JP2001284424A - Modeling method of pattern, film thickness measuring method, process state judging method, thickness measuring equipment, process state judging equipment, grinding equipment and manufacturing method of semiconductor device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for facilitating the calculation of a theoretical value in the case that the film thickness of an object is measured on the basis of a degree of the accordance of the theoretical value and the actual value of an optical characteristic. SOLUTION: An actual pattern shown in (a) is a multi-layer structure and a complicated shape. Since a TiN film 23 hardly transmits light, the structure of the rear side can be neglected, they have infinite thickness, and can be calculated even if it is supposed that these materials are entirely filled on the rear side. When such a view is taken, the pattern shown in (a) can be modeled into the pattern shown in (b). Thus, if calculation is performed after such modeling, a calculation time can be remarkably shortened.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a process for manufacturing a semiconductor, for example, when a wafer having a pattern such as a semiconductor structure or a wiring structure formed therein is polished by a polishing apparatus. Technology for optically measuring the progress of the technology and its applied technology,
More specifically, a technique for optically measuring and determining the film thickness and the progress of the polishing process by measuring the degree of coincidence between a previously calculated optical property and an actually measured optical property (for example, a spectrophotometric property) and its technology. It is related to applied technology.

[0002]

2. Description of the Related Art With the spread of the CMP process and the like, the necessity of accurately measuring the thickness of a thin film in a portion of a device wafer having a fine element or electrode structure as a base has been increasing. However, at present, it is difficult to make the light spot diameter smaller than a certain limit, and it is difficult to align the light spot with a part having a predetermined structure. Has been virtually impossible to measure. Therefore, the film thickness is measured at the blank portion other than the fine structure, and the film thickness of the portion having the fine structure is indirectly obtained by conversion, or the fine structure portion of the sample extracted by the sampling inspection is destroyed and the cross section is determined. A method of observing the film thickness by observation with an electron microscope has been generally performed.

On the other hand, the present inventor theoretically calculates the reflected light and transmitted light from the fine structure portion in the light spot as the sum of the reflected light and transmitted light from each fine portion structure portion. , The spectral characteristics of reflected light and transmitted light are determined for each of various film thicknesses, and a light spot is applied to the wafer during polishing, and the actual measured values of the spectral characteristics of reflected light and transmitted light from the wafer are obtained. A method of performing a fitting calculation with the above-mentioned spectral characteristics obtained for each film thickness and determining a film thickness that gives the best match as a measured value was invented.
No. 901 is disclosed.

According to the present invention, the film thickness of a portion having a fine structure can be accurately measured using a light spot larger than the fine structure, and the polishing end point can be determined. The method enables non-destructive measurement in the in-situ state.

[0005]

SUMMARY OF THE INVENTION Japanese Patent Application Laid-Open No. H11-3390
In the invention disclosed in Japanese Patent Application Publication No. 1 (1993), basically, a film thickness is calculated by performing a comparison operation (fitting) between an actually measured signal and a value obtained by theoretical calculation. This theoretical calculation becomes complicated and difficult as the structure of a device to be measured or the like becomes more complicated. In particular, recently, devices having a multilayer structure have been increasing,
That is what makes the theoretical calculations complicated and enormous.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and its main problem is to control the film thickness and polishing process of a substrate (object) such as a wafer having a pattern having a complicated structure therein, by using optical characteristics. It is an object of the present invention to provide a method for facilitating the calculation of the theoretical value when measuring and determining based on the degree of coincidence between the theoretical value and the actual value, and to provide a method and an apparatus applying the same. And

[0007]

The first means for solving the above problems is to measure the thickness or process state of a thin film on a substrate having a pattern formed therein by measuring the optical characteristics of the substrate, The optical characteristics of the substrate in a predetermined number of film thicknesses or process states are determined in advance by calculation, and the measurement and determination are performed based on the degree of coincidence between the measured optical characteristics of the substrate and the optical characteristics determined by the calculation. Prior to calculating the optical characteristics of the substrate in a predetermined number of film thicknesses or process states, it is necessary to model the actual pattern structure by simplifying it in consideration of the optical characteristics of the pattern. This is a method of modeling a characteristic pattern (claim 1).

[0008] In the present specification, the "process state" is a state indicating the degree of progress of the process. For example, a state such as whether a polishing end point has been reached or a change in polishing conditions has been reached. That is. As described above, even when the process state is detected, the optical characteristics of the reflected light and the transmitted light in the predetermined process state are obtained in advance by calculation,
The progress of the process is managed by measuring the degree of coincidence between the actually measured reflected light and transmitted light with the optical characteristics.

A thin film on a substrate having a pattern formed therein is, for example, a wafer substrate before polishing in which a pattern is formed near a surface thereof and a film such as an insulating film is formed thereon. Measuring the film thickness, including the thin film in what is covered, means measuring the individual film thickness of the thin film wrapped around these patterns or formed on the pattern, or the entire film thickness Means to do.

The term "optical characteristics" means those that determine the optical properties of an object, such as optical constants such as the absorption coefficient and the refractive index of each part of the pattern. On the other hand, the “optical characteristics” refer to quantities indicating the state of light, such as reflectance, transmittance, spectral reflectance, and spectral transmittance.

In this means, the method disclosed in Japanese Patent Application Laid-Open No. H11-33
Unlike the technique disclosed in Japanese Patent Application Laid-Open No. 901-901, the characteristics of the reflected light and transmitted light from each point of the fine structure are not directly calculated, but the actual pattern The model is modeled by simplification in consideration of characteristics. Therefore, the theoretical calculation for obtaining the optical characteristics becomes easy.

A second means for solving the above-mentioned problem is as follows.
In the first means, modeling is performed by simplifying in consideration of the optical characteristics of the pattern for only light that is perpendicularly incident on the surface of the substrate and that exits perpendicularly from the surface of the substrate. This is a feature (claim 2).

The first means measures the reflected light or transmitted light by making diffused light incident, or makes the reflected light or transmitted light obliquely incident on the surface of a measurement substrate (a wafer or the like) or the like. And methods for measuring the However, in actual optical measurement, light is often perpendicularly incident on the surface of the measurement substrate. In addition, reflected light and transmitted light are emitted in various directions.
If the optical system is devised so as to receive only the light emitted perpendicularly to the surface of the substrate, the measurement error is reduced and the theoretical calculation of the optical characteristics of the reflected light and the transmitted light is facilitated. Therefore, according to this means, the calculation amount of the theoretical calculation can be reduced.

A third means for solving the above-mentioned problem is as follows.
In the first means or the second means, it is assumed that a uniform object is filled on the back side of a surface on which a metal layer or an absorption layer which reaches first after light is incident on the substrate is present. By doing so, modeling is performed (claim 3).

When modeling the microstructure by simplifying it, first of all, it is conceivable that most of the miniaturized structure reflects and reflects light almost like a metal layer such as a TiN film, an Al wiring, and a W electrode. There are structures made of substances that absorb, and these dominate most of the reflection and transmission characteristics. Therefore, in this means,
Even if there is a pattern with a complicated structure on the back side of these metal layers or absorption layers that light arrives first after entering the substrate, their contribution to reflection and transmission characteristics is extremely small Focusing on this, a simplified model of the microstructure is created assuming that the back side of these metal layers or absorption layers is filled with a uniform object.

Thus, the amount of calculation for theoretically calculating the optical characteristics is reduced. What kind of optical property material is filled on the back side of the metal layer or the absorption layer depends on the optical property and thickness of the metal layer or the absorption layer, actually on the back side of the metal layer or the absorption layer Model in consideration of the pattern structure.

In the present means (claim 3), the term "back side" refers to the back side as viewed from the direction in which light is incident, that is, a portion where a shadow of light is generated, and is not necessarily a metal layer or an absorbing layer. It does not imply a direction perpendicular to the layers. For example, when light is incident obliquely, the oblique depth direction of the metal layer or the absorption layer is the back side.

A fourth means for solving the above problem is as follows.
The third means is characterized in that the uniform object is modeled on the assumption that it is the same as the metal layer or the absorption layer (claim 4).

When the thickness of the metal layer or the absorption layer is increased to some extent, light does not proceed any further, so that the optical characteristics of reflected light and transmitted light are affected no matter what. Absent. Therefore, in this case, in such a case, modeling is performed on the assumption that the metal layer or the absorption layer is clogged on the back side of the metal layer or the absorption layer. Therefore, the theoretical calculation for obtaining the optical characteristics of the reflected light and the transmitted light is further simplified.

In the present means (claim 4), the term "thickness" refers to the thickness as viewed from the light incident direction, and does not necessarily mean the thickness in the direction perpendicular to the surface of the metal layer or the absorbing layer. That is, when the light is obliquely incident, the thickness is measured along the optical path, and is larger than the thickness in the direction perpendicular to the surface of the metal layer or the absorption layer.

A fifth means for solving the above problems is as follows.
In any one of the first to fourth means, the reflectance is 0 at a depth equal to or more than a certain depth from the surface of the substrate.
Alternatively, modeling is performed assuming that an object having a certain optical characteristic is blocked (claim 5).

When light penetrates to a certain depth of the substrate having the pattern structure, the reflected light from that depth generally becomes weaker and has only a small effect on the whole. In the present means, in consideration of these, modeling is performed on the assumption that an object having a reflectance of 0 or constant optical characteristics is packed at a depth equal to or more than a certain depth from the surface of the substrate. ing. Therefore, it is not necessary to consider the structure of the pattern at a portion deeper than the predetermined depth, and the theoretical calculation for obtaining the optical characteristics of the reflected light and the transmitted light is further simplified.

A sixth means for solving the above-mentioned problem is as follows.
In any one of the first means to the fifth means, it is assumed that a pattern having a fine structure smaller than a wavelength range of light is formed of a uniform continuous object having predetermined optical characteristics. Modeling is performed (claim 6).

In general, light cannot cope with a scalar theory generally used for a structure small in comparison with its wavelength. Since the above-described modeling and the calculation of the reflected light intensity based on the modeling are based on the scalar theory, the calculation may be inappropriate for a very small uneven structure.
This phenomenon has already occurred in ultra-fine devices and multilayer devices (here, a fine structure is formed as a result of the degree of overlap of wirings).

For example, when monitoring using light having a wavelength of about 600 nm, the characteristics of a very small uneven structure are theoretically the same as those of a large uneven structure (as long as it is smaller than the optical spatial coherence length). However, as an actual phenomenon, the characteristics of a small uneven structure such as 0.13 μm require numerical calculation based on so-called vector theory. This is an enormous amount of computation to be actually performed, which is not practical.

Therefore, in this means, optically,
In such a fine structure, a material having a concavo-convex structure can be regarded as a homogeneous material (effective mass approximation). That is, when there is a fine structure, optical calculation is performed using a model in which the fine structure is replaced with a continuous and homogeneous material.

The simplest modeling is an approximation that, when there are fine irregularities compared to the wavelength, light does not enter there, that is, an approximation that ignores the structure. More realistically, it is considered that the material has been replaced with a material having different optical constants (refractive index n and absorption coefficient k) without irregularities. However, since this optical constant theoretically changes depending on the density of the microstructure, It is desirable to select an appropriate value depending on the pattern.

According to this means, since the theoretical value can be obtained by scalar calculation without using vector calculation, the calculation amount can be significantly reduced.

[0029] A seventh means for solving the above problem is as follows.
After the pattern is modeled by using any one of the first to sixth means, the optical characteristics when light is incident on the modeled pattern are considered in consideration of the modeled pattern. Measuring the film thickness of the thin film on the substrate based on the degree of coincidence between the measured optical characteristics of the substrate and the calculated optical characteristics by theoretical calculation for each predetermined type of film thickness. This is a characteristic method for measuring the film thickness (claim 7).

In this means, the above-mentioned Japanese Patent Application Laid-Open No. 11-33
As disclosed in Japanese Unexamined Patent Publication No. 901, an actual measurement value of an optical characteristic (for example, a spectral luminous intensity distribution) of reflected light or transmitted light when a measurement sample is irradiated with light is calculated and calculated in advance for each predetermined film thickness. The degree of coincidence of the optical characteristics of the reflected light or transmitted light obtained in step (1) is determined, and the film thickness is measured based on the degree of coincidence. For example, a film thickness that minimizes the sum of squares of the difference between the wavelengths of the two spectral photometric characteristics,
For example, a film thickness that maximizes the cross-correlation value between the two spectrophotometric characteristics is adopted.

In this means, in such a film thickness measuring method, when calculating optical characteristics in advance for each predetermined film thickness, a pattern which is a prerequisite for the calculation can be changed by any one of the first means to the sixth means. (It goes without saying that the use of either includes the case where some of these are used in combination. This is the same in the present specification.) The amount of calculation can be reduced accordingly, and the required memory can also be reduced.

Eighth means for solving the above-mentioned problem is:
After modeling the pattern using any one of the first to sixth means, the optical characteristics when light is incident on the modeled pattern are considered in consideration of the modeled pattern. A process of forming an insulating film or a metal electrode film on a substrate based on the degree of coincidence between the measured optical characteristics of the substrate and the calculated optical characteristics obtained by theoretical calculation for a predetermined process state. Or a process state determination method characterized by determining a process state in a step of removing these films.

In this means, when detecting the process state, when calculating the optical characteristics in advance for each predetermined process state, a pattern which is a prerequisite for the calculation can be obtained by using any of the first to sixth means. Simplification, the amount of calculation can be reduced accordingly, and the required memory can also be reduced.

A ninth means for solving the above-mentioned problem is:
An apparatus for measuring the thickness of a thin film on a substrate on which a pattern is formed, an apparatus for irradiating light to the substrate surface, an apparatus for detecting reflected light or transmitted light from the substrate, A storage unit for storing a pattern modeled by any one of the first to sixth means, and an optical system when light enters the substrate in consideration of the modeled pattern stored in the storage unit A calculation unit for obtaining the characteristic for a predetermined type of film thickness, and a film thickness determination unit for calculating a degree of coincidence between the optical characteristic calculated by the calculation unit and the actually measured optical characteristic, and then determining the film thickness. A film thickness measuring apparatus (claim 9).

When this means is used, a pattern contained in the sample to be measured is modeled by any of the first to sixth means, and the modeled pattern is stored in the storage unit. Let me do it. This means
For example, immediately before the start of measurement, the calculation unit calculates an optical characteristic when light is incident on the substrate for a predetermined type of film thickness in consideration of the modeled pattern stored in the storage unit. Keep it.

Thereafter, the surface of the substrate is irradiated with light, and reflected light or transmitted light from the substrate is detected. Then, the film thickness determination unit calculates the degree of coincidence between the optical characteristics calculated by the calculation unit and the actually measured optical characteristics using, for example, the least square method or the cross-correlation method, and determines the film thickness from the calculated values. Storing a pattern corresponds to storing numerical values that can be used to reproduce the pattern.

According to this means, the modeled pattern used for the calculation of the optical characteristics is stored, and the optical characteristics are calculated each time the measurement is performed. Requires less capacity.

A tenth means for solving the above problem is an apparatus for measuring the thickness of a thin film on a substrate having a pattern formed therein, wherein the apparatus irradiates light to the substrate surface, An apparatus for detecting reflected light or transmitted light from a substrate, and an optical element when light is incident on the substrate, calculated in consideration of a pattern modeled by any of the first to sixth means. A storage unit that stores the characteristics for a predetermined type of film thickness; and a film thickness determination unit that calculates the degree of coincidence between the optical characteristics stored in the storage unit and the actually measured optical characteristics and determines the film thickness based on the calculated degree. A film thickness measuring apparatus (claim 10), characterized in that:

According to this means, unlike the ninth means, the calculation of the optical characteristics for each film thickness is externally performed based on the pattern modeled by any of the first to sixth means. And the obtained optical characteristics are stored in an internal storage unit. Therefore, since the optical characteristics for each film thickness must be stored, a memory is required. However, unlike the ninth means, it is not necessary to calculate the optical characteristics every time the measurement is performed. I can do it.

An eleventh means for solving the above-mentioned problems is to form an insulating film or a metal electrode film on a substrate,
Or an apparatus for determining a process state in a step of removing these films, wherein the apparatus irradiates light to the substrate surface, and an apparatus detects reflected light or transmitted light from the substrate. 6. A storage unit for storing a pattern modeled by the method for modeling a pattern according to any one of 6. and a light source for the substrate calculated in consideration of the modeled pattern stored in the storage unit. A calculating unit for obtaining an optical characteristic when the light is incident on a predetermined process state; calculating a degree of coincidence between the optical characteristic calculated by the calculating unit and the actually measured optical characteristic; and determining the process state based on the calculated degree. And a process state determination device (claim 11).

The present means is different from the ninth means only in that the object to be judged is the process state, and the operation and effect are almost the same, so that the description is omitted.

A twelfth means for solving the above problems is a film forming step of forming an insulating film or a metal electrode film on a substrate,
Or an apparatus for determining a process state in a step of removing these films, wherein the apparatus irradiates light to the substrate surface, and an apparatus detects reflected light or transmitted light from the substrate. A storage unit that stores an optical characteristic when light is incident on the substrate with respect to a predetermined process state, in consideration of a pattern modeled by the pattern modeling method according to any one of 6. A process state determining unit that calculates a degree of coincidence between the optical characteristics stored in the storage unit and the actually measured optical characteristics, and determines a process state based on the calculated degree. 1
2).

The present means differs from the tenth means only in that the object to be determined is the process state, and the operation and effect are almost the same, so that the description is omitted.

A thirteenth means for solving the above-mentioned problem is that the method for measuring a film thickness according to the seventh or the method for judging a process state which is the eighth means is used in a polishing step. A polishing apparatus (claim 13).

In this means, since the method for measuring the film thickness according to the seventh aspect or the process state judging method as the eighth means is used in the polishing step, the polishing end point is determined during the polishing. In addition to making a determination, the amount of calculation at that time can be reduced, and the required memory can be reduced.

A fourteenth means for solving the above problems is a film thickness measuring device as the ninth means or the tenth means, or an eleventh means or a twelfth means.
A polishing apparatus comprising at least one of the process state determination devices as the means (14).

In this means, since the apparatus has a process state determining device, it is possible to determine the polishing end point during polishing, etc., to reduce the amount of calculation at that time, and to reduce the required memory. be able to. In many cases, the film thickness device and the process state determination device are the same, and in this case, only one device is required.

A fifteenth means for solving the above-mentioned problem comprises a step of polishing a wafer by using the thirteenth means or the fourteenth means. A method (claim 15).

According to this means, the thickness and the end point of polishing can be accurately detected during polishing, so that semiconductor devices can be manufactured with good yield.

[0050]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described with reference to the drawings. Prior to this, a film thickness measuring apparatus according to the present embodiment will be described in advance.

FIG. 1 is a view showing an example of such a film thickness measuring apparatus, and is also an embodiment of the present invention. In FIG. 1, a wafer 1 to be polished is held on a wafer carrier 2. A polishing pad 4 is provided on the surface of the polishing platen 3, and the polishing platen 3 is rotated around its central axis. The wafer carrier 2 rotates while reciprocating while pressing the wafer 1 on the polishing pad 4, and polishes the wafer 1 with the polishing pad 4. The polishing platen 3 and the polishing pad 4 have a light transmitting window 5, and a quartz plate is fitted into the light transmitting window. Light emitted from the irradiation light source 6 is projected on the surface of the wafer 1 through the light transmitting window 5. The reflected light from the wafer 1 is converted into a signal by a light receiving system / light receiving device 6 ′,
Data processing is performed by the signal processing device 7 to detect the polishing amount or the polishing end point.

FIG. 2 is a diagram showing details of an example of the optical system used in the embodiment shown in FIG. In FIG. 2, light from a xenon lamp 8, which is an irradiation light source, is converted into a parallel light beam by a lens 9, passes through a slit 10, and is condensed on a beam splitter 12 by a lens 11. The light that has passed through the beam splitter 12 is
The light beam is again converted into a parallel light beam by 3 and projected on the surface of the wafer 1 through the light transmitting window 5 in FIG.

The reflected light is transmitted to the light transmitting window 5 and the lens 13 again.
Through the beam splitter 12. In the beam splitter 12, the reflected light has its direction changed by 90 ° and is converted into a parallel light by the lens 14. Then, the light is reflected by the reflecting mirror 15 and condensed on the pinhole 17 by the lens 16. Then, noise components such as scattered light and diffracted light are removed, and are projected on the diffraction grating 19 via the lens 18,
It is split. The split light enters the linear sensor 20, and the spectral intensity is measured.

The signal processing device 7 receives the signal of the spectral intensity, and detects the thickness of the layer to be polished (uppermost layer) and the polishing end point based on a predetermined algorithm. Further, the polishing amount is determined from the initial film thickness of the layer on the wafer and the remaining film thickness of the layer being polished (uppermost layer).

According to this method, when the interlayer insulating film is polished, the metal electrode film is polished, and the STI (Sh
It is possible to measure the film thickness (remaining film thickness) or the polishing end point in any of the polishing and the polishing of the metal electrode film including the barrier layer.

When the irradiation light is applied to a portion of the wafer having no pattern, the spectral reflectance or the spectral transmittance becomes a simple waveform that can be easily predicted, and is realized by the existing film thickness detecting device. Thus, the calculation of the film thickness is also easy. However, in general, areas without patterns are very small in area, and
Since the position is not constant depending on the wafer, it is difficult to perform high-speed measurement with a simple mechanism using this method.

A feature of the apparatus shown in FIGS. 1 and 2 is that measurement can be performed not only by irradiating such a blank surface but also by irradiating an electrode pattern or an uneven portion. The reflected light from the non-blank structural part can be seen as the superposition of light waves from each layer and each part of the device (laminated thin film), and the wavelength-dependent (spectral characteristic) waveform is usually a complex interference effect, Many local maxima (peaks)
Will have.

Therefore, if the structure (two-dimensional structure and film thickness) of the device to be measured is known in advance, the waveform of the device is analyzed to determine the film thickness of the polished layer (top layer). It is possible. That is, the optical characteristics of the reflected light as a whole can be calculated by grasping the optical characteristics of the light that is incident on and reflected by the minute portions of the device structure, and adding them to the irradiated portion.

However, as described above, if the structure of the pattern becomes complicated, this calculation becomes complicated. Therefore, in the present invention, the calculation is performed after the pattern structure is modeled to be simplified. .

FIG. 3 shows an example of such a pattern structure. In the figure, reference numerals 21 and 22 denote SiO ₂ insulating layers;
Is a layer made of TiN, 24 is an Al electrode and wiring, and 25 is Si doped with impurities. In the following figure,
Although the same hatched portions indicate that they are made of the same material, the same hatching does not necessarily indicate the same material between the drawings. In the drawings showing such a pattern, the probe light is incident vertically from above, and the film thickness is measured by measuring the reflected light.

As shown in FIG. 3, the actual pattern has a multilayer structure and a complicated shape. However, since the TiN layer 23 does not transmit most of the measurement light, the structure under the TiN layer 23 can be ignored. Therefore, on the back side of the TiN layer 23, TiN
Simple modeling can be performed assuming the layers are packed. Considering this, FIG. 4 (a)
(Same as shown in FIG. 3)
It can be modeled into a pattern as shown in FIG.

FIG. 5 shows a complicated electronic device (transistor).
The modeling in the case of a structure is shown. In FIG. 5, 31 is Si
O ₂ insulating layer, 32 is a Si substrate, 33 is W, 34 is TiN
Layers, 35 are Al electrodes, 36 is an insulating layer of SiO ₂ , 37, 38
Is Si doped with impurities. About this structure,
When modeling is performed by applying the concept of modeling shown in FIG. 3, the structure of a pattern having a complicated structure as shown in FIG. 3A can be simplified as shown in FIG.

After simplifying the structure of the pattern in this way, as shown in FIG. 6, the light waves reflected from the portion irradiated with the probe light are added and calculated, and the optical characteristics (for example, the spectral characteristics) of the reflected light are calculated. (Reflectance) is calculated theoretically. In this way, the calculation can be significantly simplified. At this time, the addition is performed by a coherent, non-coherent, or partial coherent calculation depending on the definition (that is, pitch) of the pattern and the spatial coherence length of the irradiation light.

Next, an embodiment will be described in which, when the structure becomes complicated such as a multilayer wiring, calculation can be performed by a simpler method in modeling. When the structure is multi-layered, it is necessary to perform modeling up to the bottom layer for accurate modeling, but it is necessary to consider the degree of overlap of wiring viewed from above, and considerable load May be. In actual wiring, there are circumstances such as the wiring width generally increasing toward the upper layer, and thus the optical effect increases with the upper layer. Therefore, as a result of repeated studies by the inventor and the like, when handling reflected light, even if it is handled as if there is no light reflection from a certain lower layer structure, an error of a problematic degree does not occur I found that.

For example, as shown in FIG. 7, even in the case of a wiring having a four-layered electrode structure, light is not reflected from below three layers from the top, that is, modeling is performed assuming that there is nothing. Can do it. In FIG. 7, 51 is Ti
N and 52 are Al wirings and electrodes, and 53 is a SiO ₂ insulating film. However, ignoring the fourth layer of the four-layer structure pattern as shown in FIG. Consider a layered structure. Then, the above-described modeling is performed to obtain a modeled pattern as shown in (c). By employing such a method, calculation for a pattern having a multilayer structure can be significantly simplified.

Although the modeling as described above is a very effective method for simplifying the calculation, a completely different problem arises when the target fine structure becomes very small. That is, generally, light cannot cope with a structure small in comparison with its wavelength by the scalar theory of the light wave which is usually used. Since the above-described modeling and the calculation of the reflected light intensity based on the modeling are based on the scalar theory, the calculation may be inappropriate for a very small uneven structure.

This phenomenon has already occurred in an ultra-fine device or a multilayer device (in this case, a fine structure is formed as a result of the overlapping state of the wirings). For example, 600n
In monitoring using light of a wavelength of about m, the structure should theoretically be the same as the large one (as long as it is smaller than the optical spatial coherence length), but the actual phenomenon is 0.13 μm In such a small structure, numerical calculations based on so-called vector theory are required. Numerical calculations based on the vector theory of light waves are enormous and difficult to actually perform.

However, as a result of consideration by the present inventors through experiments and the like, optically, in such a fine structure, a material having an uneven structure can be regarded as a homogeneous material (effective mass approximation). Found that it can be adopted.

A specific example will be described with reference to FIG. FIG.
In the figure, 61 is a SiO ₂ insulating layer, 62 is a Si substrate doped with impurities, and 63 is an Al wiring. That is, on the doped Si substrate 62, a very fine pattern of the Al wiring 63 that needs to be calculated by vector theory is laminated. In such a case, these Al wirings 63
When there are fine irregularities as compared with the wavelength as in the pattern group of the above, an approximation that light does not enter there, that is, an approximation that ignores the structure can also be made.

However, in the present embodiment, more realistically, instead of having no fine structure, continuous (with no irregularities) having different optical constants (refractive index n and absorption coefficient k).
The model is simplified and considered as having been replaced with the substance 64. Since the optical constant theoretically changes depending on the density of the fine structure, it is desirable to select an appropriate value depending on the pattern. It is also possible to determine the value of this variable by fitting the measured value to the optical constant as a variable.

After the pattern structure is simplified according to the above-described embodiment, and the spectral reflectance when spot light is irradiated is determined based on the simplification, for example, as shown in FIG.
The film thickness is obtained by the method described with reference to FIG. Further, in an actual polishing apparatus, the film thickness is measured in-situ by a method as shown in FIG. 1, and the polishing is terminated when a target film thickness is obtained, or the polishing method is changed. Or change it.

In the above description, the film thickness measurement has been described as an example. However, the polished state can be similarly detected. That is, the pattern structure is modeled by a method similar to that described above, optical characteristics such as spectral reflectance at the polishing end point are obtained based on the model, and actual optical characteristics are measured during polishing. The two can be compared by fitting calculation or the like, and polishing can be terminated when the degree of coincidence reaches a predetermined value.

Further, the above description has been made by taking a polishing apparatus as an example, but it goes without saying that the present invention can also be used in other steps such as film thickness measurement in a film forming step. .

FIG. 9 is a flowchart showing a semiconductor device manufacturing process. When the semiconductor device manufacturing process is started, first, in step S200, an appropriate processing step is selected from the following steps S201 to S204. According to the selection, the process proceeds to any of steps S201 to S204.

Step S201 is an oxidation step for oxidizing the surface of the silicon wafer. Step S202 is a CVD step of forming an insulating film on the surface of the silicon wafer by CVD or the like. Step S203 is an electrode forming step of forming electrodes on the silicon wafer by steps such as vapor deposition. Step S204 is an ion implantation step of implanting ions into the silicon wafer.

After the CVD step or the electrode forming step, the process proceeds to step S205. In step S205, it is determined whether or not to perform the CMP process, and if so, the process proceeds to the CMP process in S206. If the CMP step is not performed, S206 is bypassed. In the CMP step, the polishing apparatus according to the present invention performs planarization of an interlayer insulating film, formation of a damascene by polishing a metal film on the surface of a semiconductor device, and the like.

After the CMP step or the oxidation step, step S
Proceed to 207. Step S207 is a photolithography step. In the photolithography process, a resist is applied to a silicon wafer, a circuit pattern is printed on the silicon wafer by exposure using an exposure apparatus, and the exposed silicon wafer is developed. Further, the next step S208 is an etching step of removing portions other than the developed resist image by etching, and thereafter stripping the resist to remove unnecessary resist after etching.

Next, in step S209, it is determined whether all necessary steps have been completed. If not, the process returns to step S200, and the previous steps are repeated to form a circuit pattern on the silicon wafer. If it is determined in step S209 that all steps have been completed, the process ends.

[0079]

EXAMPLE Actually, an attempt was made to measure the spectral reflectance of the device structure of three electrodes and to measure the film thickness (at the top layer) by comparison fitting with theoretical calculation. The measurement was performed by an optical system in which the effects of diffraction and scattering were reduced as much as possible so as to enhance the theoretical conformity of the measured values. In addition,
The inventors of the present invention have invented such a measuring method and apparatus, and have applied for a patent as Japanese Patent Application No. 289175/1998. This is as shown in FIG. 2, in which only the zero-order term is measured, excluding first-order or higher order interference light.
In this embodiment, such a spectroscopic optical system is used. The irradiation spot diameter of the measurement light was 2 mmφ.

The wavelength range used is about 400 nm to 800 nm. Here, when calculating the theoretical spectral reflectance, the structure of the lowermost layer is simply modeled as a uniform metal layer, and the measured values and the fitting operation (using the film thickness value as a variable, Is higher.)
Waveform coincidence was shown as shown in FIG. In this case, the film thickness giving the selected calculated spectral reflectance was 534 nm. On the other hand, when this sample was measured by a cross-sectional SEM, the film thickness was 530 nm, and both were in good agreement.

