JP3375338B2 - Endpoint detection technique using signal slope determination - Google Patents

Description

Description: BACKGROUND OF THE INVENTION The present invention generally relates to signal processing techniques that determine when a particular stage of a condition change has been reached, or to automatically control an event that results from a condition change. For example, in semiconductor processing technology, it relates to the monitoring of electrical signals to determine that an endpoint has been reached in connection with photoresist development or etching operations.

There are various situations in which an electrical signal is present and it is desired to monitor the electrical signal as it changes in relation to changing conditions. One group of these situations is in the course of processing a material, when it wants to determine that it has completed a particular process, ie, detecting the endpoint of a process operation. The determination of the endpoint is to monitor the progress of the process or the control of the process, for example to terminate the particular processing operation being monitored.

The semiconductor industry is one example of an industry that requires monitoring of such processes. At some stage in the fabrication of circuits on semiconductor wafers, a mask of photoresist material is formed. The manufactured mask is used to limit the processing of the layer covered by the mask and to pattern the areas. The mask is formed by exposing the photoresist layer in the desired fashion, followed by the step of developing the photoresist layer by supplying a developer thereto. In conventional photoresist materials, the exposed areas are removed in the course of development to expose the underlying layers.
When the underlying layer is exposed by removing the photoresist material, it is said to be "breakthrough" or "endpoint." The development process is allowed to continue for a certain period from the time the breakthrough is first detected, and the end of that time is the end of the development process, which is called the "process end". .

Various processes existing from batch to batch of semiconductor wafers and monitoring of the development process to determine when breakthroughs occur due to environmental changes are needed. A finite bandwidth light beam is directed onto the photoresist layer of one of the batch wafers and the reflected or transmitted light signal is detected and the electrical signal determines whether a breakthrough has occurred. Is processed for. In one form, the light reflected from the surface of the substantially transparent photoresist layer and from the bottom surface causes interference at the photodetector. In the course of development, as a portion of the photoresist layer is removed, the detected light intensity between the maximum and the minimum of the reflected light varies as the material is removed, and the relative intensity between the two interfering beams is increased. The relative phase changes. However, in breakthrough, the change in this signal is terminated and by analyzing the photodetector output signal,
Two conditions are detected. Development is continued from breakthrough detection for a fixed amount of time, at which point
Development is terminated by washing the developer or by other means.

The wet etching process uses a similar breakthrough detection process in which the photoresist material is etched away and the substantially transparent material layer is removed.
In the process of dry etching, the material is removed by bombardment bombardment in the plasma chamber and is removed from the layer being etched by detecting the emission of light in a limited wavelength band, one component of the plasma. Monitored by detecting what has been done.
This is electronically detected because the detected signal in such applications usually drops significantly during breakthrough.

One object of the present invention is to provide an improved technique for processing monitored and detected electrical signals during the processing of semiconductor wafers.

A more general object of the invention is to improve in a wide range of applications where an electrical signal is present, or is present in connection with a change in the condition under which the electrical signal is being monitored, or is obtained. It is to provide a signal processing technique.

SUMMARY OF THE INVENTION These and further objects are achieved by the present invention, but in the present invention, simply and generally speaking, the value of an individual change of an electrical signal is calculated, which Occurs during the monitoring of conditions to determine that a particular stage has been reached.

In one form, the change in the signal is monitored as the slope of the electrical signal. According to one particular feature of the invention,
A digitized electrical signal is acquired and a group of consecutive digital signal values is used to calculate the slope value of a signal. Subsequent slope values are calculated from successive groups of digital samples,
Each group utilizes at least some of the values of the most recent preceding group, except at least the first acquired sample in the most recent preceding group. The particular desired stage of change of the condition is determined when the calculated slope value is calculated to lie in a predetermined region. Rather than relying on absolute electrical oscillation values to make decisions when changes in conditions reach their particular stage, rather slopes, or other signal varying properties are used.

The embodiments of the present invention described below with reference to the accompanying drawings illustrate the application of the improved signal processing techniques according to the present invention, which relate to monitors and where appropriate control of the processing of semiconductor wafers. Used for. However, it will be appreciated that the applications of the invention also exist in various other fields. Other applications include those in monitoring electrochemical signals that are proportional to the concentration of components contained in the chemical composition, such as mass spectroscopy, electron spin resonance (ESR), nuclear magnetic resonance (NMR), temperature, pressure, Acoustic signal, refractive index change in response to changes in chemical or physical parameters, flow metering, colorimetry, strain gauge signal, light dispersion, crystal frequency change, fluoroimmunoase technology (fluoro)
-Immuno-assay-techniques), and pH monitoring of solutions, these are only some of the many possibilities.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 illustrates a process for developing a photoresist layer of a semiconductor wafer in which the present invention is utilized; FIG. 2 illustrates improved photoresist development in which the present invention is utilized; FIG. 3 is FIG. 2 is an electronic block diagram of endpoint control used in connection with any of the two photoresist developments; FIG. 4 illustrates optical monitoring of the photoresist layer to the electrical signal to be processed; FIG. 5A. 5B shows an example of the photodetector output signal obtained in the process of developing the photoresist; FIG. 5B shows a derivative waveform obtained by calculating the output of the photodetector of FIG. 5A; FIG. 6A and FIG. 6B show the process of developing the photoresist. 7A and 7B show enlarged cross-sectional views of the two stages of the etching process; and FIG. 8A shows the detector to be monitored. FIG. 8B shows an example of a digital sample of the output, FIG. 8B shows the slope values calculated from the sample of FIG. 8A; FIG. 9 generally illustrates an apparatus for performing a dry etching process through a plasma discharge. FIGS. 10A and 10B show the level of the machining signal monitored in the plasma system of FIG. 9 and its calculated slope; FIGS. 11A and 11B are the monitored increases in the plasma system of FIG. 9, respectively. FIG. 12 shows a signal and its calculated slope; FIG. 12 shows a flow chart of a signal processing technique for obtaining an endpoint; FIG. 13 shows one step of the method of the flow chart of FIG. 12 in one embodiment of the present invention. FIG. 14 shows an enlarged flow chart; and FIG. 14 is an enlarged flow chart of the same steps of the method according to the flow chart of FIG. 12, showing another embodiment of the present invention.

Description of the Preferred Embodiments Referring initially to FIG. 1, there is shown an apparatus commonly used in the development process of photoresists for forming electronic circuits on semiconductor wafers. The liquid enclosure 11 has a support 13 in it, which is supported by a shaft 15 and is rotated by a motor 17 provided outside the enclosure 11. Positioned on the rotating support 13 is a semiconductor wafer 19 being processed.
At this point in the process, the wafer structure has already been provided with a photoresist layer and exposed to a pattern of light corresponding to the desired physical pattern to be left after development of the photoresist. This physical mask pattern is utilized in subsequent steps of etching or used to treat the layer just below the photoresist layer.

Developer is typically applied to the photoresist layer by spraying, but other techniques are available as well. Nozzle 21 supplies such spray with solution from vessel 23 via an electrically controlled valve 25. Different specific techniques for intermittent spraying or continuous spraying are used in different semiconductor manufacturing processes. Once development is complete, valve 25 is closed and then another electrically controlled valve is opened to allow rinse solution from container 29 to another nozzle.
Supplied to 31.

The spray of developer and rinse is typically done when the wafer 19 is rotated at a uniform speed. The electronic process controller 33 consists of a rotary motor 17, valves 25 and 27,
And control other processing units.

Electromagnetic radiation to monitor the development process,
A beam 37 from a light source 35 in the visible or near visible region is directed at the developing photoresist layer on the developing wafer 19.
A photodetector 39 is arranged to receive the reflected light of beam 37 from wafer 19.

The electronic control system 41 drives the radiation source 35 and
The signal obtained from the photodetector 39 in 40 is processed. Since breakthrough detection is obtained as a result of light interference, some coherence of the light source 35 is generally used, as will be explained below, and a light emitting diode (LE
D) degree is used enough. At the output of the electronic system 41, what appears in the circuit 43 indicates that a breakthrough of the photoresist layer that has occurred is found. This signal often simply informs the operator that a breakthrough has occurred in the process change being monitored. In contrast, the breakthrough signal in circuit 43 is passed directly to process control 33 to stop development after a certain amount of time after a breakthrough was detected in process control 33.

Unlike spraying the developing solution onto the photoresist layer, there is a developing device that uses a tank 45, and in the tank 45, a boat 47 is immersed in a developing solution 51 while supporting a plurality of wafers, for example, a wafer 49. Has been done. The position of the front surface of the wafer 49 is monitored during the development process. When the process end is determined or believed to have occurred,
The boat 47 and its wafers are removed from the tank 45 and immersed in the rinse tank to finish the process. One endpoint controller is the controller
Substantially the same as 41, which is used concurrently with unit 53.

