Plasma processing apparatus and processing method

Abstract

The present invention provides a plasma processing apparatus and a processing method. The plasma processing apparatus includes: a first electrode on which a substrate is placed; a plasma generation source for generating plasma; a bias power supply for supplying a bias power to the first electrode; a generation source power supply for supplying a generation source power having a higher frequency than the bias power to the plasma generation source; and a control section for controlling the bias power supply and the generation source power supply, the generation source power having a first state and a second state, the control section performing control so that the first state and the second state are alternately applied in synchronization with a signal synchronized with a cycle of a high frequency of the bias power or a phase within one cycle of a reference electrical state, wherein the reference electrical state indicates any one of a voltage, a current, or an electromagnetic field measured in a power supply system of the bias power, and the control section controls the generation source power to be off at a time of a negative side peak of the phase within at least one cycle of the reference electrical state. The present invention can control ion energy distribution.

Description

Technical Field

[0001] This invention relates to plasma processing apparatus and processing methods. Background Technology

[0002] There is a known technique in which, during etching, the high-frequency power applied for ion attraction is synchronized with the on / off state of the high-frequency power used for plasma generation, so that ions reach the polysilicon layer and the etching rate of the polysilicon layer is uniform (for example, see Patent Document 1).

[0003] Existing technical documents

[0004] Patent documents

[0005] Patent Document 1: Japanese Patent Application Publication No. 10-64915 Summary of the Invention

[0006] The technical problem that the invention aims to solve

[0007] This invention provides a plasma processing device and method capable of controlling the distribution of ion energy.

[0008] Technical solutions for solving technical problems

[0009] According to one aspect of the present invention, a plasma processing apparatus is provided, comprising: a first electrode on which a substrate is mounted; a plasma generation source for generating plasma; a bias power supply for supplying bias power to the first electrode; a generation source power supply for supplying generation source power at a frequency higher than the bias power to the plasma generation source; and a control unit for controlling the bias power supply and the generation source power supply, wherein the generation source power has a first state and a second state, and the control unit controls the first state and the second state to be alternately applied in synchronization with a signal periodically synchronized with the bias power or a phase within a period of a reference electrical state, wherein the reference electrical state represents any one of voltage, current, or electromagnetic field measured in the power supply system of the bias power, and the control unit controls the generation source power to be turned off at least at the negative peak of the phase within a period of the reference electrical state.

[0010] Invention Effects

[0011] According to one aspect, it is possible to control the distribution of ion energy. Attached Figure Description

[0012] Figure 1 This is a diagram illustrating an example of a plasma processing apparatus according to one embodiment.

[0013] Figure 2This diagram illustrates a scenario where control is performed using the phase signal of a sensor installed in the power supply system, or where control is performed using a signal that is periodically synchronized with the high frequency of the bias power.

[0014] Figure 3 This is a diagram illustrating an example of the structure of the control unit in one embodiment.

[0015] Figure 4 This is a diagram used to illustrate the cause of the tilt.

[0016] Figure 5 This is a diagram showing the waveforms of the LF voltage and HF voltage used in Embodiment 1 of an implementation.

[0017] Figure 6 It means and Figure 5 The figure shows an example of the measurement results of the ion energy distribution corresponding to the pulse application time of HF.

[0018] Figure 7 This is a graph showing the correlation between the ion energy distribution and the power of the generation source in Example 2 of one embodiment.

[0019] Figure 8 This is a graph showing the relationship between the type of source power (CW, SSP phase) and the in-plane distribution of plasma electron density in Example 3 of an embodiment.

[0020] Figure 9 This is a graph showing the measurement results of the in-plane distribution of plasma electron density and its correlation with bias power in Example 4 of one embodiment.

[0021] Figure 10 This is a graph showing the time variation of bias power and plasma electron density in Example 5 of one embodiment.

[0022] Figure 11 This is a diagram illustrating the pulse application timing of the source power in Embodiment 6 of an implementation method.

[0023] Figure 12 It means and Figure 11 The figure shows an example of the measurement results of the ion energy distribution corresponding to the pulse application time of the periodic source power.

[0024] Figure 13 This is a diagram illustrating the pulse application timing of the source power in Embodiment 7 of an implementation method.

[0025] Figure 14 It means and Figure 13 The figure shows an example of the measurement results of the ion energy distribution corresponding to the pulse application time of the generation source power of the phase.

[0026] Figure 15 This is a graph illustrating the relationship between the source power and duty cycle in Example 8 of an embodiment.

[0027] Figure 16 It means based on Figure 15 The figure shows the measurement results of the ion energy distribution of the pulse duty cycle of the generation source power.

[0028] Figure 17 This is a graph illustrating the pulse application time (phase-dependent) of the source power generated in Example 9 of an embodiment.

[0029] Figure 18 It means and Figure 17 The figure shows an example of the measurement results of the ion energy distribution corresponding to the pulse application time (phase-dependent) of the source power.

[0030] Figure 19 This is a diagram used to illustrate the switching on and off of HF and the operation of ions and electrons in one embodiment.

[0031] Figure 20 It is a diagram used to illustrate the release of secondary electrons from high-energy ions.

[0032] Figure 21 This is a diagram showing the pulse application times of the bias power and the generation source power in a modified embodiment of one implementation.

[0033] Figure 22 This is a diagram illustrating the operation of electrons during the power shutdown period of a generator source in one embodiment.

[0034] Figure 23 It is a graph showing the time it takes for electrons to disappear after the power of the generating source is turned off.

[0035] Figure 24 This is a diagram schematically representing the switching on and off of the generator source power and the number of electrons reaching the substrate in the comparative example.

[0036] Figure 25 This is a diagram schematically illustrating the switching on and off of the power source and the number of electrons reaching the substrate in one embodiment.

[0037] Figure 26 This is a diagram used to illustrate the timing of the power generation source being switched on and off in one embodiment and its effects.

[0038] Explanation of reference numerals in the attached figures

[0039] 1...Plasma processing device

[0040] 10……Processing Containers

[0041] 16……Lower electrode (stage)

[0042] 20……ESC electrode

[0043] 22……ESC DC power supply

[0044] 24……Focusing ring

[0045] 26……Lower electrode cover

[0046] 34……Upper Electrode

[0047] 48... High-frequency power supply

[0048] 50…Voltage Variable DC Power Supply

[0049] 66……Processing gas supply source

[0050] 84……Exhaust device

[0051] 90... Low-frequency power supply

[0052] 100... processor

[0053] 102……Signal Generation Circuit

[0054] 200……Control Department. Detailed Implementation

[0055] Hereinafter, embodiments for carrying out the present invention will be described with reference to the accompanying drawings. In the drawings, the same reference numerals are used to label the same components, and sometimes repeated descriptions are omitted.

[0056] Hereinafter, the frequency (high frequency) of the generated source power will be referred to as "HF" (High Frequency), and the generated source power will be referred to as "HF power". Furthermore, the frequency (low frequency) of the bias power, which is lower than the generated source power, will be referred to as "LF" (Low Frequency), and the bias power will be referred to as "LF power".

[0057] [Overall structure of the plasma processing device]

[0058] First, refer to Figure 1 This describes an example of a plasma processing apparatus 1 according to one embodiment. Figure 1 This is a diagram illustrating an example of a plasma processing apparatus according to one embodiment.

[0059] One embodiment of the plasma processing apparatus 1 is a capacitively coupled plasma processing apparatus. For example, it has a cylindrical processing container 10 with a surface made of anodized aluminum. The processing container 10 is grounded.

[0060] A cylindrical support platform 14 is disposed at the bottom of the processing container 10, separated by an insulating plate 12 made of ceramic or the like. A lower electrode 16, for example made of alumina, is disposed on the support platform 14. The lower electrode 16 functions as a mounting stage, on which a substrate W, an example of a workpiece to be processed, is placed, separated by an electrostatic chuck 18.

[0061] An electrostatic chuck 18 is provided on the upper surface of the lower electrode 16 to hold the wafer W by electrostatic force. The electrostatic chuck 18 has a structure in which an ESC electrode 20 made of a conductive film is sandwiched between a pair of insulating layers or insulating sheets. The ESC electrode 20 is electrically connected to an ESC DC power supply 22. A DC voltage output from the ESC DC power supply 22 is applied to the ESC electrode 20. The substrate W is held in the electrostatic chuck 18 by electrostatic forces such as Coulomb force generated therefrom.

[0062] A conductive focusing ring 24, for example made of silicon, is disposed around the substrate W on the lower electrode 16. The focusing ring 24 is also called an etching ring. A cylindrical lower electrode cover 26, for example made of quartz, is provided on the side of the lower electrode 16 and the support stage 14.

[0063] A refrigerant chamber 28 is arranged in a ring shape inside the base support 14. A refrigerant, such as a non-reactive fluorine liquid, at a specified temperature is circulated to the refrigerant chamber 28 from an external cooling unit via pipes 30a and 30b. The processing temperature of the substrate W on the lower electrode 16 can be controlled by the temperature of the refrigerant. In addition, the refrigerant is an example of a temperature regulating medium circulated to pipes 30a and 30b, which not only cools the lower electrode 16 and the substrate W, but can also be used for heating.

[0064] Furthermore, heat-conducting gas, such as He gas, from the heat-conducting gas supply mechanism is supplied via gas supply passage 32 between the upper surface of the electrostatic chuck 18 and the back surface of the substrate W.

[0065] Above the lower electrode 16, an upper electrode 34 is disposed opposite to and parallel to the lower electrode 16. A plasma processing space is formed between the upper electrode 34 and the lower electrode 16. The upper electrode 34 forms a surface that faces the substrate W on the lower electrode 16 and contacts the plasma processing space, i.e., a facing surface.

[0066] The upper electrode 34 is supported on the upper part of the processing container 10 via an insulating shielding member 42. The upper electrode 34 includes: an upper electrode plate 36, which forms an opposing surface opposite to the lower electrode 16 and has a plurality of gas release holes 37; and an upper electrode support 38, which detachably supports the upper electrode plate 36 and is made of a conductive material such as anodized alumina. The upper electrode plate 36 may be formed of, for example, silicon or SiC. A gas diffusion chamber 40 is provided inside the upper electrode support 38, and a plurality of gas flow holes 41 communicating with the gas release holes 37 extend downward from the gas diffusion chamber 40.

[0067] A gas inlet 62 is formed on the upper electrode support 38 for introducing processing gas into the gas diffusion chamber 40. A gas supply pipe 64 is connected to the gas inlet 62, and a processing gas supply source 66 is connected to the gas supply pipe 64. A mass flow controller (MFC) 68 and an on / off valve 70 are sequentially arranged on the gas supply pipe 64 from the upstream side. Thus, processing gas for etching can be supplied from the processing gas supply source 66. The processing gas reaches the gas diffusion chamber 40 from the gas supply pipe 64, and is released into the plasma processing space in a spray-like manner through the gas flow hole 41 and the gas release hole 37. In this way, the upper electrode 34 functions as a spray head for supplying processing gas.

[0068] The upper electrode 34 is connected to the voltage-variable DC power supply 50, and a DC voltage from the voltage-variable DC power supply 50 is applied to the upper electrode 34. The polarity, current, and voltage of the voltage-variable DC power supply 50, as well as the control of the electronic switches for turning the current and voltage on and off, are controlled by the control unit 200.

[0069] The lower electrode 16 is connected to a high-frequency power supply 48 via a power supply rod 47 and an HF matching device 46. The high-frequency power supply 48 applies HF power to the lower electrode 16. The lower electrode 16 is connected to a low-frequency power supply 90 via a power supply rod 89 and an LF matching device 88. The low-frequency power supply 90 applies LF power to the lower electrode 16. This attracts ions to the substrate W on the lower electrode 16. The HF frequency is above 13.56 MHz, and the LF frequency is lower than the HF frequency, in the range of 200 kHz to 13.56 MHz. For example, 40 MHz of HF power can be output relative to 400 kHz of LF power. The HF matching device 46 matches the internal impedance of the high-frequency power supply 48 to the load impedance. The LF matching device 88 matches the internal impedance of the low-frequency power supply 90 to the load impedance. The lower electrode 16 can be connected to a filter 94 for grounding a specified high frequency. Furthermore, the HF power supplied from the high-frequency power supply 48 can be applied to the upper electrode 34.

[0070] An exhaust port 80 is provided at the bottom of the processing container 10, and this exhaust port 80 is connected to an exhaust device 84 via an exhaust pipe 82. The exhaust device 84 has a vacuum pump such as a turbomolecular pump, which can reduce the pressure inside the processing container 10 to the desired vacuum level. In addition, a substrate W feed outlet 85 is provided on the side wall of the processing container 10, and this feed outlet 85 can be opened and closed by a gate valve 86. Furthermore, a deposit shielding member 11 for preventing etching byproducts (deposits) from adhering to the processing container 10 is detachably provided along the inner wall of the processing container 10. That is, the deposit shielding member 11 constitutes the wall of the processing container. In addition, the deposit shielding member 11 is also provided on the outer periphery of the lower electrode cover 26. An exhaust plate 83 is provided between the deposit shielding member 11 on the processing container wall side at the bottom of the processing container 10 and the deposit shielding member 11 on the lower electrode cover 26 side. The deposit shield 11 and the exhaust plate 83 can be used on aluminum parts covered with ceramics such as Y2O3.

[0071] During etching in the plasma processing apparatus 1 with the above-described structure, firstly, the gate valve 86 is opened, and the substrate W is fed into the processing container 10 through the feed outlet 85 and placed on the lower electrode 16. Then, the processing gas for etching is supplied to the gas diffusion chamber 40 at a predetermined flow rate from the processing gas supply source 66, and is supplied into the processing container 10 through the gas flow port 41 and the gas release port 37. Furthermore, the processing container 10 is vented using the exhaust device 84, so that the pressure therein is a set value in the range of, for example, 0.1 to 150 Pa. As the processing gas, various processing gases that have been used in the past can be used. For example, gases containing carbon and fluorine elements, such as C4F8 gas, can be preferably used. Moreover, other gases, such as Ar gas and O2 gas, can also be used.

[0072] As described above, with etching gas introduced into the processing container 10, HF power is applied to the lower electrode 16 from the high-frequency power supply 48. Furthermore, LF power is applied to the lower electrode 16 from the low-frequency power supply 90. Additionally, a DC voltage is applied to the upper electrode 34 from the voltage-variable DC power supply 50. Furthermore, a DC voltage is applied to the ESC electrode 20 from the ESC DC power supply 22, thus adsorbing and holding the substrate W on the lower electrode 16.

[0073] The processing gas released from the gas release hole 37 of the upper electrode 34 is mainly generated into plasma through dissociation and ionization by HF power. The surface to be processed on the substrate W is etched using free radicals and ions in the plasma. Furthermore, by applying LF power to the lower electrode 16, it is possible to control the ions in the plasma to etch holes with high aspect ratios, thereby expanding the control edge of the plasma.

[0074] [Control Department]

[0075] The plasma processing apparatus 1 is equipped with a control unit 200 that controls the operation of the entire apparatus. The control unit 200 performs desired plasma processing, such as etching, according to a scheme stored in a memory such as ROM (Read Only Memory) or RAM (Random Access Memory). The scheme contains control information for the apparatus corresponding to the processing conditions, such as processing time, pressure (gas exhaust), high-frequency power and / or voltage, various gas flow rates, temperature inside the processing container, and temperature of the refrigerant output from the cooler. Furthermore, the scheme representing the above-mentioned program and processing conditions can also be stored in a hard disk or semiconductor memory. Alternatively, the scheme can be set to a predetermined location and retrieved while stored on a portable computer-readable storage medium such as a CD-ROM or DVD.

[0076] It can also be controlled so that the on / off state or high / low state of the HF power is synchronized with the high-frequency periodic signal of the bias power, or with the phase of any of the voltage, current, or electromagnetic field measured in the bias power supply system within one cycle. For example, the control unit 200 can be controlled to synchronize the on / off state or high / low state of the HF power with the phase of the voltage or current of the LF within one cycle. This allows for control of the quantity and quality of ions and free radicals. Furthermore, it can reduce the generation of IMD.

[0077] The bias power supply system refers to the low-frequency power supply 90 → LF matching unit 88 → power supply rod 89 → lower electrode 16 → (plasma) → upper electrode 34 → (ground). Any voltage, current, or electromagnetic field measured in the bias power supply system refers to the voltage, current, or electromagnetic field measured within the low-frequency power supply 90 through the LF matching unit 88, the portion from the power supply rod 89 to the lower electrode 16, and the upper electrode 34.

[0078] Additionally, the state of a signal that is periodically synchronized with the bias power, or any of the voltage, current, or electromagnetic field measured in the bias power supply system, is referred to as the "reference electrical state." The HF power (generating source power) is controlled to be applied alternately in the first and second states, described later, in phase synchronization with the reference electrical state over one cycle.

