JP6020073B2 - Method for analyzing crystal defects in semiconductor samples - Google Patents

Description

  The present invention relates to a crystal defect analysis method in a semiconductor sample in which an epitaxial layer having a lower impurity concentration is grown on a semiconductor substrate having a high impurity concentration, and more particularly to silicon carbide (hereinafter referred to as SiC) in which crystal defects are a problem. It is preferable to apply to wide band gap semiconductors such as

  Conventionally, with regard to crystal defects that are likely to affect the device characteristics existing in SiC single crystals, the crystal defects are revealed by etching the surface of the epitaxial layer with KOH (potassium hydroxide), and a transmission electron microscope (TEM) is obtained. The analysis method of clarifying the details of crystal defects by transmission electron microscope analysis is the mainstream (see, for example, Patent Document 1).

JP 2012-116732 A

  However, as shown in FIG. 9, in the SiC semiconductor sample J1 in which the epitaxial layer is grown on the SiC substrate made of SiC single crystal, in the vicinity of the interface between the SiC substrate and the epitaxial layer, for example, inclusion (introduced during crystal growth, There may be a starting point J2a of the crystal defect J2 due to a fine foreign material due to a base material or other materials that induces disorder of the crystal. In order to reduce the crystal defects J2, it is necessary to identify whether the defects were formed in the growth stage of the SiC substrate or in the growth stage of the epitaxial layer. The interface of the epitaxial layer cannot be specified. In addition, even if the thickness of the epitaxial layer is known, the thickness of the epitaxial layer is reduced by etching, and the original surface position is not determined, so it is difficult to specify the interface between the SiC substrate and the epitaxial layer, so that the defect is I can't identify where it is.

  Note that, here, an example of a semiconductor sample made of SiC has been described as an example. However, not only SiC but also semiconductor samples made of other semiconductor materials are used to identify the position of crystal defects. The same problem as above occurs.

  In view of the above points, an object of the present invention is to provide a crystal defect analysis method in a semiconductor sample that can accurately analyze the depth position of crystal defects.

  In order to achieve the above object, in the invention according to claim 1, the observation sample (3) thinned by cutting out the portion of the semiconductor sample (1) containing the starting point (2a) of the crystal defect (2). A step of acquiring a crystal defect image representing a crystal defect by performing observation with a transmission electron microscope on the observation sample, and applying a stimulus to the observation sample to change the impurity concentration from the observation sample. The origin of the crystal defect is determined based on a step of generating a corresponding light, detecting the light to obtain a light intensity mapping image of the observation sample, and a composite image obtained by superimposing the crystal defect image and the light intensity mapping image. And a step of analyzing which of the semiconductor substrate (1a) and the epitaxial layer (1b) is present.

  Thus, a crystal defect image representing a crystal defect and a light intensity mapping image representing a boundary position between the semiconductor substrate and the epitaxial layer are acquired. If a composite image in which these are superimposed is created, it is possible to identify whether the crystal defect is located in the semiconductor substrate or the epitaxial layer from the composite image. Therefore, the position where the crystal defect exists can be specified accurately.

  In addition, the code | symbol in the bracket | parenthesis of each said means shows an example of a corresponding relationship with the specific means as described in embodiment mentioned later.

1 is a perspective view of a semiconductor sample 1 that performs crystal defect analysis described in the first embodiment of the present invention. It is IB-IB 'sectional drawing of FIG. 1A. 1 is a perspective view of a semiconductor sample 1 in which crystal defects 2 are present. It is the elements on larger scale of semiconductor sample 1 which showed the cut-out range of observation sample 3 in Drawing 2A. It is an expansion perspective view of the observation sample 3 after cutting out. It is a block diagram of the analysis system which performs crystal defect analysis. It is a figure showing the transmission electron microscope analysis image used as a crystal defect image. It is a figure showing the light intensity mapping image. It is a figure showing the synthesized image which overlap | superposed the transmission electron microscope analysis image and the light intensity mapping image. It is a block diagram of the analysis system which performs the crystal defect analysis concerning 2nd Embodiment of this invention. It is a block diagram of the analysis system which performs the crystal defect analysis concerning 3rd Embodiment of this invention. It is a figure showing the transmission electron microscope analysis image.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following embodiments, parts that are the same or equivalent to each other will be described with the same reference numerals.