A block diagram of the generalized controller 41 is shown in FIG. Various elements are connected to the system bus 55, typically or for a special purpose computer. The microprocessor 57 is one of such elements, and is connected with a system dynamic random access memory (RAM) 59 and a read only memory (ROM) 61. One input / output circuit 63 has a control and display panel, such as a cathode ray tube display, or a circuit 65 with some keyboard input terminals.
Interface to. The second input / output circuit 67 receives the output analog signal of the photodetector in circuit 40 and digitizes it for processing by the control system of FIG. The third input / output circuit 69 indicates to the circuit 43 that the end point has been reached or supplies a signal to calculate the end time of the process. The control software for operating the controller of FIG. 3 includes that for determining when a breakthrough has occurred, which is stored in ROM 61 and executed by microprocessor 57.

Such an endpoint controller, light source, and photodetector assembly is sold in the Kisnik Division of Luxtron Corporation, Santa Clara. The models 2200, 2300, 2400 are wet process endpoint controllers and are widely used for this purpose. Photoresist layer breakthroughs are currently performed in these devices by computer programs through window trigger algorithms. The present invention is primarily related to an improved alternative technique for detecting breakthroughs. The improved technique of the present invention can be implemented by loading the software of the new technique into the kinematic controller memory instead of the current window trigger algorithm.

Before describing this improved breakthrough signal processing technique, the photoresist development process will be further described. FIG. 4 shows, for illustrative purposes, a semiconductor structure 71 having a photoresist layer 73 on its surface. Ray 37 '(FIG. 1) of illumination beam 37 is partially reflected at the surface layer of layer 73, shown as ray 75'. The material of layer 73 is substantially transparent so that
The 7'portion proceeds to layer 73 and is reflected as shown as ray 79 at the boundary 77 between the layer and the underlying structure.

The photodetector located in the path of rays 75 and 79
Provided that 37 has sufficient monochromaticity and interference,
A photodetector located in the path of rays 75 and 79 detects the intensity level of light resulting from the interference. Its intensity is ray 75
It depends on the difference in the optical paths that 79 passes through. As the thickness of layer 73 decreases during the development of photoresist,
The optical path length of 75 changes, and the intensity of the interference beam detected by the photodetector can be seen.

An example of the output of the photodetector (FIG. 1) during such processing is shown in FIG. 5A. Areas of the photoresist layer have been developed, causing signal oscillations at the level of exposure being removed. With the breakthrough of layer 73, however, further development does not cause a change in optical path length and vibrations subside. Thus, at 89, the oscillation of the photodetector output ceased, suggesting that a breakthrough had occurred.

The techniques of the present invention are utilized to determine when such a breakthrough has occurred.

However, this special processing is done where the electrical signal is very noisy, so the photodetector output signal is unfortunately shown in Figure 5A for brief illustration. It's not as pure as what you have.

The signal processing technique according to the invention fully takes into account the possibility that the signal to be analyzed contains noise components.

Although the breakthrough detection techniques described herein are effective when the entire material layer is removed, it is typically applied in situations where the patterned layer of material is removed. FIG. 6A is a cross-sectional view similar to FIG. 4, but showing the situation where the first breakthrough occurred in the exposed areas of photoresist layer 73 during the development process. The reflected signal in FIG. 5A shows the time 89 at which the breakthrough illustrated in FIG. 6A first occurred.
Development is further continued, usually for a fixed period of time after the breakthrough is detected, resulting in complete removal of the exposed photoresist areas as shown in Figure 6B.

Similar steps are performed in wet etching operations. Referring to FIG. 7A, the substrate structure 91 has a continuous substantially transparent layer 93 thereon, which is silicon oxide or the like, which is a mask.
Covered by 95.

When etching is done through the openings in the mask 95, a breakthrough as shown in Figure 7A eliminates signal oscillations in the photodetector output by monitoring the signal as in Figure 5A. It is determined by how high. With the development of the photoresist, the etching is continued for a period of time after the breakthrough is detected so that the larger opening is completely removed as shown in FIG. 7B.

An improved technique according to the present invention capable of detecting breakthrough from the output signal of a photodetector is shown generally in the curve of FIG. 5B.

The photodetector signal output is mathematically derived from the photodetector output signal of FIG. 5A by the controller system of FIG. 3 via the curve data of FIG. 5B. The breakthrough is determined only by the derived signal of Figure 5B, and it is unnecessary to directly use the absolute signal value of Figure 5A. Positive threshold level 97 in certain embodiments
Is set. Once it is determined that the process of development has begun, the derivative signal shown in Figure 5B is monitored for a fixed period of time, which is less than the period of noise level that occurs in the signal. And / or successive thresholds set for a period longer than the period of said oscillation
Decided by falling below 97. Such a fixed period is a fixed period such as ending at time point 99 in FIG. 5, at which point the breakthrough is understood to have occurred at time 89.

As mentioned above, it must be determined that the process has actually begun before the signal processing techniques described can locate breakthrough points. One way to do so is to detect when the vibration has begun. That is, are vertices or valleys occurring in the derived signal of Figure 5B.
Another way to determine if the process has actually started is simply for the derived signal shown in FIG. 5B to exceed threshold 97 for a period of time.

This avoids false triggers due to any other noise or the like.

Instead of just using one threshold 97, a negative threshold 101
Both of the corresponding ones can be used. In this case, the breakthrough is that the derivative signal of FIG.
Determined by the existence of only a certain period between 101. Similarly, the start of the development process is known by the fact that the derivative signal of FIG. 5B has been exceeded for a certain fixed period of time by the threshold values 97 and 101.

8A and 8B illustrate a preferred technique for calculating the derivative signal shown in FIG. 5B from the analog output signal of the photodetector shown in FIG. 5A.

As previously mentioned, the I / O circuit 67 (FIG. 2) digitizes the analog signal of the photodetector in circuit 40. Samples 103-109 of such digital signals are shown in FIG. 8A to illustrate that such a process is performed.

These values are the values 111 to 115 of the individual signal, the changed characteristics of the signal being calculated, ie its slope. Points 111-115 in FIG. 8B are derived signal points already discussed in connection with FIG. 5B.

The slope value 111 in FIG. 8B is determined from the slope model equation, in this case line 117, which is a series of digital sample points 103, 104 and 1
It conforms to 05.

Similarly, the next slope value 112 is calculated according to the same model equation, in this case another straight line 118.
, Which fits successive digital samples 104, 105 and 106.

Similarly, slope 113 corresponds to line 119, slope 114 corresponds to line 120, and slope 115 corresponds to line 121.

It can be stored that the first slope value 111 is calculated from the first group of consecutive digital samples 103, 104 and 105.

The next slope, 112, is a digital sample, calculated from the group of digital samples, excluding the first sample 103, and then adding the most recently obtained sample 106. Thus, a new digital sample of the analog signal of the photodetector obtained by the control system of FIG. 3 is obtained, the new slope value is calculated to fit the curve to the combination including the newly obtained one, and These are used to calculate the next slope, but the first digital sample is removed from the group. That is, each slope calculation is obtained by replacing the most recently acquired sample with a predetermined number of consecutive digital samples with the oldest sample in the group.

Although the number of samples in each group shown in FIG. 8A is only three, this is for ease of explanation and this group can generally be larger. The exact number of samples used in any application will depend primarily on the degree of favorable noise immunity formed by the calculation.

Also, although linear alignment is shown, again for ease of explanation, it may be possible to adapt it to more complex multi-order digital signals in some applications. This technique is an application of the well known Sobski-Gorey multi-order method. For example, ray. Sobiki and M. Jay. E. Golay, Analytical Chemistry, Volume 36, 16
Pages 27-1639 (1964), or p. A. Gory,
Analytical Chemistry, Volume 62, 570-5
See page 73 (1990). The advantage of fitting a curve to real-world data is that modeling the process allows for forward process control.

Similar signal processing techniques may find application in processing operations on a wide variety of materials or compositions.

Another suitable application from another particular semiconductor industry is outlined in FIG. Rather than etching away the layer of material with a chemical as shown in FIG. 7, etching is often done in a dry plasma process in vacuum chamber 131.

As is well known, the semiconductor wafer 133 is located below the target structure 135 in the chamber.

Depending on the particular treatment process that will be performed, the target 135 is energized by a direct current or radio frequency power source 137. A source 139 of inert or active gas selectively connects these gases to the interior of chamber 131.

A plasma 141 is formed on the semiconductor wafer 133 from the generation of gas molecules under the influence of a strong electric field. As a result, wafer 1
33 is bombarded with ions and it removes a layer of material on the semiconductor wafer 133 of material or, if such removal is masked, only removes the portion defined by the mask. To be done.

This plasma process has the same purpose as the wet etching process. That is, there is a need to accurately determine when the breakthrough first occurs and, as a result, the process of etching is stopped at a suitable time and not too long. In plasma processes, the visible or near visible electromagnetic radiation of plasma 141 is usually observed through a quartz window 142. A photodetector 143 is provided in the path of the plasma radiation, which is visible through window 142 proximate the outside of chamber 131 as shown in FIG. The radiation intensity of the plasma 141 in one or a number of selected narrow wavelength regions or lines is detected, and the electrical signal of one of the photodetectors 143 is proportional to their intensity. A detector in some limited emission wavelength range usually has a narrow bandwidth interference or other type of filter placed in front of the photodetector 143 so that it receives what is in that wavelength range. .

The wavelength region chosen to be monitored is such that its intensity changes in relation to the progress of the etching process whose intensity is being monitored.

FIG. 10A illustrates such a narrow band signal that can be monitored. Breakthrough of the layer being etched occurs at time t _B , where the change in the signal pattern is abruptly flattened.

Similarly, FIG. 11A is monitored under certain circumstances of plasma emission bandwidth, where its intensity increases as etching progresses, and the endpoint is indicated by time point t _B , From there, the increasing signal is gradually flattened.

The preferred signal processing technique for detecting when the time t _B decreases or increases, which signal occurs in FIG. 10 or FIG. 11, respectively, is the same as already described in connection with FIG.

That is, the analog output of the photodetector shown in FIG. 10A is digitized and point by point in FIG.
Processed to obtain the derived signal shown in. The negative slope threshold level 145 is set to the appropriate level. The process first detects if the signal of FIG. 10B has crossed the threshold 145 for some predetermined time, thus indicating that the etching process has begun. After the start of the process is detected, the signal in Figure 10B shows when it is at threshold level 1
It indicates that a breakthrough occurred depending on whether or not it dropped below 45 for a certain period of time. Separated threshold levels are available for process initiation and breakthrough occurrence.

Similarly, the derivative signal shown in FIG. 11B is shown in FIG. 11A.
From the photodetector output of. A threshold 147 is used with the derived signal. The derived signal shown in FIG. 11B has a threshold 147.
It is determined that the process has started when a certain time is exceeded. Breakthrough is subsequently determined when the derived signal falls below the threshold 147.

Referring to FIG. 12, there is shown the signal processing technique used in the endpoint controller of FIG. 3 to determine breakthrough from the derivative signal of FIG. 5A or 10A or 10B.

In each case, these analog signals are first digitized, and the derivative signals respectively shown in FIGS. 5B, 10B, and 11B are determined by the process shown in FIG.

The signal processing technique of the first step 151 in FIG. 12 sets various parameters and one or many counters are used. One such parameter is named "N". The consecutive data points are used once to determine each slope. In the example shown in FIG. 8, N = 3. It is preferable to let the final user select this N. The reason is that its value is uniquely determined by the amount of noise immunity required in a particular application.

The N data points are usually grouped with overlapping intervals, which are shown in FIG. 8A, but instead are grouped in stepwise intervals, where the values of each slope are different data points. Chosen from. This technique is handled in the same way in the case where overlapping cases are stepwise.

After such initialization, the next step 153 is N-
Acquire and store one subsequent data point, as illustrated in Figure 8A. Different data points are obtained and shown by step 155, where there are enough data points to calculate one slope value, which are calculated by the technique shown in FIG.
Shown at 157. Some such calculated slope values are shown in Figure 8B.

After each calculation, the next step 159 is to determine if the semiconductor process has begun. Of course, the criteria for this breakthrough do not apply until it is clearly confirmed that the process has begun. Otherwise, breakthrough decisions will be made too early, and the manufacturing process will be stopped and terminated before it begins. If the process has not been started, more data values will be used to obtain additional slope points.

If it is determined that the process has started, then the next step 161 compares the most recently calculated slope value to the target end slope. If it is lower than the target end slope, the slope end counter is incremented by 1, as shown in step 163. This is one of the counters initialized in step 151 and has the first 0 counts. The next step 165 compares the new count of the new ramp end counter with the target end count. The target end count is different from one of the parameters originally set in 151 and is preferably user selectable. That is, this set is effective in that the signal slope value is not determined before breakthrough occurs below the target value. This number is also set in relation to the amount of noise on the signal in a particular application. Thus, the target end slope operates as a value that does not exceed the threshold value.

Once the ramp end counter has been determined to be added to a value above the current target end count, it is indicated at step 167 that a breakthrough has occurred.
The process end time is then calculated by the addition of excess processing time, and the process proceeds for such a longer period as shown in step 168. If the target count is the most recent slope value and it does not reach less than the target end slope, then processing returns to step 155 to determine new data points as determined at step 165. if,
If the most recent slope value exceeds the target end slope as shown in step 161, the slope end counter is reset to zero as shown in step 169.

The special procedure is used in step 159 to determine whether the semiconductor process has been started or cannot be changed, and the corresponding type of signal to be analyzed is selected. 13 and 14 show another method for determining whether the process has started or not.

The technique shown in Figure 13 looks for slope values above a predetermined threshold and can be used to analyze the analog signals of Figures 5A, 10A and 11A. The tilt start counter is used in addition to the tilt end counter shown in FIG. The start counter in ramp, as shown in step 171, is compared to the target start counter and is also user selectable. If the ramp start counter has been pre-added to exceed the target start counter, the process continues at step 161 in FIG. If not, another step 173 (FIG. 13) occurs.

In step 173, the most recent slope value calculated in step 157 of FIG. 12 is compared to the threshold target start slope.

If the ramp value exceeds the threshold, the ramp start counter is incremented by 1 as indicated at 175. Processing then returns to step 155 (as shown in FIG. 12), continues to step 157, and then steps 159 to 159.
Proceed to 161. However, if the most recent slope value is less than the target slope, the slope start counter is reset to 0, as shown in step 177, and the processing of Figure 12 returns to steps 155 and 157.

Alternative signal processing techniques are preferably utilized for vibration signals of the type shown in FIG. 5A.

Referring to FIG. 14, step 179 is performed with the new slope values shown in FIG.
Happens after being calculated by 157 of.

In this step 179, a determination is made as to whether a signal peak has occurred. If so, it is the start of the semiconductor process and the signal processing means proceeds to the next step 161 in FIG. If it is determined that no peak is found, the next step 181 in FIG. 14 is to search for signal troughs. If such a trough is found, then the next step (Fig. 12
) 161 and, if not, return to signal processing ramp steps 155 and 157 of FIG.

In each step 179 and 181 of FIG. 14, the most recently calculated slope value is used to determine whether the most recent previous slope value and peak or valley have occurred in the derived signal. In step 179, the slope value of number P is used. According to a slope value of substantially 0, a group of a certain number of successive slope values, which was initially increasing,
It is determined whether it continues with another decreasing predetermined value.

Similarly, step 181 uses the most recently calculated slope value for the number V slope value and whether a valley is present. There are valleys, if the first slope point is reduced, followed by another number of substantially zero, followed by another number of increasing slope samples.

Although signal processing techniques have been implemented for particular semiconductor processing examples, the techniques have wide application and are widely applicable to areas where analog signals, FIGS. 5A, 10A, 11A or similar signals exist. Similarly, it can be used for the processing of articles other than semiconductor substrates for monitoring purposes, such as the beam of electromagnetic energy in question or similarly detectable electromagnetic energy that is released during processing from the material. It is possible by observing. This technique can also be used to monitor chemical reactions. In addition to this, a number of applications are also available for electrical analog signals that already exist, for example as a by-product of certain operations, for example as spectrophotometers.

Thus, the invention is to be protected within the full scope of the appended claims, not just any particular process or operation.

─────────────────────────────────────────────────── ─── Continuation of the front page (56) Reference JP-A-53-17078 (JP, A) JP-A-59-208724 (JP, A) JP-A-2-137852 (JP, A) Actual development Sho-59- 121656 (JP, U) Special Table Sho 59-500892 (JP, A) US Patent 5021362 (US, A) International Publication 91/18322 (WO, A1) (58) Fields investigated (Int. Cl. ⁷ , DB name) ) H01L 21/027 G01N 21/27 G01N 21/45

Claims (40)

(57) [Claims] 1. A method of detecting when a step of an etching process for removing a layer formed on a substrate has entered a specific stage: (a) an analog signal relating to the change of the layer caused by the process. And (b) periodically sampling the analog signal to obtain a series of digital samples, (c) for the predetermined number of groups of digital samples, the individual groups. Sample contains at least some of the samples from the most recent preceding group, but at least removes the first acquired earlier and includes recently acquired samples that do not belong to the previous group A calculation step for sequentially and repeatedly calculating the rate of change of the signal for the group; Determining when the percentage falls within a predetermined range to determine when the treatment of the layer of material has reached the particular stage; and How to determine whether you have entered the stage. 2. The method of claim 1, wherein the step of computing iteratively any time comprises fitting the curve to a group of predetermined digital samples having temporal continuity, and of the signal from said curve. Calculating one value of the change, and detecting a particular stage of change of the condition, including: 3. The method of claim 1, wherein the identifying of a change in condition includes the additional step of terminating the change in state in response to determining that the condition has reached a predetermined stage. How to detect the stage of. 4. The method of claim 1, wherein the particular stage of change in condition is detected including the additional step of treating the material whose condition has been monitored as a result of the treatment of the material. how to. 5. The method of claim 4 wherein the step of monitoring and generating the signal causes the detector to be in the path of an electromagnetic radiation signal, the changes of which correspond to changes in the properties of the material being processed. A method to detect a specific stage of change in the conditions being placed. 6. The method of claim 5, wherein the step of monitoring and generating the signal includes the step of directing electromagnetic radiation to a material, the result of which is directed to a detector, whereby the detector is of the material. A method of detecting a particular stage of a change in condition that is subject to a modified light beam due to the change in condition. 7. The method of claim 5, wherein the step of placing the detector includes the step of placing the detector in the path of electromagnetic waves emanating from the portion where the material is being treated. How to detect steps. 8. The method of claim 1 including the additional step of changing the thickness of at least a portion of a layer of a material over the surface of said layer, wherein said changing of said conditions is said material layer. A method of detecting a particular stage of a change of conditions in which the change of state is monitored by monitoring a signal generated by the thickness of the. 9. The method of claim 8 wherein the step of altering the thickness comprises the step of increasing the thickness of the layer of material to detect a particular stage of change in conditions. 10. The method of claim 8 wherein the step of varying the thickness comprises the step of reducing the thickness of the material to detect a particular stage of change in conditions. 11. The method of claim 10 wherein the step of varying the thickness of the material layer detects a particular stage of change in conditions that reduces only patterned portions of the material layer. how to. 12. The method according to claim 8, wherein the layer whose thickness of the material layer is monitored comprises a semiconductor substrate. 13. The method of claim 8, wherein the step of monitoring and generating the signal comprises the step of directing electromagnetic radiation to a layer of material and then to a detector whereby the detector is of a thickness of the layer. A method of detecting a particular stage of a change in condition that is subject to a beam modified by the change in condition. 14. The method of claim 8 wherein the step of disposing the detector comprises disposing in a path of electromagnetic radiation emitted from a portion of the process of varying the thickness of the material layer. To detect a specific stage of change in the. 15. The method of claim 8 wherein the layer whose thickness is modified is a photoresist material layer to detect a particular stage of change in conditions. 16. The method of claim 8, further comprising the additional step of terminating the modification of the thickness layer in response to determining that the modification of the thickness condition has reached the particular stage. A method for detecting a specific stage of change in conditions, including. 17. The method of claim 1, wherein the particular step of changing the condition is such that the magnitude of the digital signal sample is detected without comparing magnitude with the absolute value of a reference signal level. How to detect a specific stage of change. 18. A method for detecting when a step of etching a layer on a semiconductor substrate has entered a specific stage: detecting one electromagnetic radiation signal indicative of a rate of change associated with the change of the layer. Directing to the detector, and generating an electrical signal from the detector in the form of a continuous periodic digital value, which also indicates a rate of change, which is associated with a change in the state of the layer. A step of calculating an individual value from the digital signal of the successive group by the next step during the processing of the layer by a next step, and a first of a predetermined number of time-continuous digital signals. Creating a curve from the groups, calculating one value of the signal from the curve, and the value of the change of the signal in a plurality of different groups beforehand. Calculated from a specified number of digital values, and the individual group values are at least one of the immediately preceding immediately preceding groups removed, and the preceding curve fitting repeated sequentially in time. Determining, from the value of the change in the signal, that processing of the layer has started, and after determining that processing of the layer has started, a first preliminary step relating to the stage of the state of the layer. Determining at the time of one initially preset limit that there is a defined number of subsequent signal changes, thereby detecting that the layer state has entered the stage; A method for detecting stages. 19. The method of claim 18, wherein the step of detecting when processing of the layer has begun comprises the step of: A method for detecting a particular stage in the state of a layer that determines what has been done outside an associated second predetermined limit. 20. The method of claim 19, wherein the first
