JPH0232579B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides an imaging device using ultrasonic energy;
Particularly related to ultrasound microscopy.

In recent years, it has become possible to generate and detect ultra-high frequency sound waves of up to 1 GHz, making it possible to realize sound wavelengths of approximately 1 μm underwater, and as a result, it has become possible to obtain high-resolution sonic imaging devices with a resolution of 1 μm. Natsuta (A of Mr. RA Lemon and Mr. CF Kuwait)
IEEE entitled Scanning Aconstic Microscope
cat.No.73CH14829SU PP423−426, literature published in 1973). That is, an ultrasonic beam focused to a diameter of about 1 μm is applied to the sample using a concave lens, and the reflected waves determined by the microscopic elastic properties of the sample are detected using the same lens. By mechanically scanning the sample in two dimensions and displaying the reflected intensity as a brightness signal on a cathode ray tube in synchronization with the mechanical scanning, it is possible to enlarge the microscopic area of the sample.

By the way, the resolution of such a device includes depth resolution Δρ in the ultrasound propagation direction and azimuth resolution Δγ in a plane perpendicular to the ultrasound propagation direction, both of which are dependent on the wavelength λ of the ultrasound used and the lens. It is determined by the F number representing the brightness of , and is given by Δγ=λ・F (1) Δρ=2λ・F ² (2).

The F number of the lens that can be made is about 0.7, so if you use 1GHz ultrasonic waves, you can make it underwater (1500m/
s), Δγ ~ 1 μm and Δρ ~ 1.5 μm.

By the way, even better depth resolution is required for ICs and LSIs, which are the most important imaging targets for ultrasound microscopes. This is because, as is well known, in ICs, the pattern in the depth direction is often finer than the planar pattern. In fact, typical ICs have a multilayer structure of 1 μm to 3 μm, but it is difficult to focus on the inside of the IC from the surface and observe each of these layers independently and non-destructively. This would be impossible with a depth resolution of 2 μm. This is because the sound speed of metals such as silicon and aluminum, which are IC materials, is higher than that of water, so even if 1 GHz ultrasonic waves are used, the depth resolution is only 4 to 10 μm. In view of these circumstances, an interferometry method has been proposed as a method for greatly improving the depth resolution (Japanese Patent Application No. 58707-1983 (Japanese Patent Application Laid-Open No. 56-155847) Ultrasonic Imaging Apparatus). That is, using the same principle as an optical microscope, a reference sound wave that does not vary depending on the type of sample or mechanical scanning is generated acoustically or electrically, and this is caused to interfere with the sound waves reflected from the sample. In this way, the depth resolution becomes Δρ≒λ/5 (3) and the method described above (usually called the amplitude method)
This is a 5x improvement.

Using 1 GHz ultrasound, conventional depth resolution was only about 1.5 μm in water and 8.4 μm in silicon (8400 m/s), but according to the present invention, the depth resolution is 0.3 μm underwater and 1.7 μm in silicon (8400 m/s). The depth resolution can be improved to a high depth resolution of μm, making it possible for the first time to observe the multilayer structure of an IC at different depths.

The present invention further investigated the ability of this interferometry and found that the ability of the interferometry becomes even more effective by shifting the ultrasonic frequency used by Δf around a certain f. be. The interferometry method will be explained below, and its capabilities and problems will be discussed.

FIG. 1 is a diagram showing a schematic configuration of a probe system according to a conventional method for obtaining a reflected signal from a sample. One end surface 21 of the acoustic lens 20 (for example, a cylindrical crystal such as sapphire or quartz glass) is an optically polished flat surface, and the other end surface has a concave hemispherical hole 30 formed therein. Piezoelectric thin film 1 deposited on end face 21
An RF pulsed sound wave of a plane wave is radiated into the lens 20 by the RF pulsed electrical signal applied to the lens 20 .
This plane sound wave is focused onto a sample 50 at a predetermined focus by a positive acoustic lens formed at the interface between the hemispherical hole 30 and a medium 40 (generally water).

The ultrasonic waves reflected and scattered by the sample 50 are
Sound is collected by the same lens, converted into a plane wave, propagated through the lens 20, and finally converted into an RF electric signal by the piezoelectric thin film 10. This RF electrical signal is detected by a diode and converted into a video signal, which is then used as the input signal for the cathode ray tube.

FIG. 2a shows a detection signal in the video region when an RF pulse electric signal with a certain repetition period t _R is applied in such a conventional configuration. Here, the horizontal axis represents the time axis and the axis represents the signal strength. A shows the applied RF pulse, B shows the reflected signal from the lens interface, or C shows the reflected signal from the sample.

In conventional imaging devices, in order to distinguish this desired reflected signal C from B, the duration time _td of the applied pulse is set as short as possible so that the C and B signals do not overlap, and only the C signal is time gated. (Figure 2c)
A configuration is adopted in which samples are taken out and sampled.

Figure 2d shows an example of the interferometry method, in which the duration of the RF pulse t d is required to cause the reflected ultrasound signal C from the sample to interfere with the reflected ultrasound signal B from the lens-water interface _. This is achieved by making it longer than in the conventional example.

To explain in detail, the reflected ultrasonic signal C from the sample is greater than the reflected signal B from the interface between the lens and water.
Time for round trip propagation in water between lens and sample t _s =
2Z/V _w (where Z is the distance between the lens and the sample, V _w
is the speed of sound in water), so the PF
If the pulse duration t _d is increased to t _d >2Z/V _w (3), the two reflected signals will overlap. By extracting these overlapping time domain signals using a time gate, interference between the two reflected signals can be detected.

The reflected signal B from the interface between the lens and water is V _B (t) = Asin2πft t ₀ < t < t ₀ + t _d (4) where f is the ultrasonic frequency used t ₀ =2L/V _L , and L is the ultrasonic frequency of the lens. Assuming that the length and V _L are the sound speed in the lens material, the reflected ultrasonic signal C from the sample is Vc (t) = Bsin2πf (t +2Z/V _w ) t ₀ <t _s <t<t ₀ +t _s +t _d (5) Therefore _, under the condition of _equation ( ₃ ) _above , the two signals _overlap. ) (dashed line area in Figure 2d), and the diode 2
When multiplicative detection is performed, V(t)=A ² +B ² +2ABcos(2πf2Z/V _w ) (7) in the video domain. Using the time domain signal in equation (6), 2πf・2Z/V _w = 2πZ/(λ/2) (8) From (8), if the distance Z between the lens and the sample is changed, λ/2
The detection signal is modulated with the period. Assuming a modulation degree of 50% at which the two reflected waves interfere most, that is, a state in which the amplitudes A and B of the two reflected waves are equal, equation (7) is: V(t) = 2A ² (1 + cos θ) ... (7 ') However, θ=2πZ/(λ/2). If we normalize this so that the upper limit value of V(t) is 1 and the lower limit value is 0, we get V(t) = (1 + cosθ)/2...(7'').The amplitude is -3 dB from the maximum value. If the depth resolution is expressed by the width reduction (Rayleigh standard), then (1+cosθ ₀ )/2=1/√2, so cosθ ₀ =√2−1, and solving this gives θ ₀ = 65.5°. = 0.36πrad. That is, 2πZ/(λ/2) = 0.36π, so Z = 0.091λ≒λ/10 is obtained.The -3dB width depth resolution is -3dB on the plus side from the maximum point.
Since the width is from the point to -3 dB on the negative side, it is twice the above value, that is, 0.182λ≈λ/5.

In other words, it becomes possible to detect the unevenness of the sample and the pattern of the internal layer structure at approximately λ/5 (modulation depth of 50%). This λ/5 is the depth resolution λ in the conventional method.
(when F = 0.7), and by using interferometry, it can be improved by a factor of five.

By the way, using an F-number lens system that is the amplitude method, the intensity of the reflected signal C from the sample can be changed by changing the distance Z between the sample and the lens (0 at the focal point, negative when it is closer than the focal length). When I measured it, I found that Figure 3a
It will look like this. The width marked as Δρ in the figure, which causes periodic changes and corresponds to −3 dB with respect to the peak value, defines the depth resolution. If the same measurement is performed by interferometry, the result will be as shown in Fig. 3b. The pattern changes to a λ/2 period pattern, and its envelope is the same as in the amplitude method. The excellent depth resolution provided by such interferometry provides the ability to slice the sample at different depths. In fact, in the case of a sample where there is reflector a at the focal point and reflector b is at point R where the reflection intensity is exactly 0, as shown in Figure 3c, the total reflection intensity is the reflection from the two reflectors. Although it is the sum of waves, as shown in the figure, since the sound pressure is not irradiated to reflector b, only a can be observed separately from b. Conversely, when focusing on reflector b, the reflected wave from a disappears and only b can be observed separately. By the way, the reason why we were able to separate them so well is because the distance between the reflectors a and b exactly matches the distance between the peak position and valley position of the interference pattern (hereinafter referred to as the slice distance). be. Although this distance is λ/4, this condition does not always hold depending on the sample. Therefore, it is expected that the slicing ability of interferometry can be made even more effective by adding a means to vary the slicing distance depending on the target sample.

The present invention has been made in view of the above points, and it is an object of the present invention to provide a means for dramatically increasing the slicing ability of interferometry.

In other words, from the above study results, since the slice distance is proportional to the wavelength λ of the ultrasonic wave f used, in order to separate two layers with an arbitrary depth difference, it is necessary to make sure that the depth difference and the slice distance match. This is a method of moving the frequency used by Δf from f. When separating two reflective layers separated by Δd from the slice distance d as shown in FIG. 3d, it is sufficient to reduce the frequency by Δf. The relationship between Δf and Δd is therefore: λ/4=Vs/4(f ₀ +Δf)=d+Δd If Δf≪f ₀ , then 1/4Vs/f ₀ (1−Δf/f ₀ )=d+Δd d=1/4Vs /f ₀ =λ ₀ /4 is used, Δd=−λ ₀ /4·Δf/f ₀ (9). This change in frequency is caused by the fact that the observer
Focus on the reflective layer b from the observation screen.
You can manipulate it so that it disappears, or if the structure is known in advance like an IC, you can set it before observation. Embodiments of the present invention will be described below with reference to the drawings.

FIG. 4 is a diagram showing an example of a circuit configuration embodying the present invention, in which an RF continuous signal (for example, 1 GHz) generated by an RF continuous wave oscillator 100 is converted into an analog signal.
The switch 110 converts the signal into an RF pulse signal of duration t _d , which is then sent to the probe system 12 via the directional coupler 120.
5. This probe system 125 is composed of a lens 20 and a piezoelectric thin film 10 shown in FIG. After the reflected detection signal is amplified by the receiving amplifier 130 via the directional coupler 120, the diode detector 1
40, it is converted into a video signal (bandwidth ~10 MHz), and a time gate 150 samples the signal reflected from the sample, which is a desired signal, and uses it as an imaging signal. In this embodiment, the analog switch 110 is turned on and off.
A gate generator 160 is used. That is, the gate generator 160 uses the switch 170 to generate the gate signal for the analog switch 110 under the condition t _d >2Z/
This is a circuit that switches so that V _w or the condition t _d <2Z/V _w .

FIG. 5 shows an example of its configuration.

That is, the repetition period t _R from the pulse oscillator 161
At the rising edge of the pulse, multivibrators 162a, 162b, and 164 generate pulses with time widths of Δt ₁ , Δt ₂ , and Δt _{3 ,} respectively. Here, Δt ₁ <2Z/V _w , Δt ₂ >2Z/V _w , Δt ₃
= 2Z/V _w +t _s . The output waveforms of the multivibrators 162a and 162b (each with the second
(shown in Figures b and e) is selected by the multiplexer 163 according to the High and Low states of the switch 170,
By using the control signal of the analog switch 110 and inputting the output of the multivibrator 164 to the multivibrator 165 again, Δt ₃
A pulse delayed by
(shown in e, f). This kind of circuit is TTL
(Transistor Transistor Logic). In Figure 2 a, b, c and d, e, f, t _d <
2Z/V _w , t _d >2Z/V _w Therefore, the time charts of the control signal and gate signal in the conventional method and the present interferometry are shown.

In this embodiment, since the appearance time of the reflected signal from the sample does not change due to such mode switching, the same gate signal can be used.

Furthermore, it is possible to use the interference method and the conventional amplitude method selectively by simply switching the duration of the RF pulse signal.

That is, as is clear from the above explanation, if the duration time t _d of the RF pulse is shortened to t _d <2Z/V _w (10), the two reflected ultrasound signals B and C will not overlap.

In such a configuration, the RF continuous wave oscillator 100
Variable frequency RF that can vary the frequency as
The present invention can be implemented using an oscillator.

A commercially available variable frequency RF oscillator that moves a dial may be used as such an oscillator, or a VCO
An oscillator (voltage controlled oscillator) or a synthesizer that can be controlled by voltage or digital signals may also be used.

FIG. 6 is a diagram showing another embodiment of the present invention, in which an electrical reference wave is used, whereas the previous embodiment uses a sonic reference wave.

The output RF continuous wave electric signal (for example, 1 GHz) of the variable frequency RF oscillator 200 is sent to the analog switch 2.
10, it is converted into an RF pulse signal with a duration t _d (for example, 0.5 μs), and is sent to the probe system 2 via the directional coupler 220.
30. After the reflected ultrasound signal is amplified by the receiving RF amplifier 240 via the directional coupler 220,
RF adder 250 (signal in FIG. 7a).
The RF pulse signal is also delayed by the RF delayer 280 to the time of occurrence of the echo from the sample t ₀ +t _s , and then
It is applied to RF adder 250 via RF amplifier 290. (Signal in Figure 7b) The output of the RF adder becomes an interference signal as shown in Figure 7c, so it is converted to a video signal (bandwidth ~10MHz) by a diode detector 260, and only the interference signal is detected by a time gate 270. It is sufficient to extract the signal and use it as an imaging signal. The control circuit 201 may be the same as in the previous embodiment. In this configuration, the present invention can be realized by changing the frequency of the oscillator 200.

Note that the present invention is not limited to the embodiments described above, and the present invention can be easily applied to any device that inherently has an interference signal. Alternatively, a continuous wave oscillator that generates a chirp signal may be used to display the frequency change on the x-axis and the reflected interference signal on the y-axis on a cathode ray tube to confirm the slicing function.

As described above, according to the present invention, by changing the ultrasonic frequency used in an ultrasonic microscope having interference microscopy ability by Δf around the frequency f ₀ , the slicing ability can be significantly increased, and multilayer structures such as ICs can be It is extremely useful when performing depth-based and stratified observations, and its contribution to this industry is extremely large.

[Brief explanation of drawings]

Figure 1 shows the configuration of a conventional probe system, Figure 2 shows the configuration of a conventional probe system.
The figures are diagrams explaining the operation of a conventional ultrasound microscope, Figure 3 is a diagram explaining the principle of the present invention, Figures 4 and 5 are diagrams showing the configuration of an embodiment of the present invention, and Figure 6 is a diagram explaining the operation of a conventional ultrasound microscope. 7 is a diagram showing the configuration of another embodiment of the present invention, and FIG. 7 is a diagram explaining its operation.

Claims (1)

[Claims] 1 means for transmitting a focused ultrasonic beam toward a sample and receiving reflected waves of the focused ultrasonic beam from the sample; means for mechanically scanning the sample in two dimensions within the focal region of the beam; An ultrasonic microscope having means for generating a reference wave having the same frequency as the focused ultrasound beam and interfering with the reflected wave, comprising means for making the frequencies of the focused ultrasound beam and the reference wave variable from a center frequency. An ultrasonic microscope characterized by: