis for the method to be operated according to one of the strategies described above, in which the gas pressure and / or the triggering of a trigger pulse is influenced.

With some high-voltage capacitors, their capacitance depends strongly on the temperature. In such cases, care must be taken that the energy of the capacitor is kept constant at the time of ignition. In this case, it is not a matter of keeping the ignition voltage constant, rather the predetermined ignition voltage must be corrected using a correction calculation. For such a correction calculation, the temperature of the capacitor can be measured, or the capacitance over the duration of the charging ramp of the applied charging voltage, in order then to be corrected accordingly.

A special method is characterized in that triggering is carried out by means of a trigger electrode acting on charge carriers in an electrode gap, by reducing the blocking potential formed with respect to a cathode. That way a trigger pulse can be reached at a predeterminable point in time so as to influence the ignition delay.

In view of a high EUV luminous efficacy, the procedure can be such that the energy store is recharged without a recombination of the gas which takes place completely after the plasma has gone out until a predetermined ignition voltage is reached. As a result, the repetition frequency in particular can be increased, the energy store being able to be recharged in shorter time intervals.

It is also possible for a high-resistance plasma to be burned between the electrodes in the period between two plasma discharges forming radiation to be generated. The high-resistance plasma leads to better conditions for a starting plasma of the high-current discharge.

The feeding of stored energy into a self-breakthrough discharge occurs automatically with the breakthrough, ie with the "ignition of the plasma. However, it must be taken into account that an untriggered discharge system has only a single breakthrough point, which is determined by the conditions of the Paschen curve - This point is not stable, and if the electrode system heats up, particularly in the discharge space, the breakdown will no longer take place at the same voltage.

Breakthroughs are also repeated in rapid succession to continuously generate radiation. Between two breakthroughs, the system needs a certain time to recombine the gas in the discharge space. During this time, the gas returns to its initial state at least in part, so that the energy store can be recharged and the required voltage can be built up at its electrodes. As a result, the state of the system also depends on when the last breakthrough took place or with which repetition frequency the generation of the Radiation occurred. With a high repetition frequency, the operating point on the Paschen curve will be different than with a low repetition frequency. In practice, this means that the repetition frequency can be very limited because no stable working point can be found at all. In connection with this there are problems in that it is not possible to quickly switch from one repetition frequency to another and that it is not possible to switch on and off repeatedly even at a specific repetition frequency. Switching on and off is particularly important when operating a lithography device in which pauses have to be taken between exposure processes in order to make settings on the device.

The invention is therefore also based on the object of improving a method with the features of the preamble of claim 17 in such a way that precise control of the pulses can be achieved in methods carried out in pulse mode for generating the EUV light, in particular in a wide range Parameter field of the discharge processes in order to improve the radiation yield of EUV light in the sense of the task described above.

This object is achieved by the features of the characterizing part of claim 17.

Triggering influences the ignition conditions for the plasma. In particular, the triggering influences the distribution of charge carriers in the ignition area of the plasma and thus also the point in time at which the ignition takes place effectively. The potential of the trigger electrode before the triggering process is higher than that of the cathode. As a result, such an influence is exerted on the field formation in the discharge space that a breakthrough cannot occur. This is only possible when the potential hindering the breakthrough has been eliminated. In a special way, the method is carried out in such a way that a voltage of the trigger electrode relative to the electrode used as the cathode, the voltage at the two electrodes and the gas pressure of the discharge space are set such that the plasma is not ignited when the trigger voltage is applied. which is only initiated by switching off the trigger voltage. Switching off the trigger voltage enables the electric field in the discharge space to be designed in such a way that the breakdown conditions are met. The time of the breakdown can be precisely determined by the trigger signal, namely the switching off of the trigger voltage. It is also important that the parameter range for a discharge can be expanded considerably. The pressure in the gas space, the distance between the electrodes and the voltage at the electrodes can be selected differently depending on the trigger voltage. While the breakthrough is only determined by a single point on the Paschen curve in the untriggered case, large voltage ranges Δu or pressure ranges ΔP can be defined in the triggered case, in which there is a breakthrough after the trigger pulse.

It is possible to set the parameters so that the process is operated with repetition frequencies between> 0 Hz and 100 kHz. Good results have been shown at repetition frequencies of 10 kHz.

It is also possible to carry out the method in such a way that it is operated with long operating intervals which can be set by switching on and off, during which a fixed repetition frequency is used in each case. An operating interval begins with the switching on and ends with the switching off. For example, a waver is exposed in a partial area during an operating interval. The radiation required for the exposure is carried out according to one of the methods described above, namely with a fixed repetition frequency. After an operating interval, an adjustment of the exposure device and / or the verse, in order to then carry out the method again after re-exposure of the same wave or another wave with a predetermined repetition frequency.

The invention also relates to a device with the features of the preamble of claim 21. Such a device is to be improved, in particular for the implementation of the above-described methods, in such a way that a long service life and good coolability of the electrodes are ensured. The object described above is achieved by the features of the characterizing part of claim 21. The design of the trigger electrode as a wall ensures long durability even in the event of temperature and plasma-related removal of material and its large surfaces are easy to cool, which in turn benefits a long service life. At the same time, the arrangement of the trigger electrode at a predetermined distance from the opening of the first electrode ensures that the shape of the electric field required for field formation can be ensured by means of the first electrode.

In the above sense, it is advantageous to design the device such that the first electrode is designed as a hollow electrode and that the trigger electrode is designed as a wall or wall section in the geometry of this hollow electrode. The result is a corresponding simplification of the electrode structure.

If the trigger electrode is designed as a parallel rear wall parallel to the hollow electrode, the opening of which is opposite, the simplification of the electrode structure is particularly promoted. In particular, symmetrical configurations of the electrode system can be achieved with respect to the axis of symmetry of the bores of the electrode.

