JPH0227644A - Electron beam device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Table of Contents] Overview Industrial Application Fields Prior Art (Figs. 5 to 10) Problems to be Solved by the Invention (Fig. 11) Examples of Means and Actions for Solving the Problems One embodiment of the present invention (Figs. 1 to 4) Effects of the invention [Summary] Concerning an electron beam device, convergence can be performed even if the noise of the voltage measurement value is increased, and the The purpose of the present invention is to provide an electron beam device that can reduce the number of times of addition and averaging and shorten the measurement time.The purpose of the present invention is to provide an electron beam device that can reduce the number of times of addition and averaging and shorten the measurement time. a means for measuring the voltage at the measurement point by detecting a secondary electron signal emitted from the measurement point; and a means for adjusting the phase of the electron beam pulse so that the voltage at the measurement point matches a predetermined potential. Convergence processing means for performing convergence processing, convergence limit determining means for determining how far to continue the convergence processing, and averaging processing means for calculating accurate timing by averaging convergence values following the convergence processing, In an electron beam device that calculates in a short time the timing at which the voltage at the measurement point matches the predetermined potential, the convergence processing means is defined as a range between an upper limit and a lower limit of the timing at which the potential at the measurement point matches the predetermined potential. convergence processing means for moving and increasing/decreasing the width of the range sandwiched between the upper and lower limits by converging the upper and lower limits in different directions based on the sign of the temporal change in voltage at the measurement point; convergence limit determining means for determining whether the interval between the upper limit and the lower limit falls within a time range defined by the noise of the voltage measurement value and the slope of the voltage waveform at the measurement point; Consists of.

[Industrial application field]

The present invention relates to an electron beam device, and more particularly, it is possible to perform convergence even if the noise of voltage measurement values is increased,
The present invention relates to an electron beam device that can shorten measurement time.

As a method for quickly measuring time differences such as delay times of LSI gates using an electron beam, the rise (fall) timing value of a voltage waveform is measured by a binary search method and a convergent averaging method that uses the slope of the rise part. There is technology. Binary search method was used as a convergence processing means.

The problem with this method is that in order to perform a binary search without error, the voltage measurement error caused by the electron beam must be sufficiently small, especially if the voltage level at which the timing measurement is performed is close to the high/low level of the waveform being measured. The problem is that the measurement time becomes long.

To solve this problem, instead of the conventional convergence processing method using binary search, we have developed a convergence method based on the polarity (rising or falling) of the edge of the waveform for each of the two boundaries of the timing range where the edge exists. A new convergence processing means is provided to search for a timing range in which an edge exists by specifying a determination criterion and a convergence direction.

[Conventional technology]

5 to 10 are diagrams for explaining an example of a conventional electron beam device, FIG. 5 is a block diagram showing the configuration of an example of the conventional example, and FIG. 6 is a diagram showing the configuration of an example of the conventional example. 7 is a waveform diagram for explaining the binary search process of the conventional example, and FIG. 8(a) is a waveform diagram showing when the measured voltage of the conventional example falls within the noise error voltage range. Figure 8(b)
is a waveform diagram showing when the measurement time width of the conventional example falls within the noise error time width, FIG. 9 is a diagram for explaining an example of averaging processing, and FIG. 1O shows the configuration of an example of the conventional example. It is a flowchart.

In these figures, 40 is an LSI, 41 is an electron gun, and 4
2 is an electron lens, 43 is a deflector, 44 is an electron beam pulse, 45 is a deceleration grid, 46 is a stage, 47 is L3
1 driver, 48 a sample chamber, 49 secondary electrons, 50 a control device, 51 a secondary electron detector, 52 a signal processing device,
53 is a display device, 54 is a computer, 55 is a strobe control device, and 56 is a blanking deflector.

The electron beam device used here irradiates the measurement point of the sample with an electron beam pulse that is synchronized with the sample's operation signal with phase control, and measures by detecting the secondary electron signal emitted from the measurement point. This is an electron beam device that measures the voltage at a location and calculates the timing when the voltage at the measurement location matches a predetermined potential. A convergence processing means 1 converges the phase of the convergence process, a convergence limit determining means 2 determines whether the parameters used for the convergence process are within a preset convergence limit range, and When it is within the convergence limit range,
It is provided with an averaging processing means 3 that starts the averaging processing after finishing the convergence processing, and is configured to calculate the timing at which the voltage at the measurement point coincides with a predetermined potential in a short time.

Next, the operating principle of the electron beam device will be explained.

As shown in FIG. 6, the electron beam device irradiates the measurement location of the LSI 40 placed on the stage 46 in the sample chamber 48 with an electron beam pulse 44. Accordingly, the secondary electron detector 51 detects the secondary electrons 49 emitted from the measurement location of the LSI 40.

The LSI 40 is put into an operating state by the LSI driver 47 supplied with test data, and the LSI in this operating state
The electron beam pulse 44, which is designed to measure the voltage at the input and output points of the signal path in the SI 40, is produced by cutting the electron beam generated by the electron gun 41 into pulses by the deflector 43 and passing it through the electron lens. 42 narrows down the beam and irradiates the measurement location of the LSI 40. This electron beam pulse 44 is phase-controlled by an LSI driver 47 (stroboscopic device) and a deflector 43 that receive and receive commands from the control device 50, and is synchronized with the operation signal of the LSI 40, and is delivered to the measurement point of the LSI 40. It is designed to be irradiated.

The energy of secondary electrons 49 emitted from the measurement location of the LSI 40 by the irradiation of the electron beam pulse 44 is analyzed by the deceleration grid 45 . The secondary electrons 49 that have passed through the deceleration grid 45 are detected by a secondary electron detector 51 and sent to a signal processing device 52 . The secondary electronic signal data processed by the signal processing device 52 is supplied to the control device 50, and is further processed together with the test data by the computer 54.
Measurement results for measurement points such as input points and output points of the signal path are output to the display device 53.

Next, a binary search process used in the electron beam apparatus shown in FIG. 7 will be explained. As shown in this figure, the electron beam device uses binary search processing to minimize the number of electron beam pulses irradiated to the measurement location of the LSI, and the position where the voltage matches the threshold level Vs in the shortest time (threshold The purpose is to detect the intersection point (hereinafter also referred to as an edge) between the level Vs and the voltage waveform.

That is, in the time width (time window: T1.T!) expected from simulation data etc., first, phase 31 (time TI) and phase S8 (time Tt)
Let the phase of the midpoint be S, and it is determined whether the edge exists within the range of phases S, ~S, or within the range of phases S, ~S2.

Then, if the edge exists within the range of phase S, -Ss, let the phase of the midpoint between phase Sl and phase S be S, and if the edge exists within the range of phase S, ~S, or ,
It is determined whether the phase exists within the range from 84 to S3. Repeat the same process to perform a binary search for edges.

Next, the principle of operation when the measured voltage in the electron beam device using binary search processing falls within the noise error voltage range and when the measurement time width falls within the noise error time width will be described.

As shown in FIG. 8(a), when the voltage measured by the binary search process falls within the range of a predetermined value vb from the threshold level Vs, the electron beam device performs a binary search process. After the processing is completed, averaging processing is performed.゛Also, Figure 8(b)
As shown in FIG.
Tolerance value vb and voltage waveform slope k (=dV/dt
lv=v, ) and a predetermined time width Vb/
When it comes to exist within the range, the binary search process is terminated and the averaging process is performed. Here, the values such as the predetermined value vb are based on data from simulations and other similar measurement points. It is determined using the following.

Next, an example of the averaging process will be explained using FIG.

As shown in this figure, 1j+, shi1. By repeating the time measurement by processing +Δ1° and calculating the additive average value Tc of the values obtained by the time measurement, the error in the time measurement value due to the variation in the voltage measurement value is reduced.

Next, the conventional processing will be specifically explained using the flowchart shown in FIG.

When the timing measurement process starts, first, step 6
1, initial settings are performed. That is, the first time window (T,) that is the measurement range.

T8) Binary search determination Tb, vb, convergence coefficient, and number of additions M are set. Time window (T+, Tz
) is determined as the time width expected from simulation data, etc., and the time width will vary depending on the accuracy of the simulation used. The binary search determination Tb is for determining whether the time width of the binary search is within the time Tb, and the binary search determination b is for determining whether the measured voltage is at the threshold level VS
This is to determine whether the value is within the allowable value vb. These binary search determinations Tb, vb and convergence coefficients are determined using simulation or data from other similar measurement locations. Further, the number of additions M is defined by noise in measurement, required measurement accuracy, and the like. In this way, when the initial measurement is performed in step 61, the process proceeds to step 62 and the binary search process is started.

In step 62, both ends of the time window are set to initialized values, ie, set to initial values.

=TI, tt =T! As, the time window (t
The voltage (Vi, Vz) at i, tz) is measured. Furthermore, the process proceeds to step 63 and (Vtvs
)x (V--vs)<o is determined. In this step 63, (Vi Vs ) X (VzV
S) < O, the time window (tl, t
It is determined that an edge (the intersection of the threshold level Vs and the voltage waveform) exists within the threshold level Vs and proceeds to step 64, and in step 63 it is determined that (V, −V3)×(■2−■,) is not 0. If so, the process proceeds to step 70 to select the time window (t,).

An alarm indicating that there is no edge within tz) is output, and the timing measurement process is ended.

In step 64, i=2. a=tt, b=tz.

A=V, B=V, and proceed to step 65.
In step 65, t, ++ = (a+b)/2 and 1. The voltage V,+, at ±1 is measured. Then, proceeding to step 66, lv,+.

v, 1<vb, or l t, ++ -ti l
It is determined whether <Tb.

In step 66, lvt ++ v, l<vb
If it is determined that 1 is not directly ++ til<Tb, the process proceeds to step 73 and the binary search process is continued. In step 73, as in step 63, (A Vs ) X (Vi ++ Vs)
It is determined whether or not the value is 0. In step 73 (AVs
) x (Vi +, -VS) <o, it means that an edge exists within the time window (jl+t3), and the process proceeds to step 74 where b=t, +
, ,B=V,+,. Also, Ste Nfu”7
3T: CA-VS ) X (Vi ++ -v,
) <O, if it is determined that the time window (
tz, tx), the process proceeds to step 76 and sets a=j, ++, A=Vi +1. Furthermore, the process proceeds from steps 74 and 76 to step 75, where i=i+l and 1 is added to i, and the process returns to step 65 again.

Step 66 Nioiite, lVL++ Vs I
<Vb, or, if it is determined that -t, l<Tb, that is, the voltage V, +, measured by the binary search process is within the range of the initially set value vb from the threshold level VS. +, -1 in the time width 1 measured by binary search processing
．． When it is determined that Tb exists within the initially set time width Tb, the binary search process is ended and step 6
The process advances to step 7 and the averaging process is started.

In step 67, j=i+1. m=1 is set,
Proceeding to step 68, it is determined whether m>M. If it is determined in step 68 that m>M is not satisfied, the process proceeds to step 71, where tj+,=tJ+,-direct+(V s
VJ + @ -t) / k is set, and 1j +,
The voltage Vj+, is measured. Furthermore, the process proceeds to step 72, where m=m+1 and 1 is added to m, and the process returns to step 68 again.

In step 68, if it is determined that m>M, that is, if it is determined that the number of additions of the additive average is equal to or greater than the initially set value, the process proceeds to step 69, where the additive average value t
c is calculated by the following formula.

The measurement accuracy of the addition average value tc calculated in step 67 is defined by the number of additions M that is initially set. In this way, it is possible to calculate the timing at which the voltage waveform at the measurement point matches a predetermined potential. For example, when measuring the delay time of a gate, the voltage at the input and output points of the gate can reach the threshold level. The matching timings are calculated, and the time difference between them is calculated.

[Problem to be solved by the invention]

However, in conventional electron beam devices, the fifth
The convergence processing means 1 shown in the figure uses a binary search method, and in order to perform the binary search without error, the voltage measurement error caused by the electron beam must be sufficiently small.In particular, the voltage level at which the timing measurement is performed must be There is a problem in that the measurement time becomes long when the waveform is close to the high/low level. There are limits to the timing measurements that can be obtained using binary search, and if a certain amount of noise appears on the waveform, ambiguity will occur.

Specifically, as shown in FIG. 7, the timing at which the edge and the threshold level Vs match is determined by using binary search processing, and as shown in FIG. 1Vi between VsO
When the relationship -Vsl<Vb is established, it is determined that the convergence limit is reached, and the process moves on to the next convergence averaging process.The process of determining 0 or more is a waste of measurement time in a situation where the voltage measurement value by the electron beam contains an error due to noise. It is reasonable to set vb to the same level as the statistical magnitude of noise in the voltage measurement value Vn.

Therefore, as shown in FIG. 11(a), Vs±vb
The binary search ends if there is a measured value within the range of ), the convergence process may not be performed, and Vn<A
It is essential to increase the number of times the secondary electron signals are averaged so as to satisfy /2. In the case of FIG. 11(b), Vn
<A.

/2 (also in this case, binary search may not be able to handle it well), it is necessary to further increase the arithmetic average,
This increases the convergence processing time.

That is, if there is a Vn that is 1/4 or more of the amplitude, there is a possibility that the binary search cannot be processed properly from the beginning.

Therefore, the present invention has developed an electron beam that can converge even if the noise in the voltage measurement value is increased, can reduce the number of times of addition and averaging when detecting secondary electron signals, and can shorten measurement time. The purpose is to provide equipment.

[Means to solve the problem]

In order to achieve the above object, the electron beam device according to the present invention irradiates a measuring point of a sample with an electron beam pulse synchronized with an operation signal of the sample with phase control, and a secondary electron signal emitted from the measuring point. a means for detecting and measuring the voltage at the measurement point; a convergence processing means for converging the phase of the electron beam pulse so that the voltage at the measurement point matches a predetermined potential; and a method for determining how far the convergence processing continues. convergence limit discriminating means for determining the convergence limit, and averaging processing means for determining the accurate timing by averaging the convergence values following the convergence processing, and determining the timing at which the voltage at the measurement point matches the predetermined potential for a short time. In an electron beam device that calculates
The convergence processing means is configured to set the timing at which the potential of the measurement location matches a predetermined potential as a range sandwiched between an upper limit and a lower limit, and to perform convergence in different directions based on the sign of the time change of the voltage at the measurement location with respect to the upper limit. The convergence processing means moves the range sandwiched between the upper limit and the lower limit and increases or decreases the width by setting the lower limit, and the convergence limit determining means is configured so that the interval between the upper limit and the lower limit is noise of the voltage measurement value. and the slope of the voltage waveform at the measurement location.

[Effect]

In the present invention, the convergence processing means gives the timing at which the potential at the measurement location matches a predetermined potential as a range sandwiched between an upper limit and a lower limit, and convergence in different directions based on the sign of the time change of the voltage at the measurement location is set as the upper limit. By moving the range between the upper and lower limits and increasing/decreasing the width, the convergence limit determining means determines that the interval between the upper and lower limits is equal to the noise of the voltage measurement value and the voltage at the measurement point. It is configured to determine whether the time range is within a time range defined by the slope of the waveform.

Therefore, since it becomes possible to make separate determinations for the upper limit and the lower limit, the measurement range can be made wider than in the conventional method.

〔Example〕

Hereinafter, the present invention will be explained based on the drawings.

Figures 1 to 4 are diagrams explaining one embodiment of the electron beam device according to the present invention, Figure 1 is a block diagram showing the configuration of one embodiment, and Figure 2 explains convergence processing of one embodiment. figure, 3rd
Figures (a) and (b) are diagrams illustrating convergence limit determination processing in one embodiment, and FIG. 4 is a flowchart showing the configuration of one embodiment.

In these figures, the same reference numerals as in FIGS. 5 to 10 indicate the same or corresponding parts, and 1a is a convergence processing means, which corresponds to the convergence processing means according to the present invention.

2a is a convergence limit determining means, which corresponds to the convergence limit determining means according to the present invention.

Note that here, as the electron beam device, one having the same configuration as the conventional one according to FIG. 6 can be used.
Since the operating principle is also the same, the explanation will be omitted here. 1st
The convergence processing means 1a shown in the figure gives the timing at which the potential of the measurement point matches a predetermined potential as a range sandwiched between an upper limit and a lower limit, and convergence in different directions (upper limit The convergence limit determining means 2a determines whether the interval between the upper limit and the lower limit is the voltage This is to determine whether the time has fallen within a time range defined by the noise of the measured value and the slope of the voltage waveform at the measurement location.

Next, the principle of operation will be explained.

As shown in Figure 2, the convergence process is performed in a timing range (
Lower limit TI and upper limit ti), and lower limit T and upper limit 1. For each edge, different convergence processing is performed based on the slope of the edge (rising or falling, confirmed in the first measurement). That is, for Tt (lower limit), the measured voltage V
, becomes larger than Vs, increase Tt ++'r+ +, =T, +l ti-7,l/4), t
For i (upper limit), when the measured voltage Vi is smaller than Vs, T, +1 is reduced to L (t, +, = t mu ti - T,
l/4), and by always maintaining T, <t, the range in which edges are likely to exist is sandwiched.

As shown in Fig. 3(a), the convergence limit determination is performed by first
This is done depending on whether Tt ti l <Ta is satisfied. That is, the length of the measurement range l'r, -t, l
When becomes larger than Ta, the convergence determination is terminated because it has reached its limit. If the upper limit is T, the measured voltage V, > Vs, shift it to the right, and the lower limit is ti, the measured voltage V, <V
In the case of s, it is pinched by moving it to the left. Then, when the voltage approaches Vs, the convergence processing means 1a ends.

In addition, in Fig. 2, when the slope is positive (rising), S=O, and when the slope is negative (falling), S=
It is 1. It is desirable that Ta be equal to the time ambiguity ΔT due to noise Vn in the voltage measurement value, and Ta < ΔT.
In order to reduce unnecessary convergence operations when is specified, the third
The convergence limit determination shown in FIG. 3(b) is also performed. In FIG. 3(b), when the measured voltage V, <Vs and the measured voltage Vt>Vs occur simultaneously, the convergence determination ends. That is, the third
If the conditions shown in FIG. 3(a) or FIG. 3(b) are satisfied, the convergence determination can be completed.

As shown in Figures 2 and 3 (a) and (b), even if the noise Vn of the voltage measurement value is nearly half the amplitude of the measured waveform (when Vs is in the center), convergence operation cannot be performed. can.

Next, a more specific explanation will be given using FIG. 4.

Note that the portion surrounded by a broken line here is a characteristic portion of the present invention, and the operation of the averaging processing means 3 after the convergence limit determination means 2 (convergence limit determination) is the same as the conventional one.

When the timing measurement process starts, first step 11 is performed.
Initial settings are performed in . That is, the first time window ('r, ) that is the measurement range.

tI), the convergence determination Ta, the convergence coefficient, and the number of averaging times M are set. Next, the process proceeds to step 12, where TI, t
The voltages at I (Vl, Vl) are measured respectively. Then, proceed to step 13 and (Vl-Vs)
It is determined whether X (v, -Vs)<o. This determines whether V and V are on the opposite side with respect to Vs. In this step 13, (Vl V s
) X (vl V s ) <0, the process proceeds to step 14. Also, in step 13 (Vl
-Vs)

In step 14, if V, -v, <Q (rising), S=O is set, and if Vl vt<o (falling), S=O is set. Then proceed to step 15,
I=i=1, the width of the time window B=t, -T, is set, and the process proceeds to step 16. In step 16, (-
1)'・Vl<(-1-)' ・If Vs, then T,
++=TI +B/4 and B/4 are added to TI, and the process proceeds to step 17. Similarly, (1)' ・V+≧(1)
' ・If Vs, T++, = TI-B/4 and - to T1
B/4 is added and the process proceeds to step 17, (1)"
・vt>(-1)" ・If Vs, tl +1-tt
-B/4 is added to B/4 and t, and the process proceeds to step 17. If (-1)'-Vi≦(-1)' ・Vs, then 1
B/4 is added to ゜ + 1 = B/4 and t, and the process proceeds to step 17. In step 17, t, +, -'r
It is determined whether r ++ <Ta. If it is determined in step 17 that Ll + (Tl +1 <Ta), the process proceeds to step 21, and the operations after step 21 (additional averaging process) are the same as before.In step 17, t4
If it is determined that +, -T, +, <Ta does not hold, the process proceeds to step 18 and the voltages (v, +, , vt +,) at T'+ ++ and tt ++ are measured, respectively.

Then, proceed to step 19, and (-1)3 ・V, +,
>(-i)" ・Vs and (-1)' ・Vl +,
<(-i)" ・It is determined whether or not Vs. In this step 19, (-1)'・1+1>(-1)'
・Vs and (1)” -Vl +, <(-1)!
- If it is determined that the voltage is Vs, the process proceeds to step 21.

The operations after step 21 are the same as before. Step 1
9 (-1)" ・V, +, >(-1)' ・Vs
and (-1)" -vt +, < (-1)3 ・V
If it is determined that it is not s, the process proceeds to step 20 and I=1.
1 is added to +1 and I, 1 is added to i=i+1, and the process returns to step 16 again.

That is, in the above embodiment, since the upper limit and the lower limit can be determined separately, the measurement range can be made wider than in the conventional method. Therefore, convergence can be achieved even if the noise of the voltage measurement value is increased, and the number of additions when detecting secondary electron signals can be reduced to about 1/4 of that of the conventional method (the number of voltage measurements is smaller than that of the conventional method). (about twice as much as that required), and the measurement time can be shortened.

〔Effect of the invention〕

According to the present invention, it is possible to perform convergence even if the noise of the voltage measurement value is increased, and the number of times of addition and averaging when detecting secondary electron signals can be reduced, and the measurement time can be shortened. be.

[Brief explanation of the drawing]

1 to 4 are diagrams explaining one embodiment of an electron beam device according to the present invention, FIG. 1 is a block diagram showing the configuration of one embodiment, and FIG. 2 is a convergence process of one embodiment. FIG. 3 is a diagram explaining the convergence limit determination process of one embodiment, FIG. 4 is a flowchart showing the configuration of one embodiment, and FIGS. 5 to 10 are examples of conventional electron beam equipment. FIG. 5 is a block diagram showing the configuration of the conventional example, FIG. 6 is a schematic diagram of the device showing the configuration of the conventional example, and FIG. 7 is a waveform diagram explaining the binary search process of the conventional example. Figure 8 (a) is a waveform diagram showing when the measured voltage of the conventional example falls within the noise error voltage range, and Figure 8 (b) shows the waveform diagram when the measurement time width of the conventional example falls within the noise error time width. 9 is a diagram for explaining an example of conventional averaging processing. FIG. 10 is a flowchart showing the configuration of the conventional example.
The figure is a diagram explaining the problems of the conventional example. 1a... Convergence processing means, 2a... Convergence limit determining means, 3... Addition averaging processing means. ← 2) Illusion of the 100,000-day disaster. Figure 2 Recommendation of inferiority to ti

Claims (1)

[Claims] A measurement point of a sample is irradiated with an electron beam pulse synchronized with an operation signal of the sample in a controlled manner, and a secondary electron signal emitted from the measurement point is detected to detect the measurement point. a convergence processing means for converging the phase of the electron beam pulse so that the voltage at the measurement point matches a predetermined potential; and a convergence limit determining means for determining how far to continue the convergence processing. and an averaging processing means for calculating accurate timing by averaging the convergence values following the convergence processing, and for calculating the timing at which the voltage at the measurement point coincides with the predetermined potential in a short time. , the convergence means is given as a range between an upper limit and a lower limit, the timing at which the potential of the measurement point matches a predetermined potential, and convergence in different directions based on the sign of the time change of the voltage of the measurement point is set to the upper limit. The convergence processing means is configured to move and increase/decrease the width of the range sandwiched between the upper limit and the lower limit by causing the lower limit to move, and the convergence limit determining means is configured such that the interval between the upper limit and the lower limit is equal to noise in the voltage measurement value. 1. An electron beam apparatus comprising a convergence limit determining means for determining whether or not the time range is within a time range defined by the slope of the voltage waveform at the measurement location.