JPH0798993B2 - Film thickness control method - Google Patents

Description

TECHNICAL FIELD The present invention relates to a film thickness control method for forming a vapor deposition coating on a substrate, and more particularly to a photoelectric film thickness control method capable of controlling a desired thickness with high accuracy. .

[Prior Art] The substrate on which the coating is being formed is irradiated with a light beam having a specific wavelength λ,
When the transmitted light or the reflected light of this light beam is measured, where the refractive index of the coating film being formed is n and the geometrical thickness is d, the transmittance or reflectance increases as the thickness d increases. It is conventionally known to draw a periodic curve having an extreme value at the time when the film thickness nd becomes an integral multiple of λ / 4.
(FIG. 4) In the monochromatic side-light method, this property is used to measure the transmittance or reflectance of the substrate during the film formation at a specific wavelength λ, and the film formation is performed when the extreme value is detected. By finishing, the optical film thickness of the coating is reduced to 1/4 of the wavelength λ.
Control to an integer multiple of.

The accuracy of the film thickness control by this method largely depends on the accuracy of detecting the extreme value of the transmittance or the reflectance. As the means for detecting the extreme value, there are an expansion method, a differentiation method, and the like. The magnifying method is a method of electrically amplifying a change in transmittance or reflectance as much as possible to find an extreme value. However, since the amount of change in transmittance or reflectance with respect to a change in film thickness is extremely small especially in the vicinity of the extreme value. However, there is a problem that it is easily affected by noise, which makes it difficult to accurately control the film thickness.

The differential method is a method of differentiating the change over time in the transmittance or the reflectance and stopping the film formation when the differential value is zero, but even with this method, if there is noise, there is a large "flicker" in the differential value. Since it appears, it becomes difficult to accurately control the film thickness.

Since the enlargement method and the differential method have the above problems, they are disclosed in JP-B-57-24485 in order to enable more accurate film thickness control by using the monochromatic side light method. A method is proposed. This method includes a difference A between the average value of the reflectance or the transmittance in the vicinity of the extreme value after the start of vapor deposition and the reflectance or the transmittance in the absence of vapor deposition, and an arbitrary value after passing the extreme value of the reflectance or the transmittance. By setting the difference B between the point and the average of the extreme values and detecting that B / A indicates the target value, a film having a desired thickness is formed. The advantage of this method is that the detection of an arbitrary point after passing the extreme value of the transmittance or the reflectance has a relatively large change amount of the transmittance or the reflectance with respect to the change of the film thickness. That is, since the target value of B / A is used, the film thickness does not change even if the refractive index of the film being formed changes a little.

However, this method uses the extreme value of the transmittance or the reflectance that has passed immediately before when determining the target value as a reference, so that if the extreme value has a large error, the error is reflected as it is in the target value, This means that the high accuracy that should be obtained by this method cannot be obtained. Furthermore, it is difficult to determine the target value of the film thickness control when depositing multiple layers on the monitor substrate, and the 1/4 wavelength optical film thickness by the film thickness control method by the extreme value detection. There is also a drawback in that the "effect of naturally correcting the error in the deposited film thickness when the next coating film is deposited" obtained when the layers are laminated cannot be obtained.

As described above, in the conventional monochromatic light type film thickness monitoring method, the extreme value of the transmittance or the reflectance could not be reliably detected, so that it is difficult to control the optical film thickness of the coating accurately. There was a drawback.

[Problems to be Solved by the Invention] The present invention has hitherto been difficult. By attempting to detect the extreme value of the transmittance or reflectance of the monitor accurately with a simple method, it was attempted to eliminate the drawback that the accurate film thickness control method of the conventional film thickness control method cannot be obtained. is there.

[Structure of the Invention] That is, according to the first invention of the present invention, in the process of forming a vapor-deposited coating on a substrate, a change in transmittance or reflectance of a monitor substrate on which the coating is being formed is temporally changed from a certain point. Sampling was carried out for a fixed time up to the present time, and the 2-hour regression function P = a ₀ + a ₁ (t-tp) ² (where P is the transmittance or reflectance of the monitor substrate, a ₀ , a ₁
Is a constant, t is the elapsed time from the start of vapor deposition, tp is the extremum of the quadratic regression function, and time at which vapor deposition is started is used as a standard) until t = tp or near t = tp.
The film thickness control method in film deposition is characterized in that a value deviated by a predetermined time including zero is set as a film formation stop time with reference to a time point when t = tp.

The second aspect of the present invention is to form a vapor-deposited coating on a substrate by measuring the change over time in the transmittance or reflectance of the monitor substrate on which the coating is being formed. Sampling was performed and the quadratic regression function P = a ₀ + a ₁ (t-tp) ² (where P is the transmittance or reflectance of the monitor substrate, a ₀ , a ₁
Is a constant, t is the elapsed time from the start of vapor deposition, tp is the time at which the quadratic regression function reaches an extreme value, with the vapor deposition start as a reference). Of the film thickness control method in the film deposition, wherein a value deviated by a predetermined value based on the transmittance or reflectance of the monitor substrate on which the film is being formed is set as the film deposition stop time.

In the present invention, the vapor deposition film means a vacuum vapor deposition method, an ionization vapor deposition method, an ion plating method, an ion assisted vapor deposition method,
In the present invention, which means a film deposited by a sputtering method or the like, a ₀ , a ₁ and tp of the quadratic regression function P = a ₀ + a ₁ (t-tp) ² are the transmittance or reflectance of the monitor. The temporal change is calculated as follows from the sampling data of a certain time from a certain time to the present time.

a ₁ = α (3) N is the number of transmittance or reflectance sampling data used in the quadratic regression. Ti is the time at which the transmittance or reflectance sampling used in the quadratic regression was measured, and _N of t ₁ , t ₂ , ... t N. There is an individual. (Ie i = 1,2 ... N) P (ti) is the value of transmittance or reflectance measured at time ti, P (t ₁ ), P (t ₂ ),
…, _{N of} P (t _N ).

Means to take N sums of the terms from i = 1 to i = N.

Is.

In the present invention, it is usually preferable to set the film forming stop time when t = tp.

However, if the vapor-deposited substrate adheres a vapor-deposited film of a predetermined thickness after a certain time delay from the monitor substrate, or if it advances a certain time and a film of a predetermined thickness adheres, the maximum value of the transmittance or reflectance of the monitor substrate is reached. The film formation is not stopped when it is detected, and the time elapsed after each extreme value is measured, or when the predetermined time that has been set in advance has passed,
Or the time tp at which an extreme value is predicted and the current time t ₁
The film formation can be stopped when the difference from (t ₁ <tp) reaches the predetermined time set in advance.

When an interference filter longer or shorter than a desired wavelength for obtaining monochromatic light for measuring the transmittance or reflectance of the monitor substrate is used, the extreme value of the transmittance or reflectance of the monitor substrate is used. When the film thickness is detected, the film formation is not stopped, and the time elapsed after the extremum has been passed is measured, respectively, and when the pre-measured time comes, or when the extremum appears The film formation can be stopped when the difference between the current time t ₁ and the current time t ₁ (t ₁ <tp) is set in advance. Furthermore, when the vapor-deposited substrate is adhered to the vapor-deposited coating having a predetermined thickness with a certain time delay from the monitor substrate, the film formation is not stopped when the extreme value of the transmittance or the reflectance of the monitor substrate is detected,
It is also possible to stop the film formation when the maximum or minimum value is calculated from a regressed quadratic function and the difference between that value and the measured transmittance reaches a preset target value.

In this method, if the target value is made sufficiently large, the film formation end point will be a point sufficiently far from the extreme value, and the amount of change in the transmittance with respect to the change in the film thickness will be large, so the film thickness control accuracy will be Greatly improved.

[Operation] According to the present invention, the temporal change of the transmittance or the reflectance of the monitor substrate is obtained by sampling data from a certain time to the present time until a quadratic regression curve becomes close to the extreme value of the quadratic curve, and the extreme value is obtained. Therefore, even if there is noise in each sampling data, the time to reach the extreme value can be accurately determined, and the film formation stop time can be accurately set based on this extreme value.

Example 1 Example 1 In FIG. 1, a deposition target substrate 3 mounted on a deposition target substrate holder 2, a substrate heater 4, and a film thickness monitoring substrate (monitor glass) 5 were installed inside a cylindrical vacuum container 1. On the upper and lower sides of the vacuum container 1, monitor light transmission windows 6a and 6b are attached at positions corresponding to directly above and directly below the monitor glass 5, respectively. A film thickness monitor light source section A including a light source lamp 7 and a modulator 8 is installed below the transmission window 6a, and these are connected to a lamp lighting and modulator driving power source 9. Further, a film thickness monitor light receiving portion B including an interference filter 10 and a light receiving element 11 such as a photoelectric tube is installed above the transmission window 6b, and this light receiving portion is a photocurrent amplifier offset voltage generating portion (both shown in the figure. A / D converter 13, I / O interface 14 and computer 15
It is connected to the. Furthermore, this computer 15 has another I /
An O interface 16 is attached and a shutter plate 18 can be opened and closed via a shutter driver 17 according to a command from a computer. On the other hand, inside the vacuum container 1, the evaporation source 19
Is installed, and is opened and closed by the shutter plate 18 above it. The vacuum container 1 is connected to a vacuum pump via a conduit (not shown).

Next, the operation of the vacuum vapor deposition device will be described. Evaporation source
19 is heated by electron beam heating or resistance heating to evaporate the vapor deposition material, and when the evaporation rate of the vaporization material is stable, the shutter 18 is opened and the deposition of the vaporization material on the substrate 3 is started. At this time, the evaporated material also adheres to the monitor glass 5. It is important to stabilize the deposition rate of this coating, and it is stabilized by keeping the electric power for heating the evaporation source constant or by using a crystal vibration type deposition rate controller. The light projecting section A always irradiates the monitor glass 5 with a certain amount of light, and the light receiving section receives only the amount of light of a specific wavelength λ out of the light transmitted through the monitor glass 5 by the interference filter 10. It is measured with a photoelectric tube. The light projecting section is further provided with a modulator (chopper) 8 for modulating the monitor light at a specific frequency f.
By amplifying only the signal around the specific frequency f among the light signals received by the light receiving unit B at 12, the measured value is not affected by the stray light from the evaporation source or the like. After converting the transmitted light intensity of the monitor glass 5 into an electric signal in this way, an AD (analog / digital) converter
13, read by the I / O interface 14 into the computer 15 as a light intensity value. The transmittance of the monitor glass 5 during film formation calculated from the intensity of the sampled light theoretically draws a periodic curve as shown in FIG. 4, but this curve is usually as follows. Can be approximated by this polynomial (1), and in the vicinity of this extremum, approximation is performed even if the fourth-order term or less of the polynomial (4) is omitted.

T = a _o + a ₁ (t-tp) ² + a ₂ (t-tp) ⁴ + a ₃ (t-tp) ⁶ ＋ ……………… (4) (a _o , a ₁ , a ₂ , a ₃ ...... is constant, tp is the time appears of extrema) However, the transmittance of the monitor glass 5 is actually more complicated because of heterogeneity and the light receiving element-optical current amplifier nonlinearity of the refractive index of the coating Although it is a function, in the vicinity of the extreme value, it is sufficiently approximated to the following expression in which the fourth and subsequent terms of the polynomial (4) are omitted.

T = a _o + a ₁ (t-tp) ² (5) As shown in FIG. 3, if the change in theoretical transmittance is represented by the solid line in the figure, the quadratic regression line becomes the broken line in the figure. FIG. 2 shows a sampling value (solid line) of the transmittance of the monitor glass 5 during vapor deposition, showing that noise is superimposed, and the quadratic regression line obtained from this sampling data is shown by a broken line. Has been done. This quadratic regression line is Δ from the current time t _1.
It is calculated by the computer 15 using the equations (1), (2), and (3) based on the sampling data for a certain period of time before t. Here, if Δt is set sufficiently longer than the fluctuation cycle of the transmittance measurement value due to noise, it does not affect the accuracy of the quadratic regression line due to noise. Therefore, Δt is usually 10% or more of the time required to form an optical film having a quarter wavelength. When Δt exceeds 50% of the time required for the film formation, the obtained secondary regression line deviates from the transmittance curve.
It is set to 50% or less of the time required for the film formation. Then, until the second regression line t ₁ is equal to tp, i.e. determine the secondary regression line in extreme prompts to the current time t ₁ is tp, t ₁ = tp at immediately I / O interface A signal is sent through 16, the shutter 18 is closed and the evaporation from the evaporation source 19 is cut off.

Table 1 shows that titanium oxide was actually deposited on the monitor glass (refractive index 1.51) of BK7 and the film formation was stopped when the extreme value of the 1/4 wavelength optical film thickness was detected by the apparatus of the present invention. The difference between the optical film thickness (in vacuum) of the obtained film and 1/4 of the side light wavelength λ of the film thickness monitor is shown. As can be seen from this table, the film thickness error 1
It can be seen that it can be controlled at less than%.

When performing this experiment, the transmittance measurement value indicated by the film thickness monitor contained noise of ± 0.03% (absolute value of transmittance).
Extreme value detection by conventional expansion method (Noil level of transmittance ± 0.03% (in absolute transmittance) Optical film thickness measurement error ± 0.3% (1/4 wavelength optical film thickness is 100%
And ) The time of the data for the secondary regression was set to 300 seconds. ), The film thickness error is 2.5% (when the optical film thickness is an even multiple of 1/4 wavelength) or 2.0% (when the optical film thickness is an odd multiple of 1/4 wavelength)
Is expected to become.

As described above, according to the present invention, it is possible to accurately stop the film formation at the extreme value of the transmittance, and to obtain a thin film having an optical film thickness that is an integral multiple of 1/4 of a desired wavelength with good reproducibility. You can

Embodiment 2 In FIG. 5, the sputtering film forming apparatus 21 and the transmittance monitor units A ′ and B ′ may be known ones, and those having the same function as the vacuum vapor deposition apparatus shown in FIG. The numbers quoted in Step 2 are shown with dashes. By constantly rotating the substrate holder 2'of the sputtering film forming apparatus 21 during film formation, the monitor glass 5'and the substrate 3'pass through the bottom of the sputtering target 19 'many times to form a film. Transmittance monitor A ', B'is monitor glass 5'
Transmittance is measured only when is on the line shown in Fig. 5, that is, on the straight line connecting the light source 7 ', 8'and the light receiver 10', 11 ', and the other transmittances are measured. When there is not, the transmittance value measured immediately before by the film thickness monitor main body 12 'is held, and a signal as if continuously measured is obtained.
It is designed to be output to the AD (analog / digital) converter 13 '. By incorporating the film thickness monitoring devices 13 'to 16' similar to those shown in the first embodiment into such a sputtering apparatus, the first embodiment can be applied to the sputtering film forming method.
It was possible to perform the high-precision film thickness control similar to that obtained in.

In the above examples, the film thickness of the film was controlled by measuring the transmittance of the monitor glass, but it is clear that the film thickness can be controlled with high accuracy by measuring the film reflectance.

[Effects of the Invention] As described above, according to the present invention, the extreme value can be accurately detected even if there is some noise in the signal of the transmittance or the reflectance of the substrate during film formation. There is an effect that the film thickness can be controlled to a desired value with high precision, and in particular, high quality products can be reproducibly produced in the production of the optical multilayer film.

[Brief description of drawings]

1 shows an embodiment of the present invention. FIG. 1 is a schematic vertical cross-sectional view of a vacuum vapor deposition apparatus, FIG. 2 is transmittance sampling data (solid line) of a monitor glass and a quadratic function regressing the sampling data. Fig. 3 shows the theoretical transmittance (solid line) and the quadratic function that returns to it when a single layer of film is deposited on the monitor glass (dashed line), and Fig. 4 shows the higher refractive index on the monitor glass. FIG. 5 is a graph showing the relationship between the optical film thickness when the coating film is deposited and the transmittance (theoretical value) at a specific wavelength λ, and FIG. 5 is a schematic vertical sectional view of a sputtering apparatus showing another embodiment. Vacuum container, 2,2 '; Substrate holder, 3,3'; Substrate, 4; Substrate heater, 5,5 '; Monitor glass, 6a, 6b, 6
a ', 6b'; monitor light transmission window, 7,7 '; light source lamp, 8,
8 '; Modulator, 9; Power supply, 10,10'; Interference filter, 11,1
1 ': light receiving element, 12, 12'; film thickness monitor body, 13, 13 ';
AD converter, 14,14 '; I / O interface, 15,15; Calculator, 16,16'; I / O interface, 17,17 '; Shutter driver, 18,18'; Shutter plate, 19,19 ′; Evaporation source,
Or sputtering target, A, A '; Film thickness monitor light emitting part, B, B'; Film thickness monitor light receiving part

Claims (2)

[Claims] 1. A method of forming a vapor-deposited coating on a substrate, sampling the temporal change of the transmittance or reflectance of a monitor substrate on which the coating is being formed, from a certain point to the present time, and sampling the sample. From the data, the quadratic regression function P = a ₀ + a ₁ (t-tp) ² (where P is the transmittance or reflectance of the monitor substrate, a ₀ , a ₁
Is a constant, t is the elapsed time from the start of vapor deposition, tp is the extremum of the quadratic regression function, and time at which vapor deposition is started is used as a standard) until t = tp or near t = tp.
A film thickness control method in film deposition, wherein a value deviated by a predetermined time including zero from the time point t = tp is set as a film formation stop time point. 2. A method of forming a vapor-deposited coating on a substrate, sampling the temporal change of transmittance or reflectance of a monitor substrate on which the coating is being formed, from a certain point to the present time, and sampling the sample. From the data, the quadratic regression function P = a ₀ + a ₁ (t-tp) ² (where P is the transmittance or reflectance of the monitor substrate, a ₀ , a ₁
Is a constant, t is the elapsed time from the start of vapor deposition, tp is the time at which the quadratic regression function reaches an extreme value, with the vapor deposition start as a reference). (3) A film thickness control method in film deposition, wherein a value deviated by a predetermined value based on the transmittance or reflectance of the monitor substrate on which the film is being formed is set as the film deposition stop time.