[0081]

As described above, in the first aspect of the present invention, the actual pattern structure is
Since the model is modeled by simplification in consideration of the optical characteristics, theoretical calculation for obtaining the optical characteristics becomes easy.

In the invention according to the second aspect, since only vertical incidence, vertical reflection, and vertical transmission of light are considered, the amount of calculation in theoretical calculation can be further reduced. Claim 3
In the invention according to the above, since the back side of the metal layer or the absorption layer is made a simplified model of the fine structure assuming that the uniform body is filled, the amount of calculation when theoretically calculating the optical characteristics is reduced. , Even less.

In the invention according to claim 4, the modeling is performed on the back side of the metal layer or the absorbing layer on the assumption that the metal layer or the absorbing layer is clogged. Theoretical calculations for determining optical properties are further simplified.

According to the fifth aspect of the present invention, it is not necessary to consider the structure of the pattern at a portion deeper than the predetermined depth, so that the theoretical calculation for obtaining the optical characteristics of the reflected light and the transmitted light is further simplified. .

In the invention according to claim 6, since the theoretical value can be obtained by scalar calculation without using vector calculation, the amount of calculation can be significantly reduced. In the invention according to claim 7 and the invention according to claim 8, the amount of calculation can be reduced and the required memory can also be reduced.

In the ninth and eleventh aspects of the present invention, a modeled pattern used for calculating optical characteristics is stored, and the optical characteristics are calculated each time measurement is performed. Requires less memory capacity than storing characteristics. The invention according to claim 10, claim 1
In the invention according to the second aspect, it is not necessary to calculate the optical characteristics each time the measurement is performed, so that the calculation time can be reduced.

In the invention according to the thirteenth aspect, it is possible to determine the film thickness during polishing, the polishing end point, and the like, and at the same time, the amount of calculation and the required memory can be reduced.

According to the fourteenth aspect of the present invention, it is possible to determine the film thickness during polishing and the polishing end point, etc., to reduce the amount of calculation at that time, and to reduce the required memory.

According to the fifteenth aspect of the present invention, the film thickness and the polishing end point can be accurately detected during polishing, so that semiconductor devices can be manufactured with high yield.

[0090]

[Brief description of the drawings]

FIG. 1 is a diagram showing an example of a film thickness measuring device attached to a polishing device.

FIG. 2 is a schematic view showing an example of an optical system of the film thickness measuring device shown in FIG.

FIG. 3 is a diagram showing an example of a pattern structure to which the present invention is applied.

FIG. 4 is a diagram showing how the pattern shown in FIG. 3 is simplified and modeled.

FIG. 5 is a diagram showing modeling in the case of a complicated electronic element (transistor) structure.

FIG. 6 is a diagram illustrating a method of calculating optical characteristics of reflected light when probe light is irradiated from a simplified pattern.

FIG. 7 is a diagram illustrating an example of a procedure for simplifying and modeling a pattern of a multilayer structure.

FIG. 8 is a diagram illustrating an example in which a fine pattern is modeled and simplified.

FIG. 9 is a diagram showing a step of manufacturing a semiconductor device.

FIG. 10 is a diagram showing a correspondence between a spectral reflectance obtained by a theoretical calculation and an actually measured spectral reflectance in the example of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Wafer, 2 ... Wafer carrier, 3 ... Polishing platen, 4 ...
Polishing pad, 5: quartz transmission window, 6: irradiation light source, 6 ': optical system / light receiving device, 7: signal processing device, 8: xenon lamp, 9: lens, 10: slit, 11: lens, 12
... Beam splitter, 13 ... Lens, 14 ... Lens, 1
5: Reflecting mirror, 16: Lens, 17: Pinhole, 18:
Lens, 19: diffraction grating, 20: linear sensor, 21,
22 ... SiO ₂ insulating layer, 23 ... TiN film, 24 ... Al electrode and wiring, 25 ... Si doped with impurities, 31 ... SiO ₂ insulating layer, 32 ... Si substrate, 33 ... W, 34 ... TiN film, 35
... Al electrode, 36 ... SiO ₂ insulating layer, 36 ... Al electrode, 3
7, 38 ... Si doped with impurities, 51 ... TiN, 52
... Al wiring and electrode, 53 ... SiO ₂ insulating layer, 61 ... SiO ₂ insulating layer, 62 ... Si substrate doped with impurities, 63 ... Al wiring, 64 ... Substance

Continued on the front page F-term (reference) 2F065 AA30 BB02 BB03 BB15 BB16 BB17 BB18 CC19 CC31 DD00 DD06 DD07 FF51 GG03 GG24 HH03 HH04 HH13 JJ02 JJ09 JJ25 LL04 LL28 LL30 LL42 LL46 LL67 NN00 Q12 Q12 Q12 Q12 Q18 Q12 Q12 Q12 AQ12 DH38 DH39 DJ11

Claims (15)