[0079] When the voltage, current or electromagnetic field measured in the bias power supply system is used as the "reference electrical state", the reference electrical state is preferably the voltage, current or electromagnetic field measured from any component from the lower electrode 16 to the inside of the matching unit connected via the power supply rod 47.

[0080] As a method for measuring the reference electrical state in a bias power supply system, one example is to place voltage probes, current probes, and BZ probes (probes for measuring induced magnetic fields) near any component of the bias power supply system. Thus, the voltage, current, or induced magnetic field of each component is measured using each probe.

[0081] For example, Figure 2 This is an example of using any one of the voltage, current, or electromagnetic field measured in a biased power supply system as a "reference electrical state." For example, in... Figure 3 In (a), the processor 100 receives any one of the following from a sensor such as a VI probe installed in the power supply system: the voltage or current of HF, the voltage or current of LF, the phase signal of HF, or the phase signal of LF. The processor 100 applies the generator power alternately in a first state and a second state, synchronized with the phase of a reference electrical state within one cycle, the reference electrical state representing any one of the input voltage or current of HF, the voltage or current of LF, the phase signal of HF, or the phase signal of LF.

[0082] Alternatively, the processor 100 can generate a signal that is periodically synchronized with the high-frequency bias power output from the low-frequency power supply 90, without relying on signals from the sensor. In this case, the state of this signal can be used as a reference electrical state. Furthermore, the step of measuring the reference electrical state in the bias power supply system can be eliminated. For example, in... Figure 2 In (b), processor 100 receives a phase signal (low-power waveform) of the LF or a signal related to bias power information from low-frequency power supply 90, and generates a high-frequency periodically synchronized signal with the bias power based on the input signal. Processor 100 outputs the generated signal to high-frequency power supply 48. High-frequency power supply 48 applies the generated source power alternately in a first state and a second state based on the signal.

[0083] Alternatively, the processor 100 may generate a signal that is periodically synchronized with the low-frequency bias power, not based on the signal from the low-frequency power supply 90. In this case, the processor 100 generates, for example, a signal having... Figure 2 The periodic signal shown by LF, and generates a signal synchronized with that signal, for example... Figure 2 The on / off signal is shown in HF. The processor 100 outputs the generated signal to the low-frequency power supply 90 and the high-frequency power supply 48. The low-frequency power supply 90 outputs bias power based on this signal. The high-frequency power supply 48 applies the generated source power alternately in a first state and a second state based on this signal. That is, when the first state and the second state are applied alternately in sync with a signal that is periodically synchronized with the low-frequency bias power, the signal is not generated based on the signal from the low-frequency power supply 90, but can be a signal that is periodically synchronized with the low-frequency bias power.

[0084] Furthermore, the lower electrode 16 is an example of a first electrode on which the substrate W is mounted. The upper electrode is an example of a second electrode opposite to the first electrode. The low-frequency power supply 90 is an example of a bias power supply that supplies LF power to the first electrode. The high-frequency power supply 48 is an example of a generation source power supply that supplies HF power with a frequency higher than the LF power to either the first or second electrode. The control unit 200 is an example of a control unit that controls the bias power supply and the generation source power supply. The potential of the lower electrode 16 on which the bias power is applied is also referred to as the lower electrode potential.

[0085] [Structure of the Control Department]

[0086] Reference Figure 2 The specific structure of the control unit 200 will be explained. Figure 2 This is a diagram illustrating an example of the structure of a control unit 200 according to one embodiment. The control unit 200 includes a processor 100, a signal generation circuit 102, directional couplers 105 and 108, a reflection detector 111, and an oscilloscope 112.

[0087] In the power supply line of the high-frequency power supply 48, a directional coupler 108 is connected between the high-frequency power supply 48 and the HF matching unit 46. In the power supply line of the low-frequency power supply 90, a directional coupler 105 is connected between the low-frequency power supply 90 and the LF matching unit 88.

[0088] Directional coupler 108 sends a portion of the traveling wave power (hereinafter referred to as "Pf") and reflected wave power (hereinafter referred to as "Pr") of the HF wave to oscilloscope 112. Additionally, directional coupler 105 sends a portion of the traveling wave power of the LF wave to oscilloscope 112.

[0089] In one embodiment, the frequency of the LF displayed on the oscilloscope 112 is, for example, 400 kHz, and the frequency of the HF is, for example, 40 MHz. Therefore, the waveforms of the traveling wave power of the LF, the traveling wave power of the HF, and the reflected wave power of the HF can be observed on the oscilloscope 112. For example, Figure 5 An example of the waveform representing the traveling wave power (Pf(HF)) and the reflected wave power (Pr(HF)) of HF.

[0090] Additionally, the directional coupler 108 separates a certain proportion of the HF reflected wave and sends it to the reflection detector 111. The reflection detector 111, for example, is composed of a spectral analyzer, a power meter, etc., and measures the amount and degree of IMD (Intermodulation Distortion) generated at which wavelengths. IMD is the reflected wave power from the plasma side generated based on the frequency of the sum or difference between the fundamental and / or harmonic waves of the LF and the fundamental and / or harmonic waves of the HF. The fundamental and / or harmonic waves of the LF and the fundamental and / or harmonic waves of the HF are generated by applying HF power to the upper electrode or the lower electrode of the plasma processing device 1 and applying LF power to the lower electrode.

[0091] The directional coupler 105 sends a portion of the traveling wave power of the LF to the processor 100. The processor 100 generates a synchronization signal for HF that is synchronized with the traveling wave power of the LF. For example, the processor 100 may also generate the synchronization signal for HF in sync with the positive timing of the traveling wave of the LF. Alternatively, instead of the directional coupler 105, the waveform of the LF detected by a sensor such as a VI probe may be sent to the processor 100.

[0092] The processor 100 sends the generated synchronization signal to the signal generation circuit 102. The signal generation circuit 102 generates a control signal that is synchronized with the traveling wave power of the LF based on the sent synchronization signal, and sends it to the high-frequency power supply 48 and the low-frequency power supply 90.

[0093] There are two methods for generating the control signal. When the low-frequency power supply 90 is a normal power supply, the directional coupler 105 extracts a portion of the voltage or current of the LF output from the low-frequency power supply 90 as a waveform and inputs it to the processor 100. However, it is not limited to this; the processor 100 can directly input a portion of the LF power, etc., from the low-frequency power supply 90, or it can generate a signal to control the high-frequency power supply 48 and the low-frequency power supply 90 without relying on the signal from the low-frequency power supply 90. The processor 100 generates an on-state signal with arbitrary delay and arbitrary amplitude based on the input waveform signal or the generated signal, and sends it to the signal generation circuit 102. The on-state signal is an example of a synchronization signal.

[0094] During the period when the signal generation circuit 102 receives the turn-on signal, it issues a command signal to the high-frequency power supply 48 in order to generate HF power. The command signal can be a control signal that generates HF power during the period when the turn-on signal is received, depending on the input mode of the high-frequency power supply 48, or the turn-on signal itself.

[0095] If the low-frequency power supply 90 is an amplifier that amplifies LF power, voltage, or current, the signal from the directional coupler 105 may not be used. In this case, the signal generation circuit 102 can extract a portion of the LF power output from the low-frequency power supply 90 as a waveform, and generate an on-state signal with arbitrary delay and amplitude based on the waveform signal. The signal generation circuit 102 then sends the waveform signal and the on-state signal to the high-frequency power supply 48.

[0096] However, the above method for generating control signals is merely one example and is not limited to it. Any control signal capable of being generated—that is, a control signal that alternately applies the on / off or high / low power of the HF power synchronously with the phase of the received synchronization signal and the reference electrical state within one cycle (the phase of the LF voltage or current within one cycle, electrode potential, etc.)—is acceptable. If a control signal can be generated in this way, then it is not limited to... Figure 2 The circuitry of the control unit 200 shown can be used with other hardware or software.

[0097] The amplifier in the low-frequency power supply 90 amplifies the amplitude (AM: amplitude modulation) of the 400kHz LF modulation signal and supplies it to the lower electrode. The amplifier in the high-frequency power supply 48 amplifies the amplitude of the 40MHz HF modulation signal and supplies it to the lower electrode.

[0098] [Deflection]

[0099] Sometimes, deviation occurs when etching the film of the object being etched. Figure 4 This diagram illustrates the generation of the tilt. In the plasma processing apparatus 1 of this embodiment, bias power (LF power) and generation source power (HF power) are applied to the lower electrode 16. When the voltage of LF is negative due to the self-bias voltage Vdc, the substrate potential (electrode potential) deepens negatively. At this time, since the thickness of the sheath is proportional to the voltage, a larger voltage is applied to the substrate when the substrate potential is negative, and the sheath thickness increases. Conversely, when the substrate potential is positive, electrons are accelerated towards the substrate. Compared to ions, electrons have a smaller mass and extremely higher mobility, so a larger current flows towards the substrate, and the positive potential is mitigated. Therefore, the substrate potential does not increase in the positive direction, and the sheath thickness does not increase.

[0100] The thickness of the sheath varies approximately according to the plasma density distribution. For example, assuming the same potential is applied to the substrate W. Furthermore, assuming a non-uniform plasma density distribution, the sheath is thinner where the plasma density is high and thicker where the plasma density is low. As mentioned above, in the case of a non-uniform plasma density distribution, the sheath interface tilts towards the substrate W, and the ions are accelerated in the tilting direction.

[0101] Figure 4 An example is shown where a target film 2, such as a silicon oxide film (SiO2), is etched according to the hole pattern of a mask 3. In this example, when the sheath S is thicker, ions are incident at an angle to one side relative to the edge region of the substrate W. This results in an etch shape with an angle in the hole H formed in the target film 2. When the sheath S becomes thinner, it varies depending on the gap between the substrate W and the focusing ring 24, thus changing the ion incident angle. Consequently, a different angle is generated in the etch shape of the hole H formed in the target film 2 compared to when the sheath S is thicker. This change in the ion incident angle is repeated according to the period of the voltage of the LF. The situation where an etch shape is etched as described above is called tilting, and the etch shape with the tilting is called a tilting shape. Furthermore, Figure 4 This is just one example; the skew shape is not limited to this. Furthermore, skew shapes not only occur in holes but also include cases where tilting occurs in grooves.

[0102] To suppress the aforementioned skewness, this embodiment provides a plasma processing apparatus 1 and a processing method capable of controlling the in-plane distribution of ion energy and plasma density. This allows for control of the incident angle of ions relative to the substrate W, suppressing the generation of skewed shapes, and enabling the vertical formation of shapes such as holes.

[0103] The following describes the measurement results obtained when etching silicon oxide films using the plasma processing apparatus 1 of this embodiment under various processing conditions. Furthermore, in Embodiment 1 and other embodiments, as... Figure 2 As shown, an IEA (Ion Energy Analyzer) measuring device 113 is placed on a substrate W, with the probe in contact with the substrate W, to measure the ion energy reaching the substrate W. The IEA measuring device 113 is only configured on the substrate W when measuring the substrate potential; it is not configured for other processing. In the embodiments described below, "IEDF" refers to the measurement result of the IEA measuring device 113, indicating the number of ions reaching the substrate W with each energy. Furthermore, the in-plane distribution of plasma density was measured using a PAP (Plasma Absorption Probe) to measure the electron density.

[0104] [Example 1]

[0105] First, refer to Figure 5 and Figure 6 Example 1 of this embodiment will be described. Figure 5 This is a graph showing the waveforms of the voltage of LF (also simply LF) and the voltage of HF (also simply HF) used in Embodiment 1 of one implementation, showing the timing of the application of the HF pulse. Figure 6 It means and Figure 5 A figure showing an example of the measurement results of the ion energy distribution corresponding to the pulse application time of HF.

[0106] Figure 5 (a) shows the application of HF in the comparative example. Figure 5 In comparative example (a), HF is always on, so the traveling wave (Pf) and reflected wave (Pr) of HF are continuously output. In the following text, "CW" indicates the case where HF is continuously applied. In all embodiments and variations, "CW" serves as a comparative example.

[0107] Figure 5 (b) and (c) show the pulse application timing of the HF in Embodiment 1 of this embodiment. In all embodiments of this embodiment and the variations described later, the control unit 200 performs "SSP" control, in which the generator source power is applied intermittently at predetermined times in sync with the high-frequency period of the bias power. That is, the HF pulse is introduced in sync with the sheath's operation, hence the term "Sheath Synchronized Pulse" is simply used. Figure 5 In embodiments (b) and (c) of Example 1, the power source is turned on with a duty cycle of 40%. The reflected wave (Pr) of the HF is generated at the moments when the traveling wave (Pf) is turned on and off. A large reflected wave is generated momentarily after the HF power is turned off and then gradually decays.

[0108] In the "SSP" control, use (Phase) indicates the moment when the source power is applied synchronously with the bias power. Figure 5 (b) This indicates that the LF value is negative at the center of the pulse, which represents the period during which the power source is turned on. Figure 5 of (c) This indicates that the LF value is at its positive maximum at the center of the pulse, which represents the period during which the power source is turned on.

[0109] This indicates the situation where LF changes from positive to negative in the middle of the pulse, which represents the period during which the power source is turned on. This indicates the period during which the power source is turned on when the LF pulse changes from negative to positive in the middle of the pulse.

[0110] The processing conditions in Example 1 are as follows.

[0111] <Processing Conditions>

[0112]

[0113] Figure 6 The figure shows the measurement results of the ion energy distribution reaching the substrate, measured by the IEA measuring device 113, when plasma is generated in the plasma processing apparatus 1 under the above processing conditions. Figure 6 The horizontal axis represents ion energy, and the vertical axis represents the ion energy distribution function (IEDF). That is, it represents the distribution function of ions reaching the substrate at that ion energy. The integral value of this IEDF at each ion energy is the total number of ions reaching the wafer.

[0114] The preferred ion energy distribution has roughly two peaks: a low-energy side (around 100 eV) and a high-energy side (around 350–400 eV), with a small number of ions having energies in between. In the comparative example, "CW" (refer to...) Figure 5 In (a)), a broad peak (100–150 eV) can be observed on the low-energy side of the ion energy distribution, but not near the peak (350–400 eV) on the high-energy side.

[0115] Next, the SSP, which is used in this embodiment, was measured. ( Figure 5 (b) reference) ( Figure 5 (c) reference), The result of the ion energy distribution is that, in At 90°, a distinct peak can be observed on the high-energy side (around 350–400 eV). On the other hand, at... At 270°, the peak on the high-energy side is almost unobservable, and the maximum value of the ion energy distribution shifts to the low-energy side. Furthermore, in In this case, the amount of ions reaching the substrate is significantly reduced. A preferred ion energy distribution allows for the observation of higher and more pronounced peaks at high energies, and also allows for the observation of pronounced peaks at low energies; therefore, in this case, The optimal ion energy distribution is observed at 270°. As described above, in this embodiment, by controlling the phase of the SSP, the distribution of ion energy and the peak and amount on the high-energy side can be controlled.

[0116] Based on the results of Example 1, in the processing method of this embodiment, the timing and conditions of applying HF in a pulsed manner are appropriately controlled according to the sheath's movement (i.e., the period of LF), thereby controlling the ion energy reaching the substrate. Therefore, the phase... Further experiments were conducted to optimize the values ​​of the preferred parameters, the timing of HF pulse application, and other conditions.

[0117] [Example 2]

[0118] Below, refer to Figure 7 Example 2 of this embodiment will be described. The processing conditions for Example 2 are as follows.

[0119] <Processing Conditions>

[0120]

[0121] Figure 7 The following diagram shows the measurement results of the ion energy distribution reaching the substrate when plasma is generated in the plasma processing apparatus 1 under the above processing conditions. Figure 7 The correlation between ion energy distribution and source power is explained.

[0122] Figure 7 (a) represents the correlation between ion energy distribution and source power in the case of “CW” as a comparative example. Figure 7 (b) indicates that in Example 2, which is the embodiment of this invention... Under these conditions, the ion energy distribution is correlated with the power of the generation source.

[0123] In the case of Comparative Example "CW", increasing the generator source power from 40W to 100W to 160W resulted in a higher peak energy level on the high-energy side, but the ion energy itself shifted towards the low-energy side. This is because increasing the generator source power increases the plasma electron density (Ne), thus reducing the potential difference between the plasma potential and the substrate potential. Consequently, the energy of the accelerated ions decreases, and the ion energy distribution shifts towards the low-energy side. Based on the above, in the case of "CW", it is difficult to independently control the ion energy and the plasma electron density (Ne) proportional to the amount of ions.

[0124] In response, Figure 7 (b) In this case, when the power of the HF source is increased from 40W to 100W to 160W, the peak ion quantity on the high-energy side increases. Apart from this, there is no shift in ion energy distribution to the low-energy side, and the ion energy on the high-energy side does not decrease even when the source power is increased.

[0125] The reason is that, In this case, plasma is generated during the period when the generation source power is on, and ions are accelerated with a larger negative bias voltage during the period when bias power is applied, thus increasing the ion energy. Furthermore, when the generation source power is off, the ion sheath collapses as the plasma disappears. However, during this process, the ions hardly disappear; therefore, the ions are accelerated by the energy of the voltage applied to the substrate itself.

[0126] As can be seen from the above, according to the "SSP" processing method of this embodiment, by optimizing the phase... It can independently control the ion energy and the amount of ions reaching the substrate.

[0127] [Example 3]

[0128] Below, refer to Figure 8 Example 3 of this embodiment will be described. The processing conditions for Example 3 are as follows. In Example 3 and other examples, the plasma electron density (Ne) was measured using a PAP (plasma probe).

[0129] <Processing Conditions>

[0130]

[0131] Figure 8 The results shown illustrate the plasma electron density (Ne) measured at various locations from the center of the substrate to near the inner wall of the processing container 10, under the aforementioned processing conditions, when plasma is generated in the plasma processing apparatus 1. Figure 8 This describes the relationship between the type of source power (CW, SSP phase) and the in-plane distribution of plasma electron density (Ne) in one embodiment. Figure 8 On the horizontal axis, the center of the substrate is set to "0". 0mm on the horizontal axis is the center of the substrate, 150mm is the end of the substrate, 200mm is near the end of the focusing ring 24, and 280mm is near the inner wall of the processing container 10.

[0132] Compared to the comparative example "CW", in this embodiment... In this case, the midpoint of the HF pulse's on-state is the time when LF reaches its maximum value. In this case, the sheath is thinner, so the generated source power easily reaches the center side of the substrate. Therefore, the plasma electron density Ne becomes higher at the center side of the substrate. On the other hand, in this embodiment... In this case, the midpoint of the HF pulse's on-state is the time when LF reaches its negative maximum value. In this case, the sheath thickens, making it difficult for the generated source power to reach the center of the substrate, but making it easier to reach the outer side of the focusing ring 24. Therefore, the plasma electron density Ne becomes higher on the outer side of the focusing ring 24.

[0133] Based on the above, it can be concluded that... In this case, the sheath is relatively thin, therefore the plasma electron density Ne is highest at the center side of the substrate. In this case, the sheath thickens, making it difficult for the source power to penetrate, and the plasma electron density Ne decreases on the central side of the substrate. Therefore, in the SSP control of this embodiment, the phase of the applied source power is changed... It can change the distribution of plasma electron density Ne and control the distribution of plasma electron density Ne within the plane of the substrate.

[0134] [Example 4]

[0135] Below, refer to Figure 9 Example 4 of this embodiment will be described. The processing conditions for Example 4 are as follows.

[0136] <Processing Conditions>

[0137]

[0138] Figure 9 An example of the measurement results of the in-plane distribution of plasma electron density (Ne) at various locations from the center of the substrate to near the inner wall of the processing container 10, in relation to the bias power, is shown when plasma is generated in the plasma processing apparatus 1 under the above-described processing conditions.

[0139] What is being measured and In this situation, Under these conditions, three patterns with bias power of 50W, 100W, and 150W were measured. Furthermore, in... Under these conditions, two patterns with bias power of 50W and 100W were measured.

[0140] according to Figure 9 The measurement results, in In the case of, with Compared to the situation where the plasma electron density Ne is lower at the center of the substrate, the plasma electron density Ne is higher on the outer side of the focusing ring 24. The reason for this is that... In this case, when the HF pulse is applied, the LF bias is near the negative maximum value, so the sheath is thicker, and the power of the generated source becomes difficult to reach the substrate side but easy to reach the outside of the focusing ring 24.

[0141] exist Under these conditions, the plasma electron density Ne exhibits approximately the same distribution at any of the bias power values ​​of 50W, 100W, and 150W. Therefore, it can be concluded that by controlling the phase of the applied source power... At any given moment, the in-plane distribution of the plasma electron density Ne can be controlled. For example, by controlling the phase at which the generation source power is applied. It is possible to increase the plasma electron density Ne at the center side of the substrate, or to increase the plasma electron density Ne at the end side of the substrate, on the focusing ring 24. Thus, even if the average power of the generation source and the bias power are the same, an effective plasma electron density Ne can be obtained.

[0142] Based on the measurement results of Examples 1 to 4 above, it can be seen that in the case of the SSP in this embodiment, it is better to control the power generation source to be turned off when the negative side peak of the phase occurs within at least one cycle of the reference electrical state. That is, with In comparison, SSP is preferred. Furthermore, it is known that the plasma electron density Ne is proportional to the ion density or the amount of ions. Therefore, by controlling the generator power to be off at the negative side peak of the phase within at least one cycle of the reference electrical state, the energy and amount of ions can be independently controlled. Specifically, it is known that the ion energy can be controlled by the bias power, and the ion density can be controlled by the generator power, and these two controls are independent of each other.

[0143] [Example 5]

[0144] Next, refer to Figure 10 Example 5 of this embodiment will be described. In Example 5, the power of the generator source is controlled to be off when the voltage of LF is at its negative peak. The bias power was changed. Figure 10 This is a graph showing the time variation of bias power and plasma electron density in Example 5 of one embodiment. The processing conditions for Example 5 are as follows.

[0145] <Processing Conditions>

[0146]

[0147] Comparative example: CW (continuous wave with LF = 50W and HF = 100W)

[0148] Furthermore, the actual power of the source generated by the comparative example "CW" is the same as that of the source generated by the "SSP" of this embodiment with a duty of 40%.

[0149] Figure 10 The horizontal axis represents time, and the vertical axis represents the plasma electron density Ne. Based on the measurement results shown in Figure 10, compared to the comparative example "CW", in... In the case of [condition], the plasma electron density Ne increases when the generator power is turned on and decreases when the generator power is turned off. Furthermore, the higher the bias power, the greater the decrease in plasma electron density Ne when the HF is turned off. Moreover, during the period when the HF is turned off, an increase in plasma electron density Ne occurs later than immediately after the HF is turned off. In the case of T = 2000 ns, LF becomes a negative peak, which is the peak value of the plasma electron density Ne during the HF shutdown period. Regardless of whether HF is turned off, the plasma electron density Ne increases later than the instant of HF shutdown (T = 1000 ns). The reason for this will be explained later, but simply put, it is due to the presence of ions momentarily after HF shutdown. That is, the ions present are accelerated within the sheath and collide with the substrate, thereby releasing secondary electrons, which collide with the gas to further generate ions. Through this phenomenon, the peak value of the ion energy on the high-energy side can be increased. Next, the processing conditions for increasing the peak value of the ion energy on the high-energy side will be explained in Example 6.

[0150] [Example 6]

[0151] Reference Figure 11 and Figure 12 Example 6 of this embodiment will be described. Figure 11 This is a diagram illustrating the pulse application timing of the source power in Embodiment 6 of an implementation method. Figure 12 It means and Figure 11 A graph showing the measurement results of the ion energy distribution corresponding to the pulse application time of the periodic source power. In Example 6, for Set the on and off cycle of the generated source power to be 1 or 2 times the cycle of the high-frequency bias power.

[0152] exist Figure 11In the case of "CW" in the comparative example shown in (a), HF is always on, and the traveling wave (Pf) and reflected wave (Pr) of HF are continuously output. In this case, the substrate potential uses the voltage of LF as a reference and oscillates with the voltage of HF.

[0153] Figure 11 (b) and (c) represent Example 6 of this embodiment. The case where HF is applied. Figure 11 In (b) and (c), HF is repeatedly switched on and off with a duty cycle of 40%. Figure 11 In (b), the on and off periods of the generated source power are the same as the high-frequency period of the bias power, which is 400 kHz. Figure 11 In (c), the on and off period of the generated source power is twice the period of the high frequency of the bias power, which is 800kHz.

[0154] The processing conditions in Example 6 are as follows.

[0155] <Processing Conditions>

[0156]

[0157]

[0158] Comparative example: CW (continuous wave with LF = 50W and HF = 100W)

[0159] Furthermore, the actual power of the source generated by the comparative example "CW" is the same as that of the source generated by the "SSP" of this embodiment with a duty of 40%.

[0160] Figure 12 In the measurement results, CW (400kHz) is... Figure 11 The HF applied, as shown in (a), corresponds to the measurement results. SSP (400 kHz, Duty = 40%) is compared with... Figure 11 The measurement results corresponding to the HF application shown in (b) are the same as those corresponding to the HF application shown in (c) of Figure 11.

[0161] Figure 12The horizontal axis represents ion energy, and the vertical axis represents the ion energy distribution function (IEDF). In the case of SSP (800 kHz, Duty = 40%), there are two peaks in the ion energy distribution: one at low energy and one at high energy, and the number of ions with energies between these two is reduced. Furthermore, in the case of SSP (400 kHz, Duty = 40%), no peak is generated near 180 eV. Moreover, both SSP (800 kHz, Duty = 40%) and SSP (400 kHz, Duty = 40%) produce a peak in the ion energy distribution on the high-energy side compared to the comparative example "CW". Based on the above, under SSP control, the number of high-energy ions reaching the substrate is increased compared to the comparative example. As can be seen from the above, in SSP control, the on and off cycle of the generator power is controlled to be approximately twice the cycle of the high-frequency bias power. When the voltage of LF is at a negative peak, the generator power is controlled to be off, thereby increasing the number of high-energy ions reaching the substrate.

[0162] [Example 7]

[0163] Below, refer to Figure 13 and Figure 14 Example 7 of this embodiment will be described. FIG13 is a diagram illustrating the phase of the pulse application time of the source power in Example 7 of this embodiment. Figure 14 It means and Figure 13 The figure shows an example of the measurement results of the ion energy distribution corresponding to the phase at the moment of pulse application of the source power. In Example 7, the ion energy was measured by shifting the phase from 0° at the peak when LF was positive at the moment the HF pulse was applied (the middle of the pulse period). Figure 14 The phase that controls the power of the generated source to be on is shown. Control it relative to Offsets of 60°, 90°, 105°, 120°, and 150° were performed, and the results of IEDF were measured.

[0164] Figure 13 (a) indicates The waveforms of HF and LF, and the substrate potential. Figure 13 (b) indicates The waveforms of HF and LF, and the substrate potential.

[0165] The processing conditions for Example 7 are as follows.

[0166] <Processing Conditions>

[0167]

[0168] Make phase Compared to The optimal ion energy distribution was measured at offsets of 60°, 90°, 105°, 120°, and 150°. Furthermore, the optimal ion energy distribution refers to a distribution with two peaks—low energy and high energy—and a small number of ions with energies in between. It is desirable to more clearly and distinctly represent the high-energy peak.

[0169] according to Figure 14 The measurement results, in the ion energy distribution function, when are The peak value on the high-energy side widens slightly when it is The peak value on the high-energy side becomes quite wide. That is, it can be seen that when the peak value of the high-energy side reaches the negative maximum value of the LF bias during the HF pulse on-time, the peak value on the high-energy side becomes wider, and the height of the high-energy peak value also becomes lower. Based on this result, it is preferable to turn off the HF when the LF is near its negative maximum value. Furthermore, when the voltage of the LF changes from negative to positive, if the generator power is controlled to be on and the duty cycle of the generator power is less than 50%, it is possible to turn off the HF when the LF is near its negative maximum value.

[0170] [Example 8]

[0171] Below, refer to Figure 15 and Figure 16 Example 8 of this embodiment will be described. In Example 8, the duty cycle of the generated source power is variably controlled. Figure 15 This is a graph illustrating the relationship between the source power and duty cycle in Example 8 of an embodiment. Figure 16 It means Figure 15 The figure shows one example of the measurement results of the ion energy distribution based on the pulse duty cycle of the generated source power.

[0172] Figure 15 (a) indicates Furthermore, the waveforms of HF and LF, and the substrate potential when the duty cycle of the generated source power (HF) is 60%. Figure 15 (b) indicates Furthermore, the waveforms of HF and LF, and the substrate potential when the duty cycle of the generated source power is 20%.

[0173] The processing conditions for Example 8 are as follows.

[0174] <Processing Conditions>

[0175]

[0176]

[0177] exist In this case, the duty cycle of the generated source power pulse was set to 20%, 40%, 50%, 60%, and 80%, and the optimal ion energy distribution was measured at which duty cycle level was observed. According to... Figure 16 The measurement results, in When the duty cycle of the source power is reduced to less than 40%, the peak value on the high-energy side of the ion energy distribution becomes lower. This is because, under certain conditions... When the duty cycle of the power generation pulse is controlled to less than 40%, the on-time of the power generation pulse becomes shorter, and the pulse will be turned off before enough ions are generated, thus reducing the overall amount of ions reaching the substrate.

[0178] Therefore, it can be seen that when the duty cycle of the power generation pulse is controlled to less than 40%, the number of high-energy ions reaching the substrate decreases. Therefore, it is preferable to control the duty cycle of the power generation pulse to be above 40%.

[0179] [Example 9]

[0180] Below, refer to Figure 17 and Figure 18 Example 9 of this embodiment will be described. FIG17 is a diagram illustrating the phase of the pulse application time of the source power in Example 9 of this embodiment. Figure 18 It means based on Figure 17 The graph shows the measurement results of the phase and ion energy distribution at the moment the pulse of the power source is applied. In Example 9, the phase of HF is controlled to... and

[0181] Figure 17 (a) indicates in The duty cycle of the pulse generating the source power is 40%, and the on and off cycle of the source power is set to twice the high-frequency cycle of the bias power, i.e., 800kHz. The waveforms of HF and LF and the substrate potential are then analyzed.

[0182] Figure 17 (b) indicates that in The waveforms of HF and LF, and the substrate potential, are as follows: the duty cycle of the pulse generating the source power is 20%, and the on and off period of the source power is set to 400kHz, which is the same as the period of the high frequency of the bias power.

[0183] Figure 17 (c) indicates that in The waveforms of HF and LF, and the substrate potential, are as follows: the duty cycle of the pulse generating the source power is 20%, and the on and off period of the source power is set to 400kHz, which is the same as the period of the high frequency of the bias power.

[0184] The processing conditions for Example 9 are as follows.

[0185] <Processing Conditions>

[0186]

[0187]

[0188] according to Figure 18 The measurement results shown are: (a) SSP (800 kHz) and The case of 270° is the same as (b) SSP (400KHz) and The situation is roughly the same as the IEDF distribution, therefore it is determined that... The generation source power pulse phase contributes almost nothing to the arrival of ions on the substrate. In (c)SSP (400kHz) and In this case, the amount of ions reaching the substrate is significantly lower. Based on this result, it can be concluded that... When the power pulse phase of the generation source is applied in HF, the efficiency of ions reaching the substrate is the highest.

[0189] [The switching on and off of HF and the movement of ions and electrons]

[0190] like Figure 10 As shown, the plasma electron density Ne decreases when the HF pulse is turned off. However, the temporary decrease in electron density Ne begins to rise from around T = 1500 ns, reaching a peak around T = 2000–2250 ns. (Refer to...) Figure 19 The reasons for the increase in plasma electron density Ne are explained. Figure 19 This is a diagram used to illustrate the thickness of the sheath, the switching on and off of HF, and the movement of ions and electrons in one embodiment.

[0191] The thickness of the sheath is determined by the potential difference between the plasma potential and the substrate potential based on the bias voltage. Therefore, the sheath thickness varies approximately according to the bias period. The sheath is thickest when the LF bias is at its negative maximum value, and conversely, the sheath is thinnest during the period when the LF bias potential is positive.

[0192] The voltage of LF is Figure 19 (a) represents the state with the maximum peak value in the positive direction. It then gradually deepened into a negative state. Figure 19 The vicinity of (c) becomes exist Figure 19 (d) becomes the state with the maximum peak value in the negative direction. exist Figure 19 of (a) At that time, the power source will be switched on for a period corresponding to the duty cycle of HF. Figure 19 Between (b) and (d), the power of the generated source will be turned off.

[0193] exist Figure 19 During the period when the generator source power is turned on in (a), plasma is generated, and no electric field is applied to the electrons and ions present in the plasma. Afterwards, when the generator source power is turned off, the LF bias increases in the negative direction during the LF bias period, therefore... Figure 19 As shown in (b) to (d), the sheath S thickens accordingly.

[0194] like Figure 19 As shown in (b), due to the increased sheath thickness, when an electric field E is applied to electrons and ions, the lighter electrons accelerate towards the plasma P direction. Ions, being heavier than electrons, accelerate towards the substrate W direction with a delayed motion compared to electrons. The delayed-moving ions collide with the already-moving electrons, neutralize them, and disappear. The result is as follows... Figure 19 As shown in (c), the number of ions and electrons decreases.

[0195] Then, high-speed ions accelerated towards the substrate W collide with the substrate W, releasing secondary electrons that are accelerated towards the plasma P by the electric field. These secondary electrons, accelerated by the electric field and possessing high energy, collide with the gas, generating ions and electrons. This phenomenon can be considered a gamma-mode DC discharge.

[0196] Therefore, as Figure 19 As shown in (d), new ions and electrons can be generated regardless of whether the power of the generation source is turned off. Through this phenomenon, although... Figure 10 As shown, when the source power is turned off (T = 1000 ns), the electron density decreases, but then the plasma electron density Ne increases, and a large number of high-energy ions reach the substrate. This phenomenon can increase the number of high-energy ions reaching the substrate, thereby improving the ion generation efficiency.

[0197] On the other hand, when Figure 19 of (e) When the generator power is turned on at 180°, such as Figure 19 As shown in (f) to (h), the sheath S becomes thinner. However, when Figure 19 of (e) When it is 180°, with Figure 19 of (a) Compared to 0°, the number of ions is less. Therefore, when the generation source power is off... Figure 19In (f) to (h), the number of ions accelerated towards the substrate is small, and the sheath S gradually thins, thus suppressing ion acceleration towards the substrate. Therefore, when the generation source power is turned off... Figure 19 In (f) to (h), no power source that will be turned off will be generated. Figure 19 The ion → secondary electron → ion → secondary electron cycle in (b) to (d) makes it difficult to improve the ion generation efficiency.

[0198] As can be seen from the above, by controlling the time when the power source is turned on and the time from when the power source is turned off until the bias power reaches its negative maximum value, the energy of the ions and the amount of ions corresponding to the plasma electron density can be controlled independently, thereby improving the ion generation efficiency.

[0199] For example, this can be achieved when the sheath S is thinner, such as Figure 19 As shown in (a), the generator power is turned on, and then the voltage of the LF is applied in a deeper negative direction while the generator power is turned off. This promotes the generation of ions and electrons. Figure 19 The actions shown in (b) to (d) improve the ion generation efficiency. This increases the number of high-energy ions and enhances the plasma electron density Ne.

[0200] In particular, since the ion generation efficiency is increased by releasing secondary electrons from accelerated ions, the time for controlling the generation source power to be off is preferably less than or equal to the time during which the ion density generated when the generation source power is on is above a certain threshold. That is, it is preferable to control the generation source power to be off during the period of ion residue after the generation source power is turned off.

[0201] Reference Figure 20 The explanation covers the period of ion residue after the power of the generation source is turned off. Figure 20 It is a diagram used to illustrate the release of secondary electrons from high-energy ions. Figure 20 The source is [AV Phelps and Z Lj Petrovic] PlasmaSources Sci. Technol. 8 (1999) R21-R44. Printed in the UK.

[0202] Figure 20The horizontal axis represents the ion energy, and the vertical axis represents the number of secondary electrons released during ion collisions. Accordingly, when a 1000 eV Ar ion collides with the substrate, approximately 0.5 secondary electrons are released. As the ion energy increases, the number of released secondary electrons increases dramatically. This leads to further ion generation through collisions with the gas in the plasma processing space, further increasing the ion count. Therefore, by controlling the time the generation source power is turned off, the ion density generated when the generation source power is on is kept below a predetermined level for a specified period. Thus, regardless of whether the generation source power is turned off, a generation cycle of ion generation → secondary electron release → ion generation → secondary electron release can be formed solely through the bias power energy. Therefore, the voltage of the LF can be applied in a deeper negative direction when the generation source power is off.

[0203] Therefore, monochromaticization of ion energy on the high-energy side can be achieved. Furthermore, monochromaticization of ion energy on the low-energy side can be achieved. Additionally, the ion energy distribution can be controlled to have two peaks, one low-energy and one high-energy, with a small number of ions having ions with energies between these two peaks.

[0204] [Variation Example]

[0205] Below, refer to Figure 21 The timing of applying the bias power and the generation source power in a modified embodiment will be explained. Figure 21 This is a diagram showing the pulse application times of the bias power and the generation source power in a modified embodiment of one implementation.

[0206] In embodiments 1 to 10 above, a source power with a specified duty cycle is applied to the bias power at a specified time. Figure 21 In the modified example, the phase with the bias power is The power source is switched on synchronously. The time t1 when the power source is switched on is taken as the first state, and the time t2 when the power source is switched off is taken as the second state. The processing method of the modified example includes a first control step of alternately applying the first state and the second state synchronously with the phase within one cycle of LF.

[0207] In addition to the first control step, the modified method includes a second control step that intermittently stops the bias power and the generated source power. The second control step intermittently stops the generated source power and the bias power at a period independent of the period of the reference electrical state using the LF voltage as an example. The first and second control steps are executed repeatedly. In the modified example, the switching period of the HF is the same as or an integer multiple of the period of the LF.

[0208] When the state of the first control step is taken as the third state and the state of the second control step is taken as the fourth state, the ratio of the fourth state to (third state + fourth state) of the generated source power in the first and second control steps can be within the range of 1% to 90%.

[0209] exist Figure 22 In the example shown in (a), the DC voltage applied to the upper electrode is controlled to a high level (High) with an absolute value greater than the low level (Low) in the first control step, and to a low level with an absolute value less than the high level (High) in the second control step. The actions of electrons and ions in this case will be explained in two time periods, "A" and "B". "A" is the time when a positive bias is applied to the substrate after the HF is turned off. "B" is the time when a negative bias is applied to the substrate after the HF is turned off.

[0210] Figure 23 This graph represents the time from RF shutdown to the disappearance of electrons. The source is: T. Tsutsumi et al., Dry Process Symposium 2018E-2. The horizontal axis represents the time from RF shutdown, and the vertical axis represents the electron density (= ion density). Squares represent the case of applying -1000V RF, circles represent the case of applying -800V RF, and triangles represent the case of applying -600V RF. Based on this, approximately half of the ions remain 5 μs after HF shutdown. Approximately one-fifth of the ions remain 10 μs after HF shutdown.

[0211] For example, when the generator power is repeatedly switched on and off with a period of 2.5 μs, there is a higher ion residue during the generator power off period. When the DC voltage is controlled to a high level and applied to the upper electrode 34 during the "A" period of ion residue, as... Figure 22 As shown in (b), ions are accelerated towards the upper electrode 34. After colliding with the upper electrode 34, the accelerated ions release secondary electrons from the upper electrode 34. Figure 22 As shown in (a), the voltage of LF is positive during time period "A" and negative during time period "B". Therefore, during time period "A", the secondary electrons released from the upper electrode 34... Figure 22 As shown in (b), the lower electrode (substrate W side) with a positive bias is accelerated. As a result, during the time period “A”, the acceleration is achieved by… Figure 22 The graph in (b) shows that the vertical axis, represented by the Electron Energy Probability Function (EEPF), indicates the arrival of high-energy E1 electrons on the substrate W. Furthermore, the upper electrode is ion-sputtered, and the silicon on the surface of the upper electrode 34 reaches the substrate W.

[0212] In contrast, "B" leaves almost no residual ions during its time period. Therefore, as... Figure 22 As shown in (c), there are almost no ions accelerated towards the upper electrode 34. Furthermore, B indicates the state when LF is negative. Therefore, at time "B", fewer secondary electrons are released from the upper electrode 34. In addition, during time "B", the voltage of LF is negative, thus a negative bias is applied to the lower electrode (substrate W). Therefore, secondary electrons released from the upper electrode 34 heading towards the substrate side are reversed due to the negative electric field near the substrate W. As a result, during time "B", as... Figure 22 As shown in diagram (c), high-energy electrons do not reach the substrate W.

[0213] [The switching on and off of the power source and the number of electrons reaching the substrate]

[0214] Below, refer to Figures 24-26 The switching on and off of the power source and the number of electrons reaching the substrate W are explained. Figure 24 This is a diagram schematically representing the switching on and off of the generator source power and the number of electrons reaching the substrate in the comparative example. Figure 25 This is a diagram schematically illustrating the switching on and off of the power source and the number of electrons reaching the substrate in one embodiment. Figure 26 This is a diagram used to illustrate the timing of the power generation source being switched on and off in one embodiment and its effects.

[0215] exist Figure 24 In the comparative example shown, during the period when the LF is on, the absolute value of the DC voltage is controlled to a low negative level, and HF is continuously applied. During this period, due to the generation of plasma, secondary electrons collide with the plasma and hardly reach the substrate. That is, as shown in Figure 24, the number of electrons reaching the substrate is almost zero during the period when the LF is on.

[0216] During the LF turn-off period, the absolute value of the DC voltage is controlled to a negative high level, and no HF is applied. During this period, since no plasma is generated, during the ion residue period after LF turn-off, ions are accelerated towards the upper electrode 34 and collide with it, releasing secondary electrons that reach the substrate. Thus, as... Figure 24 As shown, during the LF turn-off period, electrons only reach the substrate during the period of ion residue of approximately 5 μs to 10 μs after the LF is turned off. At other times, electrons do not reach the substrate and become useless time.

[0217] In contrast, Figure 25In the case of the SSP of this embodiment shown, during the LF-on period, the absolute value of the DC voltage is controlled to a low negative value, and HF is applied synchronously with the phase of LF in a manner that controls HF to be off at least during the negative side peak of the phase of one cycle of the aforementioned reference electrical state. During the HF-on period, plasma is generated, and secondary electrons collide with the plasma without reaching the substrate. However, during the HF-off period, ions remain, so the ions are accelerated towards the upper electrode 34, collide with the upper electrode 34, and release secondary electrons that reach the substrate.

[0218] During the LF turn-off period, the absolute value of the DC voltage is controlled to a negative high level, and no HF is applied. During this period, no plasma is generated; therefore, during the ion residue period after LF turn-off, ions are accelerated towards the upper electrode 34, colliding with the upper electrode 34 and releasing secondary electrons, which then reach the substrate. Thus, as... Figure 25 As shown, during the LF shutdown period, electrons only reach the substrate during a period of ion residue of approximately 5 μs to 10 μs from the LF shutdown. Furthermore, ion residue occurs during the LF on and HF off period, thus electrons reach the substrate. Based on the above, in the SSP of this embodiment, the number of electrons reaching the substrate increases, the plasma electron density is higher, the number of ions increases, and the ion generation efficiency is improved compared to the comparative example.

[0219] Figure 26 (a) represents an example of etching in the comparative case. Figure 26 (b) illustrates an example of etching in the case of SSP in this embodiment. When the film 2 to be etched is etched according to the pattern of the holes H formed on the mask 3, ions and electrons reaching the substrate are incident into the holes H. Figure 26 In the comparative example (a), electrons, being lighter than ions, adhere to the sidewalls of pore H or fail to reach the bottom of pore H. Consequently, the bottom of pore H is positively charged by ions with high propagation velocity. Then, ions incident into pore H are obliquely incident due to the charging at the bottom of the pore, preventing them from reaching the bottom of pore H. As a result, as... Figure 26 As shown in (a), the perpendicularity of the etched shape of hole H deteriorates.

[0220] In contrast, Figure 26 In the case of the SSP of this embodiment shown in (b), as Figure 25 The number of electrons reaching the substrate is greater than in the comparative example. Furthermore, the HF is repeatedly switched on and off during the LF switching period. In this case, in the case of the SSP of this embodiment, as... Figure 26As shown in (b), the process of accelerating electrons and then accelerating ions to reach the substrate is repeated. This cancels the charging process before the bottom of hole H is charged. Therefore, the perpendicularity of the etched shape of hole H can be maintained.

[0221] [Effect]

[0222] Finally, the configuration and effects of this embodiment and its variations are summarized. First, when controlling the SSP, the control unit 200 controls the generation source power to be turned off at the negative side peak of the phase within at least one cycle of the reference electrical state. This SSP control involves intermittently applying the generation source power at predetermined times, synchronized with the bias power. Therefore, the distribution of ion energy can be controlled. Furthermore, the in-plane distribution of plasma electron density can be controlled. Moreover, plasma electron density and ion energy can be controlled independently.

[0223] In addition, such as Figure 25 As shown, the number of electrons reaching the substrate with the same average power increases, thus achieving an effective plasma electron density. Furthermore, monochromaticity of the ion energy on both the high-energy and low-energy sides can be achieved. Moreover, the decrease in plasma electron density can be suppressed when the voltage of the LF is negative.

[0224] Furthermore, the control unit 200 controls the time during which the generation source power is turned off, ensuring that the ion density generated when the generation source power is turned on is within a specified timeframe. In other words, the control unit 200 controls the time during which the generation source power is turned off to be shorter than the time during which the ions generated when the generation source power is turned on disappear. Therefore, regardless of whether the generation source power is turned off, the plasma electron density (ion density) can be further increased immediately after the generation source power is turned off.

[0225] Furthermore, when the generator source power is turned on, the substrate potential oscillates with the amplitude of the generator source power. Correspondingly, lighter electrons move first, followed by heavier ions. This results in ions colliding with electrons and becoming neutralized, thus reducing the number of ions reaching the substrate. In other words, when the generator source power is turned on, fewer high-energy ions reach the substrate. However, in this embodiment and its variations, the control unit 200 intermittently applies the generator source power in sync with the bias power, and controls the generator source power to be turned off at the peak of the phase within at least one cycle of the reference electrical state. Therefore, compared to the case of continuously applying the generator source power, the number of high-energy ions reaching the substrate can be increased.

[0226] In addition, the plasma generation source used to generate plasma is generated by supplying the generation source power of a high-frequency power source to the plasma processing space. However, it is not limited to this. It can also be generated by supplying the generation source power of other power sources such as microwave sources to the plasma processing space.

[0227] The plasma processing apparatus and method disclosed in this invention are illustrative in all respects and should not be considered limiting. The above embodiments can be modified and improved in various ways without departing from the scope and spirit of the appended claims. Other configurations can be added to the content described in the above embodiments without contradiction, and they can be combined without contradiction.

[0228] The plasma processing apparatus of the present invention is applicable to any of the following types of devices: Atomic Layer Deposition (ALD), Capacitively Coupled Plasma (CCP), Inductively Coupled Plasma (ICP), Radial Line Slot Antenna (RLSA), Electron Cyclotron Resonance Plasma (ECR), and Helicon Wave Plasma (HWP).

Claims

1. A plasma processing apparatus comprising: a first electrode on which a substrate is placed; a plasma generating source for generating plasma; a bias power supply for supplying a bias power to the first electrode; a generating source power supply for supplying a generating source power having a higher frequency than the bias power to the plasma generating source; and a control section that controls the bias power supply and the generating source power supply, the plasma processing apparatus characterized by: the generating source power having a first state and a second state, the control section controlling so that the first state and the second state are alternately applied in synchronization with a phase within one cycle of a reference electrical state, the reference electrical state indicating any one of a signal synchronized with a cycle of a frequency of the bias power, or a voltage, a current, or an electromagnetic field measured in a power supply system of the bias power, and the control section controlling the generating source power to be off at a negative side peak of the phase within at least one cycle of the reference electrical state, the control section making a time during which the generating source power is controlled to be off be less than a time during which an ion density generated when the generating source power is on is above a prescribed value, the control section making a cycle of on and off of the generating source power be twice a cycle of the frequency of the bias power, and controlling the generating source power to be off at a negative peak of the phase within one cycle of the reference electrical state.

2. The plasma processing apparatus according to claim 1, characterized by: the control section controlling the generating source power to be on at a time when the phase within one cycle of the reference electrical state changes from negative to positive.

3. The plasma processing apparatus according to claim 1 or 2, characterized by: the control section controlling a duty ratio of the generating source power to be 40% or more.

4. The plasma processing apparatus according to claim 1 or 2, characterized by: the control section intermittently stopping the bias power and the generating source power.

5. The plasma processing apparatus according to claim 1 or 2, characterized by: a value of the generating source power in the second state being 0.

6. A processing method implemented in a plasma processing apparatus, wherein: the plasma processing apparatus comprises: a first electrode on which a substrate is placed; a plasma generating source for generating plasma; a bias power supply for supplying a bias power to the first electrode; and a generating source power supply for supplying a generating source power having a higher frequency than the bias power to the plasma generating source, the generating source power having a first state and a second state, the processing method characterized by: alternately applying the first state and the second state in synchronization with a phase within one cycle of a reference electrical state, the reference electrical state indicating any one of a signal synchronized with a cycle of a frequency of the bias power, or a voltage, a current, or an electromagnetic field measured in a power supply system of the bias power, controlling the generating source power to be off at a negative side peak of the phase within at least one cycle of the reference electrical state, the generation source power is controlled to be turned off at a time when the ion density generated at the time when the generation source power is turned on is below a prescribed value, the period of turning on and off of the generation source power is made twice the period of the frequency of the bias power, and the generation source power is controlled to be turned off at a time of a negative peak of the phase within one period of the reference electrical state.

Images