JPH1174325A - Semiconductor surface evaluating method and device by surface photovoltage - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a surface evaluating method by means of photovoltage and device capable of evaluating the crystalline surface easily and precisely. SOLUTION: Laser beams in the transmission length sufficiently shorter than the diffusion length of the minor carrier in semiconductor crystalline while in the same as or less than that of a depletion layer formed on the surface of the crystalline surface is emitted from an Ar laser beam source to be modulated into pulse beams by an AOM(audio optical modulator) 11 for irradiating a semiconductor crystal (specimen wafer 1) to detect the generated photovoltage by a transparent electrode 3 in the capacity coupling. Finally, the crystalline characteristics of the semiconductor crystalline surface is evaluated according to the attenuation time of this surface photovoltage.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a semiconductor device manufacturing field in which a semiconductor device is formed on a silicon substrate, and whether or not the device manufacturing process is appropriate, in other words, whether the manufacturing process satisfies a desired specification. The present invention relates to a method and an apparatus for evaluating a semiconductor surface by using a surface photovoltage for inspecting whether or not the semiconductor surface is defective.

[0002]

2. Description of the Related Art Recently, high-performance semiconductor devices have a so-called planar type structure, and the performance of the device often depends on the semiconductor surface and an oxide film formed thereon. As an example, M which is widely used in large-scale integrated circuits (LSI), etc.
In the case of an OS (metal / oxide / semiconductor) transistor, the characteristics of the semiconductor surface (meaning the range from the interface between the oxide film and the semiconductor and from the crystal surface to the depth of about several μm) affect the performance of the transistor. It is widely known that it has a significant effect. Further, in a charge-coupled device widely used recently as an imaging device of a television camera or the like, the characteristics of the semiconductor surface directly affect the transport efficiency of a so-called optical carrier (excess carrier) excited by light.

More specifically, interface traps existing at the interface between a semiconductor and an oxide film and crystal imperfections near the semiconductor surface determine characteristics of the semiconductor surface. That is, when there are many interface traps, for example, the optical carriers near the semiconductor surface recombine via the interface traps and disappear. This means that in the case of the charge-coupled device, the optical carrier is lost during transportation, and the characteristics of the imaging device are significantly deteriorated. In the case of a charge-coupled device, a so-called surface potential (see FIG. 1) is generated on the semiconductor surface, and the photocarrier is stored in a potential well formed by the surface potential, that is, a depletion layer formed corresponding to the surface potential. However, if the surface properties are poor, the amount of optical carrier will decrease shortly after storage. In a MOS transistor, it becomes difficult for a current to flow between a source and a drain.
In general, the number of interface traps depends on the interface trap density D.
It is represented by the level of it (large or small).

[0004] As described above, the level of the interface trap density Dit is extremely important in many planar-type devices in that it affects device performance. The interface trap density Dit naturally depends not only on the polishing state of the semiconductor surface but also on the oxide film forming process very sensitively. For example, when an oxide film is formed by a dry oxidation method, the interface trap density Dit is extremely high, whereas in the wet oxidation method, the interface trap density Dit is extremely low.

[0005] The occurrence of interface traps depends on the state of bonding of oxygen to semiconductor atoms, and thus also depends on the thickness of the oxide film. In recent years, with the miniaturization of elements, the oxide film has been gradually reduced in accordance with the similarity rule. However, a change in the oxide film thickness naturally affects the generation of interface traps. Since the natural oxide film contains pollutants in the atmosphere, from the viewpoint of contamination,
Interface traps increase. Even when the interface trap density Dit is low at the beginning of the oxide film formation, when the semiconductor substrate (wafer) is exposed to gas plasma widely used in the semiconductor device manufacturing process, the interface trap density Di is reduced due to radiation damage or the like.
t increases significantly.

As described above, there are many causes of the increase of the interface traps in the device manufacturing process, and the increase and decrease of the interface traps are extremely fluid. Similar concerns are attached to the crystallinity of the semiconductor surface. That is, if impurities reach the crystal surface via a thin oxide film in various processes, the surface characteristics of the crystal naturally change in the direction of deterioration.
Therefore, it is clear that it is desirable to inspect the surface characteristics during the device manufacturing process in order to guarantee the integrity of the device manufacturing process.

Actually, a method for inspecting the magnitude of the interface trap density Dit and the surface characteristics has been put to practical use, and is used for evaluating a semiconductor manufacturing element process. At present, two typical methods used in a semiconductor manufacturing process for the purpose of evaluating interface traps will be described below.

(1) CV (Capacitance-Voltage) Method CV is a standard method for measuring the interface trap density.
The (capacity-voltage) method has been practically used from an early stage. In this method, the interface trap density Dit is measured using the same configuration as that of the gate electrode portion of the MOS transistor. The capacitance (capacitance; C) between the gate electrode and the semiconductor layer is changed while changing the gate voltage. And a quasi-static voltage, and the interface trap density Dit is determined from the gate voltage (V) dependence of the capacitance (C).
Since the interface trap traps (captures) the optical carrier, it forms a kind of capacitance at the interface. Since the relationship between the electric capacity and the interface trap density Dit is theoretically determined, the interface trap density Di is determined from the correspondence between theory and experiment.
t can be determined.

Although this method can accurately determine the interface trap density Dit, it is necessary to form a metal electrode on an oxide film in order to perform the measurement. As described above, the semiconductor substrate (wafer) in the course of the device manufacturing process is extracted from the process and the C-
If V (capacitance-voltage) measurement is performed, metal electrodes that are unnecessary for element manufacture are added to the wafer, so that the same wafer cannot be returned to the manufacturing process. That is, this method is basically a destructive inspection method, and it is impossible to inspect all the wafers during the manufacturing process.

Recently, a non-contact type CV in which a metal electrode is not formed on an oxide film but the electrode is opposed to the oxide film via a gap.
A law has been proposed. This method is a non-destructive inspection in that a metal electrode is not formed on a wafer.
As long as the method is used, a strong DC electric field will be applied to the oxide film. This means that the charge in the oxide film is forcibly moved, and changes the state of the wafer taken out of the process. However, when ions and the like reach the semiconductor surface, the state of the interface trap changes,
In the conventional CV method, there is a risk of changing the state of the interface during the measurement process. Therefore, non-contact C-
The V method is essentially a destructive test.

In addition, the characteristics of the semiconductor surface are not determined only by the interface traps, and the device characteristics are extremely significantly affected by the crystallinity of the depletion layer. However, C-
No information can be obtained on the crystal integrity of the depletion layer by the V method. This is also a disadvantage of the CV method.

(2) μ-wave photoconductive decay method The μ (micro) -wave photoconductive decay method has recently been used as an evaluation method of the interface trap density Dit. Originally, this method was developed to measure the minority carrier lifetime in semiconductor crystals. Although there are several definitions for the lifetime, the lifetime inside the crystal that is not affected by the crystal surface is generally called the bulk lifetime or the volume lifetime. I do.

Light is irradiated to a wafer or the like to generate an optical carrier (excess carrier), and the disappearance process of the optical carrier is detected by reflection of a microwave. The generation of the photocarrier locally increases the conductivity of the wafer, but the reflection of the microwave depends on the conductivity of the wafer. Therefore, the magnitude of the reflection of the microwave indicates the abundance of the optical carrier. The time change of the reflection intensity of the μ wave is measured, and the time at which the reflection intensity decreases to 1 / e (e indicates the base of natural logarithm) is defined as the photoconductive decay time. Often referred to as carrier lifetime.

In the process of evaluating and measuring the (effective) minority carrier lifetime of a wafer by this method, as a secondary phenomenon,
It was found that the wafer surface (including an interface when an oxide film is present) is strongly affected. In other words, the volume life time can be measured by this method in a limited case. In general, the measured photoconductive decay time is strongly affected by interface traps and the like. It is strongly influenced by the minority carrier lifetime, which is defined as the surface lifetime relative to the volume lifetime. That is, the effect that the optical carrier rapidly disappears on the surface and the effect that the optical carrier is preserved for a long time on the surface are mixed. Therefore, depending on the surface condition of the wafer, the photoconductive decay time may be longer than the volume life time, or the photoconductive decay time may be much shorter than the volume life time.

As described above, by taking advantage of the fact that the μ-wave photoconductive attenuation method is strongly affected by the surface, this method may be used for the evaluation of the interface trap density Dit in special cases.
In fact, when the interface trap density Dit increases due to irradiation damage due to plasma or the like, the microwave photoconductive decay time decreases.
This phenomenon results from a significant decrease in the surface lifetime of the minority carrier, but the same phenomenon appears not only when the interface trap density Dit increases but also when defects in crystals in the depletion layer increase. In other words, C-
Broader information than the V method can be obtained.

The measurement is completely non-contact and non-destructive because the microwave photoconductive decay method does not bring a metal electrode or the like into contact with an oxide film. Therefore, when the semiconductor surface is evaluated by the μ-wave photoconductive decay method, unlike the CV method described above, the inspection can be performed without changing the state of the wafer extracted from the device manufacturing process. It is not impossible to return the inspected wafer to the manufacturing process.

However, the observed photoconductive decay time or effective minority carrier lifetime is naturally strongly affected by minority carrier lifetime (volume lifetime) in the wafer. Therefore, in general, when the photoconductive decay time is long, it cannot be distinguished whether the cause is the crystal surface (surface life time) or the volume life time inside the wafer. The same cannot be determined when the photoconductive decay time is short.

[0018]

As described above,
When trying to evaluate the interface trap density Dit by the CV method, a strong electric field is applied to the oxide film from the outside, so that the interface state is disturbed, and the information of the interface trap density Dit before the CV method is applied is lost. There is a danger of doing it. In order to avoid this danger, it is necessary not to give a large electric field change to the interface at the measurement stage.

Although the crystallinity of the depletion layer cannot be evaluated by the CV method, what actually affects the device characteristics is the characteristics of the surface layer (depletion layer) including the interface. Therefore, an evaluation method including the state of the depletion layer is required.

Although the method of evaluating the crystal (wafer) surface by the microwave photoconductive attenuation method is simple, as described above, it is not possible to accurately evaluate the surface unless the effect of the volume life time inside the wafer is separated. Can not.

The present invention has been made in view of the above problems, and an object of the present invention is to provide a method and an apparatus for evaluating a semiconductor surface by a surface photovoltage which can easily and accurately evaluate a crystal surface.

[0022]

According to the present invention, there is provided a method for evaluating a semiconductor surface based on a surface light voltage, wherein a laser light is modulated into a pulse light by a pulse signal, and the surface of the semiconductor crystal is irradiated with the pulse light to generate a surface light voltage. Is detected by capacitive coupling,
The crystal characteristic of the surface of the semiconductor crystal is evaluated based on the decay time of the surface light voltage.

In this case, it is preferable that the transmission length of the laser beam is shorter than the diffusion length of the minority carrier in the semiconductor crystal.

In the method for evaluating a semiconductor surface by using a surface light voltage, the laser light preferably has a wavelength of 488 nm, and the pulse light has a pulse width of 100 μm.
It is preferably a rectangular wave pulse of s or less. Further, the surface light voltage is obtained by arranging two electrodes above the semiconductor crystal through a gap or an insulating film, grounding an electrode far from the semiconductor crystal, and applying a light voltage from a near electrode to the semiconductor crystal. Can be detected by capacitive coupling. In this case, the electrode may be a transparent electrode, and the pulse light may be incident on the semiconductor crystal via the transparent electrode.

A semiconductor surface evaluation apparatus using a surface photovoltage according to the present invention is an apparatus for evaluating crystal characteristics of a surface of a semiconductor crystal, which emits laser light having a light transmission length shorter than a diffusion length of a minority carrier in the semiconductor crystal. A light source, a pulse light irradiating means for modulating the laser light into a pulse light with a pulse signal, and irradiating the pulse light to the surface of the semiconductor crystal; and a surface light voltage detection for detecting a surface light voltage generated in the semiconductor crystal by capacitive coupling Means for evaluating crystal characteristics of the surface of the semiconductor crystal based on the decay time of the surface light voltage.

In the semiconductor surface evaluation apparatus using surface light voltage, the laser light preferably has a wavelength of 488 nm, and the pulse light is preferably a rectangular pulse having a pulse width of 100 μs or less. Further, the surface light voltage detecting means can be configured to have two electrodes disposed above the semiconductor crystal via a gap or an insulating film, and in this case, an electrode closer to the semiconductor crystal And the far electrode is grounded. The electrode may be a transparent electrode, and the pulse light may be incident on the semiconductor crystal via the transparent electrode.

In the present invention, the light transmission length is, for example, sufficiently shorter than the diffusion length of minority carriers in the semiconductor crystal to be evaluated, and is equal to or less than the depletion layer formed on the crystal surface. After modulating the laser light into a rectangular pulsed light, the pulsed light is applied to the semiconductor crystal, and a surface light voltage is generated from the surface of the semiconductor crystal. By detecting this, the crystal characteristics of the surface of the semiconductor crystal are detected. Can be detected non-destructively.

[0028]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be specifically described below with reference to the accompanying drawings. Note that, in this specification, p-type silicon will be described as an example of a semiconductor crystal, but this is for ease of understanding, and the present invention is not limited to application to p-type silicon. is there.

First, the principle of the present invention will be described. The present invention evaluates a semiconductor surface by surface photovoltage. In the present invention, a weak voltage generated on the semiconductor wafer surface when the semiconductor surface is irradiated with light of a specific wavelength is detected without applying an external voltage to the oxide film or the interface. Since this generated voltage is generated on the semiconductor surface by light irradiation, it is generally called “surface light voltage”, but the magnitude of the generated voltage is 1 mV or less, as will be described again later.
Compared to the value of 1 V or more applied by the CV method, the value is three orders of magnitude smaller, and the effect of the surface photovoltage on the state of the interface can be ignored.

The surface photovoltage itself was discovered by scholars in the United States in the 1950's, but because of the complexity of the phenomenon, it was hardly elucidated at that time, and the mechanism of its generation was clarified by one.
Since 970. According to reported research results, surface photovoltage is generated by oxide film charges present in an oxide film formed on the surface of a wafer or the like. The oxide film charge creates an electric field on the wafer surface, but the resulting cause of the surface photovoltage is the electric field formed on the wafer surface. Therefore, even if light is irradiated on the wafer to generate an optical carrier, no surface light voltage is generated unless an electric field exists on the wafer surface. This is a great difference from the fact that the microwave photoconductive decay method can obtain a microwave reflected signal regardless of the presence or absence of oxide film charges (surface electric field). That is, in the microwave photoconductive decay method, a signal can be obtained even when the surface electric field (oxide film charge) is zero, but no surface photovoltage is generated.

As can be seen from the surface light voltage generation mechanism described above, the surface light voltage is a phenomenon peculiar to the surface, and as long as the wavelength of the light irradiating the wafer satisfies a certain condition described later, μ
Unlike the wave photoconductive attenuation method, information inside the wafer is not mixed. Therefore, when the interface trap density Dit is evaluated using the surface photovoltage, the disadvantage of the microwave photoconductive decay method, which is hindered by the volume life time, can be eliminated as described later.

Next, the relationship between the interface trap density Dit and the surface photovoltage will be described. FIG. 1 is a diagram for explaining generation of a surface potential due to oxide film charge. A surface potential is generated on the wafer surface by the oxide film charge, and an electric field is formed. In other words, on the wafer surface, as shown in FIG.
A depletion layer (space charge layer) is formed.

There is a natural oxide film on the surface of a semiconductor wafer that has not been heat-treated, and a thermal oxide film exists on a wafer that has been heat-treated. In any case, the oxide film has an oxide film charge. This charge is often positive for p-type wafers and negative for n-type wafers. Therefore, in many cases, a surface potential is generated on the wafer surface, and there is a depletion layer in which majority carriers are depleted in proportion to the oxide film charge. For example, in the case of a p-type silicon wafer having a resistivity of 0.1 Ωm used in LSI manufacturing, when a thermal oxide film is formed, a positive charge (fixed oxide film charge) exceeding 1 mC / m ² is naturally generated in the oxide film. As a result, a depletion layer having a width of about 1 μm is formed. When this is irradiated with argon laser light, the optical carrier is almost generated in the depletion layer because the light transmission length is almost equal to the depletion layer width. As a result, the minority carriers (electrons having negative charges in this case) of the photocarriers are absorbed by the positive oxide film charges, and almost all of them remain inside the depletion layer. Conversely, the majority carriers (holes having positive charges in this case) among the optical carriers are repelled by the positive oxide film charges and are rejected to the outside of the depletion layer.

The entry of minority carriers into the depletion layer means the generation of a surface photovoltage. Because the minority carriers penetrate into the depletion layer, the effect of the positive oxide charge is negated, and the surface potential indicated by φs in FIG. 1 decreases slightly, and this is converted into a surface light voltage by an external circuit. It is because it is observed. However, the electrode that should detect the surface photovoltage cannot be in contact with the wafer surface because it is separated by an oxide film that is an electrically insulating material.
Therefore, the surface photovoltage is measured by capacitive coupling through the oxide film.

The inside of the depletion layer has a surface potential (in other words,
There are few majority carriers (positive holes in this case) depending on the amount of oxide film charge). Therefore, the minority carrier (the minority carrier portion of the optical carrier) that has entered the depletion layer has a long life because there is no partner to recombine. That is, the once generated surface light voltage does not readily decrease.

However, when the interface trap density Dit is high, the minority carriers easily recombine with holes thermally generated through the interface traps and disappear. As a result, the surface photovoltage quickly decreases. Therefore, the interface trap density Dit can be evaluated by measuring the decay time of the surface photovoltage.

Next, the relationship between the depletion layer characteristics and the surface photovoltage will be described. The (small) photocarriers sucked into the depletion layer not only disappear via interface traps, but also have other recombination processes. One of them is recombination via crystal defects in the depletion layer. That is, when foreign matter such as metal atoms or crystal defects exist in the depletion layer, recombination centers are formed, and the excess carriers are combined with holes inside the wafer and disappear through them. As described above, the decay time of the surface photovoltage reflects the state of the semiconductor surface.

Therefore, it can be said that the longer the decay time of the surface photovoltage, the closer the crystallinity of the wafer surface is to perfection.

The present invention is configured according to the above principle, and an embodiment of the present invention will be specifically described below with reference to FIG. FIG. 2 is a diagram showing a semiconductor surface evaluation apparatus using a surface photovoltage. On the grounded metal stage 1,
A sample wafer 2 is arranged, on which a laminated body of a mylar sheet 3, a transparent electrode 4, a glass plate 5, and a transparent electrode 6 is arranged. The uppermost transparent electrode 6 is grounded. Above this laminate, a lens 7
, An ND filter 8, and a mirror 9 are arranged so as to be located on an optical axis perpendicular to the surface of the sample wafer 2, and an AOM (acousto-optic modulator) is provided on the reflected optical axis of the mirror 9. 11 and an Ar laser light source 10 are arranged.
The laser light source 10 emits, for example, an argon laser beam having a wavelength of 488 nm. The AOM 11 modulates the argon laser light in synchronization with the pulse from the pulse generator 13 to obtain a pulse having a pulse width of 100 μs or less, for example. Accordingly, the laser light emitted from the argon laser light source 10 is reflected by the mirror 9 via the AOM 11, then its optical axis is changed to a direction perpendicular to the surface of the sample wafer 2, and the ND filter 8 and the lens 7 Then, the sample wafer 2 is irradiated via the transparent electrode 6, the glass plate 5, the transparent electrode 4 and the mylar sheet 3.

The surface light voltage generated on the surface of the sample wafer 2 by the incidence of the pulse light is detected by the lower transparent electrode 4 facing the sample wafer 2 via the insulating mylar sheet 3. Is input from the transparent electrode 4 to the preamplifier 15 and amplified, and then the oscilloscope 1
4 is input. On the other hand, the oscilloscope 14 receives a pulse signal from the pulse generator 13. The output pulse of the pulse generator 13 is also input to the AOM driver 12, and the driver 12 drives the AOM 11.

Next, the operation of this embodiment will be described.
The pulse generator 13 outputs a predetermined pulse signal to the AOM driver 12, and the AOM 11 modulates the laser light from the laser light source 10 with the pulse signal, and has a pulse width of 100 μs.
The following predetermined rectangular wave pulse is obtained. This rectangular wave pulse is reflected by the mirror 9, attenuated to an appropriate amount by the ND filter 8, converged by the lens 7,
Incident on the surface of.

By the irradiation of the pulse light, a light voltage is generated on the surface of the sample wafer 2 and is immediately attenuated. FIG. 3 is a voltage waveform diagram (output diagram of an oscilloscope emission line) showing generation and attenuation of the surface light voltage. 3 is obtained by converting a voltage signal waveform into a digital signal, storing the converted signal, and outputting the stored waveform to a printer. This voltage waveform shows that when a pulse light having a pulse width of 100 μs is irradiated, a light voltage is generated on the surface of the sample wafer at the rise of the pulse, and the light voltage is attenuated at the fall of the light pulse. . Note that in the voltage waveform as shown in FIG.
The time to become / e is defined as the surface light voltage decay time.

After this surface light voltage is detected by the transparent electrode 4, it is inputted to a preamplifier 15, amplified by the preamplifier 15, for example, by a factor of 100 and then output to an oscilloscope 14.
Will be displayed. The pulse signal of the pulse generator 13 is also input to the oscilloscope 14, and the oscilloscope 14 displays the pulse signal and the optical voltage simultaneously.

As described above, the decay time of the surface photovoltage reflects the state of the semiconductor surface. Therefore, the longer the decay time of the surface photovoltage, the closer the crystallinity of the wafer surface is to perfection. Thus, based on the decay time of the light voltage displayed on the oscilloscope 14, the state of the surface of the sample wafer 2, that is, the crystallinity can be evaluated.

In order to excite an optical voltage, it is necessary to excite an optical carrier in a semiconductor by light irradiation. Moreover, at this time, the generation region of the optical carrier must be near the semiconductor surface. This can be expressed as follows using mathematical expressions.

ΑL _n >> 1 (1)

In the above equation, α is the light absorption coefficient, which is determined by the type of semiconductor and the wavelength of light. L _n is the diffusion length of the minority carrier (in this case, the electron). 1 / α has a unit of length,
If this is Lα (= 1 / α), it means the transmission length of light in the semiconductor. Rewriting equation (1), L _n >> Lα ... (2) That is, the light transmission length needs to be sufficiently shorter than the minority carrier diffusion length, but in the case of a p-type silicon crystal, L
Except in the special case where n is several μm or less, blue light, for example, argon laser light (having a wavelength of 488 nm) satisfies this condition.

The conditions given by equation (1) or (2) limit the photocarrier generation to the surface area, but this makes the photocarrier unaffected by the volume lifetime. this is,
When the transmission length of light is sufficiently smaller than the diffusion length, it is apparent from the fact that light cannot distinguish the length of the diffusion length. Therefore, as long as short-wavelength light is used, the decay time of the surface photovoltage is not affected by the volume lifetime, unlike the μ-wave photoconductive decay method.

Next, the time for irradiating light to generate an optical carrier will be described. As is well known in the CV method, the charge density in the depletion layer on the semiconductor surface does not respond to a rapid voltage change. That is, the charge inside the depletion layer does not leak out when the voltage changes rapidly at about 10 kHz or more. Therefore, if the light irradiation time is about 100 μs or less, the minority carriers that have entered the depletion layer do not go out of the depletion layer and recombine with holes inside the depletion layer.

Therefore, it is desirable that the duration of the light pulse for irradiating light be about 100 μs or less. Conversely, if the pulse width exceeds 100 μs, the charge inside the depletion layer leaks from the wafer surface toward the inside, finally reaches the wafer surface, and enters the depletion layer on the back surface of the wafer. As a result, the effect of the back surface of the wafer is mixed, and the analysis becomes difficult.

The measurement of the surface photovoltage can be performed by capacitive coupling. Therefore, the surface light voltage detection electrode may be arranged on the wafer through a gap, and in principle, non-contact measurement is possible for the surface light voltage measurement.

However, if the electrodes are arranged on the wafer through the gap, the apparatus becomes expensive. In order to reduce the device cost, the detection electrode may be brought into contact with the oxide film via a thin polymer insulating film or the like. In general, the oxide film is not destroyed by the contact with the polymer insulating film. Therefore, the oxide film may be brought into contact with the oxide film via a polymer insulating film or the like in a clean environment. In general, the oxide film is not destroyed by the contact of the polymer insulating film. Therefore, even if the polymer insulating film is brought into contact with the oxide film in a clean environment, no destructive measurement is performed.

It is necessary that the electrode for detecting the surface light voltage is transparent to light or partially transmits light such as a metal net. Otherwise, light cannot be applied to the wafer. The back surface of the wafer must be brought into contact with a metal or conductive material table as large as possible, and the metal or conductive table needs to be electrically grounded. When light is applied from the back surface, the wafer may be irradiated with light from the back surface of the wafer using a transparent electrode.

The surface light voltage signal is about 1 mV or less when the optical power is about 20 μW, but this voltage signal is amplified about 100 times by a wide area amplifier, and the output of the amplifier is introduced to a wide band oscilloscope. For example, when the light irradiation is stopped after 100 μs, the surface light voltage that once rises sharply decreases.

The surface photovoltage decay time indicates how long the minority carriers of the optical carriers that have entered the depletion layer survive in the depletion layer. Conversely, the surface photovoltage decay time is the time when the photocarrier disappears due to recombination. The disappearance of the optical carrier includes a state where the charge is effectively neutralized. As described above, the recombination of minority carriers depends on the level of the interface trap density Dit and crystal defects in the depletion layer. However, the amount of majority carriers (holes) penetrating into the depletion layer from inside the wafer is also considered. Dependent.

In the 1950's, it was predicted that the surface photovoltage decay time would give the volume life time of the wafer (EO Johnson 1957: Journal of App.
liedPhysics, volume 28, Number 11 pp. 1349-1353).
However, since the decay time of the surface light voltage generally depends on the wavelength of the excitation light, it is impossible to interpret an experiment in which the wavelength is not specified. In addition, when the wavelength is not short, the surface lifetime and the volume lifetime are measured in parallel, as in the case of the μ-wave photoconductive decay method, so that an accurate volume lifetime cannot be obtained. Therefore, the interpretation of the surface light voltage decay characteristic shown in FIG. 3 is different from the interpretation of the above-mentioned literature, and is novel based on the research results of the present inventors. It has the effect that it has.

Note that there are various methods of generating pulsed light in addition to the above method. For example, pulsed light can be generated by driving a semiconductor laser and a light emitting diode with a rectangular short pulse current. Alternatively, short-wavelength light may be selected from a light source such as a halogen lamp using an appropriate filter, and short-pulse light may be formed using a mechanical shutter.

[0058]

According to the present invention, the state of the semiconductor crystal surface defined by the depletion layer and the interface can be easily evaluated without being disturbed by the carrier lifetime inside the crystal. Since the present invention can be carried out without directly contacting the metal electrode with the wafer having the oxide film, it is a non-destructive inspection, and it is possible to inspect all the wafers.

Further, according to the present invention, it is possible to detect the surface potential, that is, the presence of the oxide film charge and the sign thereof. That is, when the surface potential is positive, that is, when the oxide film charge is positive, the surface photovoltage appears in the negative direction, and when the sign of the oxide film charge is negative, the sign of the surface photovoltage becomes positive.
The sign of the charge and the surface potential can be determined. Therefore, when the lifetime of a wafer is to be measured by the μ-wave photoconductive decay method, it is possible to separately grasp whether or not the length of the decay time is due to the influence of the surface potential. As a result, it is possible to correctly interpret the measurement result obtained by the μ-wave photoconductive decay method.

In particular, in the case of a silicon crystal, a positive oxide film charge in the case of a p-type, and a negative oxide film charge in the case of an n-type substantially contribute to the generation of a surface photovoltage. Therefore, the conductivity (p or n) of the wafer can be determined from the sign of the surface light voltage. In addition, real-time measurement is possible, and the result can be determined immediately.

[Brief description of the drawings]

FIG. 1 is an energy level diagram for explaining a mechanism for generating a surface photovoltage.

FIG. 2 is a diagram showing a semiconductor surface evaluation apparatus according to an embodiment of the present invention.

FIG. 3 shows an optical voltage waveform detected by the method of the embodiment of the present invention, which is an output diagram of an oscilloscope emission line, and shows an attenuation characteristic of a surface light voltage excited by pulsed light.

[Explanation of symbols]

 1: Metal stage 2: Sample wafer 3: Mylar sheet 4, 6: Transparent electrode 5: Glass plate 10: Ar laser light source 14: Oscilloscope

Claims (11)

[Claims] A laser light is modulated into a pulse light by a pulse signal, the pulse light is irradiated on a surface of a semiconductor crystal, a generated surface light voltage is detected by capacitive coupling, and based on the decay time of the surface light voltage. A semiconductor surface evaluation method using a surface photovoltage, comprising evaluating a crystal characteristic of a surface of the semiconductor crystal. 2. The semiconductor surface evaluation method according to claim 1, wherein a transmission length of the laser light is shorter than a diffusion length of minority carriers in the semiconductor crystal. 3. The method according to claim 1, wherein the laser beam has a wavelength of 488 nm. 4. The method according to claim 1, wherein the pulse light is a rectangular pulse having a pulse width of 100 μs or less. 5. The surface light voltage is such that two electrodes are arranged above the semiconductor crystal via an air gap or an insulating film, an electrode far from the semiconductor crystal is grounded, and a closer electrode is 5. The method according to claim 1, wherein the detection is performed by extracting an optical voltage from the device to detect a capacitive voltage.
4. A semiconductor surface evaluation method using the surface photovoltage described in the item 4. 6. The surface light voltage according to claim 1, wherein the electrode is a transparent electrode, and the pulsed light is incident on the semiconductor crystal via the transparent electrode. Semiconductor surface evaluation method. 7. An apparatus for evaluating crystal characteristics of a surface of a semiconductor crystal, comprising: a light source that emits a laser beam having a light transmission length shorter than a diffusion length of a minority carrier in the semiconductor crystal; A pulse light irradiating means for modulating light into light and irradiating the surface of the semiconductor crystal with the pulse light; a surface light voltage detecting means for detecting a surface light voltage generated in the semiconductor crystal by capacitive coupling; and a decay time of the surface light voltage. Means for evaluating the crystal characteristics of the surface of the semiconductor crystal. 8. The apparatus according to claim 7, wherein the laser beam has a wavelength of 488 nm. 9. The device according to claim 7, wherein the pulse light is a rectangular wave pulse having a pulse width of 100 μs or less. 10. The surface light voltage detecting means has two electrodes disposed above the semiconductor crystal via a gap or an insulating film, and detects a light voltage from an electrode closer to the semiconductor crystal. The semiconductor surface evaluation apparatus according to any one of claims 7 to 9, wherein the electrode that is taken out and farther away is grounded. 11. The surface light voltage according to claim 7, wherein the electrode is a transparent electrode, and the pulsed light is incident on the semiconductor crystal via the transparent electrode. Semiconductor surface evaluation device.