It is preferable that the trigger electrode has a passage opening arranged in the axis of symmetry. In this way it can be avoided that particle radiation occurring during discharge and associated pulsed currents of typically a few 10 amperes undesirably flow to the trigger electronics via the trigger electrode.

For the construction of a hollow electrode, it is advantageous to design the device in such a way that the trigger electrode is cup-shaped and that a top axis which is vertical on a pot bottom is aligned with the axis of symmetry of the electrodes.

A simplified structure results from the fact that the trigger electrode is assembled with the first electrode via an insulator. The insulator allows the first electrode on the one hand and the trigger electrode on the other hand to be kept at different electrical potentials.

The above-described configuration of the device can be specified in such a way that the first electrode has an annular collar which is concentric with its opening and which overlaps the insulator overlapping the trigger electrode or engages in a ring recess of the trigger electrode, in each case while maintaining a potential-separating distance. In this way, evaporation and a short circuit of the insulator can be avoided.

The invention also relates to a device with the features of the preamble of claim 29. With such a device, ionization can occur in the discharge space. The mobile ions in the electric field hit the trigger electrode and usually have sufficient energy to knock secondary electrons out of the metallic surface of the electrode. Because of the potential difference, these electrons reach the anode. As a result, a conductive channel can be formed between the anode and the trigger electrode without the desired breakdown having already occurred in the area of the openings of the electrodes. there a noticeable part of the energy storage can be discharged via the trigger circuit, which entails the risk of this circuit being destroyed.

Furthermore, a problem can arise in that the potential of the trigger electrode drops to the level of the anode due to the formation of a conductive channel, as a result of which a high voltage is formed with respect to the cathode. As a result, undesired discharges can occur between the cathode and the trigger electrode, which likewise have a disruptive effect on the proper functioning of the device.

Ultimately, an ion or particle beam can cause it to atomize at least parts of the cathode due to its high energy. This leads to undesirable wear and deposits of atomizing particles on the surrounding surfaces.

In contrast, the invention has for its object to design a device with the aforementioned features so that a long life is achieved without disrupting the function.

The above object is achieved by the features of the characterizing part of claim 29. If the carrier electrode is arranged outside a particle beam which is formed in the axis of symmetry, the particles or ions accelerated in this axis no longer strike the carrier electrode. The malfunctions described above are therefore at least considerably reduced. The same applies if the trigger electrode has a shield that prevents the formation of a conductive channel between the trigger electrode and the anode.

An advantageous embodiment of the device is characterized in that the trigger electrode is arranged in the axis of symmetry of the openings of the electrodes and a has an insulator as a shield at least in the area of formation of the particle beam facing the openings. The trigger electrode can be arranged in the axis of symmetry in such a way that the field lines in the discharge space can be reliably influenced in a uniform manner. The isolator offers the desired protection of the trigger electrode without significantly distorting the field lines in the discharge space.

It is advantageous if the insulator is designed as a layer applied to the end face of the trigger electrode. In this case, the trigger electrode is adequately protected with a minimum of material.

The device can also be designed such that the insulator is designed as a body embedded in the end face of the trigger electrode. In this case, the trigger electrode must be assembled with the insulator using the usual mechanical production equipment.

An advantageous embodiment of the device can be characterized in that the insulator has a recess with a cross section matched to the particle beam. In this case, a particle beam can strike a bottom of the depression. Resulting atomization products are therefore mainly deposited on the inner walls of the recess and therefore hardly interfere with the other surfaces of the arrangement.

If the depression of the insulator is conically tapered, the energy of an ion beam is distributed over a larger surface and thus the local thermal heating is reduced. Correspondingly fewer atomization products are formed.

Another possibility is to design the device in such a way that the trigger electrode is completely insulated at least from the space adjacent to the first electrode is. The production of the trigger electrode for such a device can be advantageously influenced by complete insulation or coating. Inhomogeneities in the field or discharge formation on the metal surfaces of the trigger electrode in the transition region between insulated and non-insulated metal surfaces are also eliminated.

With complete insulation of the trigger electrode, it can be disadvantageous that under certain discharge conditions electrical charges accumulate on the insulated surface which could shield the trigger potential. To prevent this, the device can be designed in such a way that the shielding of the trigger electrode has a residual conductivity that dissipates surface charges, but prevents a current flow between the second electrode and the trigger electrode that influences the discharge. In this case, too, of a shield that reduces surface discharges, it is advantageous to completely isolate the trigger electrode in order to avoid additional lead paths.

If the trigger electrode is not to be located in the axis of symmetry, it is preferable to design the device in such a way that the trigger electrode is designed as a hollow cylinder surrounding the axis of symmetry.

In particular, the device can be designed such that a hollow cylindrical trigger electrode has a bottom facing away from the two electrodes, which is designed as an insulator or is a metal bottom which has the potential of one of the electrodes, for which purpose it is insulated from the trigger electrode. The insulator can then take over the functions of the insulators described above, in particular with regard to a possible particle beam. If the base is a metal base, it can either be placed on the potential of the anode so that a conductive channel is not created due to the equality of potential. The metal bo but it can also be connected to the potential of the cathode in order to suck off the charge carriers that are formed.

Furthermore, it is advantageous to design the device in such a way that the trigger electrode is an annular disk or at least one electrode pin which is / are built into the first electrode transversely to the axis of symmetry of the electrodes. The electric field in the discharge space or in the space adjacent to the trigger electrode can be influenced with the annular disk or with an electrode pin in order to influence the discharge behavior of the device. In order to achieve the goal described above, the trigger electrode is insulated and installed in the first electrode.

The device is exposed to considerable heat during its operation. It is therefore advisable to design them so that the shielding is made of temperature-resistant insulation material.

Because of the heat development described above, it is also sensible for the shield to be connected to the trigger electrode with good thermal conductivity in order to dissipate heat.

In order to intercept the predominant part of the charge carriers that reach a shield in the region of the axis of symmetry, the device is expediently designed in such a way that the shield has a diameter that corresponds at least to the diameter of the openings.

The invention is explained with reference to a drawing. It shows:

1 shows a schematic illustration of an electrode system, FIG. 2, FIG. 3 shows diagrammatic representations of the voltage profile at the electrodes of the electrode system for an ignition process of a plasma during pulse operation, 4, FIG. 5 differently configured electrode configurations, FIG. 6 a schematic representation of an electrode system, similar to FIG. 1, FIG. 7 a diagrammatic representation of the dependence of the ignition voltage of an electrode system on the pressure in a discharge space, and FIGS. 8 to 18 schematized Representations of electrode systems with differently designed trigger devices.

Fig.l shows schematically the formation of an electrode system arranged in a discharge space 10. The discharge space 10 is filled with gas of predetermined gas pressure and can be formed by suitably designed electrodes of the electrode system itself. The gas pressure is adjustable. The equipment of the discharge vessel 10 required for adjusting the gas pressure and a design of the electrode system which is matched to this are available, but not shown.

There are two electrodes 11, 12. The electrode 12 is designed as an anode with a central opening 15 which widens conically starting from an electrode gap 22.

The electrode 11 is designed as a cathode, specifically as a hollow cathode with a cavity 23 which is connected to the electrode interspace 22 via an opening 14 in the cathode. The openings 14, 15 are aligned and form an axis of symmetry 13 of the electrode system. The electrodes 11, 12 are insulated from one another. An insulator 29 serving this purpose determines the electrode spacing.

As a result of the design described above, the electrode system is capable of applying a high voltage in the range of, for example, a few 10 kV Form field lines that run in the area of the electrode gap 22 in a straight line and parallel to the axis of symmetry 13. If the voltage is increased in terms of pulses starting from a predetermined low value, a charging ramp or a voltage rise 16 results according to FIGS. 2, 3. Ionization processes occur which are concentrated in the electrode interspace 22 on account of the field strength relationships. For this purpose, the voltage rise 16 and the gas pressure are coordinated with one another in such a way that, as a result of the ionization, a gas discharge occurs on the left branch of the Paschen curve, in which a plasma channel or its plasma is not built up via a single short-term electron avalanche, but in several stages via Secondary ionization processes. As a result, the plasma distribution is already highly symmetrical in the starting phase, as the schematic representation of the plasma in FIG. 1 is intended to express. The plasma 17 that forms is a source of the radiation 17 ′ to be generated.

It goes without saying that ignition of the plasma 17 is only possible when an ignition voltage U _{z has been} reached. In the invention it is now ensured that an ignition delay 18 occurs. As a result, _{is, for} the ignition timing t despite the presence of the ignition voltage U _z correspondingly later. The size of the ignition delay 18 is regulated by controlling the gas pressure. With typical durations, the size of the ignition delay ranges from a few microseconds to a few milliseconds. The ignition delay leads to an extension of the build-up of the conductive plasma. This improves the cylinder symmetry of the plasma 17.

The plasma formed after the ignition delay can be referred to as the start plasma. It can be used to couple energy from an energy store in self-breakthrough operation. Fig.l shows a capacitor bank 21 as an energy store, which discharges after reaching the predetermined ignition voltage and ignition delay and thereby enables. Feed current pulses in the double-digit kiloampere range into the plasma. The As a result, the Lorentz forces of the magnetic field that are formed constrict the plasma, so that there is a high luminance and, in particular, the formation of extremely ultraviolet radiation and soft X-rays, which have the required wavelengths, in particular for EUV lithography.

Instead of influencing the ignition delay 18 via the gas pressure, an influence can also be exerted via a trigger electrode. With a trigger electrode 19 it can be achieved that, despite reaching a predetermined ignition voltage U _z, a breakdown between the electrodes 11, 12 does not yet occur for the discharge. A trigger delay 20 that can be achieved with a trigger electrode 19 according to FIGS. 4, 5 is shown in FIG. It is added to the ignition delay 18. Influencing an overall ignition delay by a trigger delay 20 is particularly advantageous because measurement technology can be used to achieve more precise ignition times t _z . This applies both in the event that the gas discharge operation takes place in a self-breakthrough and when a switching element is used between the electrode system and the capacitor bank. The switching element allows a voltage to be applied to the electrode system that is greater than the ignition voltage U _z required for self-breakdown operation. In the latter case, one can then work with higher gas pressures, which leads to higher intensities of the emitted radiation.

It is expedient to measure the ignition timing, especially if the charger allows a higher voltage on the electrode system than the predetermined ignition voltage U _z . The voltage applied to the electrode system, that is to say the course of a voltage rise 16, can be detected, for example by detecting the change over time in the voltage applied to the electrodes 11, 12. A dU / dt measurement is carried out. A dl / dt measurement can also be carried out, that is, a detection of the change in the discharge current over time. Current and voltage change when the Ignition point t _z suddenly. The time between reaching the predetermined ignition voltage U _z and the ignition point can be measured, for example, analogously using an integrator or digitally using a counter. This time is fed to a controller as a measured variable, which then influences the gas pressure in the sense of stabilizing the ignition delay 18. This also applies to the use of a trigger delay 20.

The measurement can be carried out, for example, with an ignition voltage integrator, which takes over the preparation of the measured variable high voltage or voltage upstream of the actual regulator at the electrode system or the capacitor bank. The ignition voltage integrator integrates the divided high voltage present at the electrodes 11, 12 and registers its end value via a sample and hold until the next charging process. The integration process begins with the charging process, that is to say with the rise in the electrical voltage applied to the electrodes 11, 12, and continues until a period of time defined by a timer. This period is usually longer than the actual charging process, so that the desired information about the size of the ignition delay can also be determined. Additional non-linear terms, such as square root extractors, can be used to improve the transmission characteristic. Information about the ignition delay, as well as about the ignition voltage, is thus obtained with the same measurement signal. In contrast to a peak detector, which detects the ignition voltage, the method is completely insensitive to interference peaks, for example from the high-voltage generator. Electronics are not required to detect the ignition timing.

