JPS61111427A - Processing method of interference signal of fourier-transform spectrometer - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for processing interference signals in Fourier spectroscopy.

The blank space Fourier spectroscopy is used, for example, to measure the thickness of an epitaxially grown layer on a silicon single crystal substrate, or to measure the carbon concentration in a silicon single crystal substrate.

In this case, transmitted light or reflected light from the object to be measured is guided to a fixed mirror and a driven mirror of the Michelson interferometer, and is converted into interference light.

The thickness of the epitaxial growth layer creates an optical path difference between the light that has passed through this epitaxial growth layer and the light that has not passed through it, and the position of the driving mirror that produces a peak in the interference light intensity of these two lights changes in response to the optical path difference. Therefore, the interference light intensity can be measured based on the change characteristics of the interference light intensity when the position of the driving mirror is changed. In the above-mentioned measurement, the interference light intensity is detected as an electrical signal by a photoelectric conversion means. Further, the position of the margin drive mirror is electrically detected.

In this case, noise caused by various disturbances is added to the electrical signal indicating the interference light intensity. Further, when temperature fluctuations occur, mechanical distortion occurs due to thermal expansion of the material constituting the interferometer, and the amount of light received by the photoelectric conversion means changes undesirably depending on the position of the driving mirror.

As a result, the electrical signal indicating the interference light intensity is distorted. Measurement errors will occur due to the above-mentioned noise and distortion.

Since the noise caused by the above-mentioned disturbances has no correlation with the change characteristics of the interference light intensity, it is necessary to use an integration circuit that repeats detection multiple times and adds up the signals from multiple times. can be removed.

However, the above distortion cannot be substantially removed even by performing the above addition.

Furthermore, when noise is removed by addition as described above, if the number of additions is increased, it will take a long time for measurement.

Accordingly, one object of the present invention is to provide a method for processing interference signals in a Fourier spectrometer that reduces signal distortion.

Another object of the present invention is to provide a Fourier spectrometer interference signal processing method that can shorten measurement time.

In this invention, the position of the driving mirror where the relative intensity of the interference light is maximum is set as a reference position, and the electric signal corrected by mutually adding the electric signal levels at positions before and after positions equidistant from this reference position is calculated. obtain.

The relative intensity of the interference light converted into a voltage signal by the photoelectric conversion means changes, for example, as shown by the solid line in FIG. 1A. In the figure, the horizontal axis corresponds to the position of the driving mirror of the Michelson interferometer, and the vertical axis corresponds to the voltage signal level. The dashed line indicates the desired signal level when the interferometer is free from distortion due to temperature fluctuations, etc.

The reference position of the voltage signal is set at the position where the maximum level is indicated by the margin below. The reference position of this voltage signal corresponds to the reference position of the driving mirror.

The distortion caused by the mechanical distortion of the Michelson interferometer is substantially opposite in front and behind the reference position.

For example, in Figure 1A, the front strain S0 and the rear strain S・'
, strain S1 and strain S, /, etc. are reversed.

Therefore, in FIG. 1A, the signal obtained by adding together the front signal and the rear signal at the same distance from the reference position (x=0) is substantially as shown in FIG. 1B. The distortion will be removed.

As shown in Figure 2A, if the signal contains noise, this noise does not show any correlation with the signal as described above, so the relative level of noise in the signal obtained by the above addition is is substantially halved as shown in FIG. 2B.

As mentioned above, in order to obtain an electric signal corrected by addition, the interference light intensity at each position of the driving mirror of the Michelson interferometer is stored in a storage means, and the stored value and the new value are stored in the storage means. For example, it is possible to add the stored value in front of the reference position and the measured value in the rear of the reference position using the measured value obtained by the second measurement.

Also, using one type of memory value stored in the storage means,
The stored values before and after the reference position can be added together. When using one type of memorized value in this way,
The time required for measurement is short.

Since the above-mentioned corrected electrical signal contains substantially no distortion and only small noise as described above, it is possible to use this corrected electrical signal and other electrical signals to emphasize a specific part. You will be able to perform arithmetic processing adequately. Such arithmetic processing is used, for example, to measure the thickness of a thin epitaxially grown layer as shown in the next example.

FIG. 8 shows a block diagram of the embodiment. Same 1
In the figure, L is a light source that outputs infrared light with a wavelength of 400 to 4000 cm. is guided to the Michelson interferometer 1 via reflecting mirrors M4 to M6.

The Michelson interferometer 1 consists of a semi-transparent mirror HM, a fixed mirror FM, and a driving mirror DM driven by a control device 4.

The interference light from the Michelson interferometer 1 is guided to a photoelectric conversion device DBT and converted into an electrical signal as an analog quantity.

The analog electrical signal of the photoelectric conversion device DET is converted into a digital signal by an analog-to-digital conversion circuit 2 in order to facilitate arithmetic processing to be described later, and is input to a computer 5 via an interface circuit 3.

The computer 5 includes an arithmetic and control device 51 and a storage device 52, and stores the digital signal from the interface circuit 3 in the storage device 52.

It also controls the control device 4 to display the calculation results on the display device 6.

As shown in FIG. 4, an epitaxial growth layer 11 having a thickness d is formed on the surface of the silicon wafer IO.

For the epitaxial growth layer 11, silicon wafer lO
When the impurity concentration is high, the light transmitted through the epitaxial growth layer 11 can be reflected on the 1° surface of the silicon wafer. The silicon wafer 10 used is, for example, an N type containing antimony of 1018 pieces/cc or more. The epitaxial growth layer 11 may be of N type or P type, but for example, it is of N type containing lo 14.1 o IS pieces/cc of phosphorus.

As shown in FIG. 4, the infrared rays 2o from the reflecting mirror Ma (FIG. 3) become reflected light 21 reflected at the surface of the epitaxial growth layer 11 on the one hand, and on the other hand at the interface between the epitaxial growth layer 11 and the silicon wafer lO. This becomes reflected light 22.

In the Michelson interferometer 1, interference light with an intensity determined by the interference between the reflected lights 21 and 22 and mutual interference between the reflected lights 21 and 22 is obtained depending on the position of the driving mirror DM.

The relative level of the voltage signal obtained from the photoelectric conversion device DET changes as shown in FIG. 5A by changing the position of the driving mirror DM.

If the optical path from the semi-transparent mirror HM to the fixed mirror is 12 and the distance is 2□, and the optical path from the semi-transparent mirror HM to the driving mirror DM is 13 and the distance is J32, then the optical path is 12 when = J32.
The intensity of the interference light due to the interference of the reflected lights 21 and 22 passing through and 18 becomes maximum.

Optical path difference 21-g+ determined by the difference between the reciprocating distance on the optical path 12 and the reciprocating distance on the optical path 18 4. When the position of the driving mirror DM is shifted from the above-mentioned maximum position so that I becomes equal to the half wavelength of the infrared light used, the intensity of the obtained interference light becomes minimum. When the optical path difference 21-gx A21 is shifted by one wavelength of the infrared light used, the intensity of the interference light becomes maximum.

Due to the mutual interference of the reflected lights 21 and 22 in the Michelson interferometer 1, the extreme and minimum levels of the interference light intensity decrease as the optical path difference 212□ --e21 increases, and then increase again.

Assuming that the refractive index of silicon is n, the refraction angle is r, and the thickness of the epitaxial growth layer 11 is d, the optical path difference between the reflected lights 21 and 22 is 2ndcosr.

When the optical path difference zl-et-e21 of the Michelson interferometer 1 becomes equal to the optical path difference 2ndcosr between the reflected lights 21 and 22, the maximum or minimum level of the interference light intensity becomes the second maximum value.

In other words, when =fi, -ndcosr, the delayed reflected light 22 passes through the optical path 13, and the advancing reflected light 2
1 passes through the optical path 12, the extremely thick or minimum level of the interference light intensity becomes maximum again.

