JPH0770516B2 - Etching end point determination method - Google Patents

Description

TECHNICAL FIELD The present invention relates to an etching end point determination method, and more particularly to an etching end point determination method suitable for etching end point determination in a dry etching apparatus.

[Conventional technology]

Conventionally, the determination of the etching end point has been performed by secondarily differentiating the emission intensity of the reaction product generated in the etching process as described in JP-A-59-94423. However, in such a conventional technique, no consideration is given to the determination of the end point when the etching time is short.

[Problems to be solved by the invention]

In the above conventional technology, when the etching time is short (for example, when the material to be etched is Poly-Si, it is about 1/10 that of SiO ₂ or Al)
There is a problem as described below.

FIG. 2 shows the configuration of the end point determination device. To determine the end point, the plasma light emission generated in the etching processing chamber 1 is guided to the photoelectric conversion element 3 through the spectroscopic device 2 that allows only a specific wavelength to pass, where it is converted into an electric signal, and the amplifier 4 and the A / D converter 5 are used.
Is passed through and is taken into the computer 6 as a digitized amount, and the end point determination is performed by the software of this computer.

Here, it is considered that the end point determination is performed by the second derivative. FIG. 3 (a) shows a temporal change in the light amount of the reaction product light. The same figure (b) differentiates it with respect to time, and the same figure (c) shows 2
It is a graph of the second derivative. In FIGS. 9B and 9C, the solid line is the actual differential value, the secondary differential value, and the broken line is the computer calculated differential value, the secondary differential value, and a delay time occurs with respect to the actual differential value. This is for the following reason.

(1) Data is sampled at regular intervals and cannot be infinitely fast.

(2) A filter is provided in the electric circuit to prevent erroneous determination due to noise. Also, in numerical processing, processing such as averaging is performed, and a delay time occurs due to this filter effect.

(3) Since the differential calculation itself is performed by calculating the difference between the past data and the present data, only the past differential value is known.

For the above reason, the first-order differential value causes a delay of t ₁ compared with the actual value, as shown in FIG. 3 (b). Since the second derivative is calculated based on the first derivative delayed by t ₁ , 2t ₁
I will be late.

The end point determination is performed using the part 1 in FIG. 3 (c).
That is, the end point is reached when the value (d ² OL / dt ² ) r having the peak of the second derivative exceeds the negative side. The first peak 2 on the negative side, which is shown in FIG. 3 (c), is generated with the rise of the emission intensity and must be ignored. In order to ignore this, from the start of etching, as shown in FIG.
The time T ₀ is provided and the end point determination is not executed within this time. This time T ₀ is hereinafter referred to as the end point determination lag time.

The above is the principle of the end point determination by the second derivative method, but this method has the following problems.

Depending on the material to be etched, the etching may be completed in a short time as shown in FIG. 1 as above
When the second and second differentials are performed, the solid line graphs in FIGS. 4B and 4C are obtained. The calculation result by the computer is also shown by the broken line in the figure, but when the second derivative of the broken line is seen, the first peak 2 and the first peak 1 on the negative side are not clearly separated due to the delay time and overlap. It becomes a wavy waveform. With this waveform, it is difficult to accurately detect the second peak, and an erroneous determination of the etching end point is likely to occur.

An object of the present invention is to provide an etching end point determination method capable of preventing an erroneous determination in the etching end point determination and accurately performing the etching end point determination even when the etching time is short.

[Means for solving problems]

The above-mentioned object is a method of determining the end point of etching by spectrally analyzing plasma emission, and when the time of the end point determination lag time at which the end point determination at the rising of the emission intensity is not performed from the start of etching, the vicinity of that time point is reached. It is assumed that the slope of the change in the amount of light emitted from the plasma is determined based on the temporal change in the amount of light emitted from the plasma, and that the amount of emitted light before the time point was on an extension line having the slope, This is achieved by performing the end point determination based on the assumed amount of emitted light and the actual amount of emitted light after the time point.

[Action]

According to the present invention, the lag time after the start of etching, that is, the light amount of the reaction product light at the time when the end point determination delay time T ₀ has elapsed, and based on the temporal change amount of the light amount in the vicinity thereof,
The amount of light before the lag time is assumed by extrapolation. That is, it is assumed that the light emission amount OL ₁ before the lag time is on an extension of a straight line having a slope α at that time from the plasma light emission amount OL (T ₀ ) at the lag time. The end point determination is performed based on the assumed emitted light amount OL ₁ and the actual emitted light amount OL ₂ after the lag time.