If the method is carried out without a trigger electrode, the ignition timing t _z can only be determined via the level of the gas pressure. In a method with a trigger electrode, the trigger delay described above can be used to determine the ignition point, if necessary in combination with a selection of the suitable gas pressure. The state of charge of the capacitor bank 21 is determined via evaluation electronics, for example with the aid of the ignition voltage integrator described above. Triggering with the aid of the trigger electrode means that despite the reaching of the ignition voltage U _z, the plasma formation causing the capacitor bank 21 to discharge is not yet produced. Only in the event of triggering is the ignition triggered, that is to say when a trigger pulse is triggered according to a predetermined trigger delay 20. The manipulated variable here can also be the gas pressure which is set, for example, via an electronic inlet valve. If the holding voltage is not reached after a predetermined readout time, the gas pressure must be reduced. Otherwise, the gas pressure must be increased if there is no ignition after a trigger pulse. The controlled variable in this process with trigger electrode is ultimately the ignition delay, i.e. the time between the triggering of the trigger pulse and the voltage breakdown. The pressure is then set so that the ignition delay is kept constant within a certain tolerance.

The trigger delay 20 shown in FIG. 3 is based there, for example, on the time at which the predetermined ignition voltage U _{z is reached} . In principle, any point in time that can be determined using suitable electronics can also be selected beforehand, for example the start of the charging process or the reaching of a predetermined value for the charging voltage.

In FIGS. 4, 5, configurations of trigger electrodes are shown, for example. The trigger electrodes 19 are adjacent to the cathode 11, on the side of the cathode 11 facing away from the anode 12. Here they are assembled with the cathode 11 via an insulator 26, means for holding the electrode 11, the insulator 26 and the trigger electrode 19 together are not shown. It is common to all embodiments of the trigger electrode that they are arranged symmetrically with respect to the axis of symmetry 13. All embodiments have an axis that is aligned with the axis of symmetry 13. The trigger electrode 19 is designed as a wall or as a wall section. It lies at a predetermined distance from the opening 14 of the electrode 11. It can at the same time be achieved that the electrode 11 is designed as a hollow electrode, for example as a hollow cathode. The trigger electrode 19 then essentially forms the rear wall of the cathode. Such a rear wall in the case of FIG. 4 is a wall 29 and in the case of FIG. 5 is a pot bottom 19 ′ of the pot-shaped trigger electrode 19.

The cup-shaped design of the trigger electrode 19 shows that this can not only be the rear wall of the electrode 11, but also the side wall of the space 23 to be bounded by this hollow electrode. It is also conceivable that the trigger electrode 19 is exclusively the side wall section of an electrode 11, which is otherwise connected to the electrode or cathode potential.

FIG. 4 illustrates that the trigger electrode 19 can be provided with a passage opening 24 which serves for the passage of particle beams which, according to the electrode formation, are formed primarily in the region of the axis of symmetry. With such a passage opening, a load on the trigger electronics can be kept within acceptable limits. The particle beams are taken up by the parts of the electrode system which are at the potential of the cathode. A passage opening 24 can also be used in the case of FIG. 5.

4 shows holes 24 'parallel to the passage opening 24. These holes 24 'can serve as gas holes, namely for the passage of gas in the sense of a gas inlet. In the sense of such a gas flow or in the sense of a gas inlet, the passage Opening 24 can be used. Both are particularly advantageous if the electrode system itself forms the discharge space 10.

In the case of the gas discharge protruding into the space 23, vaporization of the insulator 26 with metal vapor is to be expected. Such a could lead to a short circuit of the insulator 26. In order to shield it from metal vapor that occurs, the electrode 11 is provided with an annular collar 27 which is arranged concentrically with the opening 14 and overlaps the toroidal insulator 26. In addition, the trigger electrode 19 is provided with an annular recess 28. The collar 27 engages in the ring recess 28. A potential-separating distance is maintained, which, however, only needs to be small because of the normally small potential differences between the cathode 11 and the trigger electrode 19.

Trigger electrodes according to Fig. 4, 5 are also possible in connection with a hollow anode. In this case, the light from the plasma 17 would have to be coupled out from the electrode 11 or from the hollow cathode. However, it is more advantageous to decouple the light on the anode side, as shown in FIG. 1, and to operate the cathode with negative high voltage, since it is better to avoid debris from sputtering and high-frequency discharges in the part of the electrode system facing an observer.

The potential of the trigger electrode 19 is selected before triggering a trigger pulse and thus before triggering a low-resistance plasma discharge in such a way that charge carriers are withdrawn from the hollow electrode or hollow cathode and the space between the electrodes in the borehole region. This is done, for example, by applying a voltage, which is positive compared to the cathode potential, of typically a few 100 V to the trigger electrode 19. A trigger pulse is then triggered by pulling the potential of the trigger electrode down to that of the cathode or by applying it to the trigger. gerelektrode 19 a negative potential is applied. Typical time constants for a change in the potential of the trigger electrode 19 are advantageously in the range from a few nanoseconds to a few 100 ns.

In order to achieve high light yields, the aim is to keep the repetition frequency of the discharges as high as possible, namely in the range of several kHz and preferably above 10 kHz. Here the necessary re-consolidation times or recombination times of the plasma set limits. These limits depend on the type of gas with which the process is operated. With regard to a high radiation yield in the EUV area, the use of xenon is of particular interest. When operating with pure xenon, repetition frequencies above approximately 1 kHz with typical pulse energies in the range from 1 joule to 10 joule when operating in self-breakthrough can hardly be achieved. It is therefore desirable to take measures to accelerate the reconsolidation.

One possibility is to add gases. A faster recombination of the plasma after discharge of the capacitor bank can be achieved by adding gases such as air, synthetic air, nitrogen, oxygen or halogens.

In addition, the removal of charged particles from the area of the openings 14, 15 can be supported by a suitable gas flow. Flow with gas inlet via the cathode and / or through the electrode gap and with gas evacuation via the anode, which is the electrode facing the observer according to FIG. 1, is advantageous. With such a gas flow, pressure drop can be generated in the area of the anode or in the area of a hollow anode. With such pressure gradients, it is possible to displace the plasma 17 in order thereby to achieve an increased transmission for the EUV radiation in the observation path up to the user. Further measures for increasing the repetition frequency can be carried out in connection with the capacitor bank 21. It is assumed that the construction of the low-resistance Piamas takes up to several 100 microseconds, depending on the conditions. The capacitor bank 21 can now be charged faster than this build-up time for the low-resistance plasma. As a result, a complete recombination of the plasma can be dispensed with. In addition, it is even possible to have a high-resistance plasma burn in the region of the openings 14, 15 between two discharges, which can lead to better conditions for a starting plasma of the high-current discharge.

6 schematically shows the design of an electrode system arranged in a discharge space 10. The discharge space 10 is filled with gas of predetermined gas pressure and can be formed by suitably designed electrodes of the electrode system itself. The gas pressure is adjustable. The equipment of the discharge vessel 10 required for adjusting the gas pressure and a shape of the electrode system which is matched to this are available, but not shown.

The electrode 11 is designed as a cathode, specifically as a hollow cathode with a cavity 23 which is connected to the electrode interspace 22 via an opening 14 in the cathode. The openings 14, 15 are aligned and form an axis of symmetry 13 of the electrode system. The electrodes 11, 12 are insulated from one another. An insulator 29 serving this purpose determines the electrode spacing. As a result of the configuration described above, the electrode system is able to form field lines when applying an electrical high voltage in the range of, for example, a few 10 kV, which in any case run in the area of the electrode gap 22 in a straight line and parallel to the axis of symmetry 13. If the voltage is increased on a pulse basis from a predetermined low value, a charging ramp or a voltage increase results. Ionization processes occur, which are concentrated in the electrode gap 22 due to the field-strength relationships. For this purpose, the voltage rise and the gas pressure are coordinated with each other in such a way that ionization leads to a gas discharge on the left branch of the Paschen curve, in which a plasma channel or its plasma is not built up via a single short-term electron avalanche, but in several stages via secondary ionization processes. As a result, the plasma distribution is already highly cylinder-symmetrical in the start phase, as the schematic representation of the plasma in FIG. 6 is intended to express. The plasma 17 which forms is a source of the radiation '17' to be generated, an electron beam.

The plasma formed can be referred to as start plasma. It can be used for coupling energy from an energy store in self-breakthrough operation. 6 shows a capacitor bank 21 as an energy store, which discharges after reaching the predetermined ignition voltage and thereby enables current pulses in the two-digit kiloampere range to be fed into the plasma. The Lorentz forces of the magnetic field that form as a result constrict the plasma, so that there is a high luminance and, in particular, the formation of extremely ultraviolet radiation and soft X-rays, which have the required wavelengths, in particular for EUV lithography.

The electrode system shown in FIG. 6 is provided with a trigger device in the area of the electrode 11. For this purpose, the electrode 11 points in the axis of symmetry 13 a trigger electrode 19, which is held by an insulator 26 in the bottom 30 of the electrode 11. The insulator 26 serves to enable the trigger electrode 19 to be given a potential which is different from that of the electrode 11. The trigger electrode 19 has a parasitic capacitance 31 with respect to the electrode 11, measured in parallel with a switch 32, with which both electrodes 19, 11 can be brought to the same potential.

The electrode 12 is usually designed as an anode and is grounded as shown. In contrast, the cathode is at a negative potential -V while the trigger electrode 19 is at a potential -V + Vt. The potential of the trigger electrode before the start of the triggering process is therefore somewhat higher than that of the electrode 11. For the purpose of triggering, a trigger pulse is triggered by closing the switch 32, the potential of the trigger electrode 19 being pulled down to that of the electrode 11. Typical time constants for a change in the potential of the trigger electrode 19 are advantageously in the range from a few nanoseconds to a few hundred nanoseconds.

The electrode arrangement shown schematically in FIG. 6 is typically designed such that there is a distance of 1 to 10 mm between the electrodes 11, 12. The smallest passage of the openings 14, 15 is typically 1 to 10 mm. The volume of the space 23 in the electrode 11 designed as a hollow cathode is typically 1 to 10 cc. The gas pressure is between 0.01 and 1 mbar. The electrode voltage is typically 3 to 30 kV and the potential difference between the trigger electrode 19 and the electrode 11 is between 50 volts and 1000 volts.

Basically, the ignition voltage at which there is a breakdown between the electrodes 11, 12 and the pressure corresponding to the curve shown in FIG. other dependent. Figure 7 relates to the left branch of the Paschen curve.

The left curve in FIG. 7 applies to the operation of the untriggered device. There is only a single breakthrough point on this curve for V = 0, which is, for example, about 8 kV at a gas pressure of 7 Pa. Other pressures in space 23 accordingly result in different ignition voltages. The trigger voltage, ie the potential difference between the trigger electrode 19 and the electrode 1, can also deviate from 0. In this case, Vt is not equal to 0 but, for example, equal to Vi or V _2. As a result, by suitable measurements of the trigger voltage Vt, the device can be operated with different parameters. For a predetermined voltage at the electrodes 11, 12 there is the possibility of the pressure variation shown in FIG. Similarly, the voltage variation shown in FIG. 2 is possible for a predetermined pressure. Correspondingly, however, the point in time of the breakthrough can also be precisely defined with the trigger signal without thereby entering a work area in which the difficulties described above occur. In particular, repetition frequencies can be ensured, as are necessary for the required use, for example in the range from 10 to 20 kHz. Operating intervals for predetermined fixed repetition frequencies are also possible, as a result of which the energy required per se for generating the desired radiation can be saved between the operating intervals. The stability of the working point is significantly improved.

