JPS60100448A - Photovoltage image forming device - Google Patents

Abstract

PURPOSE:To improve the decrease of the resolution due to the increase in the diffusing length and the increase of a background signal by secondarily differentiating a video signal obtained on the basis of a signal generated from a semiconductor sample by emitting an optical beam to the sample, and forming a scanning image with the two differentiated signals as an intensity modulation signal. CONSTITUTION:When a semiconductor sample 1 to be observed is interposed in a sandwich state between a transparent electrode 3 and a metal electrode 4 through a transparent spacer 2 and an optical beam 5 is emitted through the electrode 3, a photovoltage is generated between the front surface and the back surface of the sample 1. When the beam 5 is pulsated, the photovoltage becomes a pulsating state. Accordingly, even if the electrode is not formed directly on the surface of the semiconductor, the photovoltage can be detected by the electrode 3 through an electric capacity formed by the spacer. The detected photovoltage is introduced to a detecting amplifier 6, and supplied to a cathode ray tube 7 for forming an image with the voltage as a DC signal. It becomes the output of the detecting amplifier 6. So-called video signal V is secondarily differentiated, and the secondary differentiated signal is supplied as an intensity modulation signal Z to a CRT7 through an intensity modulation amplifier 8.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to an apparatus for forming a scanning image using a photovoltage generated by radiation of light M as a signal.

[Background of the invention]

When forming a flying spot scanning image in a particle beam readout device or a semiconductor property measurement device, improving resolution is one of the important items. The basic concept of irradiating a point-shaped primary beam onto a sample and obtaining a scanning image using signals generated from the sample has been conventionally used in scanning electron microscopes (hereinafter abbreviated as SEM). The electron beam is realized using secondary electrons emitted from the sample as the signal source (see:
α rZeitschrift far Technisc
her PhysikJ.

1935, No. 11. , p. 470)
. As is well known, in SEM, the resolution is determined by the diameter of the electron beam as the primary beam, so several stages of electron lenses are used to reduce the diameter of the primary electron beam by several degrees (usually is used by several hundred people). Therefore, unless the sample is an electrical insulator, the resolution is not limited by the properties of the sample.

By the way, in the method of obtaining a scanning image using a photovoltage as a signal source, which is the object of the present invention, unless the diameter of the light beam as a primary beam is extremely large, the resolution is determined by the photovoltage generation rather than the diameter of the light beam. is determined by the diffusion length of minority carriers in the semiconductor that contributes to .

For example, even if the diameter of the light beam is made infinitely small, the resolution does not increase; instead, the resolution is determined by the range in which the minority carriers excited by the light are distributed. Now, if we consider the minority carrier to be a hole and express this diffusion length as R, the effective diameter is 2L.

This means that the measurement is made using the probe. If the diameter of the light beam is r, the diameter of the light beam is 2r, but in many cases it is 2r.
r << 2 L p, and the optical beam diameter is no longer an important factor with respect to resolution.

Here, the concept of the scanned image in the present invention is the SEM
It is fundamentally different from that of the scanned image.

The fact that a scanning image using a photovoltage depends on the diffusion length causes a phenomenon that is very different from that of the SEM in another way. In other words, if it changes depending on the sample,
Alternatively, if the LP varies from place to place within the same sample, not only will the resolution change, but the absolute value of the signal will also change significantly.

On the other hand, in SEM, the average value of the signal (secondary electrons) does not change significantly unless the current of the electron beam as the -order beam changes, and as a matter of course, when obtaining a scanned image, the signal There is no need to change the amplifier gain significantly.

The reason why the increase or decrease of the signal when using a photovoltage depends on the signal is as follows. That is, when the light intensity is constant, the minority carrier density ΔP that occurs is equal to the minority carrier lifetime τ1
It is well known that it depends on

A P, = K, □1 ...... given by (1). 1 is simply a proportionality constant.

The photovoltage was 6V. If the other constant of proportionality is B, then A V0= B A P ・-=-(2)= B −K
, ,・τ1 ......(2') is given. On the other hand, if LP is the number of diffusers, then LP=D, τ7... (3) Since it is given by the following equation (2), from (3), the photovoltage Δ■o is It can be seen that it is proportional to the square of the length LP.The fact that the photovoltage is proportional to the square of the diffusion length L2 is extremely important in terms of knowing the change in LP (i.e., the change in lifetime τ) in the same sample. This is an effective phenomenon.

By the way, in particle beam recording, depending on the location of the wafer,
When the LP changes, the brightness of the background of the image (portions that are not drawn, such as characters) changes depending on the location, and in extreme cases, the brightness increases locally on the cathode ray tube used to obtain the image. , which will cause halation on the photo. At the same time, the resolution is significantly reduced in that part, so reading out the record in the long part of L becomes impossible due to the above two reasons. If the gain of the signal amplifier is lowered in an attempt to prevent halation, the brightness of other parts becomes insufficient, making it impossible to form an image.

As a countermeasure against the extreme bias in the signal strength in some parts, so-called mouth'f CA with a logarithmic gain is used.
og) It is possible to use an amplifier to prevent halation. However, it is clear that this is not a fundamental solution. This is because, on the one hand, there is a demand for creating images with high contrast using minute signal differences, but if a logarithm amplifier is used, it becomes impossible to uniformly amplify minute signal differences.

In this regard, in SEM, the signal amount is determined by the electron beam intensity, so the brightness of the bank ground will not change drastically depending on the location as long as there is no change in the electron beam intensity.

Furthermore, the case of detecting the uniformity of the pn junction (a state in which there is no junction characteristic partially) is considered important. In this case, the formula (4
), the purpose is to detect that the proportionality constant B changes depending on the location.

However, if L varies greatly depending on location, even if B changes, the background change will be so large that it will be impossible to extract only the change in B.

Regarding the case of detecting a change in B, the results of the study are shown below using numerical values. FIG. 1 shows a sample in which the constant of proportionality is 80 over a range of 6ff at the longitudinal center of the sample with length Q, width W, and thickness t. And ff=
100mm, fflA=1m111, B0
=0. Is it possible to accurately detect the state of the XB?
FIG. 2 shows the light beam as a function of the position X0. In Figure 2, the vertical axis shows the signal voltage,
It is specified by the sample size Qwt, the luminous flux φ, and the proportionality constant B.

According to Figure 2, as already mentioned, it is clearly shown that as the LP becomes longer, the signal voltage increases, and at the same time, it is clearly shown that the resolution decreases and the Bo part becomes indistinguishable. has been done. L, ==0.1 mm
In the case of , it is clearly shown that Bo is smaller than B. On the other hand, when Lp is 10mm,
Not only is it impossible to identify Bo, but the signal strength is 10
This shows that the fluctuation is increased by a factor of 1, which cannot be handled by a normal image forming apparatus.

[Purpose of the invention]

The present invention has two drawbacks as mentioned above, namely: 1) L
, and 2) an increase in background signal due to an increase in the diffusion length LP.

[Summary of the invention]

In order to achieve the above object, in the present invention, a semiconductor sample is irradiated with a light beam, a video signal obtained based on a signal generated from the sample is subjected to second-order differentiation, and the second-order differential signal is converted into a luminance modulation signal. It is configured to be used as a scanner to form a scanned image.

[Embodiments of the invention]

The basic principle of the present invention will be explained below.

In image formation using photovoltage generated on a semiconductor sample by light irradiation, as mentioned above with an example,
There is a peculiarity in that the background signal increases as the resolution decreases. However, as a result of studies conducted by the present inventors, it has become clear that this obstacle can be almost completely removed by second-order differential operation. That is, it has been found that the second-order differential operation provides unprecedented advantages in image formation using photovoltage and the like.

FIG. 3 shows the result of second-order differentiation of the signal shown in FIG. 2 (horizontal and vertical axes are the same as in FIG. 2). ■1. .

In the case of ``10+n+++, it was impossible to detect the Bo part in Fig. 2, but the Bo part cannot be clearly shown in Fig. 3.Moreover, the magnitude of the signal is L,
= The range gfJt' of 10 mm to 0.1 mm is more than double as far as the overhanging portion is concerned, and it can be said that the change of 10' times shown in FIG. 2 has been almost completely eliminated. Note that image formation usually increases brightness with a positive signal, so according to the results in Figure 3, when LP is short, the signal decreases in the center, and it appears that L is /IX smaller. The image shows a tendency to darken.

As mentioned above, in image formation using photovoltage, etc., quadratic differential operation has a great effect in terms of improving image quality, and this effect is due to the special characteristics of photovoltaic force. , was discovered for the first time by the present inventors.

FIG. 4 is a diagram illustrating an embodiment of the present invention. A semiconductor sample 1 to be observed is sandwiched between a transparent electrode 3 and a metal electrode 4 with a transparent spacer 2 in between. When a light beam 5 is irradiated through the transparent electrode 3 in this state, a photovoltage is generated between the front and back surfaces of the semiconductor sample 1. When the light beam 5 is pulsed, the photovoltage also becomes pulsed, so that the photovoltage can be detected by the transparent electrode 3 via the electrical capacitance formed by the spacer without forming an electrode directly on the semiconductor surface.

Since the detected photovoltage is in the form of an alternating current, it is desirable to rectify and detect it.The detected voltage is introduced into the detection amplifier 6, where it is detected and rectified and converted into a direct current signal, which is transmitted to the cathode ray tube ( (abbreviated as CRT) 7. Normally,
This is done by introducing the output of the detection amplifier 6 into the brightness modulation amplifier 8, but in the present invention, the so-called video signal V, which is the output of the detection amplifier 6, is subjected to second-order differentiation by a circuit to be described later.
This second-order differential signal is supplied to the CRT 7 via a brightness modulation amplifier 8 as a brightness modulation signal 2.

The detection amplifier 6 is usually a synchronous detection amplifier for the purpose of improving the signal-to-noise ratio of the signal and utilizing phase information of the optical voltage as a signal. That is, it detects and amplifies a signal in synchronization with the pulse frequency of the light beam 5, and an example of this type of amplifier is a lock-in amplifier.

On the other hand, although it goes without saying that a laser beam or the like may be used as the light beam, a case where a CRT 9 is used as the light source is shown here. A CRT is used as a light source because it is easy to pulse or launch a light beam. In other words, pulsing the light beam is
It is possible to easily supply a signal of an appropriate frequency from the oscillator 11 to the brightness modulation amplifier 10 of the CRT 9.

Therefore, if the signal from the oscillator 11 is supplied to the detection amplifier 6, the optical voltage signal can be synchronously detected in synchronization with the frequency of the optical beam 5.

Now, in order to deflect and scan the electron beams of CRT9 and CRT7, it is necessary to
It is only necessary to supply appropriate deflection signals to the deflection coils 13, 13' and 14, 14' of T. In this way, a bright spot that moves in synchronization with the light beam can be formed on the screen of the CRT 7, and a scanned image by the photovoltage can be formed with high resolution on the screen of the CRT 7 using the supplied photovoltage signal. Note that the image is recorded by the camera 15. In addition, X in the figure.

Y indicates an X-direction deflection signal and a Y-direction deflection signal, respectively.

FIG. 5 shows a specific example of a circuit for obtaining the above-mentioned second-order differential signal. In the present invention, since the signal is digitally processed, the video signal Vt! as an analog quantity! :Converted into a digital quantity using the A/D converter 21. When digitizing and processing, the A/D converter is driven by the signal from the light beam deflection signal 22 in order to perform chronological addressing. The deflection signal is generated by integrating the number of pulses from the oscillator 22, and therefore, here.

A 515-decimal counter 23 is used. When the signal of this counter is converted into an analog quantity by the D/A converter 24 and outputted, this becomes an X-direction deflection signal. 515 counter 23
After counting 515 pulses, it returns to zero and starts counting again, so the X-direction deflection signal X has a sawtooth waveform.6Y
The direction is scanned in a stepwise manner every 515 pulses (one scan in the X direction). Therefore, the 515-base counter 23' is driven in response to the signal from the final stage of the counter 23 for X deflection. , when forming the X-direction deflection signal X, the D/A
A Y-direction deflection signal Y is formed by a converter 24'. signal X,
Both Y are supplied to the CRT drive circuit shown in FIG.

Now, the A/D is synchronized with the deflection signal using the deflection signal clock.
The video signal A/D converted by the converter 21 is sequentially sent to shift registers 25, 25' and 25'' connected in series. , 515 bits will be explained.

It is clear that it may be freely changed depending on the number of decomposition points of the deflection signal. The so-called depth of each bit of the shift register needs to be changed according to the gradation of the video signal, but here, the depth is 9 to 100 degrees in order to accommodate a 100 times signal change.
bit (divided into 512). - Out of the 515 points of time series signals per scan, 3 points in particular are extracted and these bits are set to 25a.

25b, 25c, 25' a, 25' b, 25' c
.

Indicated by 25'a, 25'b, and 25'c. The digitized video signal is stored in the shift register from the A/D converter 21 with a circuit delay.
6 to each shift register 25.25'.

25".

Now, when the first scan in the X direction is completed, the first signal enters bit 25a, then 25b, 25c, etc.
. . . are all stored in the shift register 25.

Then, when the second scan begins, the signal is transferred from bit 25a to shift register 25'. At the completion of the second scan, the first signal of the first scan is in the 25'a bit. Thus, when the third scan is completed, the first
The first signal of the scan is in bit 25'a, and the second
The first signal of the scan is contained in 25'a bits, and the first signal of the third scan is contained in 258 bits.

If we look at this state from a different perspective, we can see that three points of signals are lined up in the X direction and three points in the Y direction, surrounding the 25'b bit.

As is well known in the field of image processing, the second derivative of a signal is /! :I: The choice is made by multiplying the left and right signals of point I by a predetermined interval coefficient and combining the signals. That is, in the X direction, a multiplier 27a,
27b and 27c are used to multiply by a predetermined coefficient, and the adder 28 adds up the results.

Since the second derivative is also necessary in the Y direction, 2
Bit 25b above and below bit 5'b.

Using the signal of bit 25'b, multiplier 29°29'
, 29', and the adder 28 totals the results.

Thus, the second-order differentiation is completed for the second point of the second scan, and the output of the adder 28 can be impedance-converted by the amplifier 29 and sent to the brightness modulation circuit of the CRT for forming one image.

When the fourth scan starts, the differential operation moves to the second fish of the second scan, and differential signals are successively formed one after another.

In this case, it is clear that the first scanning signal and the last 515th scanning signal cannot be differentiated. Similarly, in each scan, the starting point and final point cannot be differentiated.

However, since both are 21515, from the image,
Even if this amount of information is lost, there is no particular problem.

〔Effect of the invention〕

As described in detail above, according to the present invention, it is possible to form a scanning image with a resolution higher than that of the conventional method, and its effects are great in practical use. In addition, in the above-mentioned embodiment, among the signals generated from a semiconductor sample by light irradiation, a signal caused by a photovoltage was explained in particular, but the present invention is not limited to this, and can be applied to other signals including photocurrent and photoconductive effects. It goes without saying that there is.

[Brief explanation of drawings]

FIG. 1 is a diagram explaining a sample in which the generated voltage varies depending on the location, FIG. 2 is a characteristic diagram showing the relationship between the signal voltage (normalized) and the position of the light beam on the sample, and FIG. Second
Characteristic diagram showing the result of second-order differentiation of the signal shown in the figure, No. 4
FIG. 5 is a diagram illustrating an embodiment of the present invention, and FIG. 5 is a diagram illustrating a specific example of a circuit for obtaining a second-order differential signal. 1...Semiconductor sample, 5...Light beam, 6...Detection amplifier, Fig. 1 or Q/Fig. 2 (χl-)/z)//71 No. 4 Fig. 5

Claims (1)

[Claims] A semiconductor sample is irradiated with a light beam, a video signal obtained based on a photovoltage signal generated from the sample is subjected to second-order differentiation, and the second-order differential signal is used as a brightness modulation signal to form a scanning image. 1. A photoelectric image forming apparatus characterized by comprising: