JPS63102231A - Mark signal processor using charged particle beam - Google Patents

Abstract

PURPOSE:To reduce the wear of a mask due to beam irradiation, by eliminating back ground level from a signal obtained by detecting a reflected beam, and finding a mark position on the basis of the signal in which noise component is eliminated and smoothing is done. CONSTITUTION:Charged particle beam Bi is radiated on a mask 4, generated reflection beam is received, and a signal including back ground level and noise component is output from a beam detector 5. A signal waveform resulting from elimination of the unnecessary ground component independent of the signal component can be obtained by a subtractor 6. Then the output of the subtractor 6 is input to a filter 9, and the elimination of noise and the smoothing of waveform are performed. After a trial function which best approximates a signal to the signal subjected to smoothing is obtained by a fitting circuit 10, center coodinates of the trial function are calculated by the next peak center determination circuit 11. Thus a very accurate trial function can be determined, so that a mark position is accurately obtained. Further the number of times of averaging by integrating can be extermely reduced.

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial Application Field) The present invention relates to a mark signal processing device using a charged particle beam, and more particularly to a mark signal processing device using a charged particle beam that can accurately determine mark positions. .

(Prior Art) A charged particle beam device is a device that accelerates and focuses charged particles emitted from a charged particle source and scans them two-dimensionally on a sample using a deflector to perform beam drawing or the like.

Examples of the charged particle beam device include an electron beam lithography device, a focused ion beam device, and an electron beam length measurement device. In this type of charged particle beam device, when performing beam irradiation, marks are placed on a sample (e.g., a wafer) or on a stage on which the sample is placed to determine the position of the mark. There is a need to do.

This mark positioning involves scanning a beam multiple times in the vicinity of the mark position, detecting the reflected signal with a detector, determining the peak center of the signal intensity from the integrated average of the obtained signals, and determining the center of the peak of the signal intensity as the mark position.

For example, assume that there is a mark as shown in FIG. 4(a). It is assumed that beam scanning is performed near this mark position and a signal as shown in (b) is obtained. The coordinates of the mark position are determined by setting the coordinates of the peaks PL and P2 of this signal as Xl and X2, respectively.

(Problem to be solved by the invention) In determining the mark position, as mentioned above, the S/
Integration processing is performed to improve the N ratio, but if a charged particle beam (especially an ion beam) is used to scan the sample surface, for example, 10 times, the marks will be severely worn due to the beam irradiation, and the outline of the mark will be distorted. becomes unclear. That is, if calibration and beam conditioning are repeated many times, the center of the mark cannot be positioned due to wear of the mark.

When the marks wear out, the width of the obtained signal becomes wider as shown in FIG. 5(a), and the positioning accuracy deteriorates when only statistical processing using cumulative averaging is performed. For reference, <b) shows the signal when the mark is not worn.

The present invention has been made in view of these points, and
The purpose is to realize a mark position detection signal processing device using a charged particle beam that can accurately determine mark positions.

(Means for Solving the Problems) The present invention that solves the above-mentioned problems includes: a beam detector that detects a reflected beam obtained by scanning the vicinity of a mark position with a charged particle beam; A reduction means for subtracting a background level from the obtained signal output, a filter means for removing noise and smoothing from the signal after the subtraction, and a peak center determining means for determining the center of the peak from the signal passed through the filter means. The present invention is characterized in that the mark position is determined by the output of the peak center determining means.

(Function) The two-step process of subtracting the background level from the detection signal obtained by scanning the beam near the mark position, and performing noise removal and smoothing improves the accuracy of mark detection and determines the peak center. do.

(Example) Hereinafter, an example of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of the present invention. In the figure, 1 is a vacuum space, 2 is a deflector that deflects the charged particle beam Bi, 3 is a stage (sample stage), and 4 is a positioning device placed on the sample placed on the stage 3 or on the stage. It is a mark for A beam detector 5 captures the reflected beam of the charged particle beam 13i irradiated onto the mark 4. The beam detector 5 outputs a signal as shown in FIG.

6 is a subtraction circuit that subtracts the background level from the beam detector 5 output. The output of the beam detector 5 is
Since a DC offset component (background component) based on the dark current of the beam detector 5, etc., which is unrelated to the signal component, is included, the subtraction circuit 6 subtracts this background level. This background level can be changed by adjusting the reference voltage Es using a voltage divider 7 made of a variable resistor. 8 is a soft smoothing circuit that removes noise components contained in the output of the subtraction circuit 6 and smooths the waveform by processing on software; 9 also removes noise components contained in the output of the subtraction circuit 6 and smooths the waveform. This is a hardware filter. The output of the subtraction circuit 6 is sent to the soft smoothing circuit 8. by the switches SWs and SW+'. It is connected to either one of the filters 9. 1
0 compares the signal waveform smoothed by either the soft smoothing circuit 8 or the filter 9 with a trial function created based on fitting range parameters input separately,
Fitting circuit that selects the optimal trial function, 1
Reference numeral 1 denotes a peak center determining circuit that determines the center of a peak based on the optimal trial function output from the fitting circuit 10. The peak center determining circuit 11 outputs a signal (OUTPUT) for determining the position of the mark. The operation of the device configured as described above will be explained as follows.

When the mark 4 is irradiated with the charged particle beam B1, a reflected beam is generated. This reflected beam has different reflection modes between the mark portion and other areas, and contains positional information itself. Upon receiving this reflected beam, the beam detector 5 outputs a signal containing a background component and a noise component as shown in FIG. 2(A). The background component is caused by the dark current of the beam detector 5, and has nothing to do with the signal component. Therefore, the subtraction circuit 6 removes this unnecessary background level. Specifically, a voltage value corresponding to the level of the background component to be subtracted is applied from the voltage divider 7 to the subtraction circuit 6 to remove the DC offset. As a result, a signal waveform from which the background component has been removed is obtained from the subtraction circuit 6, as shown in FIG. 2(b). Note that since the voltage corresponding to the DC offset inputted to the subtraction circuit 6 can be varied according to the background component, only the DC offset can be effectively removed without removing the signal component.

Assuming that the switch SW1.8Wt' is connected to the filter 9 side, the output of the dance reduction circuit 6 is connected to the filter 9 side.
Then, noise is removed and the waveform is smoothed, resulting in a signal waveform as shown in FIG. 2 (c). For the smoothed signal, a fitting circuit 10 inputs the fitting range parameters and uses the least squares method or nonlinear least squares method to convert a part of the signal as shown in f in Fig. 2 (d). The parameters are determined by fitting using a trial function. The part of the data to be used is set by a fitting range parameter that is input separately, so that it can be varied.

Once the trial function that best approximates the signal is found by the fitting circuit 10, the next peak center determining circuit 11 calculates the center coordinates of the trial function as shown in FIG. 2(E). The coordinates obtained in this way are regarded as the position coordinates of the mark. According to the present invention, since the background level and noise are removed before fitting, an extremely accurate trial function can be determined.
Therefore, the mark position can be accurately determined, and the number of times of averaging by integration can be extremely reduced.

In the above description, an example is given in which the hard circuit filter 9 is used for noise removal and smoothing of the detection signal, but the noise removal and smoothing may be performed by software using the soft smoothing circuit 8.

However, in this case, it is necessary to once convert the detected data into digital data using an A/D converter. As the smoothing processing method using software, for example, a quadratic equation adaptation method is used, but other methods may be used.

Note that the quadratic equation adaptation method described above can be replaced with a hardware circuit as shown in FIG. The circuit shown in the figure is a multi-input adder, and the input resistances are Ro and R as shown in the figure.
Assuming that r and R2 are equal, and resistance values of the same sign are equal, the output f is given by the following equation.

f o -α f t-2+ β f =14
− γ f self + β f;+1 + α f 1+
2 times, input f;-z, fl-1, fi,
fi+1, f'+4-2 Detection signals at each position of the hamark (in this case 5 positions), α, β, and γ are coefficients indicating gains, α-Rf/Rz and β-Rf/Rt.

γ=Rf/R.

It is.

(Effects of the Invention) As described in detail above, according to the present invention, the background level is removed from the signal obtained by detecting the reflected beam by beam detection, and the signal is further subjected to noise removal and smoothing. Since we are trying to find the mark position based on
Mark positions can be determined accurately, and the number of integrations can be reduced. Therefore, wear of the mark due to beam irradiation can also be reduced.

[Brief explanation of the drawing]

FIG. 1 is a configuration block diagram showing one embodiment of the present invention, and FIG.
The figures show signal waveforms at various parts, FIG. 3 shows an example of a filter circuit, and FIGS. 4 and 5 show signal waveforms. 1...Vacuum air 2...Deflector 3...Stage 4...Mark 5...Beam detector
6... Subtraction circuit 7... Voltage divider 8... Soft smoothing circuit 9... Filter 10... Fitting circuit 11... Peak center determining circuit SWr, SWt'... Switch patent applicant
Japan Electronics Stock Exchange Ceremony
Patent attorney Fuji Ijima 1 person

Claims (1)

[Claims] a beam detector for detecting a reflected beam obtained by scanning the vicinity of a mark position with a charged particle beam; a subtraction means for subtracting a background level from a signal output obtained by the beam detector; and a peak center determining means for determining the center of the peak from the signal passed through the filter means, and the mark position is determined by the output of the peak center determining means. A mark signal processing device using a charged particle beam characterized by: