JP5384758B2 - Plasma etching equipment - Google Patents

Description

The present invention relates to a plasma etching apparatus, and more particularly to a plasma etching apparatus that can determine the end point of an etching process using plasma discharge with high accuracy.

In the dry etching process of a semiconductor wafer, the emission intensity of light of a specific wavelength during plasma emission changes with the progress of etching of a specific film. Therefore, it is possible to detect a change in the emission intensity of light having a specific wavelength during plasma emission during the etching process of the semiconductor wafer, and to detect the etching end point of the specific film based on this detection result.

For example, Patent Document 1 acquires a plasma emission signal generated in an etching apparatus when etching a material to be etched, and obtains a linear model or a nonlinear model using a polynomial for the acquired signal. Thereafter, it is shown that an error between the signal and the calculated value of the obtained model is obtained, and when the time change rate of the error exceeds a predetermined threshold value, the end point of etching is detected.

JP 2004-79727 A

In recent years, the aperture ratio (ratio of the area to be etched with respect to the area of the semiconductor wafer) has become smaller with the miniaturization and higher integration of the processing dimensions in the semiconductor wafer. For this reason, the light emission intensity change from the plasma in the vicinity of the etching end point is very small. In addition, the accuracy of processing dimensions by etching is becoming increasingly severe.

For example, in the end point determination of etching, even if the end point determination time is delayed by only a few seconds, the film under the film to be etched is excessively etched, resulting in disorder of processing dimensions or side etching (excessive etching in the lateral direction). As a result, the processed shape deteriorates.

In order to avoid such dimensional disturbance and shape deterioration, a process of detecting a minute change in light emission intensity near the etching end point reliably and at an early stage and removing a residue of the film to be etched, which is called the next over-etching, etc. It is necessary to move to.

When the change in emission intensity becomes very small, for example, when the change in emission intensity is detected using a linear model using a polynomial, the signal component direction is detected as an error. However, it is easily affected by noise components included in the signal, and it is difficult to reliably detect the etching end point at an early stage.

The present invention has been made in view of such problems, and provides an etching end point determination technique that can detect an etching end point reliably and at an early stage to determine the etching end point.

  In order to solve the above problems, the present invention employs the following means.

A processing gas is introduced into the vacuum processing chamber via a gas introduction means, plasma is generated by supplying high-frequency energy to the introduced processing gas, and plasma is generated on the object to be processed contained in the processing chamber using the generated plasma. In an etching end point determination method for determining an end point of an etching process in a plasma etching apparatus that performs processing, light having a preset wavelength is extracted from light emission of plasma generated in the vacuum processing chamber, and the extracted specific wavelength is extracted Step of calculating a regression line based on the acquired time series data, and a time axis direction distance between the regression line obtained by the step and the time series data. And calculating an end point of the etching process based on the distance in the time axis direction obtained by the step.

Since the present invention has the above configuration, it is possible to provide an etching end point determination technique that can detect an etching end point reliably and at an early stage to determine the etching end point.

It is a figure explaining the etching end point judging device concerning an embodiment. It is a figure explaining the process of the calculating part 105 which calculates the end point of etching. It is a figure which shows the optical signal of a specific wavelength among the light emission of the plasma obtained during the etching process. It is a figure explaining a regression line. It is a figure which shows the calculation result of the distance between a regression line and each signal point. It is a figure explaining the setting method of the threshold value with respect to the waveform obtained in FIG. It is a waveform of the signal intensity direction component of the distance between the regression line and each signal point. It is a figure explaining 2nd Embodiment. It is a figure which shows the multiplication result of the maximum value of the time-axis direction component of distance, and the inclination of a regression line. It is a figure explaining 3rd Embodiment. It is a figure which shows the calculation result of the product of the difference of the area made up and down of the regression line, and the maximum value of the area enclosed by the regression line.

Hereinafter, the best embodiment will be described with reference to the accompanying drawings. FIG. 1 is a diagram for explaining an etching end point determination apparatus according to the present embodiment. The etching apparatus 101 includes a vacuum vessel 102. An etching gas is introduced into the vacuum vessel 102 through a gas introduction unit (not shown), and microwave power is supplied to the introduced etching gas to perform etching. Gas is decomposed to generate plasma, and this plasma is used to etch a sample such as a wafer accommodated in the vacuum vessel.

The light having a specific wavelength of the plasma generated in the vacuum processing chamber during the etching process is dispersed by a spectroscope, and a signal component of the specific wavelength obtained by the spectroscopy is taken into the calculation unit 105. Here, the operation unit 105 is configured by, for example, a personal computer (PC).

The regression line equation is first derived from the retrieved signal by the regression line calculator 106. It is desirable to perform gain (amplifying or attenuating signal intensity) processing for the signal captured at this time to make a change in the time axis direction stand out.

Next, the distance calculator 107 calculates the distance between the signal obtained from the most recently obtained signal up to a certain point in the past and the regression line (shown by the regression line equation) obtained by the computation. .

At this time, the number of signals used for the calculation can be arbitrarily set. In addition, a linear slope holder 108 that holds the calculated slope value of the regression line and a calculator 109 are provided. The calculator 109 holds the value (distance) obtained by the distance calculator 107 and the linear slope holder. The calculated value (slope) is calculated to obtain the time direction component of the distance. The end point determination unit 110 monitors the calculation result of the calculation unit 109, and outputs an end point determination signal when the calculation result exceeds a preset determination level. This end point determination signal is displayed on the display 111.

FIG. 2 is a diagram for explaining the processing of the calculation unit 105 that calculates the end point of etching. First, light having a specific wavelength out of plasma emission associated with the etching process is extracted as time series data by a spectroscope, and the extracted data is input to the arithmetic unit 105 (step S200).

The input time series data is first subjected to gain processing by a gain calculator. In the gain processing, for example, by reducing the signal waveform by 0.001 times, the change in the time axis direction component can be remarkably expressed. It should be noted that the number of signal points for obtaining the regression line next to the data and the number of distance calculation signals necessary for obtaining the distance from the regression line must be larger (step S201).

When the number of signals obtained exceeds the required number of signals and a predetermined determination start time has elapsed, a calculation process for obtaining a regression line is performed. The calculation for obtaining the regression line can be obtained by, for example, the least square method. In addition, the signal used for the calculation for obtaining the regression line is from the latest signal to any past point. Note that the regression line slope data obtained at this time is retained by the regression line slope retainer 108 (steps S202, S203, S204, 2).
05).

Next, using the obtained regression line, the distance from the signal obtained in the past is obtained. The signal used to determine the distance can be from the latest signal to any past point. The number of signal points M to be used is, for example, more than the number of points used to calculate the regression line.

Further, the distance d from the regression line to a point in the past can be obtained from Equation 2 if the regression line is expressed by Equation 1. In Equation 2, x1 and yl are x at a certain point in the past.
Coordinates and y coordinates are shown.

The distance obtained by Equation 2 includes both time and signal intensity components, where the horizontal axis is the time axis and the vertical axis is the signal intensity (step S206).

The signal from the plasma accompanying the etching process includes a noise component, and this noise component appears as a signal intensity component on the vertical axis. For this reason, the influence by the noise component can be reduced by extracting only the time-axis direction component dt from the signal obtained by the distance calculation. Note that the distance dt in the time axis direction between the regression line and the time series data is expressed by Equation 3
Can be calculated.

The calculation process of the distance dt in the time axis direction between the regression line and the time series data is performed for all arbitrarily set distance calculation signal points M (step S207).

Next, a power calculation of the maximum value (distance maximum value) of the time-axis direction component in the distance from the regression line calculated for all signals is performed. As a calculation method, for example, the input signal is squared, and when the calculated value exceeds a dark value that can be arbitrarily set, the end point is determined (step S209,
210). By performing the power calculation, it is possible to further amplify the signal component and reduce noise included in the signal.

FIG. 3 is a diagram illustrating an example of an optical signal having a specific wavelength among plasma emission obtained during the etching process. In this optical signal waveform, there are many signal fluctuations that are considered to be noise components due to the influence of the aperture ratio and the like, and it is difficult to catch the change in the waveform due to the progress of etching.

FIG. 4 is a diagram for explaining a regression line. The regression line is obtained by using, for example, the latest signal point and several tens of past signal points in the obtained optical signal waveform. The number of signal points used for calculating the regression line can be arbitrarily set, for example, 20 points. Next, the distance between the obtained regression line and each signal point is calculated. The distance d between the regression line and each signal point is
This is the length of a perpendicular line drawn from each signal point toward the regression line from the regression line, and its time-axis direction component dt is a signal necessary for end point determination.

FIG. 5 is a diagram illustrating a calculation result of the distance dt between the regression line and each signal point. Here, a waveform example is shown in which 20 signal points are used to obtain the regression line and 100 signal points are used to obtain the distance. The end point of etching is determined based on the obtained waveform. The threshold used for the determination can be arbitrarily set, but the threshold needs to be a value that can accurately determine the change near the etching end point at an early stage.

FIG. 6 is a diagram illustrating a threshold setting method for the waveform obtained in FIG. There is a time difference of about 3 seconds in the etching end point determination between the case where the threshold is set to the threshold 2 and the case where the threshold is set to 1. That is, when the threshold is set to threshold 1, a delay of 3 seconds occurs at the end point determination time. In this case, at a certain point in the wafer surface, the etching end point may pass and damage may be caused to the surface that is not desired to be etched during the 3 seconds.

  The waveform obtained in FIG. 5 is a waveform of the time-axis direction component of the distance between the regression line and each signal point.

As described above, the time-axis direction component of the distance between the regression line and each signal point is less affected by noise on the signal, so the threshold value can be set to a low value as indicated by threshold value 2, thereby enabling the end point to be reached early. Can be determined.

As a signal used for the determination, when a distance between the regression line and each signal point, that is, a length d of a perpendicular line drawn from each signal point toward the regression line is set, the signal representing the distance includes an intensity component. Therefore, it is difficult to reduce the threshold used for the determination.

FIG. 7 is a waveform of the signal intensity direction component dy of the distance between the regression line and each signal point. It can be seen that the waveform shown in FIG. 7 includes a larger amount of noise in the signal waveform than the waveform of FIG. 5 in which the time-axis direction component dt of the distance from the regression line is extracted. In addition, another waveform peak is observed after the waveform falls. This is observed because the optical signal waveform has fallen, and when a light emission intensity signal component is used, this component cannot be removed.

In detecting the end point, for example, a late determination using the falling portion of the waveform may be performed. For example, in the step of detecting the end point in etching, when it is desired to continue etching even if a residue or the like remains, it is effective to determine the end point later. In such a case, there is a risk of erroneous determination due to the signal component after the fall of the waveform. On the other hand, when only the time axis direction component is used, the noise component is hardly included after the waveform falls, and erroneous determination can be suppressed.

FIG. 8 is a diagram for explaining the second embodiment. In FIG. 8, the processes in steps S200 to S208 are the same as the processes in steps S200 to S208 shown in FIG.

As shown in FIG. 8, the distance d from the regression line is calculated at all signal points, and then the maximum value (maximum distance value) of the time-axis direction component dt at the distance d from the calculated regression line is calculated ( After steps S206, 207, and 208), the time-distance component dt of the calculated distance is multiplied by the slope of the regression line (step S212). The multiplied result is compared with a preset threshold value, and if the threshold value is exceeded, an end point determination detection signal is output (steps S210 and 211).

  FIG. 9 is a diagram illustrating a multiplication result of the maximum value of the time-axis direction component of the distance and the slope of the regression line.

Even in this case, since only the time-axis direction component is extracted, it is understood that the noise component included in the signal is small. That is, it is possible not only to set the threshold value 1 but also to lower the threshold value as the threshold value 2 in making the determination in this waveform, so that the determination can be performed at an earlier stage. Further, when the threshold value is not lowered, it is possible to reduce the influence of noise on the determination signal in determination, and it is possible to perform more stable determination.

FIG. 10 is a diagram for explaining the third embodiment. In FIG. 10, steps S201 to S204 are the same as those in the first embodiment. Calculate a regression line from the plasma emission signal,
The area that appears above the regression line and the area that appears below the regression line in the area surrounded by the obtained regression line and the waveform of the acquired signal are calculated (step S212). The maximum area value obtained at this time is stored (step S213). Then, the ratio of the area created above the regression line to the area created below is calculated for the area obtained. Here, since it is detected as a characteristic that the value of the area created on the upper side and the area created on the lower side are greatly different, not only the ratio but also the difference may be used. For example, the calculated value can be calculated by subtracting the area (step S214). At this time, the maximum value of the area surrounded by the regression line is used and multiplied by the obtained calculation value to obtain a final determination signal to determine the end point (step S215). The final calculation value thus obtained is used as an end point determination signal to determine the etching end point.