(First embodiment)
A method for analyzing crystal defects in a semiconductor sample according to the first embodiment of the present invention will be described with reference to FIGS.

First, as shown in FIGS. 1A and 1B, an n-type epitaxial layer 1b having a lower impurity concentration than the semiconductor substrate 1a is formed on an n-type semiconductor substrate 1a made of a semiconductor material such as SiC single crystal. A semiconductor sample 1 constituted by growing is prepared. For example, the plane orientation of the semiconductor substrate 1a is (0001), the n-type impurity concentration of the semiconductor substrate 1a is 1 × 10 ¹⁸ to 1 × 10 ¹⁹ cm ⁻³ , and the impurity concentration of the epitaxial layer 1b is 1 × 10 ¹⁵ to 1 ×. 10 ¹⁷ cm ⁻³ .

  Since the SiC semiconductor sample 1 has the crystal defect 2 as shown in FIG. 2A, the crystal defect 2 is revealed by solution etching such as KOH. Among these, as shown in FIG. 2B, for example, a region including the starting point 2a of the crystal defect 2 due to inclusion or the like is selected, and the semiconductor sample 1 is partially cut out by FIB (focused ion beam) processing, whereby the cross-sectional structure of the crystal defect Thin so that you can see. As a result, an observation sample 3 to be observed is prepared for analysis using a transmission electron microscope or the like shown in FIG. 2C. For example, the size of the observation sample 3 is set to a width of 20 μm, a height of 20 μm, and a depth of 1 μm. The crystal defect 2 existing inside is analyzed using the observation sample 3 created in this way.

  Specifically, as shown in FIG. 3, an analysis system having a transmission electron microscope apparatus 10, a first analysis computer 11, a Raman spectroscopic apparatus 12, a second analysis computer 13, and a third analysis computer 14 is prepared.

  And the observation sample 3 is arrange | positioned in the transmission electron microscope apparatus 10, and a transmission electron microscope observation is performed. That is, the observation sample 3 is irradiated with electrons accelerated at a desired acceleration voltage (for example, 200 kV) to transmit the electrons, and the state at that time is imaged by a CCD camera or the like. The imaging result is input to the first analysis computer 11, and a transmission electron microscope analysis image, that is, a crystal defect image is acquired by the first analysis computer 11. FIG. 4 is a diagram showing the transmission electron microscope image. Of the crystal defects 2 from the starting point 2a such as inclusion shown in this figure, those extending in the substantially vertical direction of the substrate are dislocations, and those extending in the substantially horizontal direction of the substrate are dislocations or stacking faults.

  In addition, the observation sample 3 is placed in the Raman spectroscopic device 12 to perform Raman spectroscopic measurement. In Raman spectroscopic measurement, when monochromatic light (laser) is irradiated to stimulate the observation sample 3, scattered light modulated by the scatterer of the observation sample 3 is emitted. In this method, the substance is evaluated based on the obtained spectrum. The representative of scatterers in Raman spectroscopy is phonons, but other examples include plasmons and magnons.

Here, attention is particularly paid to scattering due to carrier concentration: LOPC (LO phonon-plasmon coupling) mode scattering (a scattering spectrum having a Raman shift amount of about 970 cm ^{−1 in} SiC). The spectrum of LOPC mode scattering is a spectrum corresponding to the concentration of impurities contained in the semiconductor material. Therefore, by performing mapping based on the obtained spectrum of LOPC mode scattering, the observation sample 3 can be partitioned into regions having different impurity concentrations, that is, the semiconductor substrate 1a and the epitaxial layer 1b.

  Therefore, the observation sample 3 is irradiated with monochromatic light in the Raman spectroscopic device 12, and the result obtained by observing the scattered light with the spectroscope is input to the second analysis computer 13. Then, the second analysis computer 13 acquires a light intensity mapping image. FIG. 5 is a diagram showing this light intensity mapping image.

  In this way, the transmission electron microscope image acquired by the first analysis computer 11 and the light intensity mapping image acquired by the second analysis computer 13 are sent to the third analysis computer 14, and the two images are superimposed on the third analysis computer 14. By combining them, the composite image shown in FIG. 6 is created. In the composite image created in this way, the crystal defect 2 and its starting point 2a displayed in the transmission electron microscope image and the boundary position between the semiconductor substrate 1a and the epitaxial layer 1b partitioned by the light intensity mapping image are clear. The image shown in Therefore, it is possible to identify whether the starting point 2a of the crystal defect 2 is located in the semiconductor substrate 1a or the epitaxial layer 1b from this synthesized image.

  As described above, using the transmission electron microscope apparatus 10 and the Raman spectroscopic apparatus 12, the transmission electron microscope image showing the crystal defect 2 and the light intensity mapping image showing the boundary position between the semiconductor substrate 1a and the epitaxial layer 1b are shown. And a composite image is created. As a result, it is possible to identify whether the starting point 2a of the crystal defect 2 is located in the semiconductor substrate 1a or the epitaxial layer 1b from this composite image. Therefore, the depth position of the crystal defect 2 can be analyzed accurately. Note that the process of superimposing the two images may be performed by the first analysis computer 11 or the second analysis computer 13 without being performed by the third analysis computer 14.

(Second Embodiment)
A second embodiment of the present invention will be described. In the present embodiment, the Raman spectroscopic measurement is changed to the photoluminescence measurement with respect to the first embodiment, and the other parts are the same as those in the first embodiment. Therefore, only different portions from the first embodiment will be described. .

  As shown in FIG. 7, in this embodiment, an analysis system including a photoluminescence device 15 is prepared in place of the Raman spectroscopic device 12 used in the first embodiment. Then, after preparing the observation sample 3 described in the first embodiment, a composite image is created by the analysis system. That is, transmission electron microscope observation is performed with the transmission electron microscope apparatus 10 as in the first embodiment, and a transmission electron microscope analysis image is acquired with the first analysis computer 11. In addition, photoluminescence measurement is performed by the photoluminescence device 15, and a light intensity mapping image is acquired by the second analysis computer 13.

  Photoluminescence measurement is a measurement in which light is irradiated to give a stimulus to a substance, light (fluorescence) emitted when induced electrons transition to the ground state is observed, and the substance is evaluated based on the obtained spectrum. Is the method. As the light to be irradiated, light having energy larger than energy corresponding to the band gap of the semiconductor is used.

  However, since transmission electron microscope observation for obtaining a crystal defect image irradiates a high-energy electron beam, there is a problem that vacancy defects and the like are easily induced in the observation sample, leading to a decrease in carrier concentration. That is, since the light emission depending on the carrier concentration is weakened, the detection becomes difficult. At this time, attention is paid to the residual impurity concentration of the semiconductor substrate 1a and the epitaxial layer 1b. Since the epitaxial layer 1b is a high-purity crystal compared to the semiconductor substrate 1a, the residual impurity concentration is low. That is, by using 400 to 1000 nm light emission caused by residual impurities that emit light regardless of the carrier concentration, the observation sample 3 can be partitioned into regions having different residual impurity concentrations, that is, the semiconductor substrate 1a and the epitaxial layer 1b, and transmitted electrons. Measurement is possible regardless of whether the microscopic observation or the photoluminescence measurement is performed first.

  Then, the image shown in FIG. 4 is obtained as the transmission electron microscope analysis image, and the same image as FIG. 5 is obtained as the light intensity mapping image by the photoluminescence measurement. Thereafter, the third analysis computer 14 superimposes the transmission electron microscope image acquired by the first analysis computer 11 and the light intensity mapping image acquired by the second analysis computer 13 to create a composite image similar to FIG. . In the composite image created in this way, the crystal defect 2 projected in the transmission electron microscope image and the boundary position between the semiconductor substrate 1a and the epitaxial layer 1b partitioned by the light intensity mapping image are clearly expressed. It becomes an image. Therefore, it is possible to identify whether the starting point 2a of the crystal defect 2 is located in the semiconductor substrate 1a or the epitaxial layer 1b from this synthesized image.

  As described above, a composite image similar to that of the first embodiment is created using the transmission electron microscope device 10 and the photoluminescence device 15. Even using such a composite image, it is possible to identify whether the crystal defect 2 is located in the semiconductor substrate 1a or the epitaxial layer 1b.

(Third embodiment)
A third embodiment of the present invention will be described. In the present embodiment, the Raman spectroscopic measurement is changed to the cathodoluminescence measurement with respect to the first embodiment, and the other parts are the same as those in the first embodiment. Therefore, only different portions from the first embodiment will be described. .

  As shown in FIG. 8, in this embodiment, the Raman spectroscopic device 12 used in the first embodiment is eliminated, and an analysis system that performs cathodoluminescence measurement in the transmission electron microscope device 10 is prepared. Then, after preparing the observation sample 3 described in the first embodiment, a composite image is created by the analysis system. That is, transmission electron microscope observation is performed with the transmission electron microscope apparatus 10 as in the first embodiment, and a transmission electron microscope analysis image is acquired with the first analysis computer 11. Further, cathodoluminescence measurement is performed in the transmission electron microscope apparatus 10, and a light intensity mapping image is acquired by the second analysis computer 13.

  The cathodoluminescence measurement is a measurement method in which an accelerated electron beam is irradiated to give a stimulus to the substance, light (fluorescence) emitted from a sample is observed thereby, and the substance is evaluated based on the obtained spectrum. The light emission at this time is easily measured regardless of whether the observation with the transmission electron microscope or the photoluminescence measurement is performed first because the light emission caused by the residual impurity having a wavelength of 400 to 1000 nm is used regardless of the carrier concentration. Is possible.

  Then, an image shown in FIG. 4 is obtained as a transmission electron microscope analysis image, and an image similar to FIG. 5 is obtained as a light intensity mapping image by cathodoluminescence measurement. Thereafter, the third analysis computer 14 superimposes the transmission electron microscope image acquired by the first analysis computer 11 and the light intensity mapping image acquired by the second analysis computer 13 to create a composite image similar to FIG. . In the composite image created in this way, the crystal defect 2 projected in the transmission electron microscope image and the boundary position between the semiconductor substrate 1a and the epitaxial layer 1b partitioned by the light intensity mapping image are clearly expressed. It becomes an image. Therefore, it is possible to identify whether the starting point 2a of the crystal defect 2 is located in the semiconductor substrate 1a or the epitaxial layer 1b from this synthesized image.

  As described above, while using the transmission electron microscope apparatus 10, the transmission electron microscope observation and the cathodoluminescence measurement are performed in the transmission electron microscope apparatus 10 to create a composite image similar to that in the first embodiment. Even using such a composite image, it is possible to identify whether the crystal defect 2 is located in the semiconductor substrate 1a or the epitaxial layer 1b.

  In addition, since both transmission electron microscope observation and cathodoluminescence measurement can be performed simultaneously in the transmission electron microscope 10, the analysis can be simplified.

(Other embodiments)
The present invention is not limited to the embodiment described above, and can be appropriately changed within the scope described in the claims.

  Further, for example, SiC has been described as an example of a semiconductor material constituting the semiconductor sample, but the present invention can be applied to semiconductor materials such as silicon, gallium nitride, and diamond in addition to SiC. However, it is particularly preferable to apply the present invention to wide band gap semiconductors such as SiC, gallium nitride, and diamond. In a wide band gap semiconductor mainly used for power devices, an element structure differs depending on the withstand voltage, and a semiconductor sample in which the difference in impurity concentration between the semiconductor substrate 1a and the epitaxial layer 1b is small and homoepitaxial growth may be used. In the case of such a semiconductor sample, the semiconductor substrate 1a and the epitaxial layer 1b are almost the same material, so that it is difficult to partition the boundary. Therefore, it is preferable to apply the present invention to a wide band gap semiconductor in which the boundary between the semiconductor substrate 1a and the epitaxial layer 1b is difficult to be divided in this way. The crystal defects that affect the device characteristics are mainly fine crystal defects 2 such as dislocation defects, stacking faults, and inclusions. For this reason, it is preferable to apply the present invention to identify the starting point 2a of the crystal defect 2 and deal with it including these fine crystal defects 2.

  In each of the above embodiments, any of the order of observation (Raman spectroscopic measurement, photoluminescence measurement, cathodoluminescence measurement) for obtaining a transmission electron microscope observation and a light intensity mapping image may be performed first. However, when the light intensity mapping image is acquired by Raman spectroscopic measurement, if the acceleration voltage at the time of electron irradiation in the transmission electron microscope observation is set to a relatively high voltage (specifically, 400 kV or more, for example, 1000 kV), the carrier in the Raman spectroscopic measurement Cannot be visualized. For this reason, when the acceleration voltage is set to a relatively low voltage (for example, 200 kV), either transmission electron microscope observation or Raman spectroscopy measurement may be performed first. However, when the acceleration voltage is set to a relatively high voltage, Raman spectroscopy measurement is performed. Must be performed prior to observation with a transmission electron microscope.

  In each of the above embodiments, the case where the n-type epitaxial layer 1b is formed on the n-type semiconductor substrate 1a has been described. However, the present invention is not limited to this, and the present invention can be applied to the case where the epitaxial layer 1b having a lower impurity concentration is formed on the semiconductor substrate 1a having a higher impurity concentration.

DESCRIPTION OF SYMBOLS 1 Semiconductor sample 1a Semiconductor substrate 1b Epitaxial layer 2 Crystal defect 2a Starting point 3 Observation sample 10 Transmission electron microscope apparatus 11, 13, 14 1st-3rd analysis computer 12 Raman spectroscopic apparatus 15 Photoluminescence apparatus

Claims (10)

In a semiconductor sample for analyzing a depth position of a crystal defect (2) present in a semiconductor sample (1) in which an epitaxial layer (1b) having a lower impurity concentration than the semiconductor substrate is formed on the semiconductor substrate (1a) A crystal defect analysis method,
Creating a thinned observation sample (3) by cutting out a portion of the semiconductor sample containing crystal defects;
Obtaining a crystal defect image representing a crystal defect by performing a transmission electron microscope observation on the observation sample;
Producing a light according to the impurity concentration from the observation sample by giving a stimulus to the observation sample, detecting the light and obtaining a light intensity mapping image of the observation sample;
Analyzing whether the origin of the crystal defect exists in the semiconductor substrate or the epitaxial layer based on a composite image obtained by superimposing the crystal defect image and the light intensity mapping image. A crystal defect analysis method for a semiconductor sample.   2. The method for analyzing crystal defects in a semiconductor sample according to claim 1, wherein the material of the semiconductor sample is any one of silicon carbide, gallium nitride, and diamond.   3. The method for analyzing crystal defects in a semiconductor sample according to claim 1, wherein the crystal defects include any of dislocation defects, stacking faults, and inclusions.   4. The method according to claim 1, wherein the step of acquiring the light intensity mapping image is performed by Raman measurement, and light is used as the stimulus applied to the observation sample, and scattered light is generated from the observation sample. The crystal defect analysis method in the semiconductor sample as described in 1. In the step of obtaining the crystal defect image, the electron irradiation in the transmission electron microscope observation is performed at an acceleration voltage of 400 kV or more,
5. The crystal defect analysis method for a semiconductor sample according to claim 4, wherein the step of acquiring the light intensity mapping image is performed before the step of acquiring the crystal defect image.   The step of obtaining the light intensity mapping image is performed by photoluminescence measurement, and light having an energy larger than a band gap of a semiconductor material constituting the semiconductor sample is used as the stimulus given to the observation sample, and fluorescence is emitted from the observation sample. The crystal defect analysis method for a semiconductor sample according to any one of claims 1 to 3, wherein:   The crystal defect analysis method according to claim 6, wherein the light intensity mapping image is acquired using fluorescence having a wavelength of 400 to 1000 nm caused by impurities among the fluorescence.   The step of obtaining the light intensity mapping image is performed by cathodoluminescence measurement, and fluorescence is generated from the observation sample using accelerated electrons as the stimulus given to the observation sample. The crystal defect analysis method in the semiconductor sample as described in any one. In the step of acquiring the crystal defect image, the transmission electron microscope observation is performed in a transmission electron microscope apparatus,
9. The crystal in a semiconductor sample according to claim 8, wherein in the step of acquiring the light intensity mapping image, the cathodoluminescence measurement is performed simultaneously with the step of acquiring the crystal defect image in the transmission electron microscope apparatus. Defect analysis method.   The crystal defect analysis method according to claim 8, wherein the light intensity mapping image is acquired using fluorescence having a wavelength of 400 to 1000 nm due to impurities among the fluorescence.

Images