JP5720560B2 - Semiconductor substrate evaluation method - Google Patents

Description

  The present invention relates to a method for evaluating a silicon substrate using a cathodoluminescence (CL) method.

Recently, higher quality silicon wafers are required for further miniaturization and higher performance of semiconductor devices. As a method for evaluating this silicon wafer, a wide variety of methods are known as physical and chemical analysis, and the range is extremely wide.
Among such evaluation methods, the electrical property evaluation is a method close to an actual device, and is also promising from the viewpoint of sensitivity.

  Known methods for evaluating the electrical characteristics of a silicon wafer as a device material include GOI (Gate Oxide Integrity), lifetime, and DLTS (Deep Level Transient Spectroscopy). In particular, GOI is an important evaluation method with sensitivity to COP and oxygen precipitation existing in CZ silicon crystals. However, in this GOI, the outermost surface of a silicon wafer is oxidized by about 20 nm, and an electrode is formed on this to evaluate the dielectric breakdown characteristics.

  As a method for examining the insulating film quality in detail, there is a method using X-rays as disclosed in Patent Document 1. This is a method for evaluating the interface / crystal quality by changing the surface potential by irradiating an electromagnetic wave such as X-rays to generate carriers and evaluating the amount of trapped carriers. For this reason, the sample requires an insulating film, which is not sufficient for measurement and evaluation of the PN junction structure.

Although GOI can evaluate the surface layer of a silicon wafer, there is a leakage current characteristic as another method for evaluating a device active region (near the surface).
As shown in FIG. 9, a PN junction is formed by forming a diffusion layer 53 in which a dopant having a conductivity type different from that of the substrate is diffused at a position where the oxide film 52 on the surface of the silicon substrate 51 is removed. This is an evaluation method using a leakage current detected by applying a reverse voltage from the electrode 54.

  A specific example will be described. Here, a case where N type is diffused into P type will be described. When a + electric field is applied to the N type region, a depletion region (space charge region) is formed toward the P type region. Due to the presence of defects such as heavy metals in this depletion region, carriers are generated and detected as current (leakage current) by the applied voltage (see Non-Patent Document 1).

Normal leakage current is very small, and a shielded system is essential for measurement.
The leak current value is also affected by the size of the depletion region (space charge region). That is, if the substrate resistance is large, the depletion region is correspondingly large, and accordingly, the leakage current tends to be large.
As described above, the junction leakage current measurement is an effective means, but since the leakage current value is small, it is easily affected by the parasitic resistance. In addition, from the viewpoint of material evaluation, it is characterized by being easily affected by the resistivity of the silicon substrate.

  In addition, the cause of electrical characteristic deterioration cannot be determined only from electrical characteristic measurement, and physical analysis is necessary. Conventionally, for such analysis, a technique using an electron beam typified by TEM (Transmission Electron Microscopy) as a probe is used, but mainly for morphological observation and elemental analysis. On the other hand, spectroscopic techniques represented by FTIR (Fourier-Transform Infrared Spectroscopy) method, Raman, photoluminescence (PL) method and the like are important information other than elemental information, for example, bonding state of organic material, crystal stress Although information such as strain, defect, and carrier concentration can be obtained, it is not always sufficient in terms of spatial resolution.

  Since the cathode luminescence (CL) method uses an electron beam as a probe, the stress / strain distribution, defect distribution, and carrier distribution of the sample can be evaluated with high spatial resolution. CL refers to light emission in the ultraviolet, visible, and near infrared regions that is emitted when an electron beam is irradiated onto a sample.

  FIG. 10 shows a schematic diagram of electron beam transition. Characteristic X-rays are derived from inner electron transitions and mainly reflect elemental information, but CL corresponds to a transition from the vicinity of the bottom of the conduction band to the vicinity of the top of the valence band. Reflect the nature of. Reflecting the properties of this crystal is a major point, and due to the recent improvement in semiconductor substrate characteristics and miniaturization of semiconductor devices, the cause of defects in semiconductor devices is not necessarily regarded as a form, it is caused by so-called point defects The fact that the difference in crystallinity can be detected while maintaining the crystallinity is a major feature and advantage of CL.

  In this CL method, the light emission mechanism differs depending on the material, but in the case of a semiconductor, there are three types: (1) generation of electron / hole pairs, (2) carrier diffusion, and (3) emission recombination. In the case of silicon, a TO phonon line corresponding to a band gap (about 1.1 eV) is strongly observed. This is an interband transition with phonon emission because silicon is an indirect transition semiconductor. When crystal defects and impurities form energy levels in the band gap, light emission through the defects and impurities occurs in addition to interband transition light emission.

  In general, an SEM (Scanning Electron Microscopy) is used as an electron beam source as a device, and a detector / spectrometer for detecting light emission from the sample, and stage cooling for obtaining light emission intensity by suppressing lattice vibration, etc. A mechanism is needed. As can be seen from the outline of such an apparatus using an SEM as an electron beam source, the features are that it can be compared with an SEM image, an emission spectrum of a wide range of wavelengths is obtained, high resolution, by changing the acceleration voltage, Depth analysis is possible.

JP 09-162253 A

Chapter 2 of VLSI Process Control Engineering, Hideki Tsuya (Maruzen, 1995)

As described above, in the evaluation of a semiconductor substrate, it is almost impossible to identify a defect only by electric characteristics. Moreover, there is a limit in specifying the position of the defect.
On the other hand, CL, which is a physical analysis, is an evaluation using light emission characteristics using an electron beam, and has a problem that the measurable area is extremely small. For example, even if measurement is performed in a bare wafer state, the position of a defect cannot always be accurately captured in a current high-quality wafer having a low defect density.

  The present invention has been made in view of the above problems, and an object thereof is to provide a semiconductor substrate evaluation method capable of efficiently specifying and evaluating a defect position localized in a semiconductor substrate. To do.

  In order to achieve the above object, the present invention is a method for evaluating a semiconductor substrate, wherein a PN junction is formed on the surface of the semiconductor substrate, a leakage current in the PN junction is measured, and the measurement result of the leakage current is used. Select a cell that is expected to have defects, irradiate the selected cell with an electron beam, scan it, create a distribution of emission intensity for each detected emission wavelength, There is provided a semiconductor substrate evaluation method characterized in that an in-plane position of a defect existing in a cell is specified and the defect is evaluated.

  With such an evaluation method, first, by measuring the leakage current and selecting a cell based on the result, the position where the defect exists can be specified within a certain range from the entire surface of the semiconductor substrate. Then, after specifying a certain range in this way, the cell is further irradiated and scanned with an electron beam to create a light emission intensity distribution for each light emission wavelength, so that the in-plane position of the defect can be specified efficiently. Can be evaluated.

At this time, when the distribution of the emission intensity is created, it can be created for each known emission wavelength specific to the defect.
In this way, the in-plane position of the known defect in the cell and the defect type can be identified very simply.

  Further, when evaluating the defect with the specified in-plane position, the specified in-plane position is irradiated with an electron beam while changing the acceleration voltage, and the position in the depth direction is determined from the detected emission wavelength and intensity. Can be identified.

  The in-plane position of the defect is specified, and in this way, the position in the depth direction of the defect can be easily specified.

  As described above, according to the semiconductor substrate evaluation method of the present invention, even if a defect is localized in the semiconductor substrate, the defect in-plane position can be identified and evaluated more efficiently than in the past. Can do.

It is a schematic diagram of a leakage current measuring element that can be used in the present invention. It is a measurement figure which shows an example of distribution of a leakage current value. It is explanatory drawing which shows an example of the apparatus for detecting the light emission from the cell which irradiated the electron beam. It is a graph which shows the relationship between the amount of carriers inject | poured with the electron beam, and the diffusion depth for every acceleration voltage. It is a measurement figure which shows an example of in-plane distribution of emitted light intensity. It is a graph which shows the measurement result of the emitted light intensity and emitted light wavelength in an Example. 10 is an explanatory diagram showing measurement locations in a cell in Comparative Example 2. FIG. 6 is a graph showing measurement results of emission intensity and emission wavelength in Comparative Example 2. It is a schematic diagram of the element for leak current measurement in the conventional method. It is a principle figure of the measurement by a cathodoluminescence (CL) method.

Hereinafter, the semiconductor substrate evaluation method of the present invention will be described in detail as an example of an embodiment with reference to the drawings. However, the present invention is not limited to this.
As a procedure of the method for evaluating a semiconductor substrate of the present invention, for example, a semiconductor substrate to be evaluated is prepared (Step 1), a PN junction is formed on the semiconductor substrate, leakage current is measured (Step 2), and leakage current is measured. Based on the measurement result, a cell that is expected to have a defect is selected (step 3), the cell is irradiated with an electron beam and scanned to create a light emission intensity distribution for each emission wavelength (step 4). Based on the intensity distribution, the in-plane position of the defect is specified and other evaluation of the defect is performed (step 5).

Hereinafter, each process is explained in full detail.
(Step 1: Preparation of evaluation target semiconductor substrate)
First, a semiconductor substrate to be evaluated is prepared.
Although a silicon wafer is prepared here, it is not particularly limited. Moreover, there is no problem whether the substrate to be evaluated is a polished wafer or an epitaxial wafer. Further, by appropriately devising the measurement structure, an SOI wafer can be obtained. In the present invention, the form to be evaluated is not particularly limited.

(Process 2: PN junction formation and leakage current measurement)
A PN junction is formed on the prepared silicon wafer to produce a leakage current measuring element as shown in FIG. The measurement element includes a silicon wafer 1, a diffusion layer 3 formed inside the silicon wafer 1, in which a dopant having a conductivity type different from that of the wafer is diffused, an electrode 4 on the diffusion layer 3, and a diffusion layer 3. It consists of an oxide film 2 covering the other surface. A PN junction is formed between the diffusion layer 3 and the silicon wafer 1.

  As a method for manufacturing the leakage current measuring element as shown in FIG. 1, first, an oxide film is formed on the wafer surface. This oxide film is a mask for subsequent dopant diffusion, and a thermal oxide film may be formed or a CVD oxide film may be deposited. The thickness may be any thickness that can mask the dopant to be deposited thereafter. In general, the thickness is preferably 500 nm or more. This is because the dopant diffuses even in the oxide film. When applying a CVD oxide film, especially in the case of plasma CVD, attention should be paid to charge damage caused by plasma.

Next, a window opening pattern is formed on the oxide film by photolithography. Etching of the oxide film may be dry etching or wet etching based on hydrofluoric acid.
Although dry etching can process fine patterns, attention should be paid to the previous plasma damage. On the other hand, wet etching does not cause plasma damage, but is not suitable for processing fine patterns.

  The pattern itself is not particularly limited, but the window opening portion is finally considered so that the defect can be most efficiently evaluated in consideration of the range of the PN junction, the electron beam irradiation range, the working time, the cost, and the like. The size, number, etc. can be determined.

  Then, after completing the opening of the window to the oxide film, the dopant is diffused. A diffusion layer is formed by diffusing a dopant having a conductivity type different from that of the silicon wafer to form a PN junction. The dopant itself may be diffused by various methods such as ion implantation, glass deposition, and coating diffusion. Since the PN junction depth depends on the annealing conditions, the time is adjusted so that the desired depth is obtained in the preliminary experiment.

Further, an electrode made of aluminum, polycrystalline silicon, or the like can be formed on the PN junction.
Alternatively, if the outermost layer after diffusion has a high concentration of about 1 × 10 ²⁰ atoms / cm ^3, there is an advantage that the outermost layer of the diffusion layer can be used as an electrode as it is without specially forming the electrode.

  As another method of forming a PN junction, a dopant can be diffused without forming an oxide film, a pattern can be formed by photolithography, and a PN junction having a MESA structure can be formed by etching or the like.

Then, the leakage current is measured using the leakage current measuring element thus manufactured.
A probe is brought into contact with an electrode or the like formed on the PN junction, and a reverse voltage is applied to measure a leakage current. Measure the leakage current for each cell and summarize the measurement results. For example, a leakage current value distribution is created over the entire wafer surface.
FIG. 2 shows an example of the distribution of leakage current values. The level of the leakage current value is represented by shading. Here, there is a distribution in which the leakage current values repeat concentrically. The density range represented by each color is not particularly limited, and can be determined as appropriate according to the target evaluation accuracy, cost, etc. according to the request from the customer.

(Process 3: Cell selection)
A cell that is expected to have a defect is selected from the distribution of leakage current values created as described above.
More specifically, for example, a cell having a high leak current value of a certain value or more can be selected. The leak current value that serves as a reference for this selection is not particularly limited, and can be appropriately determined according to the target evaluation accuracy, cost, and the like. Here, the cell located on the right side is selected as shown by the dotted line in FIG.
The selected cell is cut out and sent to the next process.

(Step 4: Irradiation of electron beam and creation of emission intensity distribution)
Next, the selected cell is scanned by irradiating an electron beam to the PN junction interface or the depletion region, and a distribution of emission intensity is created for each emission wavelength with respect to the detected emission.
First, as described above, light emission is detected by scanning the inside of a cell with an electron beam by the CL method. At this time, light emission can be detected using, for example, an inspection apparatus as shown in FIG. The inspection apparatus includes a stage with a cooling mechanism, a filament for irradiating an electron beam, and a detector for detecting light emission from the cell.
In the measurement, a cell with a PN junction is placed on a stage with a cooling mechanism, put in a vacuum chamber (not shown), kept in a vacuum state, and cooled sufficiently. The cell is irradiated with 30 keV, and light emission from the cell is detected by a detector. Thereby, the emission spectrum can be acquired.

At this time, the carriers injected by the electron beam are diffused according to the acceleration voltage (Catholomics of Microscopy of Inorganic Solids, B. G. Yacobi and D. B. Holt p61 (1999) Plenum Press / NonL). .
In FIG. 4, the horizontal axis represents the diffusion depth and the vertical axis represents the amount of carriers injected by the electron beam for each acceleration voltage. The electron beam can be narrowed down to several μm, but the carriers generated in the semiconductor material diffuse and show a distribution of shape as shown in FIG. The evaluation method of the present invention evaluates the range in which carriers are distributed in FIG.
For example, when accelerated at 30 keV, it is known that silicon diffuses to about 10 μm with a peak of about 4 μm in depth.

In the irradiation with the electron beam, the acceleration voltage and the like can be appropriately determined according to the range to be evaluated.
In addition, the cooling temperature of the cell is not particularly limited. For example, if the cell is cooled to 77K or lower, further 30K or lower, it is not easily affected by lattice vibration, and the emission wavelength does not become broad, so that stronger light emission can be obtained. it can.

  Then, regarding the detected emission, a distribution of emission intensity is created for each emission wavelength. At this time, it is also possible to create a distribution of emission intensity for every detected emission wavelength, but in specifying the position of the defect in the cell, etc., the distribution of the emission intensity for each emission wavelength specific to the known defect Is extremely simple and efficient.

Several kinds of characteristic emission wavelengths when using silicon as a material have already been reported, and Table 1 shows examples. Such information on the known emission wavelength may be used. Of course, it is not limited to these.
FIG. 5 shows an example of in-plane distribution of emission intensity. FIG. 5 shows the distribution of emission intensity in the cell for the C line (combination of interstitial C and interstitial O) (emission wavelength: 1570 nm) in Table 1.

(Process 5: Identification of in-plane position of defect, other evaluation of defect)
Next, the in-plane position of the defect existing in the cell is specified based on the generated emission intensity distribution.
By scrutinizing this distribution, it is possible to identify locations where the emission is strong, that is, locations where defects are thought to be localized. Then, it is possible to identify which emission wavelength the distribution having a portion having a strong light emission is, identify what kind of defect the emission wavelength is associated with, and identify the defect type.
In addition, the specific value of the light emission intensity used as the specific reference | standard of the position of a defect is not specifically limited. What is necessary is just to determine suitably according to the precision etc. which make it a target.

  In the example shown in FIG. 5, it can be confirmed that there is a portion that emits light strongly at a position surrounded by a dotted line in the upper right portion. 5 is related to the C line (composite of interstitial C and interstitial O) (emission wavelength: 1570 nm) in Table 1 as described above. It can be evaluated as a defect.

In addition to the in-plane position of the defect, other evaluations can be performed.
For example, at the in-plane position specified as described above, the electron beam is irradiated with various acceleration voltages, and the position in the depth direction is specified from the wavelength and intensity of each light emission detected at that time. be able to. Estimate to some extent about the position in the depth direction of the defect from the acceleration voltage of the irradiated electron beam and the relationship as shown in FIG. 4 (the relationship between the acceleration voltage, the amount of carriers, and the diffusion depth) in identifying the in-plane position of the defect I can. However, in specifying the depth position more accurately, the electron beam can be irradiated while changing the acceleration voltage as described above, and the depth can be specified from the irradiation result.

  Of course, not only the position in the depth direction but also other evaluations can be performed. For example, it is possible to study a more detailed defect configuration by acquiring a spectrum over the entire wavelength region.

EXAMPLES Hereinafter, although an Example and a comparative example are shown and this invention is demonstrated more concretely, this invention is not limited to these.
(Example)
As a semiconductor substrate to be evaluated, a silicon wafer having a P-type conductivity, a diameter of 200 mm, and a crystal orientation <100> was used. Boron was used as a dopant for making the wafer P-type, and the substrate resistance was 10 Ω · cm.

The silicon wafer was pyrooxidized at 1000 ° C. to form a 1 μm oxide film. After that, photolithography is performed using a mask in which a large number of 0.5 mm square patterns are arranged, and windows are etched into the oxide film with buffered HF to form 0.5 mm square openings in the oxide film at intervals of 10 mm. did. Phosphor glass was deposited on this wafer using POCL ₃ as a raw material, and subsequently, nitrogen annealing was performed at 1000 ° C. for 2 hours, and then the phosphorus glass was removed with HF to form a PN junction. At this time, the diffusion depth of phosphorus was about 2 μm.
Further, an electrode was formed, and a plurality of leak current measuring elements as shown in FIG.

This wafer was placed on a prober that can move in the horizontal direction, a voltage of 3 V was applied, and the leakage current was measured at 900 points in the plane to obtain in-plane distribution data of the leakage current. An in-plane distribution similar to that in FIG. 2 was obtained.
Thereafter, a cell having a high leakage current (here, 5 × 10 ⁻¹¹ A or more) was selected and cut out from the in-plane distribution.

  After that, the selected and cut-out cell is placed as a sample on a stage with a cooling mechanism of the inspection apparatus, put in a vacuum chamber and kept in a vacuum state, and after the sample has been sufficiently cooled to 10K, The spectrum was measured by irradiating with an electron beam acceleration voltage of 30 keV to obtain the emission intensity distribution. In addition, when the distribution was created at each wavelength in Table 1, as shown in FIG. 5, the distribution of the C line in Table 1 whose emission wavelength is 1570 nm, where the emission intensity is high (location surrounded by the dotted line in FIG. 5). Was found to exist. It is considered that there is a defect at this location.

Then, the spectrum was measured at this point with an acceleration voltage of 10 keV, and the details were examined.
As a result, the position in the depth direction is considered to be about 1.5 μm, and as shown in FIG. 6 where the horizontal axis represents the emission wavelength and the vertical axis represents the emission intensity, the interband transition emission (TO line: emission wavelength). In addition to C line, H line (emission wavelength: 1340 nm) is also observed except for 1130 nm and 1170 nm), and defects related to oxygen and carbon are present. These defects become point defects and are quasi-depleted in the depletion region. It became clear that he was making a place.

(Comparative Example 1)
A silicon wafer similar to that of the example was prepared, an arbitrary part was cut out, and an spectrum was measured by irradiating an arbitrary part of the cut out surface with an electron beam in the same manner as in the example.

  As a result, except for the TO line, no light was emitted through the defect, and the position of the defect and the defect type could not be specified.

(Comparative Example 2)
A silicon wafer similar to that of the example was prepared, a leak current measuring element was prepared in the same manner, the leak current was measured, and the distribution of the leak current value was obtained. As a result, the same distribution as in FIG. 2 was obtained. And the cell similar to Example 1 was selected and cut out, and the spectrum was measured by irradiating an arbitrary part (FIG. 7) of the cut out cell with an electron beam.

  As a result, as shown in FIG. 8, except for the TO line, a notable spectrum related to defects and the like was not obtained, and the position of the defect and the defect type could not be specified. This is presumably because a portion having a high leakage current was cut out, but a characteristic signal could not be obtained by spectrum measurement because the analysis portion in the joint surface was arbitrary.

  The present invention is not limited to the above embodiment. The above-described embodiment is an exemplification, and the present invention has substantially the same configuration as the technical idea described in the claims of the present invention, and any device that exhibits the same function and effect is the present invention. It is included in the technical scope of the invention.

DESCRIPTION OF SYMBOLS 1, 51 ... Silicon substrate, 2, 52 ... Oxide film, 3, 53 ... Diffusion layer,
4, 54 ... electrodes.

Claims (3)

A method for evaluating a semiconductor substrate, comprising:
A PN junction is formed on the surface of the semiconductor substrate, the leakage current at the PN junction is measured, a cell that is expected to have a defect is selected based on the measurement result of the leakage current, and the selected cell is irradiated with an electron beam Scan and create a distribution of emission intensity for each detected emission wavelength, identify the in-plane position of the defect present in the cell based on the generated emission intensity distribution, and evaluate the defect. A method for evaluating a semiconductor substrate, comprising:   2. The method for evaluating a semiconductor substrate according to claim 1, wherein when the distribution of the emission intensity is created, the emission intensity is created for each known emission wavelength specific to the defect.   When evaluating the defect whose in-plane position is specified, the specified in-plane position is irradiated with an electron beam while changing the acceleration voltage, and the position in the depth direction is specified from the detected wavelength and intensity of emitted light. The method for evaluating a semiconductor substrate according to claim 1, wherein the semiconductor substrate is evaluated.

Images