JPH079443B2 - Electronic beam pulse width measurement method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Outline] The present invention is an electron beam apparatus that detects the wiring voltage from the intensity of secondary electrons generated by applying an electron beam to the wiring inside an integrated circuit. The present invention provides a method for measuring the pulse width of an electron beam pulse for which the time resolution for measurement is to be determined.

With the progress of integrated circuit technology, the delay time of a logic gate made of gallium arsenide or the like is very short. In order to accurately measure a logical change with an electron beam device, the pulse width of the electron beam, the electron beam The projection timing and electron beam jitter are important issues. The present invention relates to an electron beam pulse width measuring method. In order to accurately measure that an electron beam pulse is indeed short, the electron beam pulse is first applied to a microchannel plate many times to generate an electric current. Convert. Then, the pulse width of the electron beam is calculated by integrating the obtained output current and comparing the value of the DC voltage signal obtained by sample-holding with the measured value for the electron beam pulse of already known pulse width. is there. In this case, by utilizing the fact that the electron beam pulse width and the value of the DC voltage signal have a linear relationship by making the band of the system for extracting the output current of the microchannel plate very small, the DC voltage signal The electron beam pulse width is calculated in reverse from the value. Conventionally, there is a method of calculating the pulse width by deflecting the electron beam pulse at high speed and converting it into a spatial spread.
Below s, an accurate pulse width cannot be measured due to the influence of the deflection voltage jitter of the deflection plate on the electron beam pulse projection timing. According to the present invention, there is an effect that the pulse width can be accurately calculated without being affected by the pulse jitter of the electron beam.

[Industrial application field]

The present invention relates to an electron beam apparatus that applies an electron beam to a wiring inside an integrated circuit and detects the wiring voltage from the intensity of secondary electrons generated from the wiring. In particular, the time resolution is accurately measured in order to accurately measure a voltage change. The present invention relates to a method for measuring the pulse width of an electron beam pulse to be determined.

[Conventional technology]

Along with the development of large-scale integrated circuits, it has become more important to establish advanced fault detection and diagnosis technology that quickly checks whether the internal logical operation of large-scale integrated circuits is normal.

In order to know the logical state inside the integrated circuit or the change of the gate output with respect to the input change, it is necessary to know the voltage state or the voltage change of the wiring connected to the output of each gate. Therefore, electron beam devices that can detect the wiring voltage without contact have been put to practical use. This electron beam device
When an integrated circuit as a sample is mounted on a stage and an electron beam is applied to each wiring of the integrated circuit from the top of the stage, secondary electrons are emitted from the wiring. Therefore, measure the energy distribution of this secondary electron. This device detects wire voltage by means of not only observing the change of each wire voltage with the input voltage, that is, the time change, but also the voltage state of all wires in the field to which the beam is applied, that is, The spatial voltage distribution can be observed.

However, recently, when the IC to be measured is made of gallium arsenide, etc., the speed of logic change of each gate inside the IC becomes very fast, and the rise time or fall time is, for example, about 0.1 ns. Things have come to be manufactured. When measuring such a high-speed voltage change with an electron beam apparatus, it is necessary to have a sufficiently high measurement time resolution as compared with the rise time or fall time in the voltage change of the wiring voltage. There are three conditions that define this time resolution. The first is that the pulse width of the electron beam is short, and the second is the control step of the timing of projecting the electron beam, that is, how fine the electron beam is for a certain synchronization signal. The step of whether or not the timing of applying the light can be changed, and thirdly, the deviation between the clock timing for moving the DUT and the clock timing for moving the electron beam, that is, the jitter is small.

In order to obtain the required time resolution, it is necessary to find a method to accurately measure the pulse width of the electron beam pulse itself. Conventionally, there has been a method of measuring the pulse width by converting the electron beam pulse into a spatial spread by deflecting it at high speed. In this method, as shown in FIG. 2 (a), the electron beam pulse (EB pulse) 20 emitted from the pulse gate is input to the electrostatic deflector and the EB pulse and the pulse voltage applied to the deflector 21 are synchronized. Rising (down)
If the EB pulse is deflected at the part, the force received at the beginning and end of the electron beam will be different. That is, when a voltage for deflecting the EB pulse is given so as to change from a negative voltage to a positive voltage as shown in FIG. 2 (b), the force is received differently at the beginning and end of each EB pulse, It spreads out in the lateral direction and becomes a shape like EB pulse 23. By applying this to the object to be measured, the spatial spread of the electron beam becomes as shown by 24 in FIG. 2 (a), and by measuring this spread, the pulse width of the electron beam can be measured. This conventional method was a method of calculating the pulse width by deflecting the electron beam pulse at high speed and converting it into a spatial spread. However, in this method, if there is a jitter between the EB pulse and the deflection voltage as shown by the one-dot chain line and the broken line in FIG. 2 (b), the jitter corresponding to this is shown in FIG. 2 (a).
EB spatially stretched as indicated by the dashed line and the dashed line
Since the irradiation position of the pulse 24 on the object to be measured spreads, it has a drawback that an accurate pulse width cannot be measured. That is, even if the jitter is 100 ps, the pulse width is 1
If it is about n, there will be no problem, but if the pulse width becomes shorter than 1n, an error will occur, and a small pulse width cannot be measured accurately.

[Problems to be solved by the invention]

In order to eliminate the influence of jitter, which is a drawback of the conventional method, the present invention converts an electron beam pulse into an electric current and integrates this current value to obtain a DC electric signal voltage value. An electron beam pulse width measuring method for calculating a pulse width is provided.

[Means for Solving the Problems] The present invention is directed to a conversion unit that receives an electron beam pulse a number of times and converts it into an electric signal, an integrator that integrates the electric signal obtained from the conversion unit, and the integration circuit. And a sample-hold circuit connected to the output of the electron beam pulse, which is obtained by converting the electron beam pulse into an electric signal by the converting means, and then amplifying a large number of electron pulse signals to be integrated by the integrator. Calculating the electron beam pulse width from the saturation value of the DC voltage signal, and further comparing the saturation value of the DC voltage signal of the integrator with the saturation value of the integrated output for an electron beam pulse of known pulse width. Is characterized by calculating the EB pulse width by
Furthermore, the time constant of the system from the conversion means to immediately before the integrator is characterized by being larger than the EB pulse width.

[Action]

The present invention applies an electron beam pulse to a microchannel plate a number of times to convert it into an electric current, integrates it, and samples and holds it to obtain a saturation value of a DC voltage signal. The electron beam pulse width is calculated by comparison with the saturation measurement for the pulse.

〔Example〕

 Next, an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of a system based on the EB pulse width measuring method according to the present invention.

The EB pulse 10 emitted from the electron gun through the EB pulse gate in the electron beam column is projected onto the micro channel plate MCP11 in the sample chamber. MCP11 has a high gain that converts an electron beam into an electric signal. The MCP 11 may be a scintillator photo lens which is made by illuminating the scintillator. Since a voltage of several KV is applied to the MCP 11 by the DC power supply as shown in the figure and is in a high voltage state, the output current is taken out through the isolation amplifier 12. This current is amplified by the amplifier circuit 13, and its output is input to the gate circuit 14. The gate circuit 14 integrates the output voltage of the amplifier circuit 13 with an integrator 15 at substantially the same timing when the EB pulse 10 is input to the MCP 11.
The gate signal 16 synchronized with the clock timing at which the EB pulse 10 is projected is input to the other one input terminal of the gate circuit 14. Here, the gate circuit 14 is not a logical AND circuit, but an analog switch that determines whether or not the output voltage of the amplifier circuit 13 is transmitted to the output.
16 are going.

When one of the EB pulse 10 projected on the MCP11 is converted into an electric signal via the MCP11, the pulse width of the output voltage of the amplifier circuit 13 is very short.
It is impossible to calculate the pulse width of. Therefore, according to the present invention, a plurality of EB pulses 10 are continuously projected onto the MCP 11,
The output pulse voltage of the amplifier circuit 13 corresponding to each EB pulse is input to the integrator 15 via the gate circuit 14. The result integrated by the integrator 15 is input to the sample and hold circuit 16, and if the peak value of the output voltage of the integrator 15 is sampled and each sample value is held, the sample value is output to the output terminal 17. The output voltage of the amplifier circuit 13 has a waveform as shown on the upper side of FIG. 1 (b), and the waveform of the gate signal 16 which determines whether or not to transmit the output voltage to the integrator 15 is shown in FIG. 1 (b). Is a rectangular wave. The rising timing of this rectangular wave is synchronized with the timing at which the EB pulse 10 is projected from the EB pulse gate. In such an EB pulse width measuring system according to the present invention, the output voltage of the amplifier circuit 13 has a waveform as shown in the upper part of FIG. 1 (b), but the EB pulse 10 is temporarily as shown in FIG. Assuming that the rectangular wave has a height of Ip and a width of tpu, the output voltage of the amplifier circuit 13 corresponding to the rectangular wave is as shown by the V waveform in FIG. The output voltage of the amplifier circuit 13 is an exponential waveform as shown by the V waveform in FIG. 3, and it has a waveform that rises with a time constant τ until the time t reaches tpw and decreases exponentially from the time tpw. Become. Here, the time constant τ is the time constant of the system of the MCP 11, the isolation amplifier 12, and the amplifier circuit 13. The maximum value Va of the waveform V at t = tpw increases as the pulse width tpw increases. That is, the maximum output value Va of the amplifier circuit 13 is Va = AIp (1-e- ^{tpw / τ} ) (1) where Ip is the magnitude of the beam pulse current and tpw is the pulse width of the beam pulse. Where A is a coefficient and this waveform Va is tpw / τ
When is much smaller than 1, that is, when tpw / τ << 1 …… (2), it is approximated as Va = AIp × tpw / τ …… (3). That is, Va is proportional to the beam pulse tpw. Therefore, by detecting the maximum value of the output voltage of the integration circuit 15 via the sample hold circuit 16, EB
Although the pulse width of 10 pulses will be measured, in an actual system, the S / N ratio is so small that an accurate pulse width tpw cannot be measured even if one EB pulse is integrated. Therefore, according to the present invention, a plurality of EB pulses 10 are projected on the MCP 11, and the output voltage output from the amplifier circuit 13 corresponding to the plurality of EB pulses 10 is input to the integrator 15 via the gate circuit 14, and the integrator 15 is supplied. Sample and hold circuit 15 for output saturation voltage of 15
In order to obtain high S / N ratio, When a plurality of such EB pulses 10 are projected, the relationship between the number of EB pulse irradiations and the output voltage of the output terminal 17 of the sample hold circuit 16 is as shown in FIG. The output voltage is 0 when the number of EB pulse irradiations is 0. However, as the number of irradiations is increased, the output voltage is increased by the amplifier 13 as shown in the figure.
Will be saturated to a value proportional to the beak value V of the output voltage. Therefore, the electron beam pulse width tpw can be known by detecting this saturation voltage. Waveform a of Figure 4 than those already represent changes to the EB number of times of pulse irradiation of the sample and hold circuit output voltage in the case of irradiation of the pulse width known electron beam pulse to MCP 11, the output voltage V ₀ which saturation point V ₀ ∝AIp × Tpw ₀ ÷ τ ・ ・ ・ Given by (4). Here, tpw ₀ is the known pulse width of the known electron beam pulse.

On the other hand, the waveform b is a waveform showing the relationship of the output voltage Vx with respect to the number of EB pulse irradiation of unknown electron beam pulses. If the pulse width of the unknown EB pulse is tpw, the saturation voltage Vx
Is also Vx∝AIp × tpw / τ (5). Therefore, by measuring the saturation voltage Vx from the output terminal 17, the EB pulse width tpw of the unknown pulse is given as tpw = Vx / V ₀ × tpw ₀ (6), and the electron beam pulse of known pulse width is given. It is possible to calculate from the measured value V ₀ for.

According to the present invention, the current obtained by receiving the electron beam pulse 10 at the micro channel plate MCP11 in this way is amplified by the amplifier 1
If amplified by 3, the maximum output voltage Va with respect to the electron beam pulse width tpw is proportional to the pulse width tpw when tpw / τ is much smaller than 1 as shown in Fig. 3. And further integrate the amplifier output with the integrator 15, EB
Utilizing the fact that the saturation value when the integrator output voltage is saturated by sufficiently increasing the number of pulse irradiations is proportional to the maximum value Va of the amplifier output voltage. The pulse width tpw is calculated. In the system shown in FIG. 1, the EB pulse 10 to be measured is MCP11.
When projecting on, the number of electrons in the projected EB pulse is, for example, if the peak beam current is 1 nA and the pulse width is 10 ps,
On average, there is only 0.06 per EB pulse, but MC
By using P11, it will be amplified 10 ⁴ to 10 ⁶ times. Isolation amplifier 12 and amplifier circuit 13
After amplifying the output current by, the integrator 15 performs integration, and the sample-and-hold circuit 17 extracts the maximum value of the integrated output voltage as a DC voltage. Furthermore, in order to avoid that the integrated value is affected by the length of the EB pulse generation period, keep the gate signal synchronized with the EB pulse generation,
Only when the output of the amplifier circuit 13 appears
I am trying to type in. Such an electric circuit system MCP1
When the time constant of the system of 1, the isolation amplifier 12, the amplifier circuit 13, and the gate circuit 14 is τ, the present invention makes tpw / τ much smaller than 1 so that a sufficient S / N can be obtained. The output voltage proportional to the pulse width tpw is obtained from the output terminal 17 by irradiating the EB pulse multiple times. Therefore, by comparing the output voltage with the measured value of the known EB pulse, the pulse width tpw of the unknown pulse
Can be calculated.

〔The invention's effect〕

As described above, the present invention integrates the current obtained by amplifying the electron beam by the microchannel plate with respect to a large number of electron beam pulses, so that a very short EB pulse of 100 ps or less is not affected by the jitter. The effect is that the pulse width of can be measured accurately.

[Brief description of drawings]

FIG. 1 is a system configuration diagram based on the electron beam pulse width measuring method of the present invention, FIG. 2 is a principle diagram of a conventional electron beam pulse width measuring method, and FIG. 3 is based on the electron beam pulse width measuring method of the present invention. FIG. 4 is a diagram showing the output voltage waveform of the amplifier circuit 13, and FIG. 4 is a diagram showing the relationship between the output voltage of the sample hold circuit 16 based on the electron beam pulse amplification measurement method of the present invention and the number of EB pulse irradiations. 10,20 …… EB pulse, 11 …… Micro channel plate (MCP) 12 …… Isolation amplifier, 13 …… Amplifier circuit, 14 …… Gate circuit, 15 …… Integrator, 16 …… Sample hold circuit, 17 ...... Output terminal, 21 …… Electrostatic deflector.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Kazuo Okubo 1015 Kamiodanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture, Fujitsu Limited (72) Inventor Toshihiro Ishizuka 1015, Kamikodanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture, Fujitsu Limited

Claims (3)

[Claims] 1. A converting means for injecting an electron beam pulse a large number of times to convert into an electric signal, an integrator for integrating the electric signal obtained from the converting means, and a sample hold connected to an output of the integrating circuit. A circuit, which converts the electron beam pulse into an electric signal by the converting means, and then amplifies a large number of electron pulse signals obtained by the amplification by the integrator to obtain the saturation value of the DC voltage signal, An electron beam pulse width measuring method characterized by calculating an electron beam pulse width. 2. The EB pulse width is calculated by comparing the saturation value of the DC voltage signal of the integrator with the saturation value of the integrated output for an electron beam pulse with a known pulse width. The electron beam pulse width measurement method according to item 1. 3. The electron beam pulse width measuring method according to claim 1, wherein the time constant of the system from the converting means to immediately before the integrator is larger than the EB pulse width.