KR20040031082A - Method for etching structures in an etching body by means of a plasma - Google Patents

Abstract

The present invention relates to a method for etching a structure to an etch body 19 by means of a plasma 15, in particular for etching a recess to a silicon body which is precisely defined transversely by an etch mask. In this case, the high frequency power modulated at least temporarily by the high frequency alternating voltage by the high frequency alternating voltage, the low frequency modulated high frequency power is coupled, and the intensity of the plasma is modulated as a function of time.

Description

Method for etching structures in an etching body by means of a plasma}

Insufficient pocket stability problems appear several times in the plasma etching process for etching a silicon body with a precisely defined recess in the manner of DE 42 41 045 C1, for example. That is, a deviation of the desired etch profile is formed, in particular at the interface between the etch body and the dielectric interface, such as, for example, between silicon and silicon dioxide underlying it.

Application DE 199 57 169 A1 has already described a so-called double pulse technique, wherein the low frequency pulses of the high frequency modulated low frequency carrier signal of the high impulse highest power of the substrate electrode in the etching chamber of the inductively coupled plasma etching apparatus are described above. A wide process window for the plasma etch process is achieved simultaneously with the suppression of unplanned pocket formation. Thus sufficient pocket stability is achieved and some tolerance to overetching is achieved, especially when the aspect ratio of the etched structure is 5: 1 to 10: 1. However, pocket formation cannot be completely suppressed even in this process if the aspect ratio of the generated trench is still high or the overetch time is long.

As shown in DE 199 33 842 A1, the inductively coupled plasma source is also pulsed, so that negative ions generated by multiplying during plasma discharge discharge positive charge of the dielectric etch ground in structures with high aspect ratios. It helps to The big problem with this type of pulse of an ICP-plasma source (ICP = "inductively coupled plasma") is that high reflected power is generated in the assigned high frequency generator, because of the undefined conditions during which the plasma discharge is ignited in the plasma. This condition exists, making the high frequency power coupled during the transition period very difficult to match the plasma impedance. The ignition of the plasma discharge thus refers to the transition from capacitive coupling mode to inductive coupling mode, which leads to impedance mismatches and hence high reflected power.

In order to overcome this problem, in DE 199 27 806 A1, the frequency of the excitation voltage during the transient phase, via a feedback circuit according to the method of the Meissner oscillator which includes a plasma source as the frequency determining member and a high frequency generator as the amplifier of the feedback path, It is suggested that it be released. However, this method can occur outside the frequency range in which the frequency is released for industrial devices, which has the disadvantage of requiring a corresponding shield.

An unpublished application DE 100 51 831.1 has already proposed an apparatus and method for etching a substrate by inductively coupled plasma, wherein a magnetic field is provided between the substrate and the ICP-source that can be stopped or changed over time, The magnetic field is generated through magnetic coils which are flowed by current in opposite directions, arranged in at least two layers.

It is an object of the present invention to provide a method for etching a structure in an etch body with improved pocket stability, especially when the aspect ratio of the etched structure is high and the overetch time is long.

The present invention relates to a method for etching a structure into an etch body by means of a plasma according to the preamble of the independent claim, in particular for etching a silicon body with a precisely defined recess in the transverse direction.

1 is a basic diagram of a plasma etching apparatus for carrying out a method according to the invention,

2 shows a first embodiment of time-dependent modulation of plasma intensity synchronized to a high frequency pulsed and low frequency modulated high frequency power coupled to a substrate electrode,

FIG. 3 illustrates a structure of a high frequency pulse and a low frequency modulated high frequency power according to FIG.

4 shows a second embodiment of the modulation of plasma intensity and the synchronization of the modulation of the plasma intensity with the high frequency power coupled to the substrate electrode,

5 shows a third embodiment during a pulse pose clocked at low frequency;

6 shows a second embodiment during a pulse pose clocked at low frequency.

The method according to the invention has the advantage of significantly increased pocket stability compared to the prior art, for example when etching of silicon, in particular when a buried dielectric etch stop layer such as a SiO ₂ -layer is achieved and the tolerance is increased for overetching Has

Preferred refinements of the invention arise from the measures mentioned in the dependent claims.