Further, when -e2=-e1+ndcosr, the delayed reflected light 22 passes through the optical path 12, and the advancing reflected light 21 passes through the optical path 13, so that the maximum or minimum level similarly becomes maximum again.

In measuring the thickness of the epitaxial growth layer 11, first, the computer 5 operates the control device 4, and the control device 4 drives the driving mirror DM. Although not essential, the control device 4 sends one input to the computer 5 for each step of the positional displacement of the driving mirror DM.
Kao6iJliilE&1i,L.

The analog signal from the photoelectric conversion device DET is
The digital conversion circuit 2 converts the signal into a digital signal. The arithmetic and control circuit 51 of the computer 5 selects an address in the storage device corresponding to the permutation of the timing pulse in synchronization with the timing pulse, and inputs the digital signal manually into the selected address via the interface circuit 3. to remember.

Therefore, by moving the driving mirror DM by a predetermined distance, a digital signal corresponding to an analog signal as shown in FIG. 5A at each position of the driving mirror is stored in each address of the storage device 52. The distance between the storage addresses of the storage device 52 corresponds to the moving distance of the driving mirror DIVI.

Next, the storage device 52 is operated by the calculation and control device 51.
The memory addresses are scanned, and the memory address storing the maximum value of the digitized photoelectric conversion signal is detected.

Next, using the storage address of the maximum value as a reference position, the arithmetic and control device 51 mutually adds the stored information of the previous and subsequent storage addresses that are equidistant from this reference position. Each added value is stored in a new respective address of storage device 52. By this addition, a signal from which distortion has been substantially removed can be obtained as described above.

Next, the arithmetic and control device 51 scans the memory address where the above added value is stored. This scanning is performed sequentially, for example, starting from the storage address where the maximum value is stored. By this scanning, the maximum or minimum level decreases and then increases to detect the storage address where the maximum or minimum level shows the maximum value, that is, the storage address where the second maximum value is found.

Next, the calculation and control device 51 determines the distance between the storage address where the maximum value is stored and the storage address indicating the detected second maximum value, and the refractive index of silicon stored in the storage device 52 in advance. The thickness of the epitaxial growth layer 11 is calculated using the tangent angle and the oil tangent angle.

Next, the calculation results are displayed on the display device 6.

The method for measuring an epitaxially grown layer described above can be further improved by using a reference epitaxially grown layer as described below.

The improved measurement method uses a reference epitaxial growth layer that is different in thickness from the epitaxial growth layer to be measured, and subtracts the signal obtained by the reference epitaxial growth layer from the signal obtained by the epitaxial growth layer to be measured. In particular, extract only the signals necessary for thickness measurement.

If the epitaxial growth layer to be measured has a thickness of, for example, 20 μm or less, the thickness of the reference epitaxial growth layer is, for example, 40 μm or more.

For thick reference epitaxially grown layers.

The position of the driving mirror DM where the interference light intensity reaches the second maximum value is far away from the reference position. FIG. 5B shows the interference luminosity characteristics in a thick epitaxially grown layer of n thickness used as a reference. The change in the vicinity of the reference position is determined by the infrared rays used, and is the same as the change in A in the same figure.

In this improved measurement method, the measured values of the reference epitaxial growth layer are measured in advance using the same device and corrected by the addition as described above, and are stored in the storage device 52 of the computer 5.

In the measurement, first, the epitaxial growth layer to be measured is measured in the same manner as described above, and the measured value corrected by addition is stored in the storage device 52.

Next, the memory address for storing the measured value of the epitaxial growth layer to be measured and the memory address for storing the measured value for the reference epitaxial growth layer are associated one-to-one, and the measured value of the epitaxial growth layer for measurement is used to create the reference epitaxial growth layer. The measured value is subtracted, and the result is stored in the memory address where the measured value of the epitaxial layer to be measured is stored or in another address set.

The data obtained by this digital subtraction corresponds to analog data as shown in FIG. 5C. Peaks in the data at the reference location are substantially removed by the subtraction.

! Next, each memory address storing the above-mentioned subtraction data is scanned, and the memory address indicating the maximum value of the extremely thick or extremely small value is detected.

The thickness of the epitaxial growth layer to be measured is calculated based on the distance between the reference position address and the detection address, and the preset refractive index and tangent angle data as described above, and then this calculation result is calculated. It is displayed on the display device 6.

When the thickness of the epitaxial growth layer to be measured is thin, the position where the second maximum value occurs approaches the reference position, and the signal forming the maximum value at the reference position is combined with the signal forming the second maximum value. It will be done. However, the addition as described above removes signal distortion, and the subtraction as described above still allows the signal component forming the maximum value at the reference position to be removed from the signal component forming the second maximum value.

As a result, the improved measurement method measures particularly thin epitaxially grown layers with high precision and good reproducibility.

In FIG. 6, the horizontal axis shows the thickness TH of the epitaxially grown layer, and the vertical axis shows the difference R between the maximum measured value and the minimum measured value in each of the 10 thickness measurements.

In the same figure, curve A shows the characteristics obtained by the above-mentioned improved method of subtracting the measured value of the reference epitaxial growth layer from the measured value of the epitaxial growth layer to be measured corrected by addition, and curve B shows the characteristic of the measured value of the reference epitaxial growth layer without correction by addition. The characteristics obtained by subtracting the measurement value of the reference epitaxial growth layer without similar correction from the measurement value of the measurement epitaxial growth layer are shown.

As is clear from the figure, in the case of B, when the thickness of the epitaxially grown layer is 92 μm or less, measurement variations occur rapidly, making it impossible to perform measurements with good reproducibility. There are still large variations in measurement, even if it is 2 μm or more.
The measurement error is also large.

On the other hand, the improved method A has small measurement variations and exhibits sufficient reproducibility even when the thickness is 2 μm or less. Moreover, if it is 2 μm or more, only a substantially negligible measurement variation is exhibited.

The above-mentioned improved measurement method is applicable even when the epitaxial growth layer to be measured is relatively thick. In this case, similarly to the above, the epitaxial growth layer for the reference and the epitaxial growth layer to be measured are Layer thickness can be changed. For example, the thickness of the epitaxial growth layer to be measured is 20
If it is more than μm, the thickness of the reference epitaxial growth layer is 4
It is considered to be less than μm.

In the improved embodiment described above, the measured value of the epitaxial growth layer for reference is subtracted from the measured value corrected by addition, but the present invention is not limited to this.The measured value of the epitaxial growth layer for reference is subtracted first, and then Correction may be performed by addition.

[Brief explanation of drawings]

Figures 1A and B and Figures 2A and B are curve diagrams showing interference light intensity characteristics, Figure 8 is a block diagram of the measuring device of the embodiment, and Figure 4
The figure is an explanatory diagram to explain infrared reflection in silicon, Figure 5 to C is a curve diagram showing interference light intensity characteristics, and Figure 6 is a distribution diagram showing variations in thickness measurement of the epitaxial growth layer. be. L...Infrared light source, M1-M6...Reflector, FM...
・Fixed mirror, H・M...Semi-transparent mirror, DM...Driving mirror, D
ET... Photoelectric conversion device, 1... Michelson interferometer, 2... Analog-digital conversion circuit, 3... Interface circuit, 4... Control device, 5... Computer, 6...・Display device. To Fig. 1 Fig. 2 △ Fig. 1 6' Fig. 2 B Fig. 4 Fig. 6 T) ((, [m) -

Claims (1)

[Claims] 1. Obtain an interference signal corresponding to the position of the driven mirror at least twice from an interferometer having a driven mirror, and match the reference point at which the signal strength of each of the two obtained interference signals is maximum,
An interference signal processing method for a Fourier transform spectrometer, characterized in that one interference signal is reversed with respect to the point as a reference and added to the other interference signal.