Triggering is achieved by the circuit shown in Fig. 6. The capacitor bank 21 is charged by applying the electrode 11 to negative voltage while the electrode 12 is grounded. The two electrodes 11, 12 are connected to the capacitor bank 21 via a low-inductance circuit. A high-impedance circuit connects the trigger electrode 19 to the electrode. trode 11, wherein the connection can be opened by the switch 32. In the case of opening, there is a potential difference V _t at the trigger electrode 19 in relation to the electrode 11. In this case, the voltages at the electrodes 11, 12 and the gas pressure of the interelectrode space or of the space 23 of the electrode 11 are set such that when a trigger voltage Vt is applied, the plasma 17 cannot be ignited. However, if the switch 32 is closed, the potential difference V is eliminated and the trigger electrode 19 receives the potential of the electrode 11, a protective resistor 33 protecting the voltage source of the trigger voltage.

With the switch 32 open, however, it may be possible for a conductive channel with a corresponding particle beam to form between the trigger electrode 19 of FIG. 6 and the electrode 12 serving as the anode, which channel discharges the energy of the capacitor bank 21 and can also damage the trigger circuit , FIGS. 8 to 18 therefore show differently designed trigger electrodes in a schematically illustrated system of main electrodes 11, 12, which can contribute to the proper functioning of the device.

8 to 18 show trigger electrodes 19 which are arranged coaxially with the axis of symmetry 13 which is formed by the electrodes 11, 12 or their openings 14, 15. The trigger electrodes 19 of FIGS. 8 to 13 are designed such that they face the end face 34 of the opening 14. At least this end face 34 is, however, each provided with a differently designed shield 35. Each shield 35 is at least as large as the diameter of the openings 14, 15. The shield 35 is therefore present in the vicinity of the trigger electrode 19 in the training area of the particle beam.

In the case of FIG. 8, the shield 35 is an insulator in the form of an on the end face 34 of the trigger electrode 19 applied layer formed. In the case of FIG. 9, a shield 35 is also designed as an insulator, but as a body let into the end face 34 of the trigger electrode 39. The cross section of this body is, for example, circular-cylindrical in order to be inserted in a conventional manner into a bore in the trigger electrode 19 which is introduced from the end face 34 thereof. In FIG. 10 and in FIG. 11, the trigger electrode 19 is the same as that in FIG. 9. However, different shields 35 are inserted into their bore. The shield 35 of FIG. 10 is in turn a cylindrical body, which, however, has a coaxial recess 36 which is designed as a blind hole. The diameter of the blind bore is matched to the diameter of the potential particle beam. The shield 35 of FIG. 11 is formed with a recess 36 which tapers away conically from the openings 14, 15. An approximately forming particle beam strikes comparatively large areas of the shield 35, so that the beam energy is distributed over a larger surface, which prevents local thermal heating. In both cases of FIGS. 10, 11, the depressions are suitable for receiving atomization products resulting from a particle beam, which can be deposited on the inner walls of the depressions 36 and therefore hardly interfere with the other surfaces of the arrangement.

The trigger electrodes of FIGS. 12, 13 are characterized in that they are completely insulated from their shielding, at least against the space 23 adjoining the first electrode 11. The shield 35 is a coating that does not leave the surface of the trigger electrode 19 exposed at any point. As a result, there can be no inhomogeneities of any kind in the electric field that would be caused by such a release. Under certain discharge conditions, however, it can happen that 35 electrical charges collect on the surface of this shield, which can shield the trigger voltage. Shielding the trigger voltage would have failed Function of the device result. Such shields can be prevented if the shield 35 is provided with a residual conductivity that is large enough to neutralize or reduce the surface discharges that have built up. However, this residual conductivity is not large enough to allow a current to flow between the electrode 12 and the trigger electrode 19, which significantly discharges the capacitor bank 21. FIG. 13 shows such a shield 35 with a suitable residual conductivity.

In all of the above-described embodiments, the dimensions can vary within wide limits. For example, the trigger electrode 19 can also be designed as a thin wire, which is then expediently coated according to FIGS. 12, 13.

The trigger electrodes 19 of FIGS. 14 to 16 are hollow cylindrical. These trigger electrodes are also arranged coaxially with the axis of symmetry 13. As a result of their hollow-cylindrical design and the field formation, moreover, a particle beam which is formed in the region of the axis of symmetry 13 cannot reach the trigger electrode 19 and have a disruptive or destructive effect there. In FIG. 14, the trigger electrode 19 is closed by a metallic base 37 which is at ground potential and is insulated from the hollow cylindrical trigger electrode 19. No particle beam can form between the bottom 37 and the electrode 12 because this electrode is also at ground potential as an anode.

In FIG. 16, a bottom 38 is designed as an insulator and thus has a similar effect with respect to the particle beam as the shields described for FIGS. 8 to 11. In FIG. 16, the bottom 39 of the hollow cylindrical trigger electrode 19 is designed as a metal electrode which is conductively connected to the electrode 11, the cathode. Charge carriers of particle beams present in the axis of symmetry are supplied to the electrode 12 by means of the metallic base 39 via a connecting line 40.

