Measuring apparatus, film deposition apparatus, and film thickness measurement method

Abstract

To provide a measuring device, a film-forming device, and a film thickness measurement method capable of excluding or suppressing an abnormal value and improving responsiveness of the processed value relative to a true value.SOLUTION: A measuring device measures a film-forming rate on the basis of an oscillation frequency of an oscillator installed in a film-forming device having a deposition source, and includes a measurement circuit and a control circuit. The measurement circuit calculates a rate conversion value for each unit time on the basis of frequency-related information related to the oscillation frequency of the oscillator. The control circuit includes: a filtering portion for filtering a rate conversion value calculated by the measurement circuit; and a control portion for changing a filter coefficient as a degree of strength of filtering in the filtering portion on the basis of the temperature-related information related to the temperature of the deposition source.SELECTED DRAWING: Figure 6

Description

[Technical Field]

[0001] The present invention relates to a measuring device for measuring the film deposition rate based on the oscillation frequency of an oscillator installed in a film deposition apparatus, a film deposition apparatus equipped therewith, and a method for measuring film thickness. [Background technology]

[0002] Conventionally, in thin-film deposition equipment such as vacuum deposition systems, a technique called the quartz crystal microbalance (QCM) method has been used to measure the thickness and deposition rate of films deposited on a substrate. This method measures film thickness and rate by utilizing the deposition of material onto the surface of a quartz crystal oscillator placed in a chamber. This utilizes the physical phenomenon that as deposition, i.e., surface mass, increases, the resonant frequency of the quartz crystal oscillator decreases. Therefore, by observing (or using) the change in the resonant frequency of the quartz crystal oscillator, the mass of the film can be identified.

[0003] In a vapor deposition apparatus equipped with a film thickness sensor, the heating temperature of the deposition material at the deposition source is feedback-controlled based on the measured deposition rate. However, the film thickness sensor is affected by physical disturbances such as bumping (splashing) of the deposition material and changes in the surface state of the quartz oscillator 140, which can cause the observed value to fluctuate significantly instantaneously. Largely fluctuating values ​​do not reflect the actual deposited mass (true value), and therefore these values ​​must be excluded as outliers (or suppressed to within the normal range).

[0004] Conventional measuring devices that have addressed this problem perform a process called filtering on the observed values ​​and output the processed values ​​as the true values. One example of filtering is a smoothing method. For example, Patent Document 1 describes observing the resonant frequency of a quartz crystal oscillator at a certain unit time interval and outputting the increase in film thickness per unit time (rate) by taking a moving average over a certain time range of the film thickness calculated based on these observed values. It is claimed that a stable film deposition method (stable feedback control) can be realized by using the output values ​​obtained in this way.

[0005] A typical example of a smoothing method, which is an example of filtering, involves sampling observed values ​​obtained from the present to a predetermined time and applying a moving average to this sample. Generally, a simple moving average is used, but a weighted moving average may also be used. Another example is a method that uses a delay element called a low-pass filter (e.g., a first-order lag filter) to suppress outliers (outliers are suppressed because they are high-frequency components). Furthermore, processing methods that use multiple of these elements, or that involve the interaction of elements, are also known.

[0006] The measuring device using these methods could process or suppress abnormal values, which were thought to be due to noise, into the expected normal range, even if they were present in the observed values, thus stabilizing the control of the deposition source using its output value. However, the output value did not always match the true value, and in particular, response delays were a problem when the true rate value changed rapidly over time. [Prior art documents] [Patent Documents]

[0007] [Patent Document 1] Patent No. 6060319 [Overview of the project] [Problems that the invention aims to solve]

[0008] Conventional smoothing methods using moving averages fail to resolve trade-off problems. Typically, increasing the number of samples (increasing the predetermined time) yields more stable processed values, but the response performance per unit time deteriorates. Furthermore, long-period noise may be introduced depending on the surface condition of the quartz oscillator, and removing this requires a sample size of at least one period, making it unsuitable from the perspective of controlling the deposition source (i.e., in terms of production efficiency of the film to be deposited). Thus, there is a challenge in achieving both responsiveness and stability simultaneously.

[0009] In view of the above circumstances, the object of the present invention is to provide a measuring device, a film deposition device, and a film thickness measurement method that can eliminate or suppress abnormal values ​​and improve the responsiveness of the processed value to the true value. [Means for solving the problem]

[0010] A measuring device according to one embodiment of the present invention is a measuring device for measuring the film deposition rate based on the oscillation frequency of an oscillator installed in a film deposition apparatus having a deposition source, and comprises a measuring circuit and a control circuit. The measurement circuit calculates a rate-converted value per unit time based on frequency-related information related to the oscillation frequency of the oscillator. The control circuit includes a filter unit that filters the rate conversion value calculated by the measurement circuit, and a control unit that changes the filter coefficient, which is the degree of filtering strength in the filter unit, based on temperature-related information related to the temperature of the deposition source.

[0011] The control circuit may further include a calculation unit that calculates a first approximate line corresponding to the time change of the temperature-related information. The control unit may determine whether the absolute value of the slope of the first approximate line is greater than or equal to a predetermined value, and if the absolute value of the slope of the first approximate line is greater than or equal to the predetermined value, it may set the filter coefficient to a first set value, and if the absolute value of the slope of the first approximate line is less than the predetermined value, it may switch the filter coefficient to a second set value that is greater than the first set value.

[0012] The calculation unit may further calculate a second approximate line corresponding to the time variation of the real part of the admittance in the oscillator based on the frequency-related information. The control unit may further determine whether the slope of the second approximate line is within a predetermined range if the absolute value of the slope of the first approximate line is less than the predetermined value, and may switch the filter coefficient to a third setting value that is greater than the second setting value if the slope of the second approximate line is outside the predetermined range.

[0013] The second predetermined range is a range that extends between a negative threshold and a positive threshold, and the control unit may switch the filter coefficient from the third set value to the second set value after a predetermined time has elapsed when the slope of the second approximation line changes from a value less than the negative threshold to a value greater than or equal to the negative threshold.

[0014] The aforementioned filter section may include a Kalman filter.

[0015] The temperature-related information may be the temperature of the deposition source or the power used to heat the deposition source.

[0016] The calculation unit may perform preprocessing to smooth the rate conversion value calculated by the measurement circuit.

[0017] The aforementioned preprocessing may also be a median calculation process that calculates the median of the rate conversion values ​​for each unit time. Alternatively, the preprocessing may be a moving average calculation process that calculates the moving average of the rate-converted values ​​for each unit time.

[0018] Another embodiment of the present invention is a measuring device for measuring the film deposition rate based on the oscillation frequency of an oscillator installed in a film deposition apparatus having a deposition source, and comprises a measuring circuit and a control circuit. The measurement circuit calculates a rate-converted value per unit time based on frequency-related information related to the oscillation frequency of the oscillator. The control circuit includes a filter unit that filters the rate conversion value calculated by the measurement circuit, and a control unit that changes a filter coefficient, which is the degree of filtering strength in the filter unit, based on the time change of the real part of the admittance in the oscillator.

[0019] A film forming apparatus according to an aspect of the present invention includes a vacuum chamber, a deposition source, a sensor head, a measurement circuit, and a control circuit. The deposition source is disposed inside the vacuum chamber. The sensor head is disposed inside the vacuum chamber and has an oscillator that oscillates at a predetermined resonance frequency. The measurement circuit calculates a rate conversion value per unit time based on frequency-related information related to the oscillation frequency of the oscillator. The control circuit includes a filter unit that filters the rate conversion value calculated by the measurement circuit, and a control unit that changes a filter coefficient, which is the degree of filtering strength in the filter unit, based on the frequency-related information and temperature-related information related to the temperature of the deposition source.

[0020] Another film forming apparatus according to an aspect of the present invention includes a vacuum chamber, a deposition source, a sensor head, a measurement circuit, and a control circuit. The deposition source is disposed inside the vacuum chamber. The sensor head is disposed inside the vacuum chamber and has an oscillator that oscillates at a predetermined resonance frequency. The measurement circuit calculates a rate conversion value per unit time based on frequency-related information related to the oscillation frequency of the oscillator. The control circuit includes a filter unit that filters the rate conversion value calculated by the measurement circuit, and a control unit that changes a filter coefficient, which is the degree of filtering strength in the filter unit, based on the time change of the real part of the admittance in the oscillator.

[0021] A film thickness measurement method according to an aspect of the present invention is Frequency-related information relating to the oscillation frequency of the oscillator installed in the film deposition apparatus and temperature-related information relating to the temperature of the deposition source installed in the film deposition apparatus are acquired. Based on frequency-related information related to the oscillation frequency of the oscillator, a rate-converted value for each unit time is calculated. The rate conversion value calculated by the measurement circuit is filtered, The filter coefficient, which is the degree of filtering strength in the filter section, is changed based on the frequency-related information and temperature-related information related to the temperature of the deposition source.

[0022] Another embodiment of the present invention provides a method for measuring film thickness, which acquires frequency-related information relating to the oscillation frequency of a vibrator installed in a film deposition apparatus and temperature-related information relating to the temperature of a deposition source installed in the film deposition apparatus. Based on frequency-related information related to the oscillation frequency of the oscillator, a rate-converted value for each unit time is calculated. The rate conversion value calculated by the measurement circuit is filtered, The filter coefficient, which represents the degree of filtering strength in the filter section, is changed based on the time evolution of the real part of the admittance in the oscillator. [Effects of the Invention]

[0023] According to the present invention, it is possible to eliminate or suppress outliers and improve the responsiveness of the processed value to the true value. [Brief explanation of the drawing]

[0024] [Figure 1] This is a schematic cross-sectional view showing a film deposition apparatus equipped with a measuring device according to one embodiment of the present invention. [Figure 2] This is a block diagram showing one example configuration of the above measuring device. [Figure 3] This is a typical equivalent circuit for a sensor head. [Figure 4] This diagram illustrates the full width at half maximum (FWHM) of the resonant frequency. [Figure 5]This is a functional block diagram of the control circuit in the above measuring device. [Figure 6] This is a block diagram illustrating the procedure for the processing performed in the above control circuit. [Figure 7] This is a flowchart illustrating the procedure for the processing performed in the control circuit described above. [Figure 8] This figure shows the experimental results for Comparative Example 1. [Figure 9] These are experimental results explaining the operation of the above measuring device. [Figure 10] This figure shows the experimental results for Comparative Example 1. [Figure 11] These are experimental results explaining the operation of the above measuring device. [Figure 12] These are experimental results explaining the operation of the above measuring device. [Figure 13] These are experimental results explaining the operation of the above measuring device. [Figure 14] This block diagram shows another example configuration of the above control circuit. [Figure 15] This is a block diagram showing yet another example of the control circuit described above. [Modes for carrying out the invention]

[0025] Embodiments of the present invention will be described below with reference to the drawings.

[0026] Figure 1 is a schematic cross-sectional view showing a film deposition apparatus equipped with a measuring device according to this embodiment. In this embodiment, a vacuum deposition apparatus will be used as an example for the film deposition apparatus 10. First, the basic configuration of the film deposition apparatus 10 will be described.

[0027] [Film forming equipment] The film deposition apparatus 10 includes a vacuum chamber 11, a deposition source 12 located inside the vacuum chamber 11, a stage 13 facing the deposition source 12, a sensor head 14 located inside the vacuum chamber 11 as a measurement point, a vacuum pump 15 for maintaining a predetermined vacuum atmosphere inside the vacuum chamber 11, and a measurement unit 17 configured to measure the film deposition rate based on the output of the sensor head 14 and to control the deposition source 12. The sensor head 14 and the measurement unit 17 constitute a film thickness sensor 20 (measuring device).

[0028] The deposition source 12 is configured to generate vapor (particles) of the film-forming material. In this embodiment, the deposition source 12 is electrically connected to the power supply unit 18 and constitutes an evaporation source that heats and evaporates the film-forming material to release deposited particles. The type of evaporation source is not particularly limited; in this embodiment, a resistance heating type evaporation source is used, but various other types of evaporation sources such as induction heating type and electron beam heating type can also be applied. Alternatively, in the case of a sputtering deposition apparatus, a sputtering cathode including a sputtering target can be used as the deposition source. The film-forming material may be an organic material, a metallic material, a metallic compound material (e.g., a metal oxide, a metal nitride, a metal carbide, etc.).

[0029] The deposition source 12 is equipped with a temperature sensor 12a for detecting the heating temperature of the film deposition material. In the case of an evaporation source, the temperature sensor 12a is attached to the crucible containing the film deposition material. The output of the temperature sensor 12a is supplied to the measurement unit 17 as temperature-related information related to the temperature of the deposition source 12. Instead of the temperature sensor 12a, a sensor for detecting the power supplied to the deposition source 12 may be used. If the deposition source 12 is a sputter cathode, the power value supplied to the sputtering target can be used as the above temperature-related information.

[0030] Stage 13 is configured to hold the substrate W, such as a semiconductor wafer or a glass substrate, which is the target of film deposition, toward the deposition source 12.

[0031] The sensor head 14 incorporates a quartz crystal oscillator 140 (see Figure 2) having a predetermined fundamental frequency (natural frequency). The surface of the quartz crystal oscillator 140 is positioned opposite the deposition source 12 so that the deposited material can be incident on it and deposited, and is typically located near the stage 13. The output of the sensor head 14 is supplied to the measurement unit 17 as frequency-related information related to the oscillation frequency of the quartz crystal oscillator 140.

[0032] The measurement unit 17 is configured as a measuring device that measures the mass of deposits on a quartz crystal oscillator over time based on changes in the oscillation frequency (resonance point) of the quartz crystal oscillator, which is frequency-related information. In other words, the measurement unit 17 calculates a rate conversion value for each unit time calculated from the frequency-related information. The measurement unit 17 is further configured to measure the deposition rate of the evaporated film and to control the deposition source 12 via the power supply unit 18 so that the deposition rate becomes a predetermined value.

[0033] The film deposition apparatus 10 further includes a first shutter 16A and a second shutter 16B. The first shutter 16A is positioned between the deposition source 12 and the stage 13 and is configured to open or block the incident path of deposited particles from the deposition source 12 to the stage 13 and the sensor head 14. The second shutter 16B is positioned between the first shutter 16A and the stage 13 and is capable of blocking the incident path of deposited particles from the deposition source 12 to the stage 13, but does not block the incident path of deposited particles from the deposition source 12 to the sensor head 14. Depending on the specifications of the film deposition apparatus, the installation of either the first shutter 16A or the second shutter 16B, or both, may be omitted.

[0034] The opening and closing of the first shutter 16A and the second shutter 16B are controlled by a control unit (not shown). Typically, the first shutter 16A and the second shutter 16B are closed at the start of deposition until the release of deposited particles from the deposition source 12 stabilizes. The control unit then opens the first shutter 16A when it determines that the release of deposited particles has stabilized based on the output of the temperature sensor 12a, and then opens the second shutter 16B when it confirms the stable release of deposited particles based on the output of the sensor head 14. This allows the deposited particles from the deposition source 12 to reach the substrate W on the stage 13, and the film deposition process on the substrate W begins. Simultaneously, the film thickness of the deposited film on the substrate W and its deposition rate are monitored by the measurement unit 17.

[0035] [Measurement Unit] Figure 2 is a block diagram showing one example configuration of the measurement unit 17. The measurement unit 17 includes an oscillation circuit 41, a measurement circuit 42, a control circuit 43, and a storage unit 44.

[0036] The measurement unit 17 is configured to measure the mass of a substance (film-forming material) attached to the quartz crystal oscillator 140 by using the correlation between the emission to the quartz crystal oscillator 140 and the corresponding reception from the quartz crystal oscillator 140, which is obtained by electrically sweeping near the resonant frequency of the quartz crystal oscillator 140.

[0037] The oscillation circuit 41 and the measurement circuit 42 function as a network analyzer. The oscillation circuit 41 causes the crystal oscillator 140 of the sensor head 14 to oscillate by transmitting, i.e., inputting, a sinusoidal signal of a predetermined frequency to the crystal oscillator 140. The measurement circuit 42 receives the output signal of the crystal oscillator 140 and the input signal output from the oscillation circuit 41, and based on these, measures the electrical characteristics of the crystal oscillator 140, such as its resonant frequency and phase, and outputs them to the control circuit 43.

[0038] The material constituting the quartz oscillator 140 is a piezoelectric element such as an AT-cut type quartz oscillator or an SC-cut type quartz oscillator. The fundamental frequency of the quartz oscillator 140 is, for example, 3 MHz or more and 6 MHz or less, and in this embodiment it is 5 MHz. When the attached mass is minute, the fundamental frequency of the quartz oscillator is set to a higher value in order to increase the sensitivity of mass detection. For example, when detecting minute mass present in the air, a quartz oscillator with a fundamental frequency of several tens of MHz is selected.

[0039] During film deposition on the substrate W, the deposition material from the deposition source 12 also adheres to the surface of the quartz crystal oscillator 140. The deposition material adhering to the surface of the quartz crystal oscillator 140 acts as newly added mass at arbitrary time intervals, changing the vibration frequency of the quartz crystal oscillator 140. Furthermore, the mass of the deposits on the surface of the quartz crystal oscillator 140 correlates with the density of the deposits. In other words, by measuring the change in the vibration frequency of the quartz crystal oscillator 140, the thickness (film thickness) of the deposition material film adhering to the surface of the quartz crystal oscillator 140 can be determined. The measurement unit 17 excites the quartz crystal oscillator 140 by emitting vibrations, and indirectly measures the film thickness from the vibration waveform resulting from the excitation.

[0040] The measurement unit 17 uses a sinusoidal signal of a predetermined frequency as the transmission signal to perform excitation. The excited quartz crystal oscillator 140 responds as a system including the deposits attached to its surface. The measurement unit 17 receives the response of the quartz crystal oscillator 140, including the mechanical vibration phenomenon, as an electrical vibration waveform via the piezoelectric effect of the quartz crystal oscillator 140. The measurement unit 17 stores the received waveform and analyzes the stored waveform. The vibration waveform of the quartz crystal oscillator 140 is stored in the storage unit 44. The measurement unit 17 extracts and outputs the film thickness included in the waveform analysis results.

[0041] The memory unit 44 is composed of storage devices such as semiconductor memory and hard disk drives. The memory unit 44 stores the frequency response from the detection system, including the vibration waveform of the crystal oscillator 140 when a sinusoidal signal of a predetermined frequency emitted by the oscillation circuit 41 is input to the sensor head 14. The memory unit 44 also stores control programs for various processes executed in the control circuit 43, which will be described later, and various parameters necessary for calculations.

[0042] Figure 3 shows a typical equivalent circuit of the sensor head 14. As shown in the figure, the crystal oscillator 140 is represented as a parallel circuit of a series resonant circuit consisting of a momentary capacitance C1, a momentary inductance L1, and a series equivalent resistance R1, and a static capacitance C0. The series resonant circuit is an equivalent circuit that includes the mechanical vibration element of the crystal oscillator 140. The static capacitance C0 includes, for example, the capacitance between electrodes formed on the front and back surfaces of the crystal oscillator 140 and the parasitic capacitance of the holder that holds the crystal oscillator 140. The series equivalent resistance R1 represents the vibration loss components such as internal friction, mechanical losses, and acoustic losses when the crystal oscillator 140 vibrates. In general, the higher the series equivalent resistance R1, the less the crystal oscillator 140 vibrates.

[0043] Methods for measuring the resonant frequency of the crystal oscillator 140 by excitation from the oscillation circuit 41 include measuring the frequency at which the phase becomes zero, measuring the frequency at which the susceptance, which is the imaginary component of admittance, becomes zero, and measuring the frequency at which the reactance, which is the imaginary component of impedance, becomes zero.

[0044] As shown in Figure 4, the half-maximum frequencies F1 and F2 are frequencies that give half the maximum conductance at the series resonant frequency Fs. The half-maximum bandwidth is twice the half-maximum bandwidth Fw and is the difference between one half-maximum frequency F1 and the other half-maximum frequency F2. The half-maximum bandwidth Fw is half the half-maximum bandwidth. Conductance is the real part of admittance, and Figure 4 shows an example of measurement using current as the outgoing signal and voltage as the received signal. The series resonant frequency Fs can be the fundamental wave or its nth harmonics, such as the third harmonic. When selecting a single series resonant frequency Fs, it is preferable to select the fundamental wave as the series resonant frequency Fs because the conductance value is generally maximized when using the fundamental wave, and a large signal-to-noise ratio (S / N ratio) can be obtained.

[0045] The control circuit 43 is composed of a computer including a CPU and memory. The control circuit 43 comprehensively controls the operation of the measurement unit 17. The control circuit 43 acquires the amount of change in the resonant frequency and electrical information near the resonant frequency in a single measurement. The electrical information near the resonant frequency includes the values ​​of phase, resistance and reactance, conductance and susceptance, derived from the correlation between the oscillation to the crystal oscillator 140 and the reception from the crystal oscillator 140 corresponding to that oscillation. The control unit 43 repeatedly measures the amount of change in the resonant frequency over time and the displacement of the electrical information at a set time interval (repetition period). The repetition period is not particularly limited and is, for example, 100 milliseconds. When acquiring the amount of change in the resonant frequency, past measurement results may also be used.

[0046] Figure 5 is a functional block diagram of the control circuit 43. The control circuit 43 has the following functional blocks: an acquisition unit 431, a buffer unit 432, an arithmetic unit 433, a filter unit 434, and a control unit 435. These functional blocks are realized when the control circuit 43 executes a predetermined control program stored in the memory unit 44. Figure 6 is a block diagram illustrating the procedure of processing performed in the control circuit 43.

[0047] The acquisition unit 431 acquires the rate conversion value and the real part of the admittance per unit time, calculated based on the resonant frequency of the quartz oscillator 140, from the measurement circuit 42 at a predetermined sampling period. The acquisition unit 431 also acquires the temperature of the evaporation source 12 (hereinafter also referred to as the evaporation source temperature) from the temperature sensor 12a at a predetermined sampling period. The sampling period is not particularly limited, and is typically milliseconds to several seconds, and in this embodiment it is 2 seconds. The sampling period for the resonant frequency and the sampling period for the evaporation source temperature are not limited to being the same, and may be different.

[0048] The buffer unit 432 is a FIFO (First In First Out) type memory that stores and reads, for example, the resonant frequency (rate-converted value, real part of admittance) and evaporation source temperature acquired by the acquisition unit 431 in a time series. The capacity of the buffer unit 432 is not particularly limited and can be, for example, several seconds to tens of seconds.

[0049] The calculation unit 433 performs preprocessing to smooth the rate-converted values ​​of the crystal oscillator 140 for each unit time acquired by the acquisition unit 431. This reduces noise components from the rate-converted values ​​input to the filter unit 434, which will be described later, thereby improving the estimation accuracy of the calculated rate value.

[0050] Examples of the above preprocessing include a median calculation process that calculates the median Rp of the rate-converted values ​​for each unit time stored in the buffer unit 432 for a predetermined period of time. However, the above preprocessing may also include a moving average calculation process that calculates the moving average of the rate-converted values ​​for each unit time stored in the buffer unit 432 for a predetermined period of time, or a combination of the median calculation process and the moving average calculation process may be used.

[0051] The calculation unit 433 further calculates a first approximate line corresponding to the time change of the evaporation source temperature, which is temperature-related information, and a second approximate line corresponding to the time change of the real part of the admittance in the oscillator 140 based on frequency-related information. In this embodiment, the first and second approximate lines are least-squares lines calculated using the rate-converted value for a predetermined time and the evaporation source temperature stored in the buffer unit 432.

[0052] The filter unit 434 filters the rate-converted value calculated by the measurement circuit 42. The filter unit 434 reduces noise from the rate-converted value output from the measurement circuit 42 and returns it to the true rate-converted value, and in this embodiment it is composed of a Kalman filter. Furthermore, the filter section 434 can be configured using a filter composed of a state-space model in which the internal state can be estimated. A typical example is a filter that uses a system including deposits attached to the surface of a quartz oscillator as its state-space model. Using a state-space model that simulates the system being implemented is also preferable because it is expected to allow for a more accurate estimation of the internal state, and it increases the design flexibility of the state quantities used as filter coefficients (described later).

[0053] The Kalman filter is a computational method for efficiently estimating the internal, invisible "state" in a mathematical model called a state-space model. In this embodiment, information obtained from the sensor head 14 and temperature sensor 12a is used as "observed values" to estimate the current state, and the observed value is calculated from the estimated state (corresponding to the "estimated rate Rf" in Figure 6). Based on this, the deposition source 12 is controlled. In other words, the previous estimated rate is X k-1 The current rate conversion value (observed value) is Y k If we let K be the Kalman gain, then the estimated rate X is... k The estimated rate Rf is expressed as follows: X k =X k-1 +K×(Y k -X k-1 )

[0054] The Kalman gain K is based on the previous estimated state P k-1 and is expressed as follows. K = (P k-1 + σw) / (P k-1 + σw + σv) Here, if σw is a fixed value and σv is a variable value, the larger the value of σv, the smaller the Kalman gain K. As a result, the change amount from the previous estimated rate X k of the current estimated rate X k-1 becomes smaller, and the effect as a filter becomes larger (stronger). Hereinafter, σv is also referred to as a filter coefficient indicating the degree of filtering strength in the filter unit 434. In the present embodiment, a plurality of set values are prepared as the filter coefficient σv.

[0055] The control unit 435 is configured to be able to change the filter coefficient σv based on the rate conversion value calculated by the measurement unit 42 and the deposition source temperature detected by the temperature sensor 12a. That is, in the present embodiment, the filter coefficient σv is adjusted by referring to the deposition source temperature, which is an observed value other than the rate, and the real part of the admittance of the crystal oscillator 140. In the present embodiment, a first set value σv1, a second set value σv2, and a third set value σv3 are prepared as the filter coefficient σv.

[0056] As will be described in detail later, the control unit 435 determines whether or not the absolute value of the slope St of the first approximate straight line (least squares straight line) corresponding to the time change of the deposition source temperature (temperature-related information) calculated by the calculation unit 433 is equal to or greater than a predetermined value S1. In this case, when the absolute value of the slope St of the first approximate straight line is equal to or greater than the predetermined value S1, the control unit 435 sets the filter coefficient σv to the first set value σv1, and when the absolute value of the slope St of the first approximate straight line is less than the predetermined value S1, the control unit 435 is configured to switch the filter coefficient σv to a second set value σv2 that is larger than the first set value σv1.

[0057] At the beginning of the film deposition process, the observed rate tends to increase in accordance with the rise in temperature of the deposition source 12. Therefore, in this embodiment, when the absolute value of the slope St of the least-squares straight line (first approximate straight line) of the deposition source temperature is greater than or equal to a predetermined value S1, the filter coefficient σv is set to a first setting value σv1, which has a relatively weak filtering effect in the filter section 434, so that the observed rate can easily follow the rise in temperature of the deposition source. The value of the first setting value σv1 is not particularly limited and can be set arbitrarily as long as the above effect can be obtained, for example, between 1,000 and 2,000.

[0058] On the other hand, when the absolute value of the slope St is less than a predetermined value S1, the filter coefficient σv is set to a second setting value σv2, which has a relatively strong filtering effect in the filter unit 434, in order to suppress the influence of noise included in the rate observation value. This makes it possible to stabilize the rate conversion value after filtering. The value of the second setting value σv2 is not particularly limited and can be set arbitrarily as long as the above effect is obtained, for example, between 10,000 and 20,000.

[0059] Furthermore, if the absolute value of the slope St of the first approximation line is less than a predetermined value S1, the control unit 435 further determines whether the slope Sa of the second approximation line (least squares line), which corresponds to the time change of the real part (frequency-related information) of the admittance in the crystal oscillator 140 calculated by the calculation unit 433, is within a predetermined range S2. If the slope Sa of the second approximation line is outside the predetermined range S2, the control unit 435 is configured to switch the filter coefficient σv to a third setting value σv3, which is greater than the second setting value σv2.

[0060] Even when the absolute value of the slope St of the least-squares line (first approximation line) of the deposition source temperature is less than a predetermined value S1, the observed rate value may fluctuate rapidly instantaneously due to the influence of physical disturbances such as splashing of the deposition material or changes in the surface state of the quartz oscillator 140. Therefore, in this embodiment, the slope Sa of the least-squares line of the real part of admittance calculated based on the response signal from the quartz oscillator 140 is monitored, and when it is determined that this is outside the predetermined range S2, the filter coefficient σv is switched from the second setting value σv2 to the third setting value σv3 to further strengthen the filtering effect in the filter unit 434. This suppresses sudden changes in the observed rate value due to the influence of splashing, etc., and eliminates abnormal values ​​from the observed rate value. The value of the third setting value σv3 is not particularly limited and can be set arbitrarily as long as the above effect can be obtained, for example, between 50,000 and 300,000.

[0061] Here, the predetermined range S2 can be, for example, a range less than or equal to the absolute value |S2|, that is, a range including 0, which is between -S2 and +S2. The control unit 435 is configured to switch the filter coefficient σv from the third setting value σv3 to the second setting value σv2 after a predetermined time has elapsed when the slope Sa of the second approximation line changes from a value less than the negative threshold (-S2) to a value greater than or equal to the negative threshold (-S2).

[0062] This is because instantaneous fluctuations in the real part of admittance due to sudden boiling, etc., generally occur with a downward-convex peak. Therefore, even if the slope Sa of the real part of admittance recovers to above a negative threshold (-S2), it is highly likely that this is an observed value near the peak of the aforementioned peak. In this case, the slope Sa is expected to rise sharply upward again, so if the filter coefficient σv is reduced as the slope Sa recovers within the predetermined range S2, an observed value that does not reflect the true value will be obtained. Therefore, in this embodiment, even if the slope Sa of the real part of admittance recovers within the predetermined range S2, the filter coefficient σv is maintained at the third setting value σv3 for a predetermined time, and a delay is introduced before returning to the second setting value σv2. This allows the filter coefficient σv to be maintained at the third setting value σv3 until the real part of admittance stabilizes at the true value, making it easier to obtain rate observed values ​​that reflect the true value.

[0063] [Measurement unit operation] Next, the details of the control circuit 43 will be explained along with the operation of the measurement unit 17. Figure 7 is a flowchart of this process.

[0064] When the film deposition process by the film deposition apparatus 10 begins, the control circuit 43 first sweeps the frequency range near the fundamental frequency of the crystal oscillator 140 to acquire frequency-related information (rate converted value, real part of admittance) and temperature-related information (deposition source temperature) of the crystal oscillator 140 (step 101).

[0065] More specifically, the control circuit 43 acquires the rate conversion value and the real admittance portion calculated based on the resonant frequency of the crystal oscillator 140, and the deposition source temperature detected by the temperature sensor 12a.

[0066] Next, the control circuit 43 stores the acquired rate-converted value, real admittance portion, and deposition source temperature in the buffer unit 432, and calculates the median Rp of the rate-converted value, the slope St of the least-squares line (first approximation line) of the deposition source temperature, and the slope Sa of the least-squares line (second approximation line) of the real admittance portion in the quartz oscillator 140 (steps 102, 103, 104).

[0067] Next, the control circuit 43 executes a process to determine the filter coefficient σv in the filter section 434. Specifically, the control circuit 43 determines whether the absolute value of the slope St of the least squares line of the deposition source temperature is greater than or equal to a predetermined value S1 (step 105). If the absolute value of the slope St is greater than or equal to the predetermined value S1 (No in step 105), the control circuit 43 sets the filter coefficient σv to a first set value σv1 (step 106).

[0068] Furthermore, if the absolute value of the slope St is less than a predetermined value S1 (Yes in step 105), the control circuit 43 determines whether the slope Sa of the least squares line of the real part of admittance is outside the predetermined range S2 (the absolute value of the slope Sa exceeds the threshold S2) (step 107). If the slope Sa is within the predetermined range S2 (No in step 107), the control circuit 43 determines whether it has been less than a predetermined time since the slope Sa returned to being greater than or equal to the negative threshold of the predetermined range S2 (Sa ≥ -S2) (step 108). If No, the filter coefficient σv is set to the second setting value σv2 (step 109).

[0069] On the other hand, in step 107, if the slope Sa of the least squares line of the real part of admittance is outside the predetermined range S2 (Yes in step 107), the control circuit 43 sets the filter coefficient σv to the third setting value σv3 (step 110). Similarly, in step 108, if the slope Sa returns to above the negative threshold of the predetermined range S2 (Sa≧-S2) and less than the predetermined time has elapsed since then (Yes in step 108), the control circuit 43 also sets the filter coefficient σv to the third setting value σv3 (step 110).

[0070] Next, the control circuit 43 calculates the estimated rate Rf by filtering the median of the rate conversion values ​​calculated in step 102 with the filter unit 434, whose filter coefficient σv has been determined as described above (step 111). The calculated estimated rate Rf is used as the previously estimated rate input to the Kalman filter when calculating the next estimated rate Rf.

[0071] Thereafter, the film thickness of the deposited material on the substrate W is measured by repeatedly performing the above-described process. According to this embodiment, it is possible to eliminate or suppress abnormal values ​​and improve the responsiveness of the processed value to the true value.

[0072] (Comparative Example 1) Figure 8(A) shows the time evolution of the film deposition rate (left vertical axis) and the deposition source temperature (right vertical axis) at the beginning of the film deposition process (when the deposition source is started up). Figure 8(B) is a magnified view of Figure 8(A) around 160 min, when the deposition source temperature is relatively stable.

[0073] As the deposition source temperature increases, the observed rate values ​​(raw data, ave data) also increase. Raw data are the measured values ​​(raw data) for each measurement run, and ave data are the moving average of those measured values. Comparing the cases where the filter coefficient σv is set to the first setting value σv1 and the second setting value σv2, it was confirmed that with the first setting value σv1, the rate change accompanying the increase in deposition source temperature can be sufficiently tracked, but when the deposition source temperature is constant, the output value of the rate is somewhat unstable due to the influence of noise. On the other hand, when the filter coefficient σv is set to the second setting value σv2, the filtering effect on the rate conversion value becomes too strong, making it impossible to follow the rate change due to the rise in the deposition source temperature. However, it was confirmed that a stable rate with less noise is output when the deposition source temperature is constant.

[0074] (Example 1) Figures 9(A) and (B) show experimental results similar to those in Figures 8(A) and (B), illustrating the rate change when a control is performed that switches the filter coefficient σv between a first setpoint σv1 and a second setpoint σv2 according to the slope St of the deposition source temperature (steps 105 to 109 in Figure 7). According to this embodiment, by switching the filter coefficient σv to the first setpoint σv1 when the absolute value of the slope St of the deposition source temperature is greater than a predetermined value S1 (when the deposition source temperature is rising), and to the second setpoint σv2 when it is less than or equal to the predetermined value, it is possible to follow the rate change during the startup of the deposition source, and output a rate with less noise when the temperature is stable.

[0075] (Comparative Example 2) Figure 10 shows the time evolution of the film deposition rate (left vertical axis) and the real part of admittance (right vertical axis) when the deposition source temperature is constant, and represents the experimental results when the filter coefficient σv is fixed at the second setting value σv2. Because the deposition source temperature is stable, a low-noise rate can be obtained with the filtering strength of the second setting value σv2. On the other hand, around 44 min and 46.5 min, when the real part of admittance (Real) decreases due to splash or changes in the surface state of the quartz oscillator 140, the rate tends to fluctuate in response.

[0076] (Example 2) Figure 11 shows experimental results similar to those in Figure 10, illustrating the rate change when the filter coefficient σv is switched between a second setting value σv2 and a third setting value σv3 according to the slope Sa of the real part of the admittance (steps 107-110 in Figure 7). It can be confirmed that when the slope Sa of the real part of the admittance decreases significantly beyond a predetermined range, switching from the second setting value σv2 to the third setting value σv3 strengthens the filtering effect and suppresses rate fluctuations to a small extent.

[0077] (Example 3) Figure 12 shows experimental results similar to those in Figure 10, illustrating the rate change when a delay is introduced when switching the filter coefficient from the third setting value σv3 to the second setting value σv2 (steps 108 and 110 in Figure 7).

[0078] In the example shown in Figure 11, the filter coefficient was switched from the third setting value σv3 to the second setting value σv2 in accordance with the return of the slope Sa of the admittance real part to within the predetermined range S2. As a result, rate fluctuations associated with a rapid increase in the positive direction of the admittance real part could not be suppressed. In contrast, according to the experimental results in Figure 12, where a delay was introduced in switching the filter coefficient from the third setting value σv3 to the second setting value σv2, the filter can be strongly applied over the entire period of time during which the admittance real part fluctuates significantly. Therefore, rate fluctuations associated with a rapid increase in the positive direction of the admittance real part can be suppressed to a small extent.

[0079] Furthermore, it is preferable that the length of the above delay be longer than the width near the peak of the real part of admittance. Also, the delay control when switching from the third setting value σv3 to the second setting value σv2 is applied only when the slope Sa of the real part of admittance returns from the negative direction to the positive direction, as described above. This allows the filter coefficient to be quickly switched from the third setting value σv3 to the second setting value σv2 during the process in which the real part of admittance transitions from the lower peak position to a stable value (slope Sa is almost zero), thereby shortening the time it takes for the rate to return to a normal value.

[0080] Furthermore, referring to Figure 13, we compare the filtering methods used on the observed value (rate (raw)) with a moving average (rate (ave)), Kalman filter 1 (fixed filter coefficients (σv2)), and Kalman filter 2 (variable filter coefficients (σv3 / σv2 + delay)). In the example shown in Figure 13, with respect to changes in the real part of admittance, Kalman filter 1 can suppress rate fluctuations to approximately 1 / 2 and Kalman filter 2 can suppress them to approximately 1 / 5 compared to the moving average.

[0081] In the embodiments described above, control was described in which the filter coefficient σv is switched to one of the first to third set values ​​σv1 to σv3 based on both the absolute value of the slope St of the deposition source temperature and the absolute value of the slope Sa of the real part of admittance, but the control is not limited to this. For example, as shown in Figure 14, the control circuit 43 may switch the filter coefficient σv between the first set value σv1 and the second set value σv2 based only on whether the absolute value of the slope St of the deposition source temperature is greater than or equal to a predetermined value S1. In other words, the processes of steps 107, 108 and 110 in Figure 7 can be omitted.

[0082] In this embodiment, since the filter section 434 is composed of a Kalman filter, the computational load can be reduced compared to a filter composed of a state-space model that can estimate the internal state, and it has the advantage of being easy to implement in the control circuit 43. Furthermore, if the processing power of the control circuit 43 is sufficient, it is possible to implement a filter using a state-space model that simulates the system being implemented, thereby improving stability and responsiveness.

[0083] Similarly, the control circuit 43 may switch the filter coefficient σv based only on whether the absolute value of the slope Sa of the real part of the admittance of the oscillator 140 is within a predetermined range S2, as shown in Figure 15, for example. In this case, if the absolute value of the slope Sa is outside the predetermined range, the filter coefficient σv can be set to the third setting value σv3, thereby obtaining the same effects as in the above embodiment. When the absolute value of the slope Sa is within the predetermined range, the filter coefficient σv may be set to either the first setting value σv or the second setting value σv2. [Explanation of Symbols]

[0084] 10…Film deposition equipment 11… Vacuum Chamber 12...evaporation source 14...Sensor head 17…Measurement unit (measuring device) 41…Oscillator circuit 42…Measurement circuit 43...Control circuit 140...Crystal resonator 432... Buffer section 433...Arithmetic section 434...Filter section 435... Control Unit

Claims

1. A measuring device for measuring the film deposition rate based on the oscillation frequency of an oscillator installed in a film deposition apparatus having a vapor deposition source, A measurement circuit that calculates a rate-converted value per unit time based on frequency-related information related to the oscillation frequency of the oscillator, A control circuit having a filter unit that filters the rate conversion value calculated by the measurement circuit, and a control unit that changes the filter coefficient, which is the degree of filtering strength in the filter unit, based on temperature-related information related to the temperature of the deposition source. A measuring device equipped with the following.

2. A measuring device according to claim 1, The control circuit further includes a calculation unit that calculates a first approximate straight line corresponding to the time change of the temperature-related information, The control unit determines whether the absolute value of the slope of the first approximation line is greater than or equal to a predetermined value. If the absolute value of the slope of the first approximation line is greater than or equal to the predetermined value, the filter coefficient is set to a first set value. If the absolute value of the slope of the first approximation line is less than the predetermined value, the filter coefficient is switched to a second set value which is greater than the first set value. Measuring device.

3. A measuring device according to claim 2, The calculation unit further calculates a second approximate line corresponding to the time change of the real part of the admittance in the oscillator based on the frequency-related information, The control unit further determines whether the slope of the second approximation line is within a predetermined range if the absolute value of the slope of the first approximation line is less than the predetermined value, and switches the filter coefficient to a third setting value that is greater than the second setting value if the slope of the second approximation line is outside the predetermined range. Measuring device.

4. The measuring device according to claim 3, The predetermined range is a range that extends between a negative threshold and a positive threshold. When the slope of the second approximation line changes from a value less than the negative threshold to a value greater than or equal to the negative threshold, the control unit switches the filter coefficient from the third set value to the second set value after a predetermined time has elapsed. Measuring device.

5. A measuring device according to any one of claims 1 to 3, The aforementioned filter section includes a Kalman filter. Measuring device.

6. A measuring device according to any one of claims 1 to 3, The temperature-related information is the temperature of the deposition source or the power used to heat the deposition source. Measuring device.

7. A measuring device according to claim 2 or 3, The calculation unit performs a preprocessing step to smooth the rate conversion value calculated by the measurement circuit. Measuring device.

8. The measuring device according to claim 7, The aforementioned preprocessing is a median calculation process that calculates the median of the rate conversion values ​​for each unit time. Measuring device.

9. The measuring device according to claim 7, The aforementioned preprocessing is a moving average calculation process that calculates the moving average of the rate-converted values ​​for each unit time. Measuring device.

10. A measuring device for measuring the film deposition rate based on the oscillation frequency of an oscillator installed in a film deposition apparatus having a vapor deposition source, A measurement circuit that calculates a rate-converted value per unit time based on frequency-related information related to the oscillation frequency of the oscillator, A control circuit having a filter unit that filters the rate-converted value calculated by the measurement circuit, and a control unit that changes the filter coefficient, which is the degree of filtering strength in the filter unit, based on the time change of the real part of the admittance in the oscillator. A measuring device equipped with the following.

11. A measuring device according to claim 10, The control circuit further includes a calculation unit that calculates a first approximate straight line corresponding to the time change of temperature-related information related to the temperature of the deposition source, The control unit determines whether the absolute value of the slope of the first approximation line is greater than or equal to a predetermined value. If the absolute value of the slope of the first approximation line is greater than or equal to the predetermined value, the filter coefficient is set to a first set value. If the absolute value of the slope of the first approximation line is less than the predetermined value, the filter coefficient is switched to a second set value which is greater than the first set value. Measuring device.

12. A measuring device according to claim 11, The calculation unit further calculates a second approximate line corresponding to the time change of the real part of the admittance in the oscillator based on the frequency-related information, The control unit determines whether the absolute value of the slope of the second approximation line is within a predetermined range. If the slope of the second approximation line is within the predetermined range, it sets the filter coefficient to a first setting value. If the absolute value of the slope of the second approximation line is outside the predetermined range, it switches the filter coefficient to a third setting value that is greater than the first setting value. Measuring device.

13. Vacuum chamber and A vapor deposition source is placed inside the vacuum chamber, A sensor head having an oscillator that oscillates at a predetermined resonant frequency and is disposed inside the vacuum chamber, A measurement circuit that calculates a rate-converted value per unit time based on frequency-related information related to the oscillation frequency of the oscillator, A control circuit having a filter unit that filters the rate conversion value calculated by the measurement circuit, and a control unit that changes the filter coefficient, which is the degree of filtering strength in the filter unit, based on temperature-related information related to the temperature of the deposition source. A film deposition apparatus equipped with the following.

14. Vacuum chamber and A vapor deposition source is placed inside the vacuum chamber, A sensor head having an oscillator that oscillates at a predetermined resonant frequency and is disposed inside the vacuum chamber, A measurement circuit that calculates a rate-converted value per unit time based on frequency-related information related to the oscillation frequency of the oscillator, A control circuit having a filter unit that filters the rate-converted value calculated by the measurement circuit, and a control unit that changes the filter coefficient, which is the degree of filtering strength in the filter unit, based on the time change of the real part of the admittance in the oscillator. A film deposition apparatus equipped with the following.

15. Frequency-related information relating to the oscillation frequency of the oscillator installed in the film deposition apparatus and temperature-related information relating to the temperature of the deposition source installed in the film deposition apparatus are acquired. Based on frequency-related information related to the oscillation frequency of the oscillator, a rate-converted value for each unit time is calculated. The rate conversion value calculated by the measurement circuit is filtered, The filter coefficient, which is the degree of filtering strength in the filter section, is changed based on temperature-related information related to the temperature of the deposition source. Method for measuring film thickness.

16. Frequency-related information relating to the oscillation frequency of the oscillator installed in the film deposition apparatus and temperature-related information relating to the temperature of the deposition source installed in the film deposition apparatus are acquired. Based on frequency-related information related to the oscillation frequency of the oscillator, a rate-converted value for each unit time is calculated. The rate conversion value calculated by the measurement circuit is filtered, The filter coefficient, which represents the degree of filtering strength in the filter section, is changed based on the time evolution of the real part of the admittance in the oscillator. Method for measuring film thickness.

Images