When the etching end point determination is performed using the virtual light amount obtained by the extrapolation method, the first derivative, that is, the slope α becomes a constant value dOL (T ₀ ) / dt before the lag time elapses, and the second derivative value becomes It becomes 0. This eliminates the influence of the rise of the reaction product light on the first-order and second-order differential values, and the etching end point determination can be performed only by the light amount change after the lag time. Disappear.

〔Example〕

An embodiment of the present invention will be described below with reference to FIG. FIG. 1 is a principle diagram of the present invention. First, FIG. 1 (a) shows a temporal change of the light amount of the reaction product. In FIG. 1A, the virtual light quantity (OL ₁ ) is extrapolated as indicated by the alternate long and short dash line when the end point determination lag time (T ₀ ) has elapsed. With this,
The primary and secondary differentials calculated from the etching start time are also initialized as shown in FIGS. 1 (b) and (c). In the case of the first derivative, it is initialized to the amount of change in the light amount when the lag time elapses, as shown by the alternate long and short dash line in FIG. 1 (b), and the second derivative is initialized to 0 as shown in FIG. 1 (c). Be converted. After that, 1
The second and second derivatives are calculated by extrapolating past emission intensity (OL
₁ ) and the actual emission intensity after the lag time has elapsed (O
L ₂ ).

In Fig. 1 (a), the amount of light after the lag time has elapsed is O
Letting L _{(TO) be} the amount of change in the light amount, α is the extrapolation of the light amount before the lag time elapses: OL ₁ = OL _(TO) −αt. Also,
The first derivative is initialized with a value of a.

According to the present embodiment, the virtual light amount is extrapolated at the time when the end point determination lag time elapses, and the first and second derivatives are calculated using the extrapolated emission intensity and the actual light emission intensity after the elapse of the lag time. The first derivative is initialized to the amount of change in the light amount after the lag time has elapsed, and the second derivative is initialized to 0. Therefore, even if the etching time is short, the influence of the rise of the reaction product light on the 1st and 2nd derivative values is eliminated, and the etching end point can be determined only by the change in the light amount after the lag time. It can be prevented and the etching end point can be accurately determined.

〔The invention's effect〕

According to the present invention, there is an effect that it is possible to accurately determine the etching end point even for a material whose etching is completed in a short time.

[Brief description of drawings]

FIG. 1 is a schematic diagram showing the light amount of reaction product light and its first-order and second-order differential values when the end point determination according to the present invention is performed, and FIG. 2 is a configuration diagram of the end point determination device. Fig. 3 is a schematic diagram showing the light intensity of reaction product light and the time variation of its first and second derivative values when the end point is determined by the conventional technique when the etching time is long, and Fig. 4 is also the etching time. FIG. 3 is a schematic diagram showing the light amount of reaction product light and the time-dependent changes in the first and second derivative values when the value is short. T ₀ …… End point determination lag time, OL ₁ …… Extrapolated emission intensity, OL ₂ …… Actual emission intensity

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Yoshinao Kawasaki 794, Higashi-Toyoi, Shimomatsu City, Yamaguchi Prefecture Stock company Hitachi Kasado Factory (72) Inventor Katsuyoshi Kudo 794, Toyoi, Higashi-Toyo, Yamaguchi Prefecture Stock Association In-house Hitachi Ltd. Kasado Plant (72) Inventor Minoru Soraoka 794 Azuma Higashitoyo, Shimomatsu City, Yamaguchi Prefecture Stock Company In-house Hitachi Ltd. Kasado Plant (56) References JP-A-62-95829 (JP, A)

Claims (1)

[Claims] 1. A method for determining an etching end point by spectroscopically analyzing plasma emission, wherein a time point of an end point determination lag time at which the end point determination at the rise of emission intensity is not performed from the start of etching, at the time point. Obtaining the slope of the change in the light amount of the plasma emission based on the temporal change of the light amount of the plasma emission in the vicinity, assuming that the light amount of the emission before the time point was on an extension line having the slope, An etching end point determination method, characterized in that the end point determination is performed based on the assumed amount of emitted light and the actual amount of light emitted after the time point.