Is a signal change below a predetermined level, and the second predetermined limit is a signal change value greater than the predetermined level. A method for detecting stages. 21. The method of claim 18, wherein the method comprises the additional step of terminating the processing of the layer in response to the predetermined stage of the layer status. A method for detecting stages. 22. The method of claim 18, wherein the particular step of the layer state is not determined by comparing the magnitude of the digital signal value with the absolute value of the signal level. A method for detecting a particular stage. 23. A method for detecting when a step of an etching process for removing a layer on a substrate enters a specific stage,
In a method of detecting breakthrough in the process of removing material from a substantially transparent layer: to the surface of said layer of radiation reflected from it and to the surface of the layer below it during the material removal process; Directing electromagnetic waves to cause radiation from both, with one detector an electrical charge proportional to the strength of the interference between the reflection from the exposed surface of the layer and the reflection from the surface of the underlying layer. Receiving a reflected radiation whose signal is a change in the signal as the material is removed from the layer,
Obtaining digital samples of a signal, for said group of digital samples, the samples of said individual groups include at least some of the samples of the immediately preceding group, but at least the earliest preceding sample And the slope function representative of the temporal change of the signal is sequentially removed by iteratively calculating the rate of change of the signal for a group including recently acquired samples that do not belong to the previous group. After the step of calculating from the electrical signal in the process, the step of determining from the slope function that the removal process has started, and the step of determining that the process of removal of the material has started, the slope function is predetermined. Detecting that a fixed value has been reached for a predetermined time, and Method for detecting a breakthrough in the process vibrations of issue of removing material comprising a detecting step of confirming that said breakthrough is detected has occurred. 24. The method of claim 23, wherein detecting that processing of the process has begun comprises removing material that includes detecting the presence of at least one signal portion of the electrical signal. How to detect breakthroughs in a process. 25. The method of claim 23, wherein the process of determining that the process has begun has the slope function exceed the second predetermined value for the second predetermined period. A method of detecting breakthrough in a process of removing material including detecting. 26. The method of claim 25, wherein the first
And the value of the second predetermined function is as a value of the method of detecting breakthrough in the process of removing material. 27. The method of claim 23, wherein the step of calculating the slope function obtains periodic samples of digital samples of the electrical signal generated by the photodetector, the temporally continuous pre-samples. A method of detecting breakthroughs in a process of removing material comprising fitting a curve to a defined number of groups of digital samples. 28. The method of claim 27, wherein the step of detecting when the slope function matches a predetermined fixed value for a first predetermined period of time is the number of successive slope function values. Detects a breakthrough in a process of removing material that includes a period of time that determines that the first predetermined value remains below the predetermined value for the first predetermined time period. Method. 29. The method of claim 23, wherein the process of removing material comprises developing a photoresist layer that selectively dissolves according to a pattern to detect breakthrough in the process of removing material. how to. 30. The method according to claim 29, wherein the step of developing the photoresist layer includes the step of supplying a developing solution to the exposed surface to detect a breakthrough in the process of removing the material. 31. The method of claim 23, wherein the material removal process includes a process of replacing one mask and etching the material to detect breakthrough in the process of removing material. 32. The method of claim 31, wherein the layer is a process for detecting breakthrough in a process of removing a material in an etching solution to etch the exposed portion through the mask. 33. The method of claim 23, wherein the material removal process includes a process of removing material from the layer provided on one semiconductor wafer, the method of detecting breakthrough in a material removal process. 34. The method of claim 23, wherein the process of removing material includes the additional step of terminating material removal of the layer in response to the occurrence of the breakthrough being identified. How to detect. 35. The method of claim 23, wherein the breakthrough is a process of removing material that is identified without comparing the magnitude of the electrical signal to the absolute value of the reference signal level. . 36. A method of detecting when a step of an etching process for removing a layer on a substrate enters a specific stage,
In a method of detecting breakthrough in a plasma treatment process for removing material from the layer: a detecting step of detecting an element of electromagnetic radiation of the plasma to generate an electrical signal proportional to the intensity of the radiating element; and a slope of the signal. Calculating an electrical signal of the removal process of a slope function representative of the temporal change of: (a) obtaining periodic digital samples with intervals of the electrical signal detected in the detecting step; (B) creating a curve from a predetermined number of groups of time-sequential digital samples; (c) calculating a value from the slope function; (d) above (b) ) And (c) are repeated for a plurality of groups of the predetermined number of digital samples, the individual groups being small. Including an iterative step that eliminates the first one obtained immediately before the at least preceding group, determining from the slope function that the removal process has begun, and the material removal process Confirming and detecting that a breakthrough has occurred by detecting that the slope function has fallen below a predetermined value for a predetermined period after it has been determined to have begun. Breakthrough detection method in plasma processing. 37. The method of claim 36, wherein the step of detecting that the magnitude of the slope function is below the predetermined value by the predetermined period of time comprises the step of: A method of detecting breakthrough in a plasma processing process, the value being calculated from the fact that the digital signal data has become lower than the predetermined value for the predetermined period. 38. The method of claim 36, wherein the material removal process comprises a process of removing material supported by a semiconductor wafer. 39. The method of claim 36, including the step of terminating the layer material removal process in response to confirming that a breakthrough has occurred. 40. The method according to claim 36, wherein the breakthrough is not for confirming the magnitude of the electric signal by comparing with the absolute value of the certain reference signal. Detection method.