[Claims] 1. A method for measuring a film thickness or a process state of a thin film on a substrate having a pattern formed therein by measuring optical characteristics of the substrate and calculating optical characteristics of the substrate in a predetermined number of film thicknesses or process states in advance. In advance, when measuring based on the degree of coincidence between the measured optical characteristics of the substrate and the optical characteristics obtained by the above calculation, the optical characteristics of the substrate in a predetermined number of film thicknesses or process states are calculated. A pattern modeling method characterized by modeling an actual pattern structure by simplifying it in consideration of the optical characteristics of the pattern prior to obtaining the pattern. 2. The pattern modeling method according to claim 1, wherein an optical characteristic of the pattern is considered only for light that is perpendicularly incident on the surface of the substrate and emitted vertically from the surface of the substrate. A method of modeling a pattern, characterized in that the pattern is modeled by simplification. 3. The method for modeling a pattern according to claim 1, wherein a metal layer or an absorption layer which reaches first after light is incident on the substrate is provided on a back side of the surface.
A pattern modeling method characterized by performing modeling by assuming that a uniform object is full. 4. The pattern modeling method according to claim 3, wherein the uniform object is modeled on the assumption that the uniform object is the same as the metal layer or the absorption layer. How to model the pattern. 5. The method according to claim 1, wherein
The method for modeling a pattern according to the item, wherein it is assumed that an object having a reflectance of 0 or constant optical characteristics is packed at a depth equal to or more than a certain depth from the surface of the substrate. A pattern modeling method characterized by performing modeling. 6. One of claims 1 to 5
The method for modeling a pattern according to the paragraph, wherein the patterning is performed on the assumption that a pattern having a fine structure smaller than the wavelength range of light is formed of a uniform continuous object having predetermined optical characteristics. A method of modeling a pattern, characterized in that: 7. One of claims 1 to 6
After modeling the pattern using the pattern modeling method described in the section, the optical characteristics when light is incident on the modeled pattern, taking into account the modeled pattern, a predetermined type of It is obtained by theoretical calculation for each film thickness, and the film thickness of the thin film on the substrate is measured based on the degree of coincidence between the measured optical characteristics of the substrate and the optical characteristics obtained by the calculation. Measurement method. 8. One of claims 1 to 6
After performing the pattern modeling using the pattern modeling method described in the section, the optical characteristics when light is incident on the modeled pattern, a predetermined process in consideration of the modeled pattern The state is obtained by theoretical calculation, and based on the degree of coincidence between the measured optical characteristics of the substrate and the optical characteristics obtained by the calculation, a film forming step of an insulating film or a metal electrode film on the substrate, or A process state determination method, comprising: determining a process state in a film removing step. 9. An apparatus for measuring the thickness of a thin film on a substrate having a pattern formed therein, wherein the apparatus irradiates light to the substrate surface and detects reflected light or transmitted light from the substrate. An apparatus and any one of claims 1 to 6
A storage unit that stores a pattern modeled by the pattern modeling method described in the section, considering the modeled pattern stored in the storage unit, optical characteristics when light is incident on the substrate. A calculation unit for obtaining a predetermined type of film thickness, and a film thickness determination unit for calculating the degree of coincidence between the optical characteristics calculated by the calculation unit and the actually measured optical characteristics and determining the film thickness from the calculated degree. A film thickness measuring device characterized by the above-mentioned. 10. A device for measuring the thickness of a thin film on a substrate having a pattern formed therein, wherein the device irradiates light to the substrate surface and detects reflected light or transmitted light from the substrate. An optical characteristic when light is incident on the substrate is determined by considering a device and a pattern modeled by the pattern modeling method according to any one of claims 1 to 6. A storage unit that stores the film thickness of the storage unit, and a film thickness determination unit that calculates the degree of coincidence between the optical characteristics stored in the storage unit and the actually measured optical characteristics and determines the film thickness based on the calculated degree. Characteristic film thickness measuring device. 11. An apparatus for determining a process state in a step of forming an insulating film or a metal electrode film on a substrate or a step of removing these films, wherein the apparatus irradiates light to the substrate surface, An apparatus for measuring reflected light or transmitted light from a substrate, a storage unit for storing a pattern modeled by the pattern modeling method according to any one of claims 1 to 6, and a storage unit Calculated in consideration of the modeled pattern stored in the optical characteristics when light is incident on the substrate, a calculation unit for obtaining a predetermined process state, and the optical characteristics calculated by the calculation unit are actually measured. A process state determination unit that calculates a degree of coincidence with the obtained optical characteristic and determines a process state based on the calculated degree. 12. An apparatus for determining a process state in a step of forming an insulating film or a metal electrode film on a substrate, or a step of removing these films, wherein the apparatus irradiates light to the substrate surface; An apparatus for measuring reflected light or transmitted light from a substrate, and light applied to the substrate in consideration of a pattern modeled by the pattern modeling method according to any one of claims 1 to 6. A storage unit that stores the optical characteristics when the light is incident on a predetermined process state, and a process state in which a degree of coincidence between the optical characteristics stored in the storage unit and the actually measured optical characteristics is calculated, and the process state is determined from the calculated degree. A process state determination device, comprising: a determination unit. 13. A polishing apparatus, wherein the method for measuring a film thickness according to claim 7 or the method for determining a process state according to claim 8 is used in a polishing process. 14. A film thickness measuring device according to claim 9 or claim 10, or a process state judging device according to claim 11 or 12. Polishing equipment. 15. A method for manufacturing a semiconductor device, comprising a step of polishing a wafer using the polishing apparatus according to claim 13.