JPS6393548A - Monitoring system for manufacturing equipment - Google Patents

Abstract

PURPOSE:To detect the fine variation of the processing result with high sensitivity by obtaining the deviation from the management center value which is previously set for each processing result and adding said deviation onto the accumulative deviation value and comparing said deviation value with the upper and lower allowable limit values which are obtained from the number of processing results and the center value variation allowable limit width value. CONSTITUTION:A processing result is input and the deviation is obtained in the first calculator 103, and the accumulation sum is obtained in an accumulator, and the number of input of the result is obtained in a counter 102. The upper and lower allowable limit values of the accumulation sum are obtained in the second calculator 105, and the accumulation sum and the limit value are compared in a comparator 106. When the comparison result is within the both limit values, the accumulation sum and the limit value are renewed to each prescribed value, and when the comparison result is outside the both limit values, the counter 102 and an accumulator 104 are reset, and the circuit state is initialized. With such constitution, even if the variation tendency is not constant, and the average value varies more finely in comparison with the magnitude of the dispersion, the state can be detected with high sensitivity.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention generally relates to a monitoring method for manufacturing equipment, and is particularly suitable for detecting changes in wafer processing conditions from the processing results of semiconductor wafers. This relates to a manufacturing equipment monitoring method.

[Conventional technology]

As a method of grasping changes in manufacturing equipment and managing quality such as machining accuracy of workpieces so that they do not deviate from a certain tolerance range, for example, the method shown in Japanese Patent Application Laid-Open No. 15985-1985 is known. Are known.

In this method, as shown in FIG. 4, data 401 such as processing results of workpieces by manufacturing equipment, for example, processing dimensions xl (t; work start order of work lots), etc., are determined by a commonly used statistical upper control limit (UCL) 403. and the lower control limit (L
CL) 404. Furthermore, the fluctuation trend of the processing results of the workpieces up to the immediately preceding lot is approximately expressed by a regression line 405 and the variation (σm) 406, and the predicted upper limit value 4 of the workpieces to be started next time is calculated.
08 and predicted lower limit value 409 are calculated, and it is managed whether these predictions (direction) deviate from IIcL 403 or LCL 404. If a deviation from control occurs, work processing of the manufacturing equipment After the amount of adjustment of the conditions has been calculated using the data up to the previous loft and the workpiece processing conditions of the manufacturing equipment have been changed, the above method can be repeated in the same way to adjust the condition fluctuations and changes over time of the manufacturing equipment. etc. are managed.

[Problem that the invention seeks to solve]

By the way, when manufacturing semiconductor devices, for example, there are an extremely large number of variable factors that make management difficult, such as materials, environment, and even the state of semiconductor manufacturing equipment. For this reason, wafer processing results such as dimensions after pattern transfer processed by semiconductor manufacturing equipment usually have large variations, and the fluctuations often do not show a constant trend. If there is a minute change, there is a very high possibility that it will lead to an increase in defective characteristics or malfunctions of semiconductor devices.

That is, in order to monitor semiconductor manufacturing equipment, it is necessary to detect changes in the average value, which are minute compared to the magnitude of the variation width, with good sensitivity.

In the case of the above conventional technology, the fluctuation increases in a linear trend,
Or, if it decreases, it is possible to detect small changes in the average value with good sensitivity, but if the fluctuation trend is not necessarily constant, it is suitable for detecting small changes in the average value. It has become something that cannot be said to be the same.

The purpose of the present invention is to prevent fluctuation trends from being necessarily constant.
Moreover, the present invention provides a manufacturing equipment monitoring method that is capable of highly sensitive detection of minute changes in the average value compared to the magnitude of variation.

[Means for solving problems]

The above purpose is basically the processing result of the workpiece.

(Effect) Assuming that the average value of the workpiece processing result xt varies randomly according to a normal distribution with standard deviation σ, and if the average value of the workpiece processing result X increases or decreases by Δ, then the cumulative sum St is the average value t・Δk, or −t・
Δ follows a normal distribution with standard deviation tσ.6 As t increases, the average value increases in proportion to t, such as t・Δ, whereas the standard deviation increases as t・σ, etc.
Since it only increases by the square root of can also be easily detected.

〔Example〕

The present invention will be explained below with reference to FIGS. 1 to 4.

First, the principle of the present invention will be explained. If it is assumed that the processing result xt of the workpiece is randomly distributed according to the average value and a normal distribution with a standard deviation σ, then xt follows a normal distribution with a difference t·σ.

In addition, the average value of the processing result Xt of the workpiece is increased by only Δ
, or decreases, the cumulative sum S, as shown in distribution 209 or distribution 210, follows a normal distribution with a mean value of t・Δ or −t・Δ and a standard deviation of t・σ. There is. As t increases, the average value increases in proportion to t, such as t・Δ, whereas the standard deviation increases only by the square root of t, such as t・σ, so accumulation is not performed. Distribution 201 when
-203, the degree of separation between distributions is improved as the accumulation progresses. Therefore, even changes in the average value that are minute compared to the magnitude of the dispersion can be easily detected.

By the way, when applying the cumulative sum St to actual management, a criterion for determining whether or not a change in the average value has occurred is required, and as an example of this criterion, it is possible to use straight lines 204 and 205. Straight lines 204 and 205 represent the distribution 208 of the cumulative sum St when the average value of the workpiece processing results xt has not changed, and the cumulative sum S when the average value of the workpiece processing results X increases or decreases by only Δ. , the distribution of 209.2
10, and further the distributions 201 and 202.
203, respectively. Therefore, if the cumulative sum St exists between the straight line 204 and 205, it is assumed that no change has occurred in the average value of the workpiece processing result xt, and if the cumulative sum St exists above the straight line 204, then If the average value of the workpiece processing results It is something. Note that straight lines 206 and 207 indicate changes in the center of each of the distributions 209 and 210 with respect to changes in t.

Now, to explain the basic manufacturing equipment monitoring device according to the present invention, FIG. 1 shows its configuration. According to this, the entire system consists of cumulative sum calculation means 100A and fluctuation detection calculation means 100B. In the cumulative sum calculating means 100A, for example, the wafer processing result Xt is input data, and the first calculating unit 103 calculates (Xt-k), and at the same time, the cumulative sum St described above is calculated in the accumulator 104, and at the same time, the processing result is synchronized. By counting the signal P by the counter 102, the number of manpower t for the processing result Xt can be determined. Based on the number of inputs t of the processing result Xt, the upper and lower allowable limit values U of the cumulative sum St
t and Lt are the second computing unit 1 in the fluctuation detection computing means 100B.
05 is obtained as follows.

U, =-・Δ to -t ・・・・・・・・・・・・
fi+Lt”-1・Δ−t・・・・・・・・・
(2+2(11, (21) The upper and lower allowable limit values U t and L t are compared with the cumulative sum St from the accumulator 104 in the comparator 106, and are then processed. The presence or absence of a change in conditions is determined.

That is, when the cumulative sum S1 is larger than the upper allowable limit value Ut, it is determined that the average value of the wafer processing results has increased by more than Δ than the control center value, and the cumulative sum St is larger than the lower allowable limit value Lt. If it is small, it is determined that the average value of the wafer processing results is reduced by Δ or more from the control center value. The cumulative sum St is the upper allowable limit value U and the lower allowable limit value L.
If it is between t and t, it is determined that the average value of the wafer processing results has not changed. These determination results are displayed externally, and if the average value of the processing results fluctuates either upward or downward, a warning to that effect is also displayed. In addition, counter 0
If the number t of inputs from 2 and the cumulative sum St corresponding thereto are recorded, the transition of the cumulative sum SL can be easily known.

As described above, each time a processing result is input, the cumulative sum St and the allowable limit values Ut, Lt are compared, but as long as the comparison result is within the allowable limit values Ut, L, the cumulative sum St and the allowable limit are The values U,, LL are updated in a predetermined manner. However, if the value deviates from the allowable limit values Ut and Lt, a reset request from the comparator 106 is sent via the reset signal generator 101 to reset the counter 102 and the accumulator 104, so that the circuit state is initialized. It has become.

The same process is repeated again from the time when the next process result is input.

Incidentally, if the above processing is performed by a computer, the flow is shown in FIG. 3.

That is, prior to monitoring the manufacturing equipment, initial settings such as setting the management center value, center value variation allowable limit width value Δ, and setting the number of inputs t and cumulative sum SL to zero are performed in step 301. ing. After the initial setting, in step 302, the processing result xt of the wafer of the semiconductor manufacturing equipment is input as the processing result periodic signal P.

is started as an interrupt, but subsequent to process 302, the number t of input data is updated by +1 in process 303. After this, in steps 304 and 305, the deviation (xt h) and the cumulative sum Σ(xt k) are calculated.
Furthermore, depending on the process 306, a permissible limit value for the cumulative sum is determined. In process 307, the cumulative sum is compared with an allowable limit value, and depending on the comparison result, the process branches to one of processes 308, 309, and 311. That is, if the cumulative sum is within the allowable limit value, after the process for displaying no change is performed in process 311, the next process result Xt
It is now ready for input. If it is not within that range, a process is performed to display a warning to that effect in either process 308 or 309, and then the input number t and cumulative sum are cleared in process 310. . After processing 310, preparations are made for inputting the next processing result xtO.

〔Effect of the invention〕

As explained above, in the case of the present invention, even if the fluctuation trend is not necessarily constant and the average value fluctuates minutely from the control center value compared to the magnitude of the dispersion, it can be detected with high sensitivity. There is an effect.

[Brief explanation of the drawing]

Fig. 1 is a diagram showing the configuration of a basic manufacturing equipment monitoring device according to the present invention, Fig. 2 is a diagram for explaining the invention in detail, and Fig. 3 is a diagram showing the configuration of a basic manufacturing equipment monitoring device according to the present invention. FIG. 4 is a diagram illustrating a manufacturing equipment monitoring method according to the prior art. 101... Reset signal generator, 102... Counter, 103... First arithmetic unit, 104... Accumulator, 1
05...Second arithmetic unit, 106...Comparator agent Patent attorney Tadashi Akimoto Figure 1+00A Figure 2 opening, 7th Figure 3

Claims (1)

[Claims] 1. Every time a processing result for a target object is obtained, the deviation of the result from a preset control center value is calculated and added to the cumulative deviation value, and the cumulative deviation value as the addition result is imported. A manufacturing equipment monitoring method in which the results of processing on a target object are monitored while being determined based on the number of processing results and the center value variation allowable limit width value and compared with the lower limit allowable limit value.