The configurations of FIGS. 17 and 18 are alternative arrangements to FIG. 16. In all cases, charge particles located in the axis of symmetry 13 or in the space 23 are supplied to the electrode 11. 17, the trigger electrode 19 is designed as an annular disk. This ring disk is installed transversely to the axis of symmetry 13 of the electrodes 11, 12 in the first electrode 11. The upper and lower halves in FIG. 17 are conductively connected by a line 41 shown in dashed lines, and thus have the same potential. The arrangement of the trigger electrode 19 is cylindrically symmetrical with respect to the axis of symmetry 13. In the case of Fig. 18, this is no longer the case. In this embodiment, except for the line 41, the embodiment seen from the side can be, as was shown in FIG. 17. However, the axis of symmetry 13 in FIG. 18 is perpendicular to the plane of representation and FIG. 18 shows two identically designed parts 19 'and 19' 'of a trigger electrode which are arranged coaxially and transversely to the axis of symmetry 13. The parts 19 ', 19' 'represent electrode pins. Instead of the two parts 19, 19', the trigger electrode can also have several parts.

The shields 35 used in the trigger electrodes 19 consist of temperature-resistant insulation materials, such as, for example, Al ₂ O ₃ , quartz or silicon carbide. All materials used for shields 35 are connected to the trigger electrode 19 with good thermal conductivity.

Furthermore, it goes without saying that the trigger electrode 19 or its parts 19 ′, 19 ″ is / are built into the first electrode 11 in an isolated manner. The insulation 42 shown in Figures 8 to 18 perform the same functions as the insulator 26 of Figure 6. The insulation 42 in question is temperature-resistant in each case.

Claims

Claims:

1. A method for generating extremely ultraviolet radiation and soft X-rays with a gas discharge operated on the left branch of the Paschen curve, in particular for EUV lithography, in which a discharge space (10) of predetermined gas pressure and two electrodes (11, 12) are used, each having an opening (14, 15) located on the same axis of symmetry (13), and which in the course of a voltage rise (16) when a predetermined ignition voltage (U _z ) is reached, a plasma (17 ) train, which is a source of the radiation to be generated (17 '), in which the plasma (17) is ignited by influencing the gas pressure and / or by triggering, and in which the plasma (17) is ignited Energy storage by means of the electrodes (11, 12) feeds storage energy into the plasma (17), characterized in that the ignition of the plasma (17) occurs using a predetermined ignition delay (18) olgt.

2. The method according to claim 1, characterized in that the ignition delay (18) is reduced by increasing the gas pressure or is increased by reducing the gas pressure.

3. The method according to claim 1 or 2, characterized in that the ignition takes place by triggering a trigger pulse which is applied to a trigger electrode (19) influencing an ignition region of the plasma (17).

4. The method according to any one of claims 1 to 3, characterized in that the triggering to achieve a predetermined ignition delay (18) is carried out in combination with an application of a pressure interval of the gas pressure.

5. The method according to any one of claims 1 to 4, characterized in that the triggering is applied with a predetermined trigger delay (20).

6. The method according to any one of claims 1 to 5, characterized in that the voltage rise (16) and / or reaching a predetermined ignition voltage (U _z ) is / are detected by measurement, and that an influence on the gas pressure and / or on the Triggering takes into account the measurement result.

7. The method according to any one of claims 1 to 6, characterized in that the ignition timing (t _z ) is measured.

8. The method according to any one of claims 1 to 7, characterized in that the ignition point (t _z ) by measuring a voltage differential (dU / dt) of the electrode voltage (U) and / or by measuring a current differential (dl / dt) of the electrode current is measured.

9. The method according to any one of claims 1 to 8, characterized in that the time between reaching the predetermined ignition voltage (U _z ) and the ignition timing (t _z ) is measured, and that the gas pressure by means of the measurement result of the predetermined ignition delay (18) is set accordingly.

10. The method according to any one of claims 1 to 9, characterized in that a measurement of the voltage applied to the electrodes (11, 12) from the beginning of the voltage rise (16) over a predetermined The period of time that includes a putative ignition point (t ₂ ) takes place, an ignition voltage integrator preferably being used for the measurement.

11. The method according to any one of claims 1 to 10, characterized in that a measurement of the voltage (U) applied to the electrodes (11, 12) stores the ignition voltage value (U _z ) reached until the start of the subsequent voltage rise (16) includes.

12. The method according to any one of claims 1 to 11, characterized in that the state of charge of a capacitor bank (21) connected directly to the electrodes as an energy store is continuously monitored during a voltage rise (16), and that after reaching the predetermined ignition voltage (U _z ) if necessary, triggering is carried out with the predetermined trigger delay (20).

13. The method according to any one of claims 1 to 12, characterized in that the predetermined ignition voltage

(U _z ) is corrected as a function of at least one capacitance parameter of a capacitor bank (21).

14. The method according to any one of claims 1 to 13, characterized in that a triggering is carried out by means of a trigger electrode (19) acting on charge carriers of an electrode gap (22) by reducing the blocking potential formed with respect to a cathode.

15. The method according to any one of claims 1 to 14, characterized in that the energy storage is recharged without a recombination of the gas which takes place completely after the extinction of the plasma (17) until a predetermined ignition voltage (U _z ) is reached.

16. The method according to claim 15, characterized in that a high-resistance plasma is burned between the electrodes (11, 12) in the period between two plasma discharges forming radiation to be generated.

17. A method for generating extremely ultraviolet radiation and soft X-radiation with a gas discharge operated on the left branch of the Paschen curve, in particular for EUV lithography, in which a discharge space (10) of predetermined gas pressure and two electrodes (11, 12) are used, which each have an opening (14, 15) located on the same axis of symmetry (13), and which in the course of a voltage rise (16) when a predetermined ignition voltage (U _z ) is reached, a plasma located in the region of their openings (14, 15) (17), which is a source of the radiation to be generated (17 '), in which the plasma (17) is ignited by triggering, and in which, when the plasma (17) is ignited, an energy store by means of the electrodes (11, 12) feeds storage energy into the plasma (17), characterized in that the ignition of the plasma (17) takes place with a trigger electrode (19), the potential of which before the start of the triggering process is higher than which is one (11) of the electrodes (11, 12) used as the cathode.

18. The method according to claim 17, characterized in that a voltage of the trigger electrode (19) with respect to the electrode (11) used as the cathode, the voltage at the two electrodes (11, 12) and the gas pressure of the discharge space (10) are set in this way that when the trigger voltage is applied, the plasma (17) is not ignited, which is only initiated by switching off the trigger voltage.

19. The method according to claim 17 or 18, characterized in that it is operated with repetition frequencies between> 0 Hz and 100 kHz.

20. The method according to any one of claims 17 to 19, characterized in that it is operated with long operating intervals adjustable by switching on and off, during which a fixed repetition frequency is used in each case.

21. Device for generating extremely ultraviolet radiation and soft X-rays with a gas discharge operated on the left branch of the Paschen curve, in which there are a discharge space (10) of predetermined gas pressure and two electrodes (11, 12), each one on the same axis of symmetry ( 13) have an opening (14, 15), and which, in the course of a voltage rise (16) when a predetermined ignition voltage (U _z ) is reached, have a plasma (17) in the region of their openings (14, 15), which has a Source of the radiation to be generated (17 '), in which a trigger electrode (19) is present in a space (23) adjacent to a first electrode (11) for triggering the plasma (17) by triggering, in particular for carrying out a Method according to one of the preceding claims, characterized in that the trigger electrode (19) is designed as a wall (29) which has a predetermined distance at least in sections and from the opening (14) of the first electrode (11).

22. The apparatus according to claim 21, characterized in that the first electrode (11) is designed as a hollow electrode, and that the trigger electrode (19) is designed as a wall or wall section in the geometry of this hollow electrode.

23. The apparatus of claim 21 or 22, characterized in that the trigger electrode (19) as the hollow electrode parallel, the opening (14) opposite rear wall is formed.

24. Device according to one of claims 21 to 23, characterized in that the trigger electrode (19) has a passage opening (24) arranged in the axis of symmetry (13).

25. The device according to claim 24, characterized in that the passage opening (24) and / or the axis of symmetry (13) parallel bores (24 ') is / are formed as a gas inlet.

26. Device according to one of claims 21 to 25, characterized in that the trigger electrode (19) is cup-shaped, and that a top axis (25) perpendicular to a pot base (19 ') with the axis of symmetry (13) of the electrodes ( 11.12) is the same.

27. The device according to one of claims 21 to 26, characterized in that the trigger electrode (19) with the first electrode (11) is assembled via an insulator (26).

28. Device according to one of claims 21 to 27, characterized in that the first electrode (11) has an annular collar (27) which is concentric with its opening (14) and which overlaps the insulator (26) with the trigger electrode (19) or engages in a ring recess (28) of the trigger electrode (19), in each case while maintaining a potential-separating distance.

29. Device for generating extremely ultraviolet radiation and soft X-ray radiation with a gas discharge operated on the left branch of the Paschen curve, in which there is a discharge space (10) of predetermined gas pressure and two electrodes (11, 12), each having an opening (14, 15) located on the same axis of symmetry (13), and which in the course of a voltage increase (16) When a predetermined ignition voltage (U _z ) is reached, a plasma (17) is located in the region of its openings (14, 15) and is a source of the radiation (17 ') to be generated, in which one adjoins a first electrode (11) - Zenden space (23), a trigger electrode (19) for triggering the plasma (17) is present, and in which a by means of the electrodes (11, 12) storage energy into the plasma (17) is more likely to be supplied with energy , in particular for carrying out a method according to one of the preceding claims, characterized in that the trigger electrode (19) is arranged outside a particle beam which forms in the axis of symmetry (13) or a paragraph preventing the latter has shielding (35).

30. The device according to claim 29, characterized in that the trigger electrode (19) is arranged in the axis of symmetry of the openings (14, 15) of the electrodes (11, 12) and an end face (34) facing the openings (14, 15) ) has an insulator as a shield (35) at least in the training area of the particle beam.

31. The device according to claim 30, characterized in that the insulator is formed as a layer applied to the end face (34) of the trigger electrode (19).

32. Apparatus according to claim 30, characterized in that the insulator as in the end face (34) Trigger electrode (19) recessed body is formed.

33. Apparatus according to claim 32, characterized in that the insulator has a depression (36) with a cross section matched to the particle beam.

34. Apparatus according to claim 33, characterized in that the recess (36) of the insulator is conically tapered.

35. Device according to one of claims 29 to 34, characterized in that the trigger electrode (19) is at least completely insulated from the space (23) adjoining the first electrode (11).

36. Device according to one of claims 29 to 35, characterized in that the shield (35) of the trigger electrode (19) a surface charge-reducing, but a discharge-influencing current flow between the second electrode (12) and the trigger electrode (19) preventing residual conductivity Has.

37. Device according to claim 29, characterized in that the trigger electrode (19) is designed as a hollow cylinder surrounding the axis of symmetry.

38. Apparatus according to claim 37, characterized in that a hollow cylindrical trigger electrode (19). has a base facing away from the two electrodes (11, 12) which is designed as an insulator or which is a metal base which has the potential of one of the electrodes (11, 12), for which purpose it is insulated from the trigger electrode (19).

39. Apparatus according to claim 29, characterized in that the trigger electrode (19) is an annular disc or at least one electrode pin, the / the cross to the axis of symmetry (13) of the electrodes (11, 12) is / are installed in the first electrode (11).

40. Device according to one of claims 29 to 39, characterized in that the trigger electrode (19) is installed insulated in the first electrode (11).

41. Device according to one of claims 29 to 40, characterized in that the shield (35) consists of temperature-resistant insulation material.

42. Device according to one of claims 29 to 41, characterized in that the shield (35) is connected to the trigger electrode (19) with good thermal conductivity.

43. Device according to one of claims 29 to 42, characterized in that the shield (35) has a diameter which corresponds at least to the diameter of the openings (14, 15).