FIG. 11 is a diagram illustrating a calculation result of the product of the difference between the areas created above and below the regression line and the maximum value of the area surrounded by the regression line. When a determination is made using this calculation result, the calculation value changes from negative to positive, for example, so that the waveform has one peak 1101 in the negative direction and one peak 1 in the positive direction.
102 is obtained. Therefore, for example, when performing a late determination, the inflection point of the waveform is obtained as a change in the signal from negative to positive. Therefore, when the threshold value is set, for example, when the threshold value 2 is exceeded and then the threshold value 2 is exceeded. Since the sign is inverted at the time, it is possible to make a stable determination that is not easily affected by noise. At this time, as in the first embodiment and the second embodiment, since the influence of noise on the signal is small, the determination can be performed with high accuracy even in the early determination.

As described above, according to the embodiment of the present invention, the end point is determined using the time axis direction component of the distance between the regression line and each signal point. For this reason, even if a noise component is included in the signal from the plasma accompanying the etching process, the noise component appears as a signal intensity component on the vertical axis, so that the influence of the noise component can be reduced. Therefore, even if the emission intensity change from the plasma is minute, the minute emission intensity change near the etching end point can be detected reliably and at an early stage, and the etching end point can be determined quickly and reliably. . In addition, when the waveform changes steeply in the signal waveform, the inclination becomes large. Therefore, a determination waveform with less noise component can be obtained by performing the determination in the second embodiment. Also,
If it is desired to make a later determination, it is possible to easily detect an inflection point and perform a stable determination by using the third embodiment. Thus, each determination method can be selected appropriately. In addition, the accuracy of judgment differs greatly by optimizing the number of data points used for the regression line and the number of data points for calculating the distance, and the judgment accuracy is greatly improved by using appropriate values for all judgments. .

DESCRIPTION OF SYMBOLS 101 Etching apparatus 102 Vacuum container 103 Optical fiber 104 Spectrometer 105 Calculation part 106 Regression linear calculation unit 107 Distance calculation unit 108 Linear inclination holder 109 Distance calculation unit 110 End point determination unit 111 Display

Claims (3)

A vacuum processing chamber for performing plasma processing on an object to be processed;
Gas introduction means for introducing an etching gas into the vacuum processing chamber;
High-frequency supply means for supplying high-frequency power for generating the plasma in the vacuum processing chamber;
A spectroscope for spectroscopically analyzing emission of a specific wavelength of the plasma;
A calculation unit for determining an end point of etching of the object to be processed,
The calculation unit calculates a regression line based on time series data of emission intensity of a specific wavelength extracted by the spectrometer, and calculates a distance from the time series data to the regression line. A distance calculator, a calculator for calculating a distance in the time axis direction by calculating a distance from the time series data calculated by the distance calculator to the regression line and a slope of the regression line, and the calculation A plasma etching apparatus comprising: an end point determination unit that outputs an end point determination signal based on a distance in a time axis direction calculated by the chamber. A vacuum processing chamber for performing plasma processing on an object to be processed;
Gas introduction means for introducing an etching gas into the vacuum processing chamber;
High-frequency supply means for supplying high-frequency power for generating the plasma in the vacuum processing chamber;
A spectroscope for spectroscopically analyzing emission of a specific wavelength of the plasma;
A calculation unit for determining an end point of etching of the object to be processed,
The calculation unit calculates a regression line based on time series data of emission intensity of a specific wavelength extracted by the spectrometer, and calculates a distance from the time series data to the regression line. A distance calculator, a calculator for calculating a distance in the time axis direction by calculating a distance from the time series data calculated by the distance calculator to the regression line and a slope of the regression line, and the calculation A plasma etching apparatus comprising: an end point determination device that outputs an end point determination signal based on a product of a maximum value in a time axis direction distance calculated by a container and a slope of the regression curve . A vacuum processing chamber for performing plasma processing on an object to be processed;
Gas introduction means for introducing an etching gas into the vacuum processing chamber;
High-frequency supply means for supplying high-frequency power for generating the plasma in the vacuum processing chamber;
A spectroscope for spectroscopically analyzing emission of a specific wavelength of the plasma;
A calculation unit for determining an end point of etching of the object to be processed,
The calculation unit is a regression line calculation unit that calculates a regression line based on time-series data of emission intensity of a specific wavelength extracted by the spectroscope,
The maximum value among the areas surrounded on the high intensity side or the low intensity side of the emission intensity axis by the regression line and the time series data, the area surrounded on the high intensity side of the emission intensity axis, and the emission intensity A plasma etching apparatus comprising: an end point determination device that outputs an end point determination signal based on a product of a difference between an area surrounded by a low intensity side of the shaft and an area.

Images