KR102091419B1 - System of polishing substrate with light transmitting layer - Google Patents

Abstract

The present invention relates to a polishing system for a substrate, which detects feature points from reflected light reflected from a polishing layer of a substrate during a polishing process, and from values of a feature vector from a reference position or feature point for a reference wavelength to feature points of another selected wavelength. It is configured to detect the thickness of the polishing layer during the polishing process, and has the effect of accurately obtaining the thickness of the polishing layer during the polishing process while minimizing the amount of calculation.

Description

Substrate polishing system having a light-transmitting polishing layer {SYSTEM OF POLISHING SUBSTRATE WITH LIGHT TRANSMITTING LAYER}

The present invention relates to a substrate polishing system having a light-transmitting polishing layer, and more particularly, to a substrate polishing system that improves data processing speed while accurately sensing the polishing thickness of the light-transmitting polishing layer during a polishing process.

The chemical mechanical polishing (CMP) system is used for wide area flattening to remove the height difference between the cell area and the peripheral circuit area due to irregularities on the surface of the substrate generated by repeatedly performing masking, etching, and wiring processes during the semiconductor device manufacturing process, and for forming circuits. This is a system used to precisely polish the surface of a substrate in order to improve the surface roughness of the substrate due to the separation of the contact / wiring film and the highly integrated device.

In such a CMP system, the polishing head presses the substrate to perform a polishing process with the polishing surface of the substrate facing the polishing pad before and after the polishing process, and at the same time, when the polishing process ends, the substrate is directly and indirectly It moves to the next process in a state of being gripped by vacuum adsorption.

1 is a schematic diagram of a general chemical mechanical polishing system 9. As shown in FIG. 1, the chemical mechanical polishing system 9 is polished while the substrate W is pressed by the polishing head 20 on the polishing pad 11 of the polishing platen 10 to rotate 11d. At the same time, wet polishing is performed while the slurry is supplied from the slurry supply unit (not shown) on the polishing pad 11. And, in this process, while the conditioner 40 rotates (40d) and rotates, the conditioning disk surface-modifies the polishing pad 11, so that the slurry is transferred to the substrate W through the fine grooves of the polishing pad 11 To be supplied smoothly.

On the other hand, with the integration of semiconductor devices, it is necessary that the polishing layer of the substrate W is precisely polished. To this end, in the prior art, as disclosed in U.S. Patent No. 6190234, in the process of performing the polishing process, the light emitting unit 50 receives the irradiated light Li received from the light source 55, the polishing layer of the substrate W , And receives the reflected light (Lo) from the polishing layer at the light receiving unit 60, and receives the received reflected light by the spectrometer 65 as shown in FIGS. 2A and 2B. After obtaining the spectrum of, it was attempted to detect the thickness of the abrasive layer during the polishing process in a manner to grasp the thickness of the abrasive layer, in contrast to the optical interference spectrum according to the thickness stored in advance.

That is, when the irradiation light Li containing a plurality of wavelengths from the light emitting unit 50 of FIG. 1 is supplied from the light source 55 and irradiated to the polishing layer of the substrate W, the reflected light received from the light receiving unit 60 ( Lo) is also a form in which multiple wavelengths are combined. Accordingly, the light intensity distribution of the optical interference signal for each wavelength by spectralizing the reflected light Lo by the spectrometer 65, when the thickness of the polishing layer is sufficiently thick, the spacing Xi becomes dense as shown in FIG. 2A. , As the thickness of the abrasive layer becomes thinner, the gap Xo increases as shown in FIG. 2B.

Since the spectral distribution according to the wavelength follows a predetermined shape according to the variation of the thickness of the polishing layer, the distribution data of the spectrum previously stored in the memory according to the material of the polishing layer of the substrate and the spectrum obtained by the spectrometer 65 By contrasting method, the thickness of the polishing layer was detected during the polishing process.

However, since such a method actually requires optical interference signal data for continuous wavelength values, the spectrometer 65 calculates the optical interference signal according to the wavelength in real time, and the spectrum calculated during the polishing process is previously stored in the memory. There was a problem in that it takes a long calculation time to compare with the distribution data of the spectrum stored in the library form. Accordingly, there is a problem that a polishing system is expensive due to the need for a large amount of computing equipment (for example, a computer) for fast calculation, and it takes a long time to obtain a spectrum for continuous wavelengths, thus real-time As a result, there was a limit in obtaining a polishing layer thickness.

In addition, since it is necessary to store the spectrum of the data previously tested in a library, it is necessary to collect data through various experiments in advance, and there is a problem in that the accuracy of detection varies depending on the quality of the collected data. Moreover, if the pattern of the abrasive layer or the structure of the lower thin film were different, there was also the inconvenience of separately creating a library.

On the other hand, according to U.S. Patent Publication No. 6190234, there is disclosed a configuration for detecting a polishing end time of a polishing layer of a substrate by using an optical interference signal according to thickness variation for two different wavelengths. That is, the optical interference signal for the two wavelengths has a predetermined pattern that fluctuates as the thickness decreases according to a predetermined polishing layer material, so as shown in Table 3 of column 11 of the U.S. Patent Publication, the pattern fluctuating in advance It is a method that stores the data in advance and tracks the fluctuation of the measured values of the optical interference signals at two wavelengths during the polishing process to terminate the polishing process when these values reach a predetermined value.

However, with the configuration disclosed in the above-mentioned U.S. Patent Publication, the thickness of the abrasive layer is 35704 각각 only at the values (N = 4, N = 15) where the two wavelengths match each other within the tolerance range (for example, 50 Å or less). , It can be known as 10002Å, but there was a limit to not know the absolute thickness of the abrasive layer at all in the interval. Moreover, when the thickness of the polishing layer of the substrate is 35000 Å or less before the polishing process, there is only one point (N = 4) where the absolute thickness value of the polishing layer is known, so when N = 4 is reached, 10002 Å or 35704 Å Cognition is also unknown.

As described above, even in the form of following the optical interference signal according to the change of time using two or more wavelengths in the prior art, it is difficult to know the absolute value of the thickness of the polishing layer during the polishing process, and only the final polishing end point is detected. Since it is possible to do not know whether the thickness of the abrasive layer in the polishing process is close to the target thickness at which polishing is finished or whether there is sufficient margin from the target thickness, it is difficult to finish the polishing process accurately at the end of polishing unless it is an experienced worker. There was also a problem.

Accordingly, there is an urgent need for a method for sensing the absolute thickness of the polishing layer during the polishing process, without requiring a high-performance computing facility while reducing the processing time in the substrate polishing system.

The present invention was created under the above technical background, and the present invention aims to obtain the absolute thickness of the polishing layer during the polishing process.

At the same time, the present invention aims to obtain the thickness of the abrasive layer during the polishing process by minimizing computation and even with low specification computing equipment.

That is, an object of the present invention is to detect the absolute thickness of the abrasive layer using only theoretical prediction values without relying on experimental data.

In addition, an object of the present invention is to obtain the absolute thickness of the polishing layer with a relatively small amount of calculation without storing in advance a library for obtaining the thickness of the polishing layer of the substrate during the polishing process.

In order to achieve the above object, the present invention detects feature points from reflected light reflected from the polishing layer of a substrate in a polishing process, and from values of feature vectors from a reference position or feature point for a reference wavelength to feature points of another selected wavelength. The polishing layer thickness during the polishing process is sensed.

According to the present invention, there is an effect of obtaining the absolute thickness of the polishing layer during the polishing process of the substrate provided with the light-transmitting polishing layer.

At the same time, the present invention has an effect of accurately obtaining the thickness of the polishing layer during the polishing process even with a low-capacity facility within a fast computation time by minimizing the computation amount for obtaining the polishing layer thickness.

That is, the present invention does not rely on the experimental data and detects the absolute thickness of the abrasive layer using only theoretical prediction values, and thus does not require various experimental data according to specifications such as patterns or substructures of the abrasive layer prior to the polishing process. No, the effect of solving the problem of errors in detection accuracy is obtained according to the quality of the experimental data obtained in advance.

1 is a view showing the configuration of a general substrate polishing system,
Figure 2a is a graph showing the optical interference signal data according to the wavelength of the initial polishing,
Figure 2b is a graph showing the optical interference signal data according to the wavelength of the end of polishing,
Figure 3a is a front view showing the configuration of a substrate polishing system according to an embodiment of the present invention,
Figure 3b is a plan view of Figure 3a,
Figure 3c is a front view showing the configuration of a substrate polishing system according to another embodiment of the present invention,
Figure 4 is a flow chart for explaining the principle of operation of the polishing system of Figure 3,
5 is a view for explaining the principle of generation of an optical interference signal in the polishing layer of the substrate;
6A and 6B are graphs showing optical interference signal data according to the wavelengths of the beginning and the end of polishing, and for explaining the variation of the optical interference signal according to the variation of the polishing layer thickness for a predetermined wavelength;
7 is a view showing optical interference signal data according to a change in thickness of a substrate polishing layer for a plurality of wavelengths,
8 is a view showing optical interference signal data according to a polishing process of a substrate polishing layer with respect to the first wavelength of FIG. 7;
9A is a view for explaining a theoretical feature vector of a theoretical optical interference signal for a first reference wavelength;
9B is a view for explaining another theoretical feature vector of the theoretical optical interference signal for the first reference wavelength;
9C is a view for explaining a theoretical feature vector of a theoretical optical interference signal for a second reference wavelength;
10 is a view for explaining a measurement feature vector of a measurement optical interference signal;
11 is a graph showing the results of thickness calculation during the polishing process according to the present invention.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings. However, in describing the present invention, detailed description of known functions or configurations will be omitted to clarify the gist of the present invention.

As shown in the figure, the polishing system 1 of the substrate according to an embodiment of the present invention is for flat polishing the polishing layer f formed on the bottom surface of the substrate W, and the polishing pad 11 on the top surface Substrate (10) is coated and rotated (11d), the polishing platen (10), a slurry supply unit (not shown) for supplying a slurry on the polishing pad (11), and the substrate (W) placed in the lower state during the polishing process ( Measuring the thickness of the abrasive layer (f) of the substrate (W), the conditioner (40) for modifying while pressing the abrasive pad (11) while rotating (40d) during the polishing process, and pressing the W) In order for the light sensor (100, 100 ') and the light sensor (100, 100') equipped with a light receiving unit for irradiating the irradiation light (Li) and a light receiving unit for receiving reflected light (Li) reflected from the polishing layer (f) It comprises a control unit 200 for obtaining the thickness of the substrate polishing layer f during the polishing process from the reflected light received by the light receiving unit.

A transparent window 111a is formed on the polishing platen 10 and the polishing pad 11 to emit light from the optical sensor 100 on the polishing surface of the substrate W that is being polished from the bottom of the polishing platen 10. The light Li is irradiated from the part, and it becomes possible to receive the reflected light Lo reflected from the polishing layer f of the substrate W at the light receiving part of the optical sensor 100. In this case, since the light-receiving portion of the optical sensor 100 is fixed at a certain position with respect to the polishing pad 11, if the oscillation motion of the substrate W is ignored, the reflected light Lo following a constant trajectory of the substrate W ) Signal. In the drawing, a configuration in which the light emitting part and the light receiving part are formed as one body is illustrated for convenience, but the light emitting part and the light receiving part may be formed as separate bodies.

On the other hand, the optical sensor 100 is disposed below the transparent window 11a penetrating the polishing platen 10 and the polishing pad 11 together, and the light emitted from the optical sensor 100 irradiates the transparent window 11a. It may be disposed to reach the substrate through, and the reflected light reflected from the substrate polishing layer may pass through the transparent window and be received by the optical sensor. In parallel with or instead of this, while forming a groove on the top surface of the polishing platen 10 while penetrating through the polishing pad 11, or forming a groove through a part of the polishing pad 11, an optical sensor ( 100 ') may be arranged to irradiate the substrate polishing layer f with the irradiated light emitted from the optical sensor 100' and receive the reflected light reflected from the substrate polishing layer f. In this case, the light receiving portion of the optical sensor 100 'receives the reflected light Lo from the substrate W along the rotational trajectory of the polishing pad 11 passing through the lower side of the substrate W.

Hereinafter, for convenience, a light sensor formed by a light emitting unit and a light receiving unit as one body, for example, irradiating light (Li) from the light emitting unit of the light sensor and receiving the reflected light (Lo) from the light receiving unit 130 of the light sensor. Explain.

The substrate W is formed of a light-transmitting material through which the light is transmitted by the polishing layer f in the process of manufacturing a semiconductor device. Here, the 'light transmissive material' includes not only that all of the irradiated light Li irradiated from the light emitting unit is transmitted, but also that only a portion of 1% or more of the light Li irradiated from the light emitting unit 120 is transmitted. Includes. For example, the polishing layer f may be formed of an oxide layer, and accordingly, as illustrated in FIG. 5, a part of the irradiated light Li irradiated is reflected light reflected from the surface of the polishing layer f ( Loe), and the other part of the irradiated light Li passes through the polishing layer f and becomes reflected light Loi reflected from the surface of the non-transmissive layer wo.

The polishing head 20 may be formed in a variety of well-known shapes or structures, and the polishing surface of the substrate W is continuously attached to the polishing pad 11 in a state where the substrate W is positioned downward during the polishing process. It serves to maintain contact.

For example, the polishing head 20 is formed by receiving a rotational driving force from the outside, a rotating body, a base rotating together with the main body, and a bottom plate in the form of a disk in the shape of a substrate W and fixed to the base. The formed flexible membrane and the ring shape surrounding the periphery of the substrate W are maintained in close contact with the polishing pad during the polishing process so that the substrate W is pushed out of the polishing head 20 during the polishing process. It can be configured to include a retainer ring to suppress the.

Here, in the membrane, the end of the fixed flap in the form of a ring extending upward from the flexible base plate is fixed to the base, thereby forming a plurality of pressure chambers between the membrane base plate and the base. Then, as each pressure chamber of the polishing head is supplied with air pressure from the pressure adjusting unit and the pressure is controlled independently, the substrate W located at the lower side of the membrane bottom plate is pressurized during the polishing process with different pressing forces for each pressure chamber. It can be polished.

For example, the light-receiving units of each of the optical sensors 100 and 100 'respectively receive reflected light Lo reflected at the position of the polishing layer f of the substrate W to which the irradiated light is emitted from the light emitting unit. When the thickness of the polishing layer of the substrate W is obtained during the polishing process for each position of the substrate polishing layer f on which the reflected light Li is reflected, using the received reflected light, the pressure adjusting unit has a relatively high polishing layer thickness. The upper pressure chamber at the substrate position detected as being increased to a higher pressure to increase the pressing force on the substrate to increase the removal rate per unit time (Removal Rate), and the upper pressure chamber at the substrate position detected as a relatively low polishing layer thickness It can be controlled to lower the pressing rate on the substrate by adjusting to a lower pressure to lower the polishing rate per unit time.

The conditioner 40 performs a reciprocating sweep motion across the conditioner so that the conditioning disk is in contact with the polishing pad 11 and has a radial component of the polishing pad 11. At this time, the pressing force of the conditioning disk may be kept constant, but according to the thickness information of the polishing layer of the substrate, the pressing force is applied to the polishing pad 11 passing through the substrate position where the thickness of the polishing layer of the substrate is detected to be relatively high. By lowering the height of the polishing pad and adjusting the height of the polishing pad, the polishing pad (such as adjusting the height of the polishing pad low by increasing the pressing force) for the polishing pad 11 passing through the substrate position where the thickness of the polishing layer of the substrate is sensed to be relatively low. The height of 11) can also be adjusted to intentionally deviate.

The optical sensor 100 includes a light emitting unit that irradiates the irradiation light Li toward the polishing layer f of the substrate W, and a light receiving unit that receives the reflected light Lo reflected from the polishing layer f. As shown in FIG. 5, according to another embodiment of the present invention, the light-emitting portion irradiates the substrate polishing layer (f) obliquely with irradiation light (Li), and the light-receiving portion obliquely reflected by the substrate polishing layer (f) ( Lo). However, as shown in FIG. 3B, in order to eliminate the complexity of calculation by the incident angle and the reflected angle of the irradiated light Li and the reflected light Lo, the light emitting part of the optical sensor 100 irradiates the irradiated light Li with the substrate polishing layer It is preferable to irradiate vertically to (f) and to receive the reflected light (Lo) vertically reflected from the substrate polishing layer (f) by the light-receiving portion of the optical sensor (100).

Here, the irradiated light Li irradiated by the light emitting unit is defined as light having two or more wavelengths, and may be preferably light containing about 5 to 15 wavelengths. In this aspect, the irradiation light (Li) may be a white light that combines light of a plurality of consecutive wavelengths, for this purpose, the light source 105 of the irradiation light (Li) irradiated to the light emitting portion of the optical sensor 100 is an LED ( LED). When the irradiated light Li is white light, as described later, two or more selected wavelengths are selected from a plurality of consecutive wavelengths, and the optical interference signal for the selected wavelength is processed to process the substrate during the polishing process. The absolute thickness of the abrasive layer can be obtained.

In order to obtain the absolute thickness of the polishing layer without a time delay during the polishing process, the more the number of selected wavelengths, the more advantageous. The absolute thickness of the layer can be obtained almost continuously. In addition, the selected wavelength may be selected in a wavelength band of 4000 kHz to 7000 kHz, and the selected wavelengths may be selected at uniform intervals from each other (for example, 100 to 700 kHz), for example, determined by an interval of 400 kHz. In the case, it is preferable to set a uniform interval (that is, a wavelength interval of 320 Hz to 480 Hz) within an error range of 80 Hz, which is 20% of 400 Hz, which is a wavelength interval between each other.

On the other hand, the irradiation light (Li) irradiated to the polishing layer (f) of the substrate, the substrate polishing layer (f) through the light emitting portion of the optical sensor 100 from the light source 105 for outputting white light having a plurality of continuous wavelengths Irradiated to, the reflected light having a plurality of wavelengths is received through the light receiving unit of the optical sensor 100.

Here, as shown in Figure 3a, the irradiation light (Li) generated from the light source 105 is transferred to the main optical fiber 101 extending from the light source 105 to the optical sensor 100 in the irradiation light path to polish the substrate The reflected light (Lo) irradiated to the layer and reflected by the substrate polishing layer (f) is received by the light receiving unit and transmitted through the main optical fiber, and then transmitted along the optical fiber 102 for branching (103) in the Y-shape from the main optical fiber 102 It is received by the control unit 200.

On the other hand, as shown in Figure 3c, according to another embodiment of the present invention, instead of using a Y-shaped branch 103, the half mirror (half) between the light source 105 and the spectrometer 230 The mirror 104 is disposed, and the irradiated light Li emitted from the light source 105 is reflected by the half mirror 104 to reach the polishing layer f of the substrate, and the reflected light reflected from the substrate polishing layer f (Lo) may be configured to pass through the half mirror 104 and be transmitted to the spectrometer 130.

The control unit 200 is a control device of the polishing system 1 of the substrate, using a refractive index (n) according to the material of the polishing layer (f) of the substrate W for a plurality of selected wavelengths for the thickness of the polishing layer The calculation unit 210 for calculating the theoretical optical interference signal, and the normalization module 220 for normalizing the reflected light Lo received from the light receiving unit of the optical sensor 100 to obtain the thickness of the substrate polishing layer at a constant light intensity And, during the polishing process based on the spectrometer 230 to generate the reflected light (Lo) received from the light receiving unit of the optical sensor 100 as an optical interference signal for each wavelength, and the optical interference signal generated by the spectrometer 230 It comprises a thickness calculating unit 240 for sensing the thickness of the substrate polishing layer.

The calculation unit 210 calculates the theoretical optical interference signal data according to a predetermined selected wavelength in consideration of the refractive index n according to the material of the polishing layer f of the substrate W (S10). Here, the theoretical optical interference signal data refers to obtaining the position information of the waveform of the optical interference signal according to the thickness of the polishing layer according to the material of the polishing layer and the characteristic values of the waveform for each wavelength in the form of a feature vector.

In other words, when the material of the polishing layer of the substrate is determined, the refractive index of the polishing layer is determined, and accordingly the period of the thickness of the polishing layer is determined by (λ / 2n) (where λ is the wavelength of light and n is the refractive index Im), it is possible to obtain a period of the thickness of the polishing layer according to the wavelength (λ) of light, through which the theoretical light interference signal (light intensity, intensity) increases from the point where the thickness of the polishing layer (f) of the substrate (W) is 0 ).

More specifically, in the light intensity signal of the optical interference signal of the reflected light reflected from the light-transmitting polishing layer, when the thickness of the polishing layer fluctuates by polishing, the intensity of light reflected from the polishing layer is the remaining thickness of the polishing layer. Depending on the periodic change, the changing period is determined by the wavelength (λ) of light and the refractive index (n) of the thin film.

Here, the light intensity (Intensity) is proportional to cos (4π * n / λ * t). Here, t is the thickness of the polishing layer, n is the refractive index of the polishing layer, and λ is the wavelength of light. From this, the polishing layer thickness period is determined as (λ / 2n), and the optical interference signal of the same pattern is repeated for each λ / 2n, and the time period T1 required for one cycle of the polishing layer thickness is (λ / (2n *). RR)). Here, RR is a removal rate per unit time. In addition, the phase of the reflected light at the interface where the thickness of the thin film is 0 is 0, and the light intensity at this time is 1, which is the maximum.

According to the characteristics of the light intensity (optical interference signal) according to the light interference in the light-transmitting polishing layer f, the calculating unit 210, although the polishing rate RR per unit time cannot be known in advance, the thickness of the polishing layer The optical intensity (optical interference signal) according to the optical interference for change can be obtained for each wavelength.

According to one embodiment of the present invention, since the white light is applied as the light source 105 and countless wavelength light that is continuously included in the reflected light Lo, the operation unit 210 includes a plurality of predetermined selected wavelengths λ1 and λ2 , λ3, λ4, λ5, λ6, the theoretical optical interference signal for the thickness t of the abrasive layer is calculated as shown in FIG. 7 in consideration of the refractive index n depending on the material of the abrasive layer.

Here, the optical interference signal data according to the abrasive layer thickness calculated by the calculator 210 in advance to obtain the abrasive layer thickness during the polishing process, as described above, 5 to 15 preselected in the wavelength band of 4000Å to 7000Å It is sufficient to calculate only the selected wavelengths of the degree (λ1, λ2, λ3, λ4, ....). In addition, the preselected wavelength may be selected at substantially uniform intervals (for example, 100 to 700 Hz).

For reference, FIG. 7, for convenience of explanation, selects the first wavelength (λ1), the second wavelength (λ2), the x wavelength (λx), and the y wavelength (λy) having non-uniform spacing from each other as a selection wavelength. Thus, the data of the 'theoretical optical interference signal' calculated and calculated by the calculating unit 210 is displayed according to the thickness of the polishing layer.

That is, the 'theoretical optical interference signal' and similar terms described in the present specification and claims are calculated by the calculating unit 210 of the control unit 200, and the optical interference signal according to the thickness of the polishing layer for the selected wavelength and data related thereto. Is defined as

On the other hand, if the computational unit 210 knows the theoretical optical interference signal for the selected wavelengths in a very short time by a simple calculation if the refractive index n of the polishing layer is known, there is no need to store it in the form of a library in memory. It is sufficient to calculate the theoretical steel tube interference signal in advance by the calculator 210 before the substrate W is supplied to the polishing head 20 and the polishing process is performed. Alternatively, the theoretical light interference signal is calculated in real time by the operation unit 210 after the polishing process of the substrate W is started, but it may be compared with the measured light interference signal extracted from the reflected light Lo received by the light receiving unit.

In this way, for a plurality of preselected wavelengths, a 'theoretical feature vector' is pre-calculated. Here, the theoretical feature vector is position data of feature values such as a peak value and a valley value of the theoretical optical interference signal, and among the selected wavelengths, the selected wavelength is different from the feature value of the reference wavelength. It means the vector about the relative distance and direction to the characteristic values of.

Hereinafter, for convenience of description, a method of obtaining the absolute thickness value of the abrasive layer by previously determining four selected wavelengths will be described.

For example, referring to FIG. 9A, the first wavelength λ1 is set as a reference wavelength, and the first peak value P1a of the theoretical optical interference signal of the first wavelength λ1, that is, the first characteristic value described in the claims With respect to the position where the polishing layer reaching) becomes the thickness of t1 as a reference position, it relates to the direction and distance from the other selected wavelengths (λ2, λx, λy) to the surrounding valley values (P2, Px, Py). As a vector, theoretical feature vectors are defined as [e2, ex, ey], respectively. (In the following, for convenience, [] is a vector.) Here, ex, ey, and e2 are in different directions, so if they are converted to relative positions, they can be expressed as (e2, ey, -ex).

In addition, a polishing layer having a first wavelength λ as a reference wavelength and reaching a first peak value P1a of the theoretical optical interference signal of the first wavelength λ1 (that is, the first characteristic value described in claims) A vector relating to the direction and distance from the position where the thickness of t1 becomes the reference position to the peripheral peak values (P2 ', Px', Py ') of other selected wavelengths (λ2, λx, λy), The theoretical feature vectors are defined as [e2 ', ex', ey '], respectively. Here, ex, ey, and e2 are in the same direction, so if they are converted to relative positions, they can be expressed as (e2, ey, ex).

On the other hand, referring to Figure 9b, the first wavelength (λ1) as a reference wavelength, the first wavelength (λ1) of the first peak value (P1a) of the theoretical light interference signal and the first valley value (V1a) of the intermediate value (Reference position, R1, i.e., the first intermediate value of the claims) A position where the polishing layer reaches the thickness of t2 is used as a reference position, from which the surrounding valley values of different selected wavelengths (λ2, λx, λy) As a vector about the direction and distance to (P2, Px, Py), the theoretical characteristic vector is defined as [E2, Ex, Ey]. Here, since E2, Ex, and Ey are all in the same direction, conversion to a relative position can be expressed as (E2, Ex, Ey).

Similarly, with the first wavelength λ1 as the reference wavelength, the intermediate value (reference position, R1, of the first peak value P1a and the first valley value V1a of the theoretical optical interference signal of the first wavelength λ1) That is, the position at which the polishing layer reaching the first intermediate value of the claims becomes the thickness of t2 is used as a reference position, from which the peripheral peak values P2 ', Px' of other selected wavelengths λ2, λx, λy , Py '), and the theoretical feature vectors are defined as [E2', Ex ', Ey'], respectively. Here, since E2, Ex, and Ey are all in the same direction, conversion to a relative position may be expressed as (E2 ', Ex', Ey '). As described above, if the reference position is determined as the intermediate value between the peak value and the valley value, the distance between the peak value and the valley value of the optical interference signal of a selected wavelength other than the reference wavelength is shortened, and the direction of the feature vector is also constant. It is possible to obtain an effect of reducing errors in contrasting the measurement feature vector of the measurement optical interference signal measured during the process.

Even in the case of using the intermediate value as the reference position, similar to using the feature value as the reference position, the theory is based on the reference wavelength of at least one of the selected wavelengths other than the first wavelength (for example, the second wavelength). Feature vectors can be obtained.

Meanwhile, the reference wavelength need not be limited to one, and as shown in FIG. 9C, the second wavelength λ2 is set as another reference wavelength, and the second valley of the theoretical optical interference signal of the second wavelength λ2 ( The position where the polishing layer reaching P2, i.e., the second feature value described in the claims, becomes the thickness of t3 as a reference position, from which the peripheral valley values V1a of different selected wavelengths λ1, λx, λy, As a vector about the direction and distance to Px, Py), the theoretical feature vectors are respectively defined as [f1, fx, fy]. Here, since f1, fx, and fy are in different directions, conversion to a relative position may indicate (f1, -fx, -fy).

In this way, the operation unit 210 has a first wavelength (λ1), a second wavelength (λ2), an x-th wavelength (λx), and y for four selected wavelengths (λ1, λx, λy, and λz). The wavelength (λy) is set as a reference wavelength, and for each, a feature value such as a peak value or a valley value or a median value thereof is used as a reference position, up to a feature value such as a peak value and a valley value of other selected wavelengths adjacent thereto. The theoretical feature vectors are [e2, ex, ey], [e2 ', ex', ey '], [f1, fx, fy], [E2, Ex, Ey], [E2', Ex ', Ey'], .... In each theoretical feature vector, the absolute thickness values of the substrate abrasive layer f are t1, t2, t3, ..., etc., respectively.

That is, the measurement feature vectors representing the direction and the relative distance between the characteristic values of the measured optical interference signals of the selected wavelengths (λ1, λ2, λx, λy) represent the direction and the relative distance between the characteristic values of the theoretical optical interference signal. If a theoretical feature vector matching the theoretical feature vectors below the tolerance is found, the theoretical feature vector is obtained for the absolute thickness value of each known abrasive layer, and thus the thickness value of the abrasive layer corresponding to the found theoretical feature vector ( For example, the theoretical feature vectors of [e2, ex, ey] and [e2 ', ex', ey '] shown in FIG. 9A correspond to a polishing layer thickness of t1, and [Ee2, Ex] shown in FIG. 9B. , Ey], [E2 ', Ex', Ey '] the theoretical feature vector corresponds to the thickness of the abrasive layer t2, and the theoretical feature vector [f1, fx, fy] shown in FIG. Can be obtained during the polishing process as an absolute thickness value of the polishing layer.

Here, with respect to t1, which is one thickness value of the substrate polishing layer f, two theoretical feature vectors can be obtained in both directions in [e2, ex, ey], [e2 ', ex', ey ']. . Further, if the number of selected wavelengths is further increased, information that allows the absolute thickness values (t1, t2, t3, ...) of the abrasive layer to be known more can be obtained in the form of a theoretical feature vector.

The normalization module 220 normalizes the average value of the reflected light intensity to be constant with respect to the reflected light Lo received from the light receiving unit of the optical sensor 100 while the polishing process is started and progressed. This, the reflected light (Lo) reflected from the substrate polishing layer (f) may vary somewhat depending on the intensity of the light generated from the LED, the light source 105, the intensity of the ambient light, etc., there may be fluctuations due to errors in individual wavelength bands , Since fluctuations due to the light intensity of the light source 105 or the intensity of the ambient light are independent of the thickness of the substrate polishing layer f, the average value of the reflected light intensity (intensity) so that the reflected light intensity is displayed only for the thickness of the substrate polishing layer f Normalize it to be constant.

The normalization module 220 may normalize the reflected light to have the same average value as the immediately preceding average value of the reflected light before the reflected light Lo received from the light receiving unit of the optical sensor 100 is transmitted to the spectrometer 230. For example, the signal processing so that the total intensity (integrated area of the graph) integrated with the optical interference signal according to the wavelength of the initial polishing shown in FIG. Can do

On the other hand, the normalization module 200, the reflected light (Lo) received from the light-receiving unit of the optical sensor 100 is transmitted to the spectrometer 230, the spectrometer 230 for the optical interference signal spectroscopically selected for the selected wavelength You can also normalize it. Here, since the spectrometer 230 calculates the optical interference signal for the selected wavelength, instead of integrating the optical interference signal for the entire wavelength, the sum of the intensity (light intensity, intensity) of the optical interference signal obtained for the selected wavelength The intensity of the total optical interference signal is adjusted in proportion to equal the sum of the intensity of the optical interference signal obtained immediately before. That is, the normalization of the optical interference signal spectrometered by the spectrometer 230 is a constant adjustment of the sum of the calculated optical interference signals, so that the integral of the optical interference signal normalized before spectrometering by the spectrometer 230 is normalized. It is done in the same way as the principle.

As a result, it is possible to prevent the reflected light signal from distorting the reflected light signal that allows the thickness of the substrate polishing layer to be known due to ambient brightness or variation in light intensity of the white light generated by the light source 105.

The normalization module 220 may configure a separate external device, or may be configured as software for signal processing as part of the control unit (or control device). Meanwhile, the normalization module 220 is for more reliably obtaining the thickness of the polishing layer f of the substrate, and may be excluded from the configuration of the present invention as necessary.

The spectrometer 230 spectralizes the reflected light Lo received by the light receiving unit of the optical sensor 100 as a light interference signal for each wavelength.

More specifically, referring to FIG. 5, in which the irradiation light Li and the reflected light Lo are inclined for convenience, the polishing surface of the substrate W has an oxide layer f and light through which light can pass. Since it is made of the impermeable non-transmissive layer (Wo), a part of the light (Li) irradiated from the light emitting unit 120 is reflected (Loe) on the surface (Sx) of the oxide layer (f), and the light emitting unit ( A part of the light Li irradiated from 120 passes through the oxide layer f and is reflected (Loi) in the opaque layer (Wo). Accordingly, the reflected light Lo received from the light receiving unit 130 is reflected light (Loe) reflected from the surface of the oxide layer (f) and reflected light (Loi) reflected through the oxide layer (f) and reflected by the non-transmissive layer (Wo). It includes, and since the light paths differ as the reflected light (Loe, Loi) is proportional to the thickness of the oxide layer (f) at a fine interval (d), the optical interference signal (X) similar to the sine wave form while interfering with each other Is included in the reflected light Lo for each wavelength. That is, the optical interference signal is the intensity of each wavelength of the reflected light (Lo), which is the reflection light (Lo) reflected from the non-transmissive layer (Wo) and the reflected light (Loe) reflected from the surface (Sx) of the polishing layer (f). ).

Then, the spectrometer 230 spectralizes the optical interference signal X for each wavelength included in the reflected light Lo. That is, in the initial thickness to which the thickness t of the polishing layer f of the substrate is sufficiently thick, the optical interference signal according to the wavelength is shown in FIG. 6A, but the thickness t of the polishing layer f of the substrate Becomes thinner, and at the end of polishing, the optical interference signal according to the wavelength becomes the form shown in FIG. 6B.

In addition, the optical interference signals at the respective wavelengths (λ1, λ2, λ3, λ4, λ5, λ6) shown in FIGS. 6A and 6B have a tendency to fluctuate up and down, respectively, according to the fluctuation of the thickness of the polishing layer f, As shown in FIG. 7, one point S1 of the optical interference signal for any one wavelength λ1 forms a sine wave-like waveform according to the variation of the polishing layer thickness, and accordingly, shown in FIG. 8 As described above, the optical interference signal according to the elapse of the polishing time forms a waveform similar to a sine wave. Here, the slope as of the optical interference signal of the sine wave is varied according to the removal rate per unit time.

As such, the spectrometer 230 separates the optical interference signals according to a plurality of wavelengths (λ1, λ2, λ3, λ4, λ5, λ6, ...) from the reflected light Lo received from the light receiving unit of the optical sensor 100. Order. Since the substrate polishing system 1 according to an embodiment of the present invention uses white light as a light source 105 and prepares for a measurement optical interference signal and a theoretical optical interference signal of a plurality of predetermined wavelengths, the spectrometer 230 The wavelength light, which is spectroscopically, may be obtained only for the preselected wavelengths, and is obtained by being spectroscopically determined for the predetermined selected wavelengths (λ1, λx, λy, and λ2 in FIG. 7) from the reflected light Lo received from the light receiving unit. The optical interference signal is extracted as a 'measurement optical interference signal'.

That is, the 'measurement optical interference signal' and similar terms described in the present specification and claims are defined as optical interference signals for selected wavelengths spectrometer or the like spectrometer reflected from the light received from the light receiving unit (Lo).

On the other hand, if the intensity (strength) of the reflected light is normalized by the normalization module 220 before extracting the measured optical interference signal by the spectrometer 230, the amount of light generated from the light source fluctuates unexpectedly during the polishing process or the polishing process This is a result of preliminarily filtering the deviation according to the brightness of the surroundings, so that more accurate measurement optical interference signal data can be obtained. That is, as shown in FIG. 9, the measured optical interference signal has a maximum value of 1, a minimum value of -1, and a value between the peak value and the valley value of 0 is obtained.

The thickness calculation unit 240, the substrate polishing process begins to be performed (S20), the irradiation light (Li) irradiated from the light-emitting portion of the optical sensor reflected light (Lo) reflected from the polishing layer (f) of the substrate Received by the light receiving unit (S30), and the intensity of the reflected light Lo currently received by the normalization module 220, if necessary, the average value of the previous reflected light intensity (for example, the integral value of the optical interference signal of FIG. 6A). After normalizing in the same manner as when the reflected light (Lo) transmitted from the light receiving unit extracts a measurement optical interference signal for at least selected wavelengths by the spectrometer 230 (S40), the measured light received from the spectrometer 230 The absolute thickness of the polishing layer is calculated during the polishing process by comparing the interference signal with the theoretical optical interference signal calculated by the calculator 210 (S50).

More specifically, the thickness calculator 240 detects feature values such as peak values and valley values of each wavelength value from a measurement optical interference signal of selected wavelength values. Here, the measured optical interference signal is displayed with respect to time.

Since the signal received from the spectrometer 230 includes measurement optical interference signals of selected wavelengths at the same time, as shown in FIG. 9, the thickness calculation unit 240 may select a predetermined number of measurement optical interference signals with the selected wavelength. Obtained over the course of the polishing time.

Then, as illustrated in FIG. 10, the thickness calculator 240 includes a peak value and a valley value in real time from the optical interference signals for the selected wavelengths λ1, λ2, λx, and λy that are continuously obtained. Obtain feature values. At the same time, a 'measured feature vector' for the distance and direction from a feature value or an intermediate value to another feature value is obtained.

For example, if the reference wavelength is the first wavelength (λ1) and the intermediate value between the peak value and the valley value is the reference position (Rm1), from this to adjacent peak values of other selected wavelengths (λ2, λx, λy) The measurement feature vector of is obtained by [r2, rx, ry]. Although not shown in the drawing, similarly, as the reference position Rm1, the intermediate value between the peak value and the valley value is used as a reference feature Rm1, and a vector from adjacent valley values of different selected wavelengths can be obtained as a measurement feature vector. In addition, the measurement feature vector may be obtained by using the intermediate value between the peak value and the valley value as a reference position, and similar to the theoretical feature vector, a feature value such as a peak value or a valley value may be used as a reference position, or a peak value or valley of another selected wavelength. It can also be obtained as a vector up to a value.

That is, the measurement feature vector is in the same manner as the theoretical feature vector, with one of the selected wavelengths being the first reference wavelength, and the optical interference signal of the first reference wavelength reaching any one first characteristic value, The first feature value may be determined as a vector to a feature value of the optical interference signal for other selected wavelengths except the first reference wavelength. In addition, the measurement feature vector, in the same manner as the theoretical feature vector, sets the other of the selected wavelengths except the first reference wavelength as the second reference wavelength, and the optical interference signal of the second reference wavelength is any one of the second. In a state in which the feature value has been reached, the second feature value can be determined as a vector up to the feature value of the optical interference signal for selected wavelengths other than the second reference wavelength using the reference location as the reference position.

However, the theoretical feature vector and the measurement feature vector of the present invention, according to a preferred embodiment of the present invention, it is preferable to use all the vectors obtained as above as a feature vector, but according to another embodiment of the present invention, among the vectors obtained as above, Vectors other than some vectors may be used as feature vectors.

In addition, the thickness calculator 240 includes [e2, ex, ey], [e2 ', ex', ey '], [f1, fx, fy], [theoretical feature vectors] calculated by the calculator 210, [ E2, Ex, Ey], [E2 ', Ex', Ey '], .... In addition, the theoretical feature vector and measurement are prepared against the measurement feature vectors [r2, rx, ry] obtained immediately before during the polishing process. Find the theoretical feature vector that the feature vector satisfies within the tolerance. Preferably, a theoretical feature vector in which the deviation between the measured feature vector obtained immediately before the polishing process and the theoretical feature vectors previously calculated by the polishing unit 210 is minimized is found.

At this time, since the theoretical feature vectors know the thickness t of the substrate polishing layer f in advance as t1, t2, t3, ..., by finding the theoretical feature vector with the smallest deviation from the measurement feature vector, The absolute thickness of the substrate polishing layer f can be obtained during the polishing process. The absolute thickness value Tm of the substrate polishing layer f thus obtained is shown in FIG.

Particularly, since the theoretical feature vectors are provided two for each direction toward the valley value and the peak value with respect to one thickness of the substrate polishing layer f already known, the absolute thickness value of the substrate polishing layer in real time during the polishing process It is also possible to determine whether the absolute thickness value of the abrasive layer already obtained from the value of the measurement feature vector obtained subsequently is correct.

In addition, in the configuration illustrated with reference to the drawings, as the optical interference signals for four selected wavelengths are compared with the theoretical value and the measured value, the separation distance of the known abrasive layer thickness data is relatively large according to the variation of the abrasive layer thickness. When, in advance, it is determined to be approximately 10 selected wavelengths, each of the 10 selected wavelengths is set as a reference wavelength, and the thickness of the abrasive layer is determined through a theoretical feature vector based on the characteristic value and the median of the optical interference signal of the 10 selected wavelengths. Since it can be seen, it is possible to obtain an effect that can be seen during the polishing process in real time with a thickness variation of 50 mm to 100 mm.

Moreover, since the present invention requires only calculations for a selected number of selected wavelengths, the computational speed is much higher than in the prior art where the absolute thickness value of the polishing layer was known only when signal processing for hundreds to thousands of wavelengths was required. It is possible to quickly and accurately know the absolute thickness value of the abrasive layer, and also has the advantage of being able to construct equipment such as a computer for calculation at a lower cost and cheaply.

In addition, since the absolute thickness value of the polishing layer can be reliably obtained during the polishing process by the above-described method, the polishing end point is predicted in advance compared to the prior art, so that the polishing end point is not missed, and the deviation according to the position of the polishing layer on the substrate By accurately detecting in real time, it is possible to obtain an effect that can be utilized for control of a polishing head or the like.

On the other hand, when the polishing rate per unit time (RR) is kept constant during the polishing process, since the deviation between the theoretical feature vector and the measurement feature vector occurs only at a constant rate, the absolute thickness of the polishing layer during the polishing process by the above method You can get the value accurately.

However, if the pressing force of the polishing head or the pressing force for each sweep position of the conditioner fluctuates during the polishing process, a deviation between the theoretical feature vector and the measured feature vector may occur according to the variation of the polishing rate RR per unit time.

Accordingly, the thickness calculator 240 calculates the polishing rate RR per unit time from the thickness data of the abrasive layer obtained as described above. Here, the polishing layer thickness period is determined as (λ / 2n), and the optical interference signal of the same pattern is repeated for each λ / 2n, and the time period T1 required for one cycle of the polishing layer thickness is (λ / (2n * RR). )), The polishing rate per unit time (RR) can be obtained from the equation of the measured time period (T1) = λ / (2n * RR).

Then, the thickness calculator 240 is shown in FIG. 11 from the polishing rate per unit time (RR) calculated over time in the polishing process and the absolute thickness value (Tm) of the polishing layer obtained over time. As shown, curve fitting is performed to generate a thickness fitting curve (Tr) of the absolute thickness of the abrasive layer.

Here, the slope in the variation curve of the absolute thickness of the polishing layer represents the polishing rate RR per unit time, from which the time remaining until the target thickness Te can be calculated. Then, while the absolute thickness value Tm of the substrate polishing layer obtained over time by the thickness calculator 240 is displayed in FIG. 11, the shift curve Tr of the absolute thickness value of the polishing layer is the newly displayed absolute thickness value ( Tm) is reflected and is continuously corrected as the polishing process progresses.

At this time, among the absolute thickness values obtained as described above, the absolute thickness value data Ei spaced apart from the variation curve Tr of the absolute thickness value is regarded as an incorrect measured thickness value. Then, the time remaining from the absolute thickness value variation curve Tr of the polishing layer to the target thickness of the polishing layer is calculated and displayed.

When the absolute thickness value of the polishing layer is obtained at a value close to the target thickness of the polishing layer, and the time remaining to reach the target thickness is less than the tolerance, the polishing process is terminated to reach the target thickness Te. Here, the target thickness Te may not be a feature value or an intermediate value. In this case, since the theoretical feature vector does not exist in the target thickness Te, the thickness of the polishing layer closest to the target thickness Te is reached. If it is detected, the polishing process is terminated after the polishing process is continued for only the calculated remaining time.

When the polishing process is performed in this way, even if the polishing rate (RR) per unit time of the polishing layer varies during the polishing process, the thickness value of the substrate polishing layer f can be accurately known in real time. In addition to accurately sensing, even when a variation in the amount of polishing of the substrate polishing layer occurs during the polishing process, the polishing profile of the polishing layer is intended by adjusting the pressing force for pressing the pressing force of the polishing head for each substrate area or by adjusting the pressing force of the conditioner In one form, it is possible to obtain an advantageous effect that can be more accurately controlled.

The present invention has been exemplarily described through preferred embodiments, but the present invention is not limited to such specific embodiments, and various forms within the scope of the technical idea, specifically, the claims described in the present invention. Can be modified, changed, or improved.

W: substrate f: oxide layer
Li: irradiation light Lo: reflected light
d: spacing of reflected light t: oxide layer thickness
X: Optical interference signal 20: Polishing head
30: conditioner 40: conditioner
1: polishing system 100, 100 ': light sensor
105: light source 200: control unit
210: operation unit 220: normalization module
230: spectrometer 240: thickness calculator

Claims (40)

A polishing system for a substrate having a light-transmitting material abrasive layer formed on its bottom surface,
A polishing head for positioning the substrate in a state where the polishing layer of the substrate contacts the polishing pad;
A light emitting unit that irradiates the polishing layer with irradiation light having a plurality of wavelengths;
A light receiving unit receiving the reflected light reflected from the polishing layer;
Calculate the theoretical optical interference signal for the thickness of the polishing layer according to the polishing layer material for a plurality of predetermined wavelengths, and extract the measurement optical interference signal for the selected wavelengths from the reflected light received from the light receiving unit Thus, measurement feature vectors representing a direction and a relative distance between feature values of the measured optical interference signal of the selected wavelengths are allowed with theoretical feature vectors representing a direction and a relative distance between feature values of the theoretical optical interference signal. A control unit for finding a theoretical feature vector having an error or less and obtaining a thickness value of the polishing layer already obtained and known for the found theoretical feature vector as an absolute thickness value of the polishing layer during the polishing process;
Characterized in that it comprises a polishing system for a substrate.
According to claim 1,
And the controller calculates the theoretical optical interference signal data in advance before performing the polishing process of the substrate.
According to claim 1,
And the controller simultaneously calculates the theoretical optical interference signal data during the polishing process of the substrate.
According to claim 1,
The selective wavelength is set to 5 to 15, characterized in that the polishing system of the substrate.
The method of claim 4,
The selective wavelength is a substrate polishing system, characterized in that selected from the wavelength range of 4000 kHz to 7000 kHz.
The method of claim 4,
The selective wavelengths are the polishing system of the substrate, characterized in that the wavelength spacing between each other is determined at uniform intervals within an error range of 20%.
According to claim 1,
The irradiation light is a white light polishing system of a substrate, characterized in that.
The method of claim 7,
The light source of the irradiation light is a substrate polishing system, characterized in that the LED.
The method of claim 8,
The light emitting portion and the light receiving portion is a polishing system for a substrate, characterized in that irradiating the irradiation light perpendicular to the polishing layer and receiving the reflected light vertically.
The method of claim 9,
The light emitting portion and the light receiving portion are formed as one body;
A polishing system for a substrate, characterized in that a main optical fiber forming an irradiation light path extending from the light source to the light emitting unit is formed with an optical fiber for transmitting reflected light branched and extended toward the control unit.
The method of claim 9,
The polishing pad is provided with a transparent window, the light emitting portion and the light receiving portion is a polishing system for a substrate, characterized in that disposed under the transparent window.
The method of claim 9,
A polishing system for a substrate, characterized in that a recessed portion which penetrates at least a portion of the polishing pad and does not contact the substrate is provided, and the light emitting portion and the light receiving portion are disposed in the recessed portion.
According to claim 1,
The polishing layer is a polishing system for a substrate, characterized in that the oxide layer.
According to claim 1,
The control unit includes a spectrometer for extracting an optical interference signal for the selected wavelength from the reflected light.
The method of claim 14,
The control unit includes a normalization module for normalizing the reflected light received from the light receiving unit so that the average value of the reflected light intensity is constant.
The method of claim 15,
The normalization of the reflected light by the normalization module is performed before the reflected light is transmitted to the spectrometer.
The method of claim 15,
The normalization of the reflected light by the normalization module is a polishing system for a substrate, characterized in that the reflected light is performed with respect to the optical interference signal of the selected wavelength that is spectrometered by the spectrometer.
delete According to claim 1,
The characteristic value is at least one of a peak value and a valley value.
According to claim 1,
Obtaining the absolute thickness value of the polishing layer with respect to the measurement feature vector is a polishing system for a substrate, characterized by finding a theoretical feature vector in which the deviation between the measurement feature vector and the theoretical feature vectors is minimal.
According to claim 1,
The theoretical feature vector,
Set one of the selected wavelengths of the theoretical optical interference signal as a first reference wavelength, and reference the first characteristic value in a state where the optical interference signal of the first reference wavelength reaches any one first characteristic value A position is determined as a vector up to a feature value of the optical interference signal for selected wavelengths other than the first reference wavelength;
The measurement feature vector,
With respect to the reflected light received by the light-receiving unit, in the state in which the optical interference signal of the first reference wavelength reaches any one of the first feature values, the first reference wavelength is set by using the first feature value as a reference position. A polishing system for a substrate, characterized in that it is determined by a vector up to a characteristic value of the optical interference signal for the other selected wavelengths.
The method of claim 21,
The theoretical feature vector,
Among the selected wavelengths of the theoretical optical interference signal, the other one except the first reference wavelength is used as a second reference wavelength, and the optical interference signal of the second reference wavelength reaches any second characteristic value. , Set as a vector up to a feature value of an optical interference signal for selected wavelengths other than the second reference wavelength using the second feature value as a reference position;
The measurement feature vector,
With respect to the reflected light received by the light-receiving unit, in the state in which the optical interference signal of the second reference wavelength reaches any one of the second characteristic values, the second reference wavelength is set by using the second characteristic value as a reference position. A substrate polishing system characterized in that it is determined by a vector up to a characteristic value of the optical interference signal for other selected wavelengths.
According to claim 1,
The control unit includes measurement feature vectors representing a direction and a relative distance from an intermediate value to a feature value of a peak value and a valley value of the measured optical interference signal of the selected wavelengths of a peak value and a valley value of the theoretical optical interference signal. The theoretical feature vectors representing the direction and relative distance from the median to the feature values and the theoretical feature vectors below the tolerance are found, and the thickness value of the polishing layer corresponding to the found theoretical feature vector is the absolute thickness of the thickness of the polishing layer. A polishing system for a substrate, characterized in that it is obtained during the polishing process by value.
The method of claim 23,
The characteristic value is at least one of a peak value and a valley value.
The method of claim 23,
Obtaining the absolute thickness value of the polishing layer with respect to the measurement feature vector is a polishing system for a substrate, characterized by finding a theoretical feature vector in which the deviation between the measurement vector and the theoretical feature vectors is minimal.
The method of claim 23,
The theoretical feature vector,
In the state where one of the selected wavelengths of the theoretical optical interference signal is a first reference wavelength, and the optical interference signals of the first reference wavelength reach a first intermediate value of adjacent valley values and peak values, A median of up to a characteristic value of an optical interference signal for selected wavelengths other than the first reference wavelength is determined as a reference position as a median value;
The measurement feature vector,
With respect to the reflected light received by the light-receiving unit, the optical interference signal of the first reference wavelength reaches the first intermediate value of adjacent valley values and peak values, and the first intermediate value is used as the reference position. A vector up to a feature value of the optical interference signal for other selected wavelengths except one reference wavelength;
Characterized in that the polishing system of the substrate.
The method of claim 26,
The theoretical feature vector,
Of the selected wavelengths of the theoretical optical interference signal, the other one except the first reference wavelength is used as a second reference wavelength, and the second intermediate value of the valley value and the peak value where the optical interference signals of the second reference wavelength are adjacent to each other. In the state of reaching, the second intermediate value is set as a vector to a characteristic value of the optical interference signal for selected wavelengths other than the second reference wavelength, using the reference position as a reference position;
The measurement feature vector,
With respect to the reflected light received from the light-receiving unit, the optical interference signal of the second reference wavelength reaches the second intermediate value of adjacent valley values and peak values, and the second intermediate value is used as a reference position. A polishing system for a substrate, characterized in that it is determined by a vector up to a characteristic value of an optical interference signal for selected wavelengths other than the second reference wavelength.
The method according to any one of claims 1 to 17 or 19 to 27,
The control unit, the polishing system of the substrate, characterized in that for obtaining the polishing rate per unit time (RR) from the period of the optical interference signal of the reflected light during the polishing process.
The method of claim 28,
The control unit generates a variation curve Tr of the absolute thickness value of the abrasive layer by a curve fitting method over time based on the polishing rate per unit time (RR) and the thickness value of the abrasive layer. Substrate polishing system.
The method of claim 29,
A polishing system for a substrate, characterized in that a value obtained by the measured thickness value is regarded as an erroneous measured thickness value if it is outside a predetermined range and a variation curve (Tr) of the absolute thickness value of the polishing layer.
The method of claim 29,
The control unit calculates and displays the remaining time in consideration of a variation curve (Tr) of the absolute thickness of the polishing layer and a polishing rate per unit time to a target thickness.
The method of claim 29,
The control unit terminates the polishing process when the remaining time elapses even if the absolute absolute thickness value is not obtained for the remaining time in consideration of the variation curve Tr of the absolute thickness value of the polishing layer and the polishing rate per unit time to the target thickness. A polishing system for a substrate, characterized in that.
A control device used in a polishing process of a substrate on which a polishing layer of light transmissive material is formed,
A computation unit for pre-calculating a theoretical optical interference signal for the thickness of the polishing layer according to the polishing layer material for a plurality of predetermined wavelengths of reflected light reflected from the polishing layer and received by the light receiving unit;
A spectrometer extracting a measurement optical interference signal for the selected wavelengths from the reflected light reflected and received by the polishing layer;
The measurement feature vectors representing a direction and a relative distance between the characteristic values of the measured optical interference signal of the selected wavelengths are less than a tolerance with the theoretical feature vectors representing a direction and a relative distance between the characteristic values of the theoretical optical interference signal. A thickness calculator for finding a theoretical feature vector and obtaining a thickness value of the polishing layer already obtained and known for the found theoretical feature vector as an absolute thickness value of the polishing layer during the polishing process;
Control device for a substrate polishing system comprising a.
The method of claim 33,
A normalization module that normalizes the reflected light received from the light receiving unit so that an average value of reflected light intensity is constant;
A control device for a polishing system for a substrate, further comprising.
The method of claim 34,
The normalization of the reflected light by the normalization module is performed before the reflected light is transmitted to the spectrometer.
The method of claim 34,
The normalization module is a control device for a polishing system of a substrate, characterized in that the reflected light is performed on the optical interference signal of the selected wavelength, which is spectroscopically measured by the spectrometer.
The method of claim 33,
The thickness calculator may calculate a thickness value in which feature vectors between measured feature values of the measured optical interference signal of the selected wavelengths and feature vectors between theoretical feature values of the theoretical optical interference signal are below an allowable error. Finding and obtaining the thickness of the polishing layer as a measured thickness value during the polishing process Control device for a substrate polishing system.
The method of claim 33,
A polishing rate (RR) per unit time is obtained during a polishing process from the period of the optical interference signal of the reflected light, and curve fitting over time is based on the polishing rate per unit time (RR) and the thickness value of the polishing layer. A control unit generating a shift curve Tr of the absolute thickness value of the polishing layer in a manner;
It characterized in that it comprises a control device for a substrate polishing system.
The method of claim 33,
A control apparatus for a polishing system for a substrate, characterized in that the remaining time is calculated in consideration of a variation curve (Tr) of the absolute thickness of the polishing layer and a polishing rate per unit time to a target thickness.
The method of claim 38,
Considering the variation curve (Tr) of the absolute thickness of the polishing layer and the polishing rate per unit time to the target thickness, the polishing process is terminated when the remaining time elapses even if the absolute thickness is not obtained for the remaining time. A control device for a substrate polishing system.

Images