JPH02136772A - Optical probing method - Google Patents

Abstract

PURPOSE:To execute the measurement with high S/N and to contrive shortening of the measuring time by monitoring a light beam from a means for turning on and off a light beam, and inputting measuring data of an optical exciting current by synchronizing therewith. CONSTITUTION:In a method for measuring a state of the inside of a semiconductor device 30 by turning on and off a light beam by an optical modulating means (acousto- optical modulation element 42) and irradiating the semiconductor device 30 thereby, and detecting a variation of a power source current of the semiconductor device 30 based on a light exciting current generated in the semiconductor device 30 by a light irradiation at the time of turn-on, in a light exciting current detecting circuit 60, an output light of the acousto-optical modulation element 42 is monitored by a monitoring circuit 47, and based on a monitor output, by synchronizing with a light irradiation to the semiconductor device 30, a detection signal of the power source current is inputted as measuring data. According to such a constitution, a timing of the light irradiation to the semiconductor device 30, and a detection timing of a variation of the power source current by the light exciting current generated in the semiconductor device 30 by the light irradiation are synchronized. Accordingly, the light exciting current is measured exactly, and the measuring time is shortened.

Description

[Detailed description of the invention]

[Industrial Field of Application] The present invention provides a method for manufacturing a semiconductor device using light such as a laser.
The present invention relates to an optical probing method suitable for use in inspecting whether a semiconductor device is good or bad by measuring the internal state of the device. [Conventional technology 1: Remarkable semiconductor devices in recent years! 2. Description of the Related Art With advances in manufacturing technology, semiconductor devices, such as LSIs, are becoming increasingly denser and more highly integrated. Along with this, failure diagnosis of LSIs has also become highly complex, and it is necessary to measure the electrical state of the internal circuits of LSIs. Conventionally, electrical measurements of the internal circuits of this LSI have been performed by bringing a metal probe into contact with the AI wiring pattern of the LSI. However, with the miniaturization of LSI, this method
Aj! It is difficult to make precise mechanical contact with the wiring, ■ The measurement accuracy of the capacitance of the metal probe deteriorates, ■ Aj! There were drawbacks such as the possibility of destroying wiring or internal circuits. To address these problems, an electron beam tester that applies a scanning electron microscope has been developed and put into practical use. This method irradiates the LSI surface in a vacuum sample chamber with an electron beam accelerated to several kV, detects the secondary electrons emitted from the LSI surface, and records them as a secondary electron image or internal waveform.
I! It has the advantage of being compatible with the miniaturization of LSIs. However, with this method, it is necessary to vacuum the sample chamber, which takes time for measurement and is inefficient.
Since the surface of the LSI is irradiated with an electron beam, there are problems such as the characteristics of the LSI or the transistor elements constituting the LSI changing due to damage caused by the electron beam, and (2) the device being expensive. In addition, to eliminate the drawbacks of the above-mentioned measurement method using a metal probe, the LSI surface is irradiated with a laser beam in the atmosphere, and the optical excitation generated within the LSI substrate is detected. A method has been proposed for non-contact analysis of the state of
079-56>. The principle by which the logic state of a transistor inside an LSI can be analyzed by this method will be explained below with reference to FIGS. 5 and 6. The following explanation is based on an example of an 0MO3 inverter, which is an N-type substrate 1 and a P-type well 2 as shown in FIG. In the figure, the left part is an n-channel transistor,
The right part is an n-channel transistor, 3 and 6 are source regions, 4 and 7 are drain regions, 5 is a substrate potential contact, and 8 is a well potential contact. Vcc is a power supply voltage, and GIID is a ground terminal. Further, Vg is an input voltage. As shown in FIGS. 5A and 5B, when the drain region 7 of the n-channel transistor is irradiated with the laser 9, the laser 9
If the wavelength of is shorter than the absorption edge of the semiconductor, light absorption occurs,
There, electron-hole pairs are generated. Semiconductor or silicon
When the laser 9 is an Ar laser (488nn), the laser 9
The depth of penetration is approximately 1. Since holes are minority carriers in the drain region, the holes among the electron-hole pairs generated in the drain region 11117 cause a gradient in the concentration distribution of holes in the drain region 7 . Therefore, the holes diffuse, and most of them reach the PN junction at the boundary of the drain region 7. The PN junction has a potential distribution as shown in FIG. 6, and holes are thereby accelerated outside the drain region 7. Since electrons are majority carriers, they hardly diffuse. On the other hand, since electrons are minority carriers in the P-well 2, among the electron-hole pairs generated in the P-well 2 below the drain region, electrons cause a gradient in the electron concentration distribution in the P-well 2. There, the electrons diffuse, and most of them also reach the PM junction at the boundary of the train region 7, and are accelerated into the train region 7 by the same potential distribution as described above. As a result,
The inside of the drain region 7 is negatively charged, and the inside of the P well 2 is positively charged. Therefore, a potential distribution occurs in each of the drain region 7 and the P well 2, and a current tends to flow. For example, as shown in FIG. 5A, when the input voltage vg of the CMOS inverter is at a high level, the n-channel transistor is on and the n-channel transistor is off, so the arrow line from the train region 7 A current flows as shown in 10. Further, in the P well 2, a current flows as indicated by the arrow 11. Since these two currents are closed in a loop, they cancel, and therefore, these currents do not flow outside the LSI. Conversely, when the input voltage vg of the CMOS inverter is at a low level as shown in FIG. A current flows through. Since the current flow in the P-well 2 is the same as in the case of FIG. 5A, the photoexcitation current flows out of the LSI. It can be easily inferred that a similar phenomenon occurs in the drain region 4 of the P-channel transistor due to laser irradiation. From the above, by irradiating the drain region 4.7 with a laser and detecting the change in the power supply current of the LSI at that time, it is possible to know the logic state of the transistor at the portion where the loop is irradiated. The above is a case of a CHOS device, but in the case of devices with other structures such as NHOS and bipolar devices, the logic state of the internal circuit can be determined by light irradiation based on the same principle. [Problems to be Solved by the Invention] As described above, the photoexcitation current can be measured by detecting the change in the @ current flowing through the power supply line of the LSI when the laser is irradiated. In this case, generally the power supply current is A
/D conversion is carried out as a digital value and this is taken in as measurement data, and this is processed by a computer to detect the change and inspect the semiconductor device.Here, the change in the power supply current due to the photoexcitation current is extremely small on the order of microamperes. Since the current is small, it is necessary to detect it separately from current changes other than this photoexcitation current. Therefore, if the sampling point of the power supply current during A/D conversion does not coincide with the generation timing of the photoexcitation current due to light irradiation, there is a risk that changes in the power supply current other than those caused by the photoexcitation current may be captured as measurement data. , the measurement S/N becomes worse. In addition, the above method can be applied to LSI, especially ASIC (Appl.
cation 5special Integra
When attempting to verify the design of a ted C1rcuit (special purpose integrated circuit), the test pattern signal applied to the LSI must be input at a speed that matches the design specifications of the LSI. do not have. By the way, as mentioned above, since the method described above detects extremely small changes in the current superimposed on the LSI's electric current, changes in the current other than those caused by the photoexcitation current during the measurement period should be as small as possible. For this reason, measurements are generally made after waiting for the transition of the state of the internal circuit due to a change in the input signal of the LSI to be completed. However, the maximum operating speed of an LSI is usually determined by the maximum time for the state transition of the internal circuit to complete due to a change in the input signal, so when the LSI is operated at the maximum speed, each state of the input signal almost all the time,
This means that the internal circuit of the LSI is operating. In other words,
The power supply current is changing all the time during each state, and if this continues, it will not be possible to take time for measurement. Therefore, a method is adopted in which the input test pattern signal is slowed down or stopped during the period to be measured. However, in this case, if the speed of the test pattern signal is temporarily slowed down or stopped for a long time, the meaning of the test will be diminished in terms of testing against the design specifications of the LSI mentioned above. Therefore, in order to measure the photoexcitation current by irradiating the laser, it is possible to temporarily slow down or stop the test pattern signal, in other words, to shorten the measurement time as much as possible so that the measurement can be performed in a state close to the design specifications. It is required to do so. In view of the above-mentioned problems, the present invention aims to provide a light blowing method that can measure a good S/H and shorten the measurement time.

[Means to solve the problem] The present invention turns on and off light using a light modulation means to irradiate the semiconductor device, and detects a change in the power supply current of the semiconductor device based on the photoexcitation current generated in the semiconductor device by the light irradiation when the light is on. In the method for measuring the internal state of the semiconductor device, The optical probing method monitors the output light of the optical modulation means, and captures the detection signal of the power supply current as measurement data based on the monitor output in synchronization with the irradiation of the semiconductor device with light. [Effect]

The timing of irradiating the semiconductor device with light and the timing of detecting a change in the power source t a due to the photoexcitation current generated in the semiconductor device by the irradiation with light are synchronized. Therefore, the photoexcitation current can be measured accurately. In addition, when measuring,
Since there is no need to wait an unnecessary amount of time, the measurement time can be shortened.

Embodiment 1 An embodiment of a semiconductor device inspection apparatus to which the optical broaching method according to the present invention is applied will be described below with reference to the drawings. FIG. 1 is a diagram showing the overall configuration of the inspection apparatus. The inspection apparatus of this example includes a laser optical system 40 for scanning and converging and irradiating a laser beam onto a semiconductor device 30 to be inspected, and a semiconductor device 30 to be inspected. A semiconductor device fixture 31 that fixes the device 30 at a measurement position and connects a signal line for providing a test pattern, a power supply device 51 that provides a power supply voltage to the semiconductor device 30, and a power supply for this power supply device 51 and the semiconductor device 30. A photoexcitation current detection circuit 60 that is installed in the middle of the power supply line 52 between the pins and detects changes in the power supply current, and a controller 70 that controls the entire inspection device and processes and displays the data of the measured photoexcitation current. Equipped with. The controller 70 includes a CPtJ 71, a storage device 72, a work area memory 73, an optical system controller 74, an XY stage controller 75, a timing signal generation circuit 76, and a semiconductor device power supply controller 7.
7, a display controller 78, and a display 79. 41 is a laser light source, and in this example, a laser is used. Since the Ar laser has a relatively short wavelength, one laser is used.
It is possible to condense light to a diameter of , and it can also be miniaturized and correspond to S■, etc. Laser light from this laser light source enters the laser optical system 40. The laser light incident on the laser optical system 40 is controlled in intensity by the optical modulation means, an acousto-optic modulation element 42 in this example, and turned on and off.
Off control or not. The acousto-optic modulator 42 has a 7e0
2. PbHo0. , i Consists of an acousto-optic medium 81 made of fused silica or the like and piezoelectric elements 82 and 83. Piezoelectric element 8
When a high frequency signal is applied to 2.83, the acousto-optic medium 81 vibrates, an ultrasonic wave propagates through the acousto-optic medium 81, and a diffraction grating is formed in the acousto-optic medium 81 by the elastic wave 84. Therefore, when a laser beam 85 is incident on this as shown in the figure, the light is subjected to black diffraction, and the output light is a laser beam 86 that passes through the incident light without being diffracted, and a laser beam 87 that is diffracted and bent and then output. and 0 occur alternately depending on the frequency of the ultrasonic wave.For example, the solid line laser beam 8
7 is the light that irradiates the semiconductor device, and the broken line is the laser light 86.
The laser can be turned on and off by disabling it. By changing the frequency of the high-frequency signal supplied to the piezoelectric elements 82 and 83, the on/off frequency of the laser beam can be changed. Changing the intensity of the ultrasonic waves changes the amount of light that is bent, making it possible to modulate the intensity of the light. In this example, the optical system controller 7 of the controller 70
A high frequency signal modulated from 4 is applied to piezoelectric elements 82 and 83, and the router light is turned on and off, and its intensity is modulated. In this example, the frequency of the high-frequency signal is, for example, several hundred 8II
It is said to be Z. The on/off switching frequency of this laser beam is sufficiently higher than the frequency of state changes of the test pattern signal. The laser beam 87 when turned on from the modulation element 42 is transmitted to the acousto-optic modulation element 4 for use in the X direction and Y direction, which are orthogonal to each other.
3.44 in sequence. Acousto-optic deflection element 43.44
are controlled by the optical system controller 74 of the controller 70, and the laser beams are deflected in the X direction and the X direction with predetermined scanning widths, respectively. The laser light that has passed through the deflection elements 43 and 44 is focused by an objective lens 46 placed on an XY stage 45, and the semiconductor device 30 is irradiated with the laser light. In this case, the entire semiconductor device 30 can be irradiated with a laser beam by an objective lens 46 placed on the XY stage 45 and a mirror optical system. The X-Y optical system 45 is controlled by the X-Y stage controller 75 of the controller 70 in response to deflection control of the deflection elements 43 and 44. Here, when a 50x objective lens is used as the 46° objective lens, the laser scanning area in the semiconductor device 30 is 25°.
6 diagonal angles, and the laser spot diameter and laser irradiation position resolution are 1 sting. The semiconductor device 30 is connected to the performance port 3 using a socket 32 such as a zero insertion force socket.
It is installed on 3. Performance port 33 is connected to fixture 31 via pogo pins (not shown). This fixture 31 has a coaxial cable 5.
3, a test pattern signal is supplied from a test pattern signal generator 154. This test pattern signal generating means 54 can use a pattern generator or an LSI tester. In this example, a dedicated fixture 31 is used, but it goes without saying that a test head such as an LSI tester can be directly placed in the device and measurements can be made. The power supply device 51 is configured with a stabilized power supply with little ripple, and supplies power supply voltage to the semiconductor device 30 . The power supply device 51 is preferably a variable voltage power supply,
In this example, the output voltage of this power supply device 51 can be controlled by a controller 70. A photoexcitation current detection circuit 60 inserted into the power supply line 52 between the power supply device 51 and the power supply bin of the semiconductor device 30 is
In this example, the configuration is as shown in FIG. That is, the pulse transformer 63 is connected between the power output terminal 61 of the power supply device 51 and the power supply bin 62 of the semiconductor device 30 . The pulse transformer 63 has a predetermined impedance of, for example, about 50Ω at several tens of MHz, which is the frequency at which the laser beam is turned on and off, and is set so that the impedance is sufficiently low at frequencies lower than this. In other words, in the frequency range from DC (direct current) to several MHz, no voltage drop occurs in this pulse transformer 63. Therefore, the test pattern signal is supplied to the semiconductor device 30 from the test pattern signal generation circuit 54, and the test pattern signal is Even if the current consumption inside the semiconductor device 30 changes for each state, no voltage drop occurs in the pulse transformer 63, so the voltage of the power supply bin 62 remains almost constant regardless of changes in the current consumption inside the semiconductor device 30. The semiconductor device 30 will not malfunction due to voltage drop. On the other hand, in the semiconductor device 30, the modulation element 42 has several tens of M
Since a laser beam that is switched on and off at tlz is irradiated, a photoexcitation current is generated at the timing of this laser beam irradiation (on timing). Therefore, the pulse transformer 63 obtains a voltage that changes depending on the presence or absence of this photoexcitation current. The voltage generated in the pulse transformer 63 is supplied to a bandpass filter 65 via a coupling capacitor 64. This bandpass filter 65 has a pass band of several 7 H11z, which is the ON/OFF switching frequency of the laser beam in the modulation element 42. Therefore, a voltage that changes depending on the presence or absence of the photoexcitation current is obtained from the bandpass filter 65, and is supplied to the A/D converter 67 via the high-gain high-frequency amplifier 66. This A/D converter 67 is supplied with a sampling pulse synchronized with the timing at which a photoexcitation current is generated by laser irradiation from the timing signal generation circuit 76 of the controller 70, and this sampling pulse causes the amplifier 66 outputs are sampled and each sampled value is converted to a digital signal. Since the sampling timing is synchronized with the light irradiation, the light irradiation generates a photoexcited current, and at the timing when the electric current changes, the A/D converter 67 samples and digitizes the current. be. This digital signal is taken into the work area memory 73 of the controller 70. The photoexcitation current measured as described above is stored in the memory 73.
Based on the data taken in, a timing chart or optical excitation is displayed as a flow image on the display 79 of the controller 70. Then, this is compared with, for example, simulation data output from a Li design CAD, or data measured for a good semiconductor device and a defective semiconductor device are compared to find a defective location. Next, the timing signal generation circuit 7 in this inspection device
The timing control in 6 will be explained below. Here, first, the deviation in the actual on/off timing of the laser beam with respect to the on/off control signal in the acousto-optic modulation element 42 is taken into consideration. That is, in the modulation element 42, the laser is turned on/off or the intensity is modulated so that the modulated ultrasonic waves are transmitted to the acousto-optic medium 8.
Occurs when the laser beam position in 1 is reached. The rise and fall of the laser beam are determined by the laser beam diameter and the speed of the ultrasonic wave in the acousto-optic medium 81. Also, the delay time τ from when the electric signal to turn on the laser is given to when the laser actually turns on depends on the distance At d between the piezoelectric element 82, 83 and the incident position of the laser beam, and the speed of the ultrasonic wave. It's decided. This time is, for example, about 600 nanoseconds. This delay time corresponds to, for example, 6 clock cycles in a semiconductor device operating with a tarock of 10 Htlz. Recent CHO8 devices are 20MHz to 408)1
Since it operates at a high speed of 2, it is necessary to perform verification at high speed. Therefore, it is not possible to wait for the above-mentioned 600 nanoseconds before performing the measurement, and accurate timing adjustment is required. In addition, the pattern controller or LSI tester that generates the test pattern is used in external clock mode to adjust the timing, but generally it takes several clocks from when the external clock is applied until the test pattern signal is output. There is a delay of +α. Therefore, it is also necessary to perform timing control on this test pattern signal generation circuit 54. Therefore, in this example, as shown in FIG.
The optical path is changed at , the light enters the monitor circuit 47 , and is detected by the monitor circuit 47 . This monitor circuit 47 is composed of a high-speed photodiode and an amplifier, and monitors the laser on/off timing. The monitor signal MO from this monitor time #I47 is supplied to the timing signal generation circuit 76. In addition, the test pattern reference clock RE from the test pattern signal generation circuit 54
FC is supplied to this timing signal generation circuit 76. In place of the control by the CPU 71, the timing signal generation circuit 76 determines the timing to output eggplant 1 to pattern, the timing to turn on and off the laser, and the A/D control in the photoexcitation current detection circuit 60 based on the monitor signal HO. Adjust the timing of conversion. The general method for aligning these three timings is to first use one as a reference, and use delay means to delay the timing of the other two within the necessary range and with the necessary resolution. It is true. FIG. 4 is a diagram showing an example of the timing signal generation circuit 76 and a circuit portion for timing adjustment. In the same figure, 91 is a tarokk signal generation circuit, which includes a sampling clock SPC, a laser timing clock LAC, and a test pattern generation tarokku PC.
You can get something like that. The sampling clock SPC is used as a reference in this example, and is supplied as is to the A/D converter 67 of the detection circuit 60. The laser timing clock [80] is supplied to the laser optical system controller 74 via a delay circuit 92. Test pattern generation clock TPO is supplied to test pattern signal generation circuit 54 via delay circuit 93. The monitor signal H from the monitor circuit 47 of the laser optical system 40 is input to this timing signal generating circuit 76.
O and the test pattern reference clock REFC from the test pattern signal generation circuit 54 are each sampled by the sampling clock SPC in the D flip-flop circuit 194.95. The sampling output is sent to the CPU 7 via the data bus.
1. The CPtJ 71 refers to these sampling outputs, determines the lead or lag, and sends the delay circuit 92
, 93, and the sampling clock SPC, monitor signal MO, and test pattern reference clock RE.
The delay amount R92, 93 is set so that the FC has a predetermined relationship.
Set the delay amount of A/D change timing and laser
Make sure that the on/off timing matches the test pattern signal generation timing. By the way, as mentioned above, the change in the a current due to the photoexcitation current handled by this device is an extremely small current, and moreover,
It is a high frequency signal of several hundred MHz. Currently, there is no delay line that passes such high-frequency signals and has a relatively long delay time of 600 nanoseconds in the modulation element 42 described above.Therefore, in this example, the delay circuit uses a reference clock. It consists of a counter for counting and a delay line of minute time (120 cycles or less) that can control the delay time with precision of, for example, 1 nanosecond, and the amount of delay of each is set by the CPU 71. To measure the number of clock cycles of delay: That is, the reference clock is set in a state where it can be counted by a counter, and counting is started when the laser clock is started. Then, the monitor circuit 47 stops counting the counter when the monitor signal HO indicating that the laser is detected in the OFF state arrives at the timing signal generation circuit. The count value of the counter at this time is the number of delayed clock cycles. After that, the timing is adjusted even more precisely by controlling the minute delay time. As described above, when the three timings match, t&,
Add any additional delay needed between each signal. This is, for example, a delay in which, after the laser is turned on, the power supply current increases and the timing of the A/D conversion is delayed by the time it takes for the signal to be amplified and reach the A/D converter. Note that in the above example, other timing signals were adjusted using the sampling clock SPC as a reference.
Of course, you can use either as the standard. Further, in the above example, an acousto-optical modulator was used as the light modulator, but it goes without saying that the light modulator is not limited to this. Further, in the above example, as the laser scanning means, although an acousto-optic deflection element is used, a galvano mirror or a polygon mirror may be used, or a combination of these scanning means may be used. Furthermore, it goes without saying that analysis in the depth direction of a semiconductor device can be performed by using various laser light sources with different wavelengths as the laser light source. [Effects of the Invention] According to the present invention, since the light from the means for controlling the light on and off is monitored and the measurement data of the photoexcitation current is taken in in synchronization with this, it is possible to monitor the light from the means for controlling the light on and off. Measurement data can be obtained at any time. Therefore, the photoexcitation current can be measured accurately, and since there is no need to wait an unnecessary amount of time during measurement, the measurement time can be shortened.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an embodiment of a semiconductor device inspection apparatus to which the optical broaching method according to the present invention is applied, and FIG.
The figure shows an example of a light modulation element, FIG. 3 is a circuit diagram of an example of a partial circuit of the example in FIG. 1, FIG. 4 is a circuit diagram of an example of the main part of the example in FIG. 1, and FIGS. FIG. 6 is a diagram for explaining the generation of photoexcitation current in a semiconductor device. 30. Semiconductor device 40. Laser beam system 41; Laser light source 42; Light intensity modulation element 43; Light deflection element 47; Monitor circuit 51; Power supply device
52; power supply line 54; test pattern signal generation means 60; optical excitation@flow detection circuit 74; optical system controller 76: timing signal generation circuit 92.93; delay means spc, sampling clock LAC, laser timing clock TPO, test pattern generation Clock NO, optical monitor signal REFC, test pattern reference Kronk's agent Patent attorney Masami Sato Fig. 4 (OFF) (ON) 4.7 Turn on the first page (OFF)

Claims (1)

[Claims] By irradiating the semiconductor device with light by turning it on and off using a light modulation means, and detecting the change in the power supply current of the semiconductor device based on the photoexcitation current generated in the semiconductor device by the light irradiation when the light is on, In the method of measuring the internal state, the output light of the light modulation means is monitored,
An optical probing method that captures a detection signal of the power supply current as measurement data in synchronization with light irradiation to the semiconductor device based on the monitor output.