JP6310676B2 - Electron beam application apparatus and defect classification method using the same - Google Patents

Description

  The present invention relates to an electron beam application apparatus for observing minute defects generated in a manufacturing process of a device or the like with an electron beam application apparatus and classifying them by elemental composition, and a defect classification method using the electron beam application apparatus.

  A technique for forming a fine pattern such as a semiconductor device, a liquid crystal device, and a hard disk is constituted by a process having a very large number of steps including several hundred processes. If a defect occurs on a fine pattern due to a process defect, a large number of device manufacturing defects may occur.To maintain and improve the yield of device manufacturing, the cause of the defect is continuously identified. Countermeasures need to be implemented. Therefore, the defect of the device is monitored by inspecting the foreign matter and the disconnection of the wiring for each main process. In general, an inspection apparatus uses an optical application apparatus such as a foreign substance inspection or an appearance inspection, but has a limitation in throughput. For this reason, defects are monitored by monitoring the number and density of defects by narrowing down the process to be inspected or by using a spot inspection.

  In addition, defect analysis and classification may be performed by a defect coordinate review device detected by an inspection device. In particular, in a review apparatus using a scanning electron microscope, a high-magnification observation image of a defect and elements are analyzed. In general, the element analysis uses a technique for estimating an element based on a combination of wavelengths of characteristic X-rays generated during electron beam irradiation, and a defect element can be easily estimated without destroying a wafer. Furthermore, estimating the element of the defect is very important in identifying the cause of the occurrence of the defect in the manufacturing process and taking measures to prevent the occurrence of the defect.

  The review apparatus includes a stage for holding a sample, an electron optical column that focuses an electron beam on the sample, a secondary electron detector that converts secondary electrons emitted from the sample into an electrical signal, and the secondary electrons. A scanning electron microscope (SEM) having an image generation system that converts an electrical signal output from a detector into a digital image, and an image processing system that analyzes the coordinates and shape features of an observation object from the digital image Consists of. The review device can acquire inspection data detected by the foreign matter inspection device and the appearance inspection device, and can automatically obtain a high-definition observation image of the foreign matter / defect.

  In Patent Document 1, the review apparatus includes a deflection system that irradiates the electron beam to the coordinates of the observation target, an X-ray detector that converts X-rays emitted from the sample into electrical signals, and an X-ray detector. A waveform generation system that converts the output electrical signal into a digital waveform of the X-ray energy spectrum, a peak detection system that extracts the wavelength of the characteristic X-ray from the digital waveform, and the element is estimated by collating the characteristic X-ray with a database Add element search system. The X-ray detector is a semiconductor detector (such as a silicon detector or a silicon drift detector) and a micro calorimeter. Waveform generation systems that convert an electrical signal output from an X-ray detector into a digital waveform of an X-ray energy spectrum include an energy dispersion type and a wavelength dispersion type. The X-ray counting speed and energy resolution differ depending on the combination of the X-ray detector and the waveform generation system.

  The peak detection system extracts the characteristic X-ray wavelength from the digital waveform of the X-ray energy spectrum. The element search system stores the combination of characteristic X-ray wavelengths for each element in the periodic table in a database, and collates the database with the characteristic X-ray wavelengths extracted from the X-ray spectrum generated when the sample is irradiated with an electron beam. In addition, the function of displaying the corresponding element and the peak position of the spectrum is provided, and the operator can estimate the element / composition contained in the sample portion that has emitted X-rays. As a result, the review apparatus can be provided with an elemental analysis method for detecting characteristic X-rays generated at the time of electron beam irradiation and examining the element and concentration in a minute region from the energy and intensity.

  When analyzed at an acceleration voltage of 5 kV, light elements such as carbon and fluorine are efficiently detected because the characteristic X-ray is 0.3 to 0.7 kV, but heavy metals such as copper and iron have a characteristic X-ray of 6 to Since it is 8kV, it cannot be detected. On the other hand, when analysis is performed at an acceleration voltage of 15 kV, heavy elements are detected efficiently, but light elements decrease the relative intensity of characteristic X-rays. Patent Document 2 discloses a method of detecting heavy elements by performing elemental analysis once at a low acceleration voltage to detect light elements, increasing the acceleration voltage to high acceleration, performing elemental analysis again, and detecting heavy elements.

JP 2003-107022 A Japanese Patent Laid-Open No. 09-304600

  Along with the progress of miniaturization of semiconductor devices, a function for more easily examining the elements and concentrations of finer defects is required for the defect review apparatus detected by the inspection apparatus. However, when a minute defect is irradiated with an electron beam, it is transmitted and the relative intensity of characteristic X-rays decreases. On the other hand, if the acceleration of the electron beam is reduced in order not to reduce the relative intensity of characteristic X-rays, high-energy characteristic X-rays are not generated. Furthermore, even when estimating elements using only low-energy characteristic X-rays, the energy resolution that depends on the combination of the X-ray detector and waveform generation system is insufficient, and the elements cannot be identified due to the overlap of characteristic X-ray peaks. There is a case. As described above, it is necessary to appropriately set the acceleration of the electron beam in consideration of the element of the characteristic X-ray to be detected and the size of the defect, but it is impossible to estimate all the elements before the characteristic X-ray detection.

  On the other hand, Patent Document 2 describes that a light element is detected by elemental analysis at a low acceleration voltage, and a heavy element is detected by elemental analysis at a high acceleration by increasing the acceleration voltage. However, in the method described in Patent Document 2, the wavelength of the characteristic X-ray extracted from the X-ray spectrum and the database each time the observation point on the sample is moved and each time switching between low acceleration and high acceleration is performed. Using the function of matching and displaying the corresponding element and the peak position of the spectrum, the operator must estimate the element contained in the sample site that emitted X-rays.

  The present invention solves the above-described problems of the prior art and newly adds a function of simply examining the element and the concentration of fine defects in the defect coordinates detected by the inspection apparatus to the review apparatus.

In order to solve the above-described problems, in the present invention, an electron beam application apparatus is mounted on a stage unit on which a sample in which a defect is previously detected by another inspection apparatus can be placed and moved in a plane. An electron optical system for focusing and irradiating an electron beam on the placed sample; a secondary electron detector for detecting secondary electrons generated from the sample irradiated with the electron beam focused by the electron optical system; The secondary electron detector receives a signal for detecting secondary electrons, generates a secondary electron image of the sample, and processes the generated secondary electron image, and the electron optical system focuses the image. X-ray detector for detecting X-rays generated from defects on the specimen irradiated with the electron beam, and a peak of characteristic X-rays from a signal obtained by detecting X-rays generated from the sample with this X-ray detector Defect element that generated X-rays using wavelength information It comprises an element search unit for searching, a control unit for controlling the stage unit, the electron optical system, the image generation / processing unit, and the element search unit. The control unit controls the element search unit and the control electrode, and the defect detected on the sample is detected by irradiating the electron beam controlled to the first acceleration energy by the control electrode. Among the defects irradiated with the electron beam controlled to the first acceleration energy by processing the line detection signal by the element search unit, defects having characteristic X-ray peak wavelengths overlapping between elements are detected by the control electrode. An electron beam controlled to the second acceleration energy is irradiated and extracted as a defect for detecting X-rays.

Further, in order to solve the above-described problems, in the present invention, a sample in which a defect is detected in advance by another inspection apparatus is placed on a stage unit movable in a plane, and the sample placed on the stage unit is placed on A secondary electron detector that detects a secondary electron by detecting a secondary electron by detecting a secondary electron generated from a sample irradiated with the focused electron beam by an electron optical system. A secondary electron image of the sample is generated in response to a signal from the vessel, the generated secondary electron image is processed to detect a defect, and X generated from the defect on the sample irradiated with the focused electron beam The element of the defect which generated the X-ray using the information of the peak wavelength of the characteristic X-ray from the signal obtained by detecting the X-ray with the X-ray detector and detecting the X-ray generated from the sample with this X-ray detector in defect classification method using an electron beam apparatus for searching for, focusing of And the electron beam processing the first X-ray detection signal obtained by detecting the X-rays generated from the defect by irradiation control to the defects on the specimen to the acceleration energy, control the first acceleration energy of the defects was irradiated with electron beam, a defect having a peak wavelength of characteristic X-rays overlap between elements, as defects by irradiating an electron beam is controlled to the second acceleration energy detecting the X-rays was so that issuing extraction.

  According to the present invention, it is possible to easily check the element and concentration of fine defects detected by other inspection apparatuses in a review apparatus. As a result, it has become possible to provide an effective analysis means for identifying the cause of the occurrence of defects in the manufacturing process and taking measures to prevent the occurrence of defects.

It is a block diagram which shows the whole block diagram of the review apparatus based on the Example of this invention. It is a block diagram which shows the whole block diagram of the modification of the review apparatus based on the Example of this invention. It is a flowchart which shows the flow of the whole process which classify | categorizes the composition of the defect which concerns on the Example of this invention. It is a flowchart which shows the detailed flow of the process which classifies the composition of the defect which concerns on the Example of this invention. It is a flowchart which shows the flow of the process which produces the observation database which concerns on the Example of this invention. It is sectional drawing of the data which show the conception which irradiates an electron beam to a micro defect and a comparatively large defect on a board | substrate, respectively. It is a figure which shows the beam acceleration dependence of the signal amount of the silicon | silicone micro defect in characteristic X-ray measurement. It is a figure which shows the beam acceleration dependence of the signal amount of the micro defect of aluminum in characteristic X-ray measurement. It is a figure which shows the beam acceleration dependence of the signal amount of nitrogen (N) in the micro defect of the titanium nitride in characteristic X-ray measurement. It is a figure which shows the beam acceleration dependence of the signal amount of titanium (Ti) in the micro defect of titanium nitride in characteristic X-ray measurement. It is a graph which shows the beam acceleration dependence of the detectable area | region for calculating the threshold value with respect to a defect size. FIG. 6 is a diagram showing a list of element candidates extracted by performing overlap processing with a characteristic X-ray having a width in consideration of energy resolution, both in the case where the acceleration energy of the electron beam is 3 kV and 15 kV The case where an element candidate is extracted is shown. FIG. 5 is a diagram showing a list of element candidates extracted by performing overlap processing with a characteristic X-ray having a width in consideration of energy resolution. When the acceleration energy of an electron beam is 3 kV, element candidates are extracted. In the case of 15 kV, the case where an element candidate was not extracted is shown. It is a figure which shows the combination of the conditional branch of element estimation. It is a figure which shows the structure of the observation database produced | generated by the present Example. It is a figure which shows the result of having classified the defect based on the observation database produced | generated by the present Example, and is a graph which shows the relationship with the number of detected defects classified according to defect type for every shape feature-value of a defect. It is a figure which shows the result of having classified the defect based on the observation database produced | generated by the present Example, and is a graph which shows the relationship with the number of detected defects classified according to the size of a defect for every composition of a defect. It is a front view of the display screen which concerns on a present Example.

  The present invention relates to a review apparatus, which can take in inspection data detected by a foreign substance inspection apparatus and an appearance inspection apparatus, and can automatically obtain a high-definition observation image of a foreign substance / defect with an SEM. The present invention relates to an elemental analysis method for detecting characteristic X-rays generated during electron beam irradiation using an electron microscope and examining the element and concentration in a minute region from the energy and intensity.

  The review apparatus according to the present invention includes a stage for holding a sample, an electron optical column for focusing an electron beam on the sample, a secondary electron detector for converting secondary electrons emitted from the sample into an electric signal, SEM equipped with an image generation system that converts the electrical signal output by the secondary electron detector into a digital image, and an image processing system that analyzes the coordinates and shape features of the observation object from the digital image acquired by the SEM The inspection data detected by the foreign substance inspection apparatus and the appearance inspection apparatus can be taken in, and a high-definition observation image of the foreign substance / defect can be acquired.

  Further, the review device includes a deflection system that irradiates the observation target coordinates with an electron beam, an X-ray detector that converts X-rays emitted from the sample into an electrical signal, and an electrical signal output by the X-ray detector. , Equipped with a waveform generation system that converts X-ray energy spectrum into a digital waveform, a peak detection system that extracts the characteristic X-ray wavelength from the digital waveform, and an element search system that collates the characteristic X-ray with the database to estimate the element .

  In this way, the element search system stores the combination of the characteristic X-ray wavelengths for each element in the periodic table in the observation database, and the characteristic X-ray wavelength extracted from the X-ray spectrum generated when the electron beam is irradiated onto the sample. A function of collating with the database and displaying the corresponding element and the peak position of the spectrum is provided, and the operator can estimate the element / composition contained in the sample region that emitted X-rays.

  The review apparatus includes means for setting the electron beam to a first acceleration, and stage control means for automatically moving to a plurality of observation points and irradiating the electron beam, and is constituted by the digital image and the digital waveform. A list of measurement data for each observation point is created and maintained, and for each observation point, a list of shape features of an observation target such as a defect size and a defect type analyzed from the digital image by the image processing system; A first element list extracted from the digital waveform by the peak detection system and the element search system is created and held. Furthermore, each observation is performed by means for changing the electron beam to the second acceleration, stage control means for automatically moving to a plurality of observation points and irradiating the electron beam, the peak detection system, and the element retrieval system. Create and maintain a second element list for the points. For each observation point, an observation database in which a measurement data list, a shape feature amount list, a first element list, and a second element list are linked is created.

  The first element list and the second element list are read from the observation database for each observation point, the element is estimated in consideration of the detection limit of characteristic X-rays depending on the electron beam acceleration and the defect size, and the third element. Create a list. The third element list is associated with each observation point and added to the observation database.

  In particular, at an observation point where an undetected element that cannot be detected by increasing electron beam acceleration exists between the first element list and the second element list, the defect size is read from the shape feature amount list, and characteristic X-rays can be detected. An electron beam acceleration threshold value is calculated, and if either the first acceleration or the second acceleration exceeds the threshold value, an undetected element is added to the third element list.

  The first acceleration is set to a lower acceleration than the second acceleration, and when checking the characteristic X-ray wavelength extracted from the X-ray spectrum generated when the sample is irradiated with the low acceleration beam, the peak detection system The characteristic X-ray wavelength is extracted in consideration of energy resolution, and the element search system extracts element candidates that have the possibility of overlapping characteristic X-ray peaks. The observation point from which the above element candidate is extracted is automatically selected, the wavelength of the characteristic X-ray extracted from the X-ray spectrum generated during irradiation of the second acceleration sample and the database are collated, and the corresponding element is selected. Extract to create a second element list.

  Further, when capturing the defect position and size data on the sample detected by the inspection apparatus, defects having a defect size smaller than the threshold are extracted to narrow down the observation target for the first acceleration, and further to the observation target for the second acceleration. It is also effective to add a defect larger than the threshold.

  When using three or more multi-acceleration electron beams, create an element list corresponding to each acceleration and link the element list for each acceleration, the list of measurement data, and the shape feature list for each observation point. The attached observation database is formed. In this observation database, the combination of accelerations of the element list is different for each observation point.

  Read the element list for each observation point from the observation database, create the estimated element list by estimating the elements considering the characteristic X-ray detection limit depending on the electron beam acceleration and defect size, etc. Each observation point is linked and added to the observation database.

  In particular, the observation points where undetected elements that cannot be detected when the electron beam acceleration is increased are present in the element list, read the defect size from the shape feature list, and calculate the threshold value for electron beam acceleration that can detect characteristic X-rays. If the acceleration exceeds the threshold, an undetected element is added to the estimated element list.

  When collating the database with the characteristic X-ray wavelength extracted from the X-ray spectrum generated when the sample is irradiated with the low acceleration beam, the peak detection system extracts the characteristic X-ray wavelength in consideration of the energy resolution, The element search system extracts element candidates that have the possibility of overlapping characteristic X-ray peaks. The observation point from which the above element candidate is extracted is automatically selected, the wavelength of the characteristic X-ray extracted from the X-ray spectrum generated when the sample is irradiated with the high acceleration beam and the database are compared, and the corresponding element is selected. Extract and create a list of elements for each acceleration.

  Further, when the data of the position and size of the defect on the sample detected by the inspection apparatus is taken in, the defect having a defect size smaller than the threshold is extracted to narrow down the observation object for the first acceleration, and the observation object after the second acceleration Add a defect larger than the threshold to.

The review apparatus according to the present invention will be described below with reference to the drawings.
FIG. 1A is a schematic diagram illustrating an overall configuration diagram of a review apparatus 100 according to the first embodiment. The review apparatus 100 includes an SEM including an electron optical column 104 having an internal stage 102 and an electron optical system 150 for holding a sample 101, and an image generation / processing unit 160, an element search unit 170, a control / input unit. An output unit 180 is provided.

  The electron optical system 150 of the SEM electron optical column 104 accelerates the electron source 151 that generates electrons, the extraction electrode 152 that