It is therefore highly desirable that the intensity of the plasma is modulated or pulsed so that the plasma discharge in the "discharge pose" does not immediately disappear but remains in the inductively coupled mode, i.e. as high frequency power as needed to maintain the minimum discharge immediately during this time. Supplied to the plasma source or inductively coupled plasma. The plasma is not completely extinguished in the discharge pose or pulse pause, thereby preventing the occurrence of high reflected power each time the plasma continues to rise to maximum intensity, because the capacitive coupling of plasma discharge This is because the start phase is largely prevented and simultaneously started in the inductively coupled phase of the plasma discharge.

In this case the high frequency power coupled to the substrate electrode made according to the double pulse technique described in DE 199 57 169 A1 is correlated or synchronized with the modulation of plasma intensity over time.

Also in this connection during the discharge pose the plasma dominated by the positively charged ions and electrons is transferred to a so-called "bipolar" plasma consisting of positive and negatively charged ions, i.e. so called "after-glow". In the) -phase, it is preferred that free electrons be trapped by recombination with positively charged ions or by the capture of neutral particles. Due to the numerical overweight of the neutral particles surrounding the electrons, the generation of negative ions by electron capture is the dominant reaction in this case. The number of negative charge carriers having a mass corresponding to several times the quantum mass in the "normal" plasma is thus three to four times smaller than the number of positive charge carriers having a mass corresponding to several times the quantum mass, while in this phase The number of negative and positive charge carriers is about the same. In addition, since the amount of free electrons formed smaller than the ions of the plasma also eliminates the effects of unequal charge carrier mass and charge carrier mobility, the plasma potential of the positive value in the range of several ten volts previously had a value of zero volts. By approximating to, positive and negative charge carriers can reach the etching body to be processed, such as a silicon wafer, in the same way, which allows for optimal charge balance even when the aspect ratio is high.

Although the inductively coupled plasma cannot be maintained completely without electrons, in this case the smaller the density of the electrons, the more positive and negative charge carriers become identical and the more neutral the charging impedes. In this respect it is very desirable for the method according to the invention that the electron density is as small as possible or that the electron density can be kept small when the method according to the invention is carried out.

The modulation of the plasma intensity, implemented as a function of time, alternatively or additionally in addition to the high frequency power coupled to the plasma from the corresponding coil generator, which, in turn, is particularly preferably time-dependently alternatingly or in addition to a magnetic field, such as DE It can also be achieved by means of a periodically changing magnetic field strength of the magnetic field of the device according to the scheme of 100 51 831.1.

The invention is explained in more detail by the figures and the following description.

1 shows a plasma apparatus 5 known from DE 100 51 831.1, in which an anisotropic plasma etching process, for example, is carried out to produce a trench in silicon in the manner of DE 42 41 045 C1. In detail, for this purpose there is provided an etching chamber 10, a substrate electrode 18 on which a substrate 19 is placed, for example a silicon wafer. The substrate electrode 18 is also electrically connected to the second match box 21 and the substrate power generator 22 for impedance matching.

In the upper region of the etching chamber 10 is provided a coil 11 surrounding the etching chamber 10, which is connected to the coil generator 13 via a first match box 12 for impedance matching. The high frequency power is coupled to the etching chamber 10 by the coil 11 via the coil generator 13 and the first match box 12, whereby an inductively coupled plasma 15 is formed. Also according to FIG. 1, the etching chamber 10 has a gas inlet 14 in its upper region, and in its lower region a gas outlet 20 for supplying or discharging process gases, such as alternating etching and passivation gases. Has

Finally, the etching chamber 10 between the generating region of the inductively coupled plasma 15 and the substrate electrode 18 is surrounded by two magnetic field coils 16, where two corresponding spacers 17 are formed. Inserted into the side wall of 10), the spacer receives the coil 16.

See also the embodiment of DE 100 51 831.1 in connection with the detailed structure of the plasma etching apparatus 5.

In order to modulate the intensity of the plasma 15 as a function of time by means of the coil generator 13 and the first match box 12, a device known from DE 199 27 806 A1 or preferably DE 199 33 842 A1 is provided, The device is integrated into the first match box 12 or the coil generator 13, for example as described in the above document.

In addition, as described in DE 199 57 169 A1, the high frequency frequency pulsed at a high frequency and modulated at a low frequency by the substrate power generator 22, the second match box 21 and the substrate electrode 18 is applied to the substrate 19. Is coupled to.

Figure 3 shows the high frequency pulsed, low frequency modulated high frequency power, and at the so-called "duty cycle" of pulse packets 30 clocked at low frequency periodically and at 20% to 80%, preferably 50% A substrate having a pulse pause 31 clocked at a low frequency with a frequency of, for example, 1 Hz to 500 Hz, preferably 10 Hz to 250 Hz, such as 100 Hz, and preferably having an average power of 5 Watt to 20 Watt, such as 10 Watt Is coupled to the electrode 18. In this case, the low frequency clocked pulse packet 30 according to FIG. 3 is in the order of the alternating high frequency clocked pulse 32 and the high frequency clocked pulse pose 33, and the frequency of the period is 10, for example. kHz to 500 kHz, such as 100 kHz, and the "duty cycle" is preferably 2% to 20%, such as 5%. The average power coupled to the substrate electrode 18 in this case is, for example, 5 Watts to 40 Watts, in particular 20 Watts, at an average time during the high frequency clocked pulse 32.

Finally, in Figure 3 it can be seen that the pulse 32, clocked at an individual high frequency, consists of a high frequency carrier signal having a frequency of 13.56 MHz and preferably a high frequency power of 100 Watts to 1 kWatt, such as 400 Watts. See also application DE 199 57 169 A1 in connection with further detail to FIG. 3.

In the formation of the signal at the substrate electrode 18 according to FIG. 3, in particular, a pulse pose 31, which is caused by a sufficiently long, low frequency clocking, is maintained and discharges in the region of the dielectric boundary layer of the trench etched during this pulse pose. It can be noted that what can be done. While these long pulses reduce the process stability itself, thus causing a narrow process window, high frequency carriers with as low a pulse-to-pose-ratio ("duty cycle") as possible, such as a ratio of 1: 10 or 1: 20 Further high frequency modulation of the signal 34 results in a very high substrate electrode voltage when a low current flows into the substrate electrode 18 at the same time, which nevertheless sets a very wide, permissible process window. The “duty cycle” in this case controls the current voltage relationship and observes possible ohmic resistance of the plasma 15 from the substrate electrode 18.

Figure 2 shows a first embodiment of the method according to the invention, wherein the high frequency pulsed, low frequency modulated high frequency power of the substrate electrode 18 is immediately synchronized with the modulation of plasma intensity so that the minimum power (plasma is nearly Plasma excitation, i.e., the first plasma intensity minimum 41, falls with the pulse pose 31 clocked at a low frequency, respectively. This causes an amplified discharge of the generated trenches during the low frequency clocked pulse pose 31, not only because the self-discharge of the trenches occurs, but also the negative ions are attracted to the structural ground several times, and there are This is because the positive charge is neutralized.

The clock ratio of 1: 1, ie, the ratio of the duration of the first plasma intensity maximum 40 and the duration of the first plasma intensity minimum 41, shown in FIG. 2 is shown by way of example only. Rather, due to the plasma excitation effect, it is preferable that the plasma 15 is excited as long as possible and posed for as short a time as possible, i.e., if the excitation or plasma intensity is set to a duty cycle, it is desirable that a ratio of less than 1: 1 is preferably set. This prevents the required impulse peak power of the high frequency power coupled to the plasma 15 from becoming too large. For example, in order to achieve the desired time average, an impulse peak power of 6 kW to 10 kW is already needed for the average power of the coil 11 of 3 kW to 5 kW in a 1: 1 duty cycle.

Finally, in FIG. 2, the duration of the first plasma intensity maximum 40 and the subsequent first plasma intensity minimum 41 is the duration of the pulse packet 31 clocked at low frequency or the pulse pose 31 clocked at a subsequent low frequency. The same as time can be presented. Also, the intensity of plasma 15 during the first plasma intensity minimum 41 is chosen so small that the plasma 15 does not immediately disappear during the plasma intensity minimum.

FIG. 4 shows a second embodiment of the modulation of plasma intensity and the synchronization of high frequency pulsed and low frequency modulated high frequency of electrode substrate 18. In this case, in order to keep the clock ratio as high as possible, for example, as shown, two or more high frequency clocked pulses 32 are respectively applied to the substrate electrode 18 by a second plasma intensity maximum 40 '. Connected, the plasma 15 is switched to " Low " -mode during the subsequent high frequency clocked pulse pose 33, ie, the plasma intensity reaches a second plasma intensity maximum 41 ', The second plasma intensity maximum is chosen so small that plasma 15 does not immediately die at this time. In this regard, before the high frequency clocked pulse pose 33 drops again with the second plasma intensity minimum 41 ', always two or more high frequency clocked pulses 32 are second plasma intensity maximums. Descend to 40 '.

As an advantage of the embodiment according to FIG. 4 over the embodiment according to FIG. 2, in the case of FIG. 4, a small number of charges may accumulate in the generated trenches for a relatively long “one” time of relatively low frequency modulation of plasma intensity. Each time, for example, up to 20 relatively few high frequency clocked pulses 32 drop with the second plasma intensity maximum 40 ', before discharge occurs again during the subsequent second plasma intensity minimum 41', Only relatively few charges accumulate in the trench during this time.

In addition, the highest anion concentration is present in the plasma 15 only for a short time after the second plasma intensity maximum 40 'is destroyed and immediately after the destruction, i.e., the maximum discharge effect is determined by the second plasma intensity minimum ( 41 ') and only for a short time thereafter. Only a clearly reduced anion concentration can then be used in the plasma 15, which is advantageous in the use of the shortest plasma intensity minimum 41 ′ possible and in the corresponding high discharge effect. Also in the case of the embodiment according to FIG. 4, it is not advisable to proceed with a duty cycle much smaller than 1: 1 at the time of plasma excitation or high frequency power coupled to plasma 15, because otherwise the impulse required for plasma excitation is required. This is because the peak power is uneconomically high.

Specifically, the high frequency power coupled to the inductively coupled plasma 15 again lies between 3 kW and 5 kW, and the frequency of the high frequency clocked pulse 32 and the subsequent high frequency clocked pulse pose 33 is 5%. In the "duty cycle", for example 100 kHz, the average high frequency power coupled to the substrate electrode 18 is, for example, 20 Watts during the low frequency clocked pulse packet 30, and also in FIG. 4 is a separate, low frequency clock according to FIG. Only the modulation of the plasma intensity is shown during the given pulse packet 30. This low frequency clocked pulse packet 30 is followed by a low frequency clocked pulse pose 31 according to FIG. 2, and simultaneously during the low frequency clocked pulse pose the intensity of the plasma 15 is determined according to FIG. 2. It falls to one plasma intensity minimum 41 and is maintained there for the time of the pulse pose 31 clocked at low frequency. This is shown fully in FIG. 6 to illustrate precisely.

5 shows a third embodiment of the modulation of the intensity of plasma 15 and the temporal synchronization of high frequency power coupled to substrate electrode 18, pulsed at high frequency and modulated at low frequency. In this case, as a difference from FIG. 6, the modulation over time of the intensity of the plasma 15 according to FIG. 4 is maintained even during the pulse pause 31 clocked at a low frequency.

In this way, a relatively long (here "relatively long" meaning the decay duration of the anion concentration in plasma 15 after the destruction of the plasma or after the plasma intensity has transitioned to the second plasma intensity minimum 41 ') As the plasma 14 rises and falls again again during the low frequency clocked pulse pose 31, the plasma disruption phase is repeated by the elevated anion concentration associated therewith. In this respect, the low frequency clocked pulse pose 31 is used more effectively, where the plasma breakdown at the start of the pulse pose rises to an anion concentration (the anion concentration is then measured, the low frequency clocked pulse pose 31). Not only generated once, but again, this phase is always provided again for the discharge of the trenches generated in the etch body. In this respect, the low frequency clocked pulse pose 31 is no longer only used for self-discharge of the trench, but additionally the "anion concentration peak" which accelerates the discharge process during the low frequency pulsed pulse pose 31. Is provided. This measure also mitigates the problem of small "duty cycles" for plasma generation, because the plasma generation is now separated from the low frequency modulation of the high frequency power of the substrate electrode 18, which is now more than 1: 1 in a simple manner. This is because a "duty cycle" can also be achieved.

In order to set and stabilize the intensity of the plasma 15 to the intensity minimums 41 and 41 'near the extinction of the plasma 15, the power reflected by the coil generator 13 leaps when the plasma 15 is extinguished. Ascending, i.e., transitioning from inductive mode to capacitive coupling mode. The forward power of the coil generator 13 can be immediately adjusted to a high degree so that the state can be captured and maintained at the boundary of the inductively coupled operating mode so that the power of the coil generator 13 raised in this manner is increased. The electron density in the plasma 15 is offset again to the value of the stable operating state.

In this case, the forward power P _Forward of the coil generator 13 is coupled to the power P _Reflected from the plasma intensity minimums 41, 41 ′ to the coil generator 13 according to the following equation:

P _Forward = P _Soll + _{VP Reflected}

Where V is the amplification factor of the regulating circuit, where preferably V >> 1 is applied. V = 1 corresponds to the usual “Load” -regulation in current high frequency generators, ie the forward power is the difference between the forward power and the backward power when the reflected power is generated, ie actually the plasma 15. The high frequency power coupled to is adjusted to remain constant corresponding to the preset value. This conventional "Load" -regulation is considerably insufficient to stabilize the plasma 15 at the critical mode boundary, because in this case the plasma, effectively necessary for the capture of the destroyed plasma 15, This is because the high frequency power coupled to 15 does not rise. In this respect the adjustment factor V is here preferably set to a value much greater than 1, for example between 5 and 10, and also as a set value P _soll for the boundary mode of the plasma 15, ie It is set to approximately less than or nearly close to the value needed for the intensity slightly greater than the disappearance of the plasma 15.

The pulse strategy, i.e. the modulation of the high frequency power coupled to the substrate electrode 18 as a function of plasma intensity and time, is also carried out during the deposition cycle and during the etching cycle in the process according to DE 42 41 045 C1. Can be used. In general, however, it is desirable to be limited to an etching cycle, since there is a risk of pocket formation only during the etching cycle. Also, full generator power is used during the deposition cycle. It is also highly desirable that the coupling of high frequency power to the substrate electrode 19 is completely interrupted during the deposition cycle.

As described in application DE 100 51 831.1 and shown in FIG. 1, a very simple modulation of the intensity of plasma 15 is achieved by the use of an inductively coupled plasma source with a magnetic coil arrangement. In this case, at least two magnetic coils 16 are arranged between the ICP source, i.e., the inductively coupled plasma 15 and the substrate 19, the upper magnetic coil 16 is directed to the ICP source and the lower magnetic coil is the substrate ( 19), the magnetic field coils are perfused by currents of generally different magnitudes pointing in opposite directions to each other, thereby generating magnetic fields having generally different strengths pointing in opposite directions to each other.

In this case, in detail, the upper magnetic field coil 16 directed to the ICP source is set to the magnetic field strength necessary for optimal plasma generation, while the lower magnetic field coil 16 directed to the substrate 19 is directed in the opposite direction. To generate a set magnetic field, the strength of the magnetic field is set to be necessary for optimal uniformity of etching, ie to optimally divide the energy application across the surface of the substrate 19.

By using the magnetic field coil 16, first of all, in the case of a plasma boundary that does not immediately disappear, the plasma 15 with lower excitation density and electron concentration can be maintained than is possible without the magnetic field coil. This reduces the wall loss in the source region so that the generated magnetic field raises the "lifetime" of the electrons present in the plasma 15, and thus the desired "bipolar" with the minimum free electron density during the plasma intensity minimums 41 and 41 '. "It appears that the plasma can be kept very good.

In order to achieve modulation of the intensity of the plasma 15, in addition to or alternatively to the high frequency power connected to the plasma 15 via the coil 11 from the coil generator 13 in the plasma apparatus 5 according to FIG. 1. Alternatively, the strength of the magnetic field generated by the magnetic field coil 16 in the chamber 10 is also used. Thus for the setting of the first plasma intensity maximum 40 a coil current in the magnetic field coil 16 is first set, for example, which coil current corresponds to the current target value of the process according to DE 100 51 831.1, ie the upper magnetic field coil. 10 Ampere is set for (16) and 7 Ampere is set for the lower magnetic field coil 16 of opposite polarity.

The current is then reduced such that, for example, two currents in the magnetic field coil 16 return to zero or clock at the same time in order to switch to the plasma intensity minimums 41, 41 ′. Alternatively, however, an intermediate value can also be achieved, for example 3 Ampere for the upper magnetic field coil 16 and 2 Ampere for the lower magnetic field coil 16 are set.

In this regard, in the simplest embodiment, the two coil currents are each switched back and forth between high and low extreme values at the same time, resulting in the same effect as the effect caused by the return of the power of the coil generator 13, ie Plasma density breaks down when the magnetic field coil current decreases to reach the plasma intensity minimums 41 and 41 ', resulting in high anion densities consisting of recombination of electrons and neutral gas particles in a short time. In this case, however, it should be noted that the magnetic coil current cannot be modulated as quickly as the high frequency power coupled to the plasma 15. In particular, due to the inductance of the magnetic field coil 16, only the clock frequency may be less than 10 kHz. On the other hand, the change in rectification is much simpler and without problems than the change in the high frequency alternating voltage by the coil generator 13.

In addition, in order to prevent the reflected power, it is highly desirable that the change in the current in the magnetic field coil 16 is not instantaneous and have a final speed of change, which of course also applies to the change in the high frequency power of the coil generator 13, The impedance matching by the first match box 12 in the coil generator can generally be made easier by the longest possible power variation.

This final rate of change can be designed very simply by modulation of the magnetic coil current, since only constant voltage or rectification should overlap by the modulation voltage and the final edge slope.

Thus in the scope of this embodiment, for example, an alternating voltage or alternating current is used in the two magnetic field coils 16, which change corresponding to the following equation:

_{U 1 (t) = U 01} · sin (ωt) U 2 (t) = -U 02 · sin (ωt)

In this case the alternating voltage or current of the two magnetic field coils 16 is in phase out at each time point, where U ₀₁ and U ₀₂ are each the voltage amplitude or current amplitude at one of the two magnetic field coils 16.

It is also possible to operate by alternatively rectified alternating voltage or current, and the indication sin (ωt) can be replaced by the absolute value abs (sin (ωt)) respectively.

Finally, as mentioned above, it is highly desirable that the magnetic coil current does not return to zero. The coil voltage or coil current passing through the two magnetic field coils 16 follows, for example, the following equation:

_{U 1 (t) = U offset} , 1 + U 01 · abs (sin (ωt))

_{U 2 (t) = -U offset} , 2 - U 02 · abs (sin (ωt))

In this case, the offset-current or offset-voltage U _{offset, 1} or U _{offset, 2} , respectively, is still effectively suppressed in the so-called "Beaking-effect" generation in the edge region of the substrate 19, so as to cover the entire surface of the substrate. Sized so that still homogeneous etching results are achieved.

If the frequency ω according to the equation is also relatively small, i.e. from 10 Hz to 50 Hz, and the speed of the first match box 12 used in impedance matching is high enough to achieve modulation of this plasma intensity, this way As a result, the generation of the reflected power in the coil generator 13 can be completely prevented, and nevertheless undesirable pocket formation can be significantly suppressed. In this case only, it is important that the density of the plasma 15 is modulated and that the desired periodic phase of the anion concentration raised by this modulation is used, which discharges the trench with a high aspect ratio.

Claims (15)

At least temporarily pulsed at a high frequency to the etching body by a high frequency alternating voltage so that a low frequency modulated high frequency power is combined, A method for etching a structure into an etch body by means of a plasma, in particular in etching a recess defined precisely transversely by an etching mask into the silicon body, The intensity of the plasma (15) is modulated as a function of time. The method of claim 1, The intensity of the plasma 15 is modulated at least temporarily, in particular between a maximum value 40, 40 ′ maintained for a first duration and a minimum value 41, 41 ′ maintained for a second duration, and The minimum value (41, 41 ') is characterized in that the plasma (15) is chosen such that it does not disappear for a second duration. The method of claim 2, The minimum (41, 41 ') is selected so small that the plasma (15) is not immediately extinguished. The method of claim 1, Wherein the intensity of the plasma (15) is modulated to at least temporarily, in particular periodically, the intensity drop to zero for a second duration with a maximum (40, 40 ') for a first duration. The method according to any one of claims 1 to 4, The intensity of the plasma 15 is continuously modulated between a maximum value 40, 40 ′ maintained for a first duration and a minimum value 41, 41 ′ maintained for a second duration, in particular a rectangular pulse. And the minimum (41, 41 ') is selected such that the plasma does not disappear for a second duration. The method according to any one of claims 1 to 5, The intensity of the plasma 15 in the first time slot is modulated between a maximum value 40, 40 ′ that continues to be periodically maintained for the first duration and a minimum value 41, 41 ′ that is maintained for the second duration, in particular Rectangularly pulsed and modulated, the minimum values 41 and 41 'are chosen such that the plasma does not disappear for a second duration, In the second time slot the intensity of plasma 15 drops to a reduced intensity 41 relative to zero or a maximum of 40 or 40 'for a third duration, in particular to an intensity at which plasma 15 does not immediately disappear. A method characterized by descending. The method of claim 6, Said first and second time slots being repeated periodically alternating, in particular immediately successively. The method according to any one of claims 1 to 7, The high frequency pulsed and modulated plasma intensity, pulsed at high frequency, is given the maximum intensity 40 of the plasma 15 during the time that the low frequency clocked pulse packet 30 is coupled to the etching body 19. Characterized in that they are correlated with each other in a fashion. The method according to any one of claims 1 to 8, The high frequency pulsed, low frequency modulated high frequency power and the modulated plasma intensity are mutually correlated in such a way that the minimum intensity of plasma 15 is given during the time that the low frequency clocked pulse pose 31 is applied to the etching body 19. Correlated. The method according to any one of claims 1 to 9, The low frequency clocked pulse packet 30 includes a number of alternating, high frequency clocked pulses 32 and pulse poses 33, the duration of pulse 32 and the duration of pulse pose 33. Defines a first pulse-to-pose-ratio, The first time duration where the intensity of plasma 15 has a maximum value 40 'and the second time duration where the intensity of plasma 15 has a minimum value 41' define a second pulse-to-pose-ratio and , And the second pulse-to-pose-ratio is greater than the first pulse-to-pose-ratio. The method according to any one of claims 1 to 10, The frequency at which the high frequency clocked pulse 32 and the high frequency clocked pulse pose 33 are repeated is the maximum 40, 40 'and minimum 41, 41' of the intensity of the plasma 15 repeated. Characterized in that greater than the frequency. The method according to any one of claims 1 to 11, And wherein the frequency at which the low frequency clocked pulse packet (30) and the low frequency clocked pulse pose (31) are repeated is approximately equal to the frequency at which the first and second time slots are repeated. The method according to any one of claims 1 to 12, And wherein the frequency at which the low frequency clocked pulse packet (30) and the low frequency clocked pulse pose (31) are repeated is approximately equal to or less than the frequency at which the first and second time slots are repeated. The method according to any one of claims 1 to 13, The modulation of the intensity of the plasma 15 is characterized in particular through the high frequency power coupled to the plasma 15, and / or in particular the intensity of the magnetic field acting on the plasma, which changes periodically over time. How to. The method of claim 1, The etching body (19) is connected to the substrate electrode (18), and a high frequency power is applied to the substrate electrode.