JP5100853B2 - Plasma processing method - Google Patents

Description

  The present invention relates to a plasma processing method for performing plasma processing on a substrate to be processed, and more particularly to improvement of a system for supplying a required high frequency to a high frequency electrode for plasma generation via a variable matching device.

  In a manufacturing process of a semiconductor device or an FPD (Flat Panel Display), a plasma processing apparatus that performs processing such as etching, deposition, oxidation, and sputtering using plasma is often used. Generally, in a plasma processing apparatus, a high-frequency electrode is disposed inside or outside a processing container or chamber, and a high-frequency power is supplied to the high-frequency electrode from a high-frequency power supply unit. The high-frequency power supply unit includes not only an oscillator or high-frequency power source that outputs high frequency, but also a matching unit for matching between the impedance on the load side (electrode, plasma, chamber) and the impedance on the high-frequency power source side. Used.

In general, this type of matching device includes one or a plurality of variable reactance elements such as a variable capacitor or a variable inductance coil, and each step position or position within a variable range is selected by a step motor or the like, so that and it is configured to impedance thus load impedance Z _in as a variable matching circuit which can variably adjusted. Then, during plasma processing, the plasma impedance changes depending on the pressure variation, matching the impedance positions thereof variable reactance element variable adjusted automatically load impedance Z _in and correcting the alignment points (50 [Omega) It is like that. Therefore performing automatic matching circuit and to measure the load impedance Z _in, controller or the like for variably controlling the impedance positions of the variable reactance element through step motor to match the measured value of the load impedance matching point (50 [Omega) Is used.

JP-A-11-61456

As described above, when auto-matching is performed _in the variable matching device, the impedance position of the variable reactance element is variably controlled by feedback control so that the measured value of the load impedance Zin matches the matching point. Here, the matching point Z _MP is usually set to a pure resistance value 50Ω (Z _MP = 50 + j0) equal to the output impedance of the high frequency power supply.

However, if the measurement accuracy of the load impedance measuring circuit there is an error, even it shows _{50Ω (Z in = 50 + j0} ) measurements the apparent net resistance value of the obtained load impedance measuring circuit load impedance Z _in, Actually, it may not coincide with the matching point (Z _MP = 50 + j0). In addition, even when the waveform distortion of the high frequency is large, the measurement accuracy of the load impedance measurement circuit becomes unstable, and even if the apparent measurement value of Z _in = 50 + j0 is obtained, it matches the matching point (Z _MP = 50 + j0). There may not be. In such a case, if the feedback control for impedance matching as described above is performed, the result is that the load impedance is matched to the impedance deviated from the matching point Z _MP , and the RF transmission characteristic is deteriorated to affect the plasma process. There is a fear.

  The present invention solves the problems of the prior art as described above. Even if there is an error in the measurement accuracy of the load impedance or the high frequency waveform distortion during the actual process, Provided is a plasma processing method which compensates for a decrease in load impedance measurement accuracy and improves process stability and reproducibility.

  In the plasma processing method according to the first aspect of the present invention, a dummy substrate to be processed is disposed at a predetermined position in a depressurizable chamber, and a predetermined degree of vacuum is set in the chamber under substantially the same conditions as a desired process recipe. The plasma is generated in the chamber by supplying a predetermined processing gas into the chamber and supplying a predetermined high frequency with a predetermined power from a high frequency power source to a predetermined high frequency electrode via a variable matching device. A first step of variably controlling the impedance of the variable matching unit so that the measured value of the load impedance matches a plurality of preselected reference impedances, and obtaining the high frequency power source side under each of the reference impedances A second step of measuring the power of the reflected wave and recording the measured value, and the measured value of the reflected wave power is the minimum value among the plurality of reference impedances Or a third step of registering a reference impedance when the value is close to the minimum value, and a regular substrate to be processed is disposed at the predetermined position in the chamber, and a predetermined vacuum is formed in the chamber under the conditions of the process recipe. The pressure is reduced each time, a predetermined processing gas is supplied into the chamber, and the high-frequency power is supplied from the high-frequency power source to the high-frequency electrode through the variable matching unit with a predetermined power to generate plasma in the chamber. And a fourth step of performing plasma processing on the substrate by variably controlling the impedance of the variable matching unit so that the measured value of the load impedance matches the registered reference impedance.

  In the plasma processing method according to the second aspect of the present invention, a dummy substrate is disposed on a lower electrode in a depressurizable chamber, and the chamber is subjected to a predetermined process under substantially the same conditions as a desired process recipe. The pressure is reduced to a vacuum level, a predetermined processing gas is supplied into the chamber, the processing gas is discharged in the chamber with a first high frequency to generate plasma, and a high frequency of a second frequency is generated from a high frequency power source. The variable electrode is supplied to the lower electrode with a predetermined power through a variable matching unit, and ions in the plasma are drawn into the substrate, and the variable matching is performed so that the measured value of the load impedance matches a plurality of preselected reference impedances. A first step of variably controlling the impedance of the device, and measuring the power of the reflected wave obtained on the high-frequency power source side under each of the reference impedances A second step of recording the measured value, a third step of registering a reference impedance when the reflected wave power measurement value is at or near the minimum value among the plurality of reference impedances, A regular substrate to be processed is disposed at the predetermined position in the chamber, the chamber is depressurized to a predetermined degree of vacuum under the conditions of the process recipe, a predetermined processing gas is supplied into the chamber, The processing gas is discharged with the first high frequency to generate plasma, and the second high frequency is supplied from the high frequency power source to the lower electrode with a predetermined power via the variable matching unit. The impedance of the variable matching unit can be adjusted so that the measured value of the load impedance matches the registered reference impedance. Control to, and a fourth step of performing plasma processing on the substrate.

  In the plasma processing method of the present invention, a dummy substrate is used and plasma is generated under substantially the same conditions as the actual process, or ions in the plasma are drawn into the substrate, and the power of the reflected wave at that time is generated. Correct the matching point of the matcher in the direction of zero or minimum. Since the correction of the matching point by auto-learning is performed by changing the process condition for each process recipe, a corrected matching point or reference impedance is obtained for each process recipe.

  According to a preferred aspect of the present invention, in the first step, every time an operation for matching the measured value of the load impedance with each reference impedance is performed, a high frequency from a high frequency power supply is supplied to the high frequency electrode again. Generate plasma.

  According to another preferred aspect, in the second step, the measured value of the reflected wave power is taken in at the time when matching is established by the timer function.

  According to the plasma processing method of the present invention, due to the configuration and operation as described above, in the variable matching device, the load impedance measurement accuracy is reduced due to measurement errors of the load impedance measurement circuit, high-frequency waveform distortion during the actual process, and the like. Compensation can improve process stability and reproducibility.

It is a longitudinal cross-sectional view which shows the structure of the plasma etching apparatus which can apply the plasma processing method in one Embodiment of this invention. It is a partial expanded sectional view which shows the structure of the high frequency electric power feeding part in the plasma etching apparatus of FIG. It is a block diagram which shows the structure of the high frequency electric power feeding system in a 1st technique. It is a flowchart figure which shows the procedure of the OFF preset value optimization auto-learning of the matching device impedance position by a 1st technique. It is a figure which shows an example of the procedure which moves the impedance position of a variable capacitor sequentially in the auto learning of the 1st technique. It is a figure which shows the mapping of the _Vpp measurement value obtained in each impedance position of a variable capacitor in the auto learning of the 1st technique. It is a figure which shows an example of _Vpp measured value distribution obtained by the auto learning of the 1st technique with a characteristic curve. It is a flowchart figure which shows the main procedures of the plasma processing in a 1st technique. It is a block diagram which shows the structure of the high frequency electric power feeding system in embodiment of this invention. It is a flowchart figure which shows the procedure of the auto learning for a matching point correction | amendment function in embodiment. It is a figure which shows an example of the procedure which moves the point (selection value) of a reference impedance sequentially in the auto learning of embodiment. It is a figure which shows an example of the measurement result obtained at each point of reference | standard impedance in the auto learning of embodiment by mapping. It is a block diagram which shows the structural example of the main control part in embodiment.

  Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings.

  FIG. 1 shows a configuration of a plasma etching apparatus to which a plasma processing method according to an embodiment of the present invention can be applied. This plasma etching apparatus is configured as a capacitively coupled parallel plate plasma etching apparatus, and has, for example, a cylindrical chamber (processing vessel) 10 made of aluminum whose surface is anodized (anodized). The chamber 10 is grounded for safety.

  A cylindrical susceptor support 14 is disposed at the bottom of the chamber 10 via an insulating plate 12 such as ceramic, and a susceptor 16 made of, for example, aluminum is provided on the susceptor support 14. The susceptor 16 constitutes a lower electrode, on which, for example, a semiconductor wafer W is placed as a substrate to be processed.

  An electrostatic chuck 18 is provided on the upper surface of the susceptor 16 to hold the semiconductor wafer W with an electrostatic attraction force. The electrostatic chuck 18 is obtained by sandwiching an electrode 20 made of a conductive film between a pair of insulating layers or insulating sheets, and a DC power source 22 is electrically connected to the electrode 20. The semiconductor wafer W is attracted and held on the electrostatic chuck 18 by a Coulomb force by a DC voltage from the DC power supply 22. A focus ring 24 made of, for example, silicon is disposed on the upper surface of the susceptor 16 around the electrostatic chuck 18 to improve etching uniformity. A cylindrical inner wall member 26 made of, for example, quartz is attached to the side surfaces of the susceptor 16 and the susceptor support base 14.

  Inside the susceptor support 14, for example, a refrigerant chamber 28 extending in the circumferential direction is provided. A refrigerant having a predetermined temperature, such as cooling water, is circulated and supplied to the refrigerant chamber 28 via pipes 30a and 30b from an external chiller unit (not shown). The processing temperature of the semiconductor wafer W on the susceptor 16 can be controlled by the temperature of the coolant. Further, a heat transfer gas such as He gas from a heat transfer gas supply mechanism (not shown) is supplied between the upper surface of the electrostatic chuck 18 and the back surface of the semiconductor wafer W via the gas supply line 32.

  Above the susceptor 16 is provided an upper electrode 34 facing the susceptor in parallel. The space between both electrodes 16 and 34 is a plasma generation space. The upper electrode 34 faces the semiconductor wafer W on the susceptor (lower electrode) 16 and forms a surface in contact with the plasma generation space, that is, an opposing surface. The upper electrode 34 is arranged in a ring-shaped or donut-shaped outer upper electrode 36 that is disposed to face the susceptor 16 at a desired interval, and is insulated from the outer upper electrode 36 in the radial direction. And a disc-shaped inner upper electrode 38. The outer upper electrode 36 and the inner upper electrode 38 have the former (36) as the main and the latter (38) as the auxiliary relationship for plasma generation.

FIG. 2 shows the configuration of the upper high-frequency power feeding unit in this plasma etching apparatus. As clearly shown in FIG. 2, an annular gap (gap) of, for example, 0.25 to 2.0 mm is formed between the outer upper electrode 36 and the inner upper electrode 38, and a dielectric 40 made of, for example, quartz is formed in this gap. Is provided. In addition, a ceramic 96 can be provided in the gap. A capacitor is formed between the electrodes 36 and 38 with the dielectric 40 interposed therebetween. The capacitance C40 of the capacitor is selected or adjusted to a desired value according to the size of the gap and the dielectric constant of the dielectric 40. A ring-shaped insulating shielding member 42 made of, for example, alumina (Al ₂ O ₃ ) is airtightly attached between the outer upper electrode 36 and the side wall of the chamber 10.

  The outer upper electrode 36 is preferably made of a low-resistance conductor or semiconductor having a low Joule heat, such as silicon. An upper high-frequency power source 52 is electrically connected to the outer upper electrode 36 via an upper matching unit 44, an upper power feed rod 46, a connector 48, and a power feed tube 50. The upper high frequency power supply 52 outputs a frequency of 13.5 MHz or higher, for example, a high frequency of 60 MHz for plasma generation. The upper matching unit 44 is for matching the load impedance with the internal (or output) impedance of the upper high-frequency power source 52, and the output impedance and load impedance of the upper high-frequency power source 50 when the plasma is generated in the chamber 10. It works to match in appearance. The output terminal of the upper matching unit 44 is connected to the upper end of the upper feeding rod 46.

  The power supply tube 50 is made of a conductive plate having a cylindrical shape, a conical shape, or a shape close thereto, such as an aluminum plate or a copper plate, and a lower end is continuously connected to the outer upper electrode 36 in a circumferential direction, and an upper end is connected to the upper power supply rod by a connector 48 The lower end of 46 is electrically connected. Outside the power supply tube 50, the side wall of the chamber 10 extends upward from the height position of the upper electrode 34 to constitute a cylindrical ground conductor 10a. The upper end portion of the cylindrical ground conductor 10 a is electrically insulated from the upper power supply rod 46 by a cylindrical insulating member 54. In such a configuration, in the load circuit viewed from the connector 48, a coaxial line having the former (36, 50) as a waveguide is formed by the feeding tube 50, the outer upper electrode 36, and the cylindrical ground conductor 10a.

  In FIG. 1 again, the inner upper electrode 38 has an electrode plate 56 made of a semiconductor material such as Si or SiC having a large number of gas vent holes 56a, and a conductive material that detachably supports the electrode plate 56 such as alumite. And an electrode support 58 made of treated aluminum. Inside the electrode support 58, there are provided two gas introduction chambers, that is, a central gas introduction chamber 62 and a peripheral gas introduction chamber 64 which are divided by an annular partition member 60 made of, for example, an O-ring. A central shower head is constituted by the central gas introduction chamber 62 and a large number of gas ejection holes 56a provided on the lower surface thereof, and the peripheral gas introduction chamber 64 and a large number of gas ejection holes 56a provided on the lower surface thereof constitute a peripheral area. Shower head is configured.

A processing gas from a common processing gas supply source 66 is supplied to these gas introduction chambers 62 and 64 at a desired flow rate ratio. More specifically, the gas supply pipe 68 from the processing gas supply source 66 branches into two in the middle and is connected to the gas introduction chambers 62 and 64. The flow control valves 70a and 70b are respectively connected to the branch pipes 68a and 68b. Is provided. Since the conductances of the flow paths from the processing gas supply source 66 to the gas introduction chambers 62 and 64 are equal, the flow rate ratio of the processing gas supplied to the gas introduction chambers 62 and 64 is arbitrarily adjusted by adjusting the flow rate control valves 70a and 70b. It can be adjusted. The gas supply pipe 68 is provided with a mass flow controller (MFC) 72 and an opening / closing valve 74. In this way, by adjusting the flow rate ratio of the processing gas introduced into the center gas introduction chamber 62 and the peripheral gas introduction chamber 64, the gas vent hole 56a in the center of the electrode corresponding to the center gas introduction chamber 62, that is, the center shower head. the ratio between the flow rate F _E of the gas ejected from the gas-passing holes 56a, i.e. near the shower head electrode peripheral portion corresponding to the flow rate F _C and the peripheral gas introduction chamber 64 of the gas to be more ejected (F _C / F _E) It can be adjusted arbitrarily. It is also possible to vary the flow rate per unit area of the processing gas ejected from the central shower head and the peripheral shower head. Furthermore, it is also possible to select the gas type or gas mixture ratio of the processing gas ejected from the central shower head and the peripheral shower head independently or separately.

  The upper high-frequency power source 52 is electrically connected to the electrode support 58 of the inner upper electrode 38 via the upper matching unit 44, the upper power feed rod 46, the connector 48, and the lower power feed cylinder 76. A variable capacitor 78 capable of variably adjusting the capacitance is provided in the middle of the lower feeding cylinder 76.

  Although not shown, the outer upper electrode 36 and the inner upper electrode 38 are also provided with appropriate refrigerant chambers or cooling jackets (not shown) so that the temperature of the electrodes can be controlled via the refrigerant by an external chiller unit. May be.

  An exhaust port 80 is provided at the bottom of the chamber 10, and an exhaust device 84 is connected to the exhaust port 80 via an exhaust pipe 82. The exhaust device 84 has a vacuum pump such as a turbo molecular pump, and can depressurize the plasma processing space in the chamber 10 to a desired degree of vacuum. A gate valve 86 that opens and closes the loading / unloading port for the semiconductor wafer W is attached to the side wall of the chamber 10.

  In this plasma etching apparatus, a lower high frequency power supply 90 is electrically connected to a susceptor 16 serving as a lower electrode via a lower matching unit 88. The lower high frequency power supply 90 outputs a frequency within a range of 2 to 27 MHz, for example, a high frequency of 2 MHz for ion attraction. The lower matching unit 88 is used to match the load impedance to the inside or output impedance of the lower high-frequency power source 90. When the plasma is generated in the chamber 10, the output impedance and load impedance of the lower high-frequency power source 90 apparently appear. Works to match.

  The inner upper electrode 38 is electrically connected with a low pass filter (LPF) 92 for passing the high frequency (2 MHz) from the lower high frequency power supply 90 to the ground without passing the high frequency (60 MHz) from the upper high frequency power supply 52. Yes. The low-pass filter (LPF) 92 may be preferably composed of an LR filter or an LC filter, but it provides a sufficiently large reactance with respect to the high frequency (60 MHz) from the upper high frequency power supply 52 even with only one conductor. You can do that. On the other hand, the susceptor 16 is electrically connected to a high pass filter (HPF) 94 for passing a high frequency (60 MHz) from the upper high frequency power supply 52 to the ground.

  The main control unit 100 is composed of a computer system including a CPU, a memory, etc., and individual operations and overall operations of each unit in the apparatus, particularly the high-frequency power sources 52 and 90, the processing gas supply source 66, the matching units 44 and 88, and the like ( Sequence). The configuration within the main controller 100 will be described later with reference to FIG.

  In order to perform etching in this plasma etching apparatus, first, the gate valve 86 is opened, and the semiconductor wafer W to be processed is loaded into the chamber 10 and placed on the susceptor 16. Then, a DC voltage is applied from the DC power source 22 to the electrode 20 of the electrostatic chuck 18 to fix the semiconductor wafer W to the susceptor 16. Next, the inside of the chamber 10 is exhausted by the exhaust device 84, and an etching gas (generally a mixed gas) is introduced into the gas introduction chambers 62 and 64 from the processing gas supply source 66 at a predetermined flow rate and flow rate ratio. The pressure in 10, that is, the etching pressure is set to a set value (for example, several mTorr). Next, a high frequency (2 MHz) is applied from the lower high frequency power supply 90 to the susceptor 16 with a predetermined power, and then a high frequency (60 MHz) is also applied from the upper high frequency power supply 52 to the upper electrode 34 (36, 38) with a predetermined power. . The etching gas discharged from the gas vent hole 56a of the inner upper electrode 38 is turned into plasma in a glow discharge between the upper electrode 34 (36, 38) and the susceptor 16, and radicals or ions generated from the plasma cause the semiconductor wafer W to be plasma. The surface to be processed is etched.

  As described above, between the high frequency (60 MHz) output from the upper high-frequency power supply 52 and the high frequency (2 MHz) output from the lower high-frequency power supply 90, the high frequency for ion attraction (2 MHz) is inserted in the lower electrode. 16, the upper high frequency (60 MHz) for plasma generation is applied to the upper electrode 34 later, and there is a relationship on the power supply start time.

  In this plasma etching apparatus, by applying a high frequency in a high frequency region (5 to 10 MHz or more at which ions cannot move) to the upper electrode 34, the plasma is densified in a preferable dissociation state, and high even under lower pressure conditions. A density plasma can be formed. In addition, the impedance position off-preset value optimization function in the upper matching unit 44 on the first-in side, which will be described in detail later, allows various conditions of the process recipe (pressure in the chamber, upper and lower high-frequency power, gas flow rate, etc.). The plasma can be reliably ignited even under low pressure (or even if deposition is deposited in the chamber) without requiring special selection or switching. Further, in auto-matching after plasma ignition, even if there is a measurement error in load impedance monitoring, a matching state can be established with high accuracy by a matching point correction function described later.

  The above-described plasma etching apparatus is an example of an apparatus for etching a silicon oxide (SiO 2) film, and a lower high frequency (2 MHz) for ion attraction is applied to the lower electrode 16 in a first-in manner, and a plasma generating apparatus is used. In this order, the upper high frequency (60 MHz) was applied to the upper electrode 34 later. However, for example, in a plasma etching apparatus for etching a polysilicon film, a high frequency (for example, 60 MHz) is first applied to the upper electrode depending on circumstances or conditions, and then a high frequency (for example, 13.56 MHz) is applied to the lower electrode. There are many cases of order. Incidentally, the plasma processing apparatus for etching the polysilicon film is desired to have a low pressure condition in the chamber.

  In the following, for the sake of convenience, a description will be given of a case of high-frequency insertion into the upper electrode and high-frequency post-insertion into the lower electrode, which are frequently used in a plasma processing apparatus for etching a polysilicon film.

FIG. 3 shows the configuration of a high-frequency power feeding system according to the first technique applicable to this plasma etching apparatus. The upper matching unit 44 is a variable matching unit with variable impedance, and includes a matching circuit 102 including at least one variable reactance element, and a controller for variably controlling the impedance position of each variable reactance element of the matching circuit 102. 104 and an RF sensor 106 having a function of measuring the load impedance Z _in including the matching circuit 102.

In the illustrated example, the matching circuit 102 is configured as a T-shaped circuit including two variable capacitors C ₁ and C ₂ and one inductance coil L _1, and the controller 104 passes through the variable capacitors (C ₁ , The impedance position of C ₂ ) can be variably controlled. The RF sensor 106 has, for example, a voltage sensor and a current sensor for detecting an RF voltage and an RF current on the transmission line at the installation position, and displays a measured value of the load impedance Z _in from the measured voltage value and the measured current value in a complex number. Ask for. Controller 104, for example, a microcomputer, receives the measurement value of the load impedance Z _in from the RF sensor 106, receives various setting values and commands from the main control unit 100.

Further, according to the first technique, the V _pp measuring circuit 112 for measuring the peak value (peak-to-peak value) V _pp of the high-frequency voltage on the transmission line or the waveguide on the output side of the upper matching unit 44 is also the upper matching. It is provided in the vessel 44. The measured value of V _pp obtained by the V _pp measuring circuit 112 is taken into the main controller 100.

  FIG. 4 shows an auto-learning procedure (particularly a processing procedure of the main control unit 100) performed before the operation of the apparatus for optimizing the off-preset value of the impedance position in the upper matching unit 44 and the lower matching unit 88.

Hereinafter, the auto-learning procedure for the upper matching unit 44 will be described. First, in initialization (step A ₁ ), various setting values for auto-learning are set in a predetermined register, and a mechanical movable part is set at an initial position. Next, the exhaust device 84 is operated to bring the inside of the chamber 10 into a vacuum state of a predetermined pressure (Step A ₂ ), and then the upper high frequency power supply 52 is turned on (Step A ₃ ). Then, the high frequency output from the upper high frequency power source 52 onto the transmission line is supplied to the upper electrode 34 in the chamber 10 through the upper matching unit 44. Here, the output of the upper high-frequency power source 52 is set to a low power that does not cause discharge between the upper electrode 34 and the lower electrode 16. The lower high frequency power supply 90 is left off.

In the upper matching unit 44, the impedance position of both variable capacitors (C ₁ , C ₂ ) is set to the initial position by the previous initialization, and the V _pp measuring circuit 112 is set to (C ₁ , C 2 ₂ ) Measure the peak value V _pp of the high-frequency voltage obtained on the output-side waveguide under the impedance position. Here, the peak value V _pp of the high-frequency voltage on the waveguide measured by the V _pp measuring circuit 112 is proportional to the peak value of the high-frequency voltage at the upper electrode 34. V _pp measurement value obtained by the V _pp measuring circuit 112 is taken into the main control unit 100 via the controller 104, it is stored in the memory of the main control unit 100 (Step A _4).

Thus, when the measurement of V _pp is completed under the initial (C ₁ , C ₂ ) impedance position, the main controller 100 then passes the variable capacitors (C ₁ , C ₁ , C) via the controller 104 and step motors 110 and 112. The impedance position of C ₂ ) is moved by one step (step A ₅ ), and the V _pp measurement value obtained from the V _pp measurement circuit 112 under the destination impedance position is captured (step A ₆ → A ₇ → A ₄ ). After that, a series of processes of moving the impedance position of the variable capacitors (C ₁ , C ₂ ) to the next position and capturing the V _pp measurement value obtained from the V _pp measurement circuit 112 at the destination impedance position. Are repeated (steps A ₅ → A ₆ → A ₇ → A ₄ → A ₅ ...).

FIG. 5 shows an example of a procedure for sequentially moving the impedance positions of the variable capacitors (C ₁ , C ₂ ). In this example, the variable movement position of the capacitor C ₁ is pre-variable range (0% to 100%) m (100%) 0 (0%) in the up (m + 1) pieces selected, the movement position of the variable capacitor C ₂ in advance N + 1 items are selected from 0 (0%) to n (100%) within a variable range (0% to 100%).

In the procedure shown in FIG. 5A, the position of the variable capacitor C ₂ is _first swung from 0 to n while keeping the position of the variable capacitor C ₁ at 0, and (C ₁ , C ₂ ) = (0 , 0), (0, 1),... Measure V _pp under each impedance position (0, n) and record the measured value. Then, the position of the variable capacitor C ₁ is transferred to 1, the position of the variable capacitor C ₂ waving from 0 to _{n, (C 1, C 2} ) = (1,0), (1,1), ·· Measure V _pp under each (1, n) impedance position and record the measured value. Thereafter, the position of the variable capacitor C ₁ is sequentially shifted to 2, 3,... And V _pp measurement and recording (logging) similar to the above are repeated, and finally (C ₁ , C ₂ ) = (m, 0), (M, 1), ... _Vpp is measured under each impedance position of (m, n) and the measured value is recorded. In addition, the pitch at which the positions of the variable capacitors C ₁ and C ₂ are shaken is not limited to 1, and may be 2 or more.

In the procedure shown in FIG. 5B, when the position of the variable capacitor C ₁ is an even number (0, ₂ , 4,...), The positions of the variable capacitor C ₂ are arranged in ascending order (0, 1, 2, 3,. When the position of the variable capacitor C ₁ is an odd number (1, 3, 5,...), The position of the variable capacitor C ₂ is descended (n, n−1, n−2,. 5 is the same as FIG. 5A. In FIG. 5B, it is not necessary to rewind the step motor 108 on the variable capacitor C ₂ side every time the position of the variable capacitor C ₁ is moved by one step, and the efficiency is higher.

As a result of the above-mentioned logging, as shown by the mapping in FIG. 6, the memory within the main control unit 100 is selected within the variable range of the variable capacitors (C ₁ , C ₂ ) (as a whole). All V _pp measurements corresponding to all impedance positions covered are available.

FIG. 7 shows an example of the V _pp measurement value distribution with a characteristic curve. This characteristic curve is obtained by varying the variable capacitors C ₁ and C _{2 in} the range of 0 to 1500 steps, in which the position of C ₁ is swung at 100 pitches and the position of C ₂ is swung at 75 pitches. It is. In this example, it can be seen that the measured V _pp value (about 640 volts) obtained at the position of (C ₁ , C ₂ ) = (1500, 1050) is the maximum value or a value in the vicinity thereof.

The main control unit 100 turns off the upper high-frequency power supply 52 after the end of logging (step A ₈ ), and determines the maximum one of all the V _pp measurement values stored in the memory by the maximum value determination process (comparison calculation). (Step A ₉ ), the impedance position of the variable capacitor (C ₁ , C ₂ ) corresponding to the maximum value of V _{pp is} set to the position at the start of the upper RF input for plasma processing, that is, the optimum off-preset for plasma ignition · to register as a position (step a _10).

  The lower matching unit 88 is also a variable matching unit having the same configuration or function as the upper matching unit 44, and the auto-learning procedure for optimizing the off-preset value of the impedance position in the lower matching unit 88 is also basically. This is the same as the auto-learning procedure related to the upper matching unit 44 described above. However, in the auto-learning related to the upper matching unit 44 on the first-in side, the off-preset value of the impedance position optimum for the plasma ignition is determined (identified), whereas the auto-learning related to the lower matching unit 88 on the second-in side is determined. This is not necessary for learning. Rather, it is desirable to determine (identify) the off-preset value of the impedance position so that the already ignited plasma does not fluctuate as much as possible when power supply to the lower electrode 16 is started.

In view of the above, in the auto-learning of the lower matcher 88, after obtaining the V _pp logging information in the same procedure as the auto-learning of the upper matcher 44 (Step A ₁ to A _8), all V _The smallest _pp measurement value is determined by the minimum value determination process (comparison operation) (step A ₉ ), and the impedance position of the lower matching unit 88 corresponding to this V _pp minimum value is determined at the start of lower RF input. The position is registered as the optimum off-preset position that has the least influence on the plasma after ignition (step A ₁₀ ).

FIG. 8 shows the main procedure of plasma processing (particularly, the processing procedure of the main control unit 100) in the first technique. In the initialization (step B ₁ ), setting values for various conditions used in the actual process (setting values for the process recipe) are set in a predetermined register, and the mechanical movable part is set at the initial position. Among them, the main control unit 100 sets the upper matching unit 44 and the lower matching unit 88 at the impedance position registered by the auto-learning. More specifically, the off-preset value of the upper matching unit 44 is set to the registered impedance position corresponding to the V _pp maximum value, and the off-preset value of the lower matching unit 88 is set to the registered impedance position corresponding to the V _pp minimum value. To match.

After initialization, the semiconductor wafer W to be processed is loaded into the chamber 10 and placed on the susceptor 16 (step B ₂ ). Then, a DC voltage is applied from the DC power source 22 to the electrode 20 of the electrostatic chuck 18 to fix the semiconductor wafer W to the susceptor 16.

Next, the inside of the chamber 10 is exhausted by the exhaust device 84 (step B ₃ ), the etching gas is introduced into the gas introduction chambers 62 and 64 from the processing gas supply source 66 at a predetermined flow rate and flow rate ratio, and the chamber 10 is exhausted by the exhaust device 84. Is set to a set value (for example, several mTorr) (step B ₄ ).

Next, under the initial state in which the upper matching unit 44 is set to the registered impedance position corresponding to the V _pp maximum value as described above, the upper high frequency power supply 52 is turned on to output a high frequency with a preset RF power. (Step B ₅ ). The high frequency output from the upper high frequency power supply 52 is supplied to the upper electrode 34 via the upper matching unit 44, and a high frequency electric field is formed between the upper electrode 34 and the lower electrode 16. Here, even if impedance matching is not established between the upper high-frequency power source 52 and the load, the V _pp (and thus the high-frequency electric field between the electrodes) of the upper electrode 34 is the maximum voltage wave possible under the upper RF power. Since it is at a high value (and consequently the maximum electric field strength) or a value in the vicinity thereof, the etching gas starts to discharge quickly, and gas plasma is generated.

Next, the main control unit 100 cancels the off-preset value with respect to the upper matching unit 44 after a predetermined time from the start of the application of the upper RF power using, for example, a timer function, and auto-matching in the upper matching unit 44 (Step B ₆ ). In this auto-matching, the controller 104 receives the measured value of the load impedance Z _in from the RF sensor 106, and adjusts the measured value to the reference impedance for matching or the matching point Z _MP through the variable motors through the pulse motors 108 and 110. The impedance position of (C ₁ , C ₂ ) is variably controlled by feedback control. Thus, the high frequency from the upper high frequency power supply 52 is fed to the upper electrode 34 with the maximum transmission efficiency and the set power, and high density plasma is generated. In order to control switching from the off-preset value to auto-matching, a sensor that optically confirms that the plasma has ignited can be used instead of the timer function.

Next, the main control unit 100 turns on the lower high frequency power supply 90 to output a high frequency with a preset lower RF power, for example, when the plasma is stabilized by the same timer function as described above (step B ₇ ). When the lower RF is started, the lower matching unit 88 is set to the registered impedance position corresponding to the V _pp minimum value as described above. For this reason, the V _pp of the lower electrode 16 corresponding to the lower high frequency becomes the minimum voltage peak value possible in the vicinity of the lower RF power or a value in the vicinity thereof, and the influence on the plasma generated or generated is minimal. Can be suppressed. Then, for example, the timer function similar to the above is used to release the off-preset value for the lower matching unit 88 after a predetermined time, and switch to auto matching in the lower matching unit 88 (step B ₆ ). Although not shown in the auto-matching of the lower matching unit 88, the impedance position of the variable reactance element is passed through the pulse motor so that the controller inside the matching unit matches the load impedance measurement value obtained from the RF sensor with the matching point. Is variably controlled by feedback control. Thus, the high frequency from the lower high frequency power supply 90 is supplied to the lower electrode 16 with the maximum transmission efficiency and the set power, and plasma etching is performed (step B ₉ ).

When the desired plasma etching process is completed, the upper and lower high-frequency power sources 52 and 90 are turned off (step B ₁₀ ), and the supply of the etching gas to the processing gas supply source 66 is stopped (step B ₁₁ ). Thereafter, the processed semiconductor wafer W is unloaded from the chamber 10 (step B ₁₂ ).

  In the first technique described above, the off-preset value or initial value of the impedance position in the upper matching unit 44 and the lower matching unit 88 is not performed after the start of the process without switching process conditions (pressure, RF power, etc.). By setting the value to a specific value (optimum value), the plasma can be reliably ignited and maintained stably even under a low pressure, and fine etching using high-density plasma can be performed.

In the above first technique, the main control unit 100, V _pp measurement of capture and storage, V _pp maximum value of the measured value (minimum value) determining process, changes to the registration impedance positions off preset value, The off / preset value is canceled by the timer function. However, the present invention is not limited to this. For example, the timer function and the release of the off-preset value by the timer function may be performed by the controller 104 in the matching unit 44, and in fact, this is often preferable. Also, V _pp measurement of capture and storage, V _pp maximum value of the measured value (minimum value) determining process, may be performed by controller 104 changes or the like to the registration impedance positions off-preset value.

FIG. 9 shows a configuration of a high-frequency power feeding system according to an embodiment of the present invention. In this embodiment, there are two different configurations from the first technique described above. One is that the circuit configuration of the matching circuit 102 is changed from a T shape to a ladder shape, and the controller 104 variably controls the impedance position of the variable capacitor C ₃ and the variable inductance coil L ₂ through the step motors 108 and 110, respectively. It is supposed to be. Note that the capacitor C ₄ and the inductance coil L ₃ in the subsequent stage are a fixed capacitance and a fixed inductance coil, respectively. As described above, the matching circuit 102 in the matching unit can be configured by an arbitrary variable matching circuit having one or a plurality of variable reactance elements. Another characteristic configuration is a reflected wave measurement circuit 114 that receives reflected waves propagating on the transmission line from the matching units 44 and 88 to the output terminals of the high frequency power supplies 52 and 90 and measures the power of the reflected waves. , 116 is provided.

Similar to the first technique, when auto-matching is performed in each of the matching units 44 and 88, the controller 104 makes the measured value of the load impedance Z _in obtained by the RF sensor 106 coincide with the matching point Z _MP. variably controls the variable capacitor C ₃ and a variable inductance impedance positions of the coil L ₂ by a feedback control through the pulse motor 108, 110. Here, the matching point Z _MP is usually set to a pure resistance value 50Ω (Z _MP = 50 + j0) equal to the output impedance of the high frequency power supply.

However, if the measurement accuracy of the RF sensor 106 there is an error, even shows 50Ω measurements apparently pure resistance value of the resulting load impedance Z _in at the RF sensor _{106 (Z in = 50 + j0} ), actually It may not coincide with the matching point (Z _MP = 50 + j0). Further, even when the high-frequency waveform distortion is large, the measurement accuracy of the RF sensor 106 becomes unstable, and even if the apparent measurement value of Z _in = 50 + j0 is obtained, it coincides with the matching point (Z _MP = 50 + j0). There may not be. In such a case, if the feedback control for impedance matching as described above is performed, the result is that the load impedance is matched to the impedance deviated from the matching point Z _MP , and the RF transmission characteristic is deteriorated to affect the plasma process. There is a fear.

  In this embodiment, as described below, a matching point correction function is provided to compensate for a decrease in load impedance measurement accuracy in the RF sensor due to a measurement error of the RF sensor, a high-frequency waveform distortion during an actual process, or the like.

  FIG. 10 shows an auto-learning procedure (particularly a processing procedure of the main control unit 100) performed before the apparatus is operated for the matching point correction function in this embodiment.

  This auto-learning for the matching point correction can be performed for each of the upper matching unit 44 and the lower matching unit 88 under substantially the same conditions as a specific actual process, or both can be performed simultaneously. Is possible.

For convenience of explanation, an auto-learning procedure in the upper matching unit 44 will be described. First, in initialization (step C ₁ ), a process recipe identical to a specific actual process is set in a predetermined register, and a mechanical movable part is set in an initial position. However, for auto-learning, the reference impedance Z _s (Z _s = R + jX), which is the target point of the load impedance Z _in the auto-matching of the upper matching unit 44, is centered on the standard matching point 50Ω (Z _MP = 50 + j0). Many are selected in advance, and one of them (initial value) is set.

After the initialization, a dummy semiconductor wafer W that is not an actual processing target is loaded into the chamber 10 and placed on the susceptor 16 (step C ₂ ). Then, a DC voltage is applied from the DC power source 22 to the electrode 20 of the electrostatic chuck 18 to fix the dummy wafer W to the susceptor 16.

Next, the inside of the chamber 10 is exhausted by the exhaust device 84 (step C ₃ ), and the etching gas (generally mixed gas) is supplied from the processing gas supply source 66 under the same conditions as the actual process at a predetermined flow rate and flow rate ratio, 64, and the pressure in the chamber 10 is set to a set value by the exhaust device 84 (step C ₄ ), and then the upper high-frequency power source 52 and the lower high-frequency power source 88 are turned on with preset RF power (step C ₅ ). . In this case, similarly to the first embodiment described above, the upper RF may be turned on by first-in and the lower RF may be turned on by last-in.

In upper matcher 44, automatic matching for matching the load impedance Z _in the upper RF to reference impedance Z _s is performed. More specifically, the controller 104 allows the variable capacitor C ₁ and / or variable through the pulse motors 108 and 110 so that the load impedance measurement obtained by the RF sensor 106 matches the reference impedance Z _s (Z _s = R + jX). variably controlling the impedance positions of the inductor L ₁ by feedback control. Here, the reference impedance Z _s is an initial value set in the previous initialization (step C ₁ ).

The main control unit 100, for example, sure to allow suitable time establishing matching the timer function, and stores in the memory takes in the measurement values of the obtained reflection wave power P _r with the reflected wave measurement circuit 114 (Step C _6). Thereafter, the upper high frequency power supply 52 is temporarily turned off (step C ₇ ). Then, the reference impedance Z _s is moved to the next point (step C ₇ ), the upper high frequency power supply 52 is turned on again, and RF power feeding is resumed (step C ₉ → C ₁₀ → C ₅ ). to perform the automatic matching, taking measurements of the resulting reflected wave power P _r with the reflected wave measurement circuit 114 (step C _6). Thereafter, the reference impedance Z _s is sequentially moved to a preset position, and a series of processes for capturing the _Pr measurement value obtained from the reflected wave measurement circuit 114 under the reference impedance Z _s at each movement destination is repeated (steps). C ₈ → C ₉ → C ₁₀ → C ₅ → C ₆ .

FIG. 11 shows an example of a procedure for sequentially moving the points (selected values) of the reference impedance Z _s (Z _s = R + jX). In this example, the real part R is selected at multiple points with a pitch of 5 (Ω) within the range of 30 (Ω) to 70 (Ω) around the standard 50 (Ω), and the imaginary part X is the standard 0 ( A plurality of points are selected at a pitch of 5 (Ω) within a range of −20 (Ω) to 20 (Ω) with respect to (Ω).

In the illustrated procedure, first, the real part R is fixed to 50 (Ω), the imaginary part X is swung from 0 (Ω) to 20 (Ω), and (R + jX) = (50 + j0), (50 + j5),. (50 + j20) to perform the automatic matching upper matcher 44 under each reference impedance Z _s of, record the measured value by measuring the P _r. Next, with the real part R kept at 50 (Ω), the imaginary part X is swung from 0 (Ω) to −20 (Ω), and (R + jX) = (50−j0), (50−j5), ·· (50-j20) to perform the automatic matching upper matcher 44 under each reference impedance Z _s of, record the measured value by measuring the P _r. Next, the real part R is moved or changed to 45 (Ω), and the imaginary part X is changed from 0 (Ω) to 20 (Ω) and from 0 (Ω) to −20 (Ω) in the same manner as described above. waving, the upper matcher 44 under each reference impedance Z _s to perform the automatic matching, record the measured value by measuring the P _r. Next, the real part R is moved or changed to 55 (Ω), and the imaginary part X is changed from 0 (Ω) to 20 (Ω) and from 0 (Ω) to −20 (Ω) in the same manner as described above. waving, the upper matcher 44 under each reference impedance Z _s to perform the automatic matching, record the measured value by measuring the P _r. Thereafter, the movement of the reference impedance Z _s and the _Pr measurement and recording (logging) are repeated as described above.

As a result of the above-described logging, as a result of auto-matching of the upper matching unit 44 for all preselected reference impedances Z _s on the memory in the main control unit 100 under the same conditions as the actual process. It appears all together P _r measurements obtained from the reflected wave measuring circuit 116 total. Actually, as shown in FIG. 12, for example, depending on the value of the reference impedance Z _s , matching cannot be performed at all, reflection is too large, or the reflected wave causes hunting and measurement is impossible (disengagement from the interlock). Since there are many cases, logging may be narrowed down from trend analysis.

The main control unit 100 turns off the processing gas supply source 66 after logging is finished and stops the supply of the etching gas (step C ₁₁ ), and is the smallest (ideal) of all the _Pr measurement values stored in the memory. Specifically, P _r = 0) is determined by the minimum value determination process (comparison operation) (step C ₁₂ ), and the reference impedance Z _s when this P _r minimum value is obtained is used as the matching point corresponding to the actual process. registered as Z _MS (step C _13). For example in FIG. 12, since the P _r minimum at point R + jX = 60-j10 is obtained, matching point Z _MS of upper matcher 44 for the actual process is set to Z _MS = 60-j10 Is done.

The auto-learning for the matching point correction in the lower matching unit 88 may be performed by the same procedure (FIG. 10) as described above among the main control 100, the lower matching unit 88, and the reflected wave measurement circuit 116. However, the lower radio frequency power feeding system and the upper high-frequency power supply system is different even load impedance itself not only frequency and RF power, the value of the correction matching point Z _MS resulting auto-learning is also different from the usual. Further, since the auto-learning of this embodiment is performed by changing the process conditions for each process recipe, an independent correction matching point _ZMS is obtained for each process recipe. In automatic matching in actual process (e.g. Step B _6, B ₈ in FIG. 8), the correction matching point Z _MS corresponding to the process recipe is applied to the controller 104 from the main control unit 100.

  According to this embodiment, even when there is an error in the measurement accuracy of the RF sensor 106 of the upper matching unit 44 or the RF sensor (not shown) of the lower matching unit 44, or when high-frequency waveform distortion is large during the actual process. However, since the matching point correction function by auto-learning as described above performs auto-matching of the matching units 44 and 88 in a direction that makes the power of the reflected wave zero or minimum, process stability and reproducibility can be improved.

In the above embodiment, the main control unit 100, capture and storage of the reflected wave power P _r, the minimum value determination processing of P _r measurement is performed to register the Z _MS, and the like. However, similar to the first technique, the controller 104 may perform the above function without being limited to this.

  FIG. 13 shows a configuration example of the main control unit 100. The main control unit 100 of this configuration example includes a processor (CPU) 122, a memory (RAM) 124, a program storage device (HDD) 126, a disk drive (DRV) 128 such as a floppy drive or an optical disk connected via a bus 120. An input device (KEY) 130 such as a keyboard and a mouse, a display device (DIS) 132, a network interface (COM) 134, and a peripheral interface (I / F) 136.

  The processor (CPU) 122 reads a code of a required program from a storage medium 138 such as an FD or an optical disk loaded in the disk drive (DRV) 128 and stores it in the HDD 126. Alternatively, a required program can be downloaded from the network via the network interface 76. Then, the processor (CPU) 122 develops the code of the program necessary for each stage or each scene from the HDD 126 onto the working memory (RAM) 124, executes each step, performs necessary arithmetic processing, and performs the peripheral interface 78. Each unit in the apparatus (matching units 44 and 88, high-frequency power sources 52 and 90, processing gas supply source 66, exhaust apparatus 84, etc.) is controlled via the. All of the programs used in the auto-learning and actual processes described in the first and second embodiments are executed by this computer system.

Although one embodiment has been described above, various modifications can be made within the scope of the technical idea of the present invention. For example, in a parallel plate type, a method of supplying high frequency only to one electrode and not supplying power to the other electrode, or a method of supplying two types of high frequency to one electrode simultaneously is possible. However, the present invention can be applied to a variable matching device connected between each high-frequency power source and each electrode. In addition, the configuration inside the matching unit can be variously modified not only in the matching circuit 102 but also in the RF sensor and the controllers, and a configuration in which the V _pp measurement circuit 112 is provided outside the matching unit is also possible.

  The above-described embodiment relates to a capacitively coupled parallel plate plasma processing apparatus in which a high-frequency electrode is provided in the chamber. However, the present invention relates to a helicon wave plasma method, an ECR plasma method, or the like in which a high-frequency electrode is provided outside the chamber. It is also applicable to the plasma processing apparatus. The present invention is not limited to plasma etching but can be applied to various plasma processes such as plasma CVD, plasma oxidation, plasma nitridation, and sputtering. The substrate to be processed in the invention is not limited to a semiconductor wafer, and various substrates for FPD, a photomask, a CD substrate, a printed substrate, and the like are also possible.

10 chamber 16 susceptor (lower electrode)
34 upper electrode 44 upper matcher 52 upper radio frequency power source 66 the processing gas supply source 88 lower matcher 90 lower radio frequency power source 100 main control unit 102 matching circuit C _1, C _2, C ₃ variable capacitor L ₁ variable inductance coil 104 controller 106 RF Sensor 108, 110 Step motor 112 V _pp measurement circuit
114, 116 Reflected wave measuring circuit 122 Processor (CPU)
124 Memory (RAM)
128 Program storage device (HDD)
138 storage media

Claims (4)

A dummy substrate to be processed is disposed at a predetermined position in a chamber capable of depressurization, the chamber is depressurized to a predetermined vacuum degree under substantially the same conditions as a desired process recipe, and a predetermined processing gas is supplied into the chamber. A high-frequency power source is used to generate a plasma in the chamber by supplying power to a predetermined high-frequency electrode with a predetermined power from a high-frequency power source through a variable matching unit, and a plurality of load impedance measurement values that are selected in advance. A first step of variably controlling the impedance of the variable matching device so as to match a reference impedance;
A second step of measuring the power of the reflected wave obtained on the high frequency power source side under each of the reference impedances and recording the measured value;
A third step of registering a reference impedance when a measured value of the reflected wave power becomes a minimum value or near a minimum value among the plurality of reference impedances;
A regular substrate to be processed is disposed at the predetermined position in the chamber, the chamber is depressurized to a predetermined degree of vacuum under the conditions of the process recipe, a predetermined processing gas is supplied into the chamber, and the high-frequency power source The high frequency is supplied to the high frequency electrode with a predetermined power via the variable matching unit to generate plasma in the chamber, and the variable value is measured so that the measured value of the load impedance matches the registered reference impedance. And a fourth step of performing plasma processing on the substrate by variably controlling the impedance of the matching unit. A dummy substrate is placed on the lower electrode in the depressurizable chamber, the chamber is depressurized to a predetermined degree of vacuum under substantially the same conditions as a desired process recipe, and a predetermined process is performed in the chamber. A gas is supplied, and the processing gas is discharged with a first high frequency in the chamber to generate plasma, and a high frequency of a second frequency is supplied from a high frequency power source to the lower electrode with a predetermined power via a variable matching device. A first step of energizing and drawing ions in the plasma to the substrate, and variably controlling the impedance of the variable matching unit so that a measured value of load impedance matches a plurality of preselected reference impedances;
A second step of measuring the power of the reflected wave obtained on the high frequency power source side under each of the reference impedances and recording the measured value;
A third step of registering a reference impedance when a measured value of the reflected wave power becomes a minimum value or near a minimum value among the plurality of reference impedances;
A regular substrate to be processed is disposed at the predetermined position in the chamber, the pressure in the chamber is reduced to a predetermined degree of vacuum under the conditions of the process recipe, a predetermined processing gas is supplied into the chamber, The plasma is generated by discharging the processing gas with the first high frequency and supplying the second high frequency from the high frequency power source to the lower electrode with a predetermined power via the variable matching unit. A fourth step of performing plasma processing on the substrate by drawing ions therein into the substrate, variably controlling the impedance of the variable matching unit so that the measured value of the load impedance matches the registered reference impedance, and A plasma processing method.   In the first step, each time an operation for matching the measured value of the load impedance with each of the reference impedances is performed, the high frequency from the high frequency power source is supplied to the high frequency electrode to generate plasma. The plasma processing method according to claim 1 or 2.   The plasma processing method according to any one of claims 1 to 3, wherein, in the second step, the measured value of the reflected wave power is taken in at a timing when matching is established by the timer function.

Images