JPH06295883A - Dry etching end point detection method and end point condition automatic determination method - Google Patents

Abstract

(57) [Summary] [Objective] When detecting the end point of the process by detecting the change in the emission intensity of light generated during the process in dry etching using plasma, which is one of the semiconductor manufacturing processes, the end point detection conditions are set. It is an object of the present invention to provide a dry etching end point detection method that is automatically determined and accurately detects the end point even when there is a waveform deformation or sudden noise due to a change in atmosphere in the chamber. [Configuration] When the end point is indicated on the graph of the light emission intensity of the monitored light to determine the end point detection condition, the end point detection condition is defined by the shape surrounded by the time axis and the graph up to the end point. For the light that is being monitored for detection, the end point detection is the end point when the shape of the part surrounded by the time axis and the graph up to the present point in the graph of emission intensity vs. time and the shape of the end point detection condition match. Method.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an end point detection method for detecting the end of the state change of an object which emits a wavy signal due to the state change, and more specifically, an etching process which is one of semiconductor manufacturing processes. More particularly, the present invention relates to a method for detecting an end point and a method for determining an end point detection condition in dry etching using plasma or the like.

[0002]

2. Description of the Related Art Generally, in dry etching, which is one of the semiconductor manufacturing processes, the end point of the process is detected by detecting the change in the emission intensity of light generated during the process. The end point was detected based on the first derivative value and the second derivative value of the emission intensity.

A method for determining the end point detection condition in this end point detection method, and the end point detection condition is automatically determined by designating one point on the graph of the emission intensity versus the time of the process for which the condition should be determined. As a method, the first derivative value or the second derivative value of the emission intensity at each sampling point is calculated, and the first derivative value or the second derivative value at the designated point and the appearance of the value up to the designated point The number of times was used as the end point detection condition.

[0004]

By the way, when many kinds of processing are performed in one dry etching apparatus, the end point detection condition must be determined for each processing.
In order to save labor in this work, the current dry etching apparatus is capable of automatically determining the end point detection condition by designating one point on the graph of time-emission intensity displayed on the CRT of the apparatus. However, when the end point detection condition is determined by the conventional method, there is a problem that the end point cannot be detected or is erroneously detected even by a slight change in the waveform or a small sudden noise.

Next, the first derivative will be described as an example. The waveform sampled to determine the end point detection condition is shown in FIG.
Suppose it was like (a). If point A is specified as the end point, the waveform of 1 will be displayed in the conventional method as shown in FIG.
The secondary differential was calculated, and "the secondary differential value becomes a for the second time after the processing starts" was used as the end point detection condition. However, under these conditions, when the atmosphere in the chamber was changed during processing many wafers and the waveform for detecting the end point was deformed as shown by the solid line in FIG. 6 (the broken line is the graph in FIG. 5A). )
The primary differential waveform is as shown in FIG. 7, and it is desired to detect the end point at point B, but it cannot be detected. When sudden noise is added to the waveform for detecting the end point as shown in FIG. 8, the first-order differential waveform becomes as shown in FIG. 9, and the end point is desired to be detected at point C, but it is erroneously detected at point C '. It is an object of the present invention to provide a dry etching endpoint detection method and endpoint detection condition determination method that solve these problems.

[0006]

In order to achieve the above object, the present invention comprises the following two steps. In the first step, when the user indicates the end point on the monitored light emission intensity vs. time graph to determine the end point detection condition,
The shape enclosed by the graph and the time axis up to the designated point is the end point detection condition. For example, when the point A is designated as the end point in FIG. 5A, the shape of the end point detection condition is as shown by the shaded area in FIG. In the second step, the light monitored to detect the end point is made into a graph of emission intensity versus time, and the end point is detected by the pattern matching between the shape surrounded by the graph and the time axis and the shape of the end point detection condition. .

[0007]

According to the present invention, by the method described above, the work for determining the end point detection condition can be saved, and the end point can be detected without causing a slight change in the waveform or a small sudden noise that has not been detected or erroneously detected. .

[0008]

Embodiments of the present invention will be described with reference to FIGS. Let FIG. 2 be the waveform monitored to determine the end point detection conditions. The endpoint detector stores this waveform in the form of emission intensity at each sampling point. When the user designates point C as the end point in this waveform, the end point detector detects each end point by the following procedure, using each sampling point up to point C and the emission intensity at that sampling point as the end point detection condition. For the sake of explanation, it is assumed that the emission intensity at the _nth sampling point is cn and the point C is the Nth sampling point in the monitored waveform for determining the end point detection condition.

(N, c_n) (N = 1, 2, 3, ..., N) is the waveform monitored to determine the end point detection condition.
The shape surrounded by the graph and the time axis up to the point indicated by
This form is the end point detection condition in the present invention. Well
Also, the nth sub-waveform of the waveform being monitored to detect the end point.
The emission intensity at the sampling point is d_nAs the M-th service
It is assumed that the sampling point is sampled. At this time
The end point is determined by the following procedure. N and M are generally different
So c _nAnd d_nYou can't compare them directly. So
Here d ’_nTo introduce. d ’_nIs defined by the following equation.

[0010]

[Equation 1]

However, d'which is not determined by this equation is determined by linear interpolation. For example, when N = 100 and M = 10

[0012]

[Equation 2]

[0013] The waveform monitored to detect the end point is finally as shown in FIG. 3, and N = 30 in FIG.
In the case of M = 10, M = 20, and M = 30, the graphs of d ′ at the time points of M = 10, M = 20, and M = 30 are shown by the solid lines in FIGS. 4 (a), (b), and (c), respectively (broken lines up to point C in FIG. 2). Of the graph and the shaded area will be described later).

[0014]

[Equation 3]

Represents the shape of the portion surrounded by the time axis and the graph. This form of end point detection conditions _(n, c n)
(N = 1, 2, 3, ..., N) to determine the end point. The judgment is

[0016]

[Equation 4]

Is calculated and the end point is reached when the value of s becomes smaller than a certain value set by the user. s corresponds to the shape of the end point detection condition and the area of the non-common part of the shape up to the time when the waveform at which the end point should be detected is monitored (FIG. 4A).
(Hatched portion of (c)).

[0018]

According to the present invention, by using the method of detecting the end point by capturing the global property of the waveform, the end point detection condition is automatically determined, but it is not affected by the change of the waveform or the influence of the sudden noise. This has the effect of enabling end point detection with little detection or erroneous detection.

[Brief description of drawings]

FIG. 1 is a diagram for explaining a pattern focused on as an end point detection condition by the present invention.

FIG. 2 is a monitored diagram for determining an end point detection condition according to an embodiment of the present invention.

FIG. 3 is a diagram showing a waveform being monitored to detect the end point.

FIG. 4 is a diagram showing a state when M = 20, 30, 40 in FIG.

FIG. 5 is a diagram showing a conventional end point detection method.

FIG. 6 is a waveform diagram for detecting the same end point.

FIG. 7 is a diagram showing a derivative of the graph of FIG.

FIG. 8 is a diagram showing a state where noise is added to a waveform to be detected.

9 is a diagram showing a waveform obtained by differentiating the graph of FIG.

Claims (4)

[Claims] 1. An end point detecting method for detecting an end point of a process by detecting a change in light emission intensity of light generated during a process such as dry etching, wherein the emission intensity is sequentially monitored at a short sampling time and its waveform is detected. End point detection method for dry etching, which detects the end point by capturing the global property of. 2. A graph of time vs. emission intensity of the light being monitored as a measure of the global property, showing the time axis, the shape of the part surrounded by the graph, and the shape at a preset end point. The method for detecting an end point of dry etching according to claim 1, wherein pattern matching is used. 3. The end point detection method for dry etching according to claim 1, wherein attention is paid to an area of a non-common portion when a scale in a time axis direction is matched as a pattern matching method. 4. The end point detection condition according to claim 3, which is a method for determining an end point detection condition in an end point detection method of detecting a change in emission intensity of light generated during processing such as dry etching to detect the end point of the processing. A method for automatically determining a dry etching end point condition in which the end point detection condition is automatically determined by designating one point on the graph of the emission intensity versus time of the process for which the end point detection condition is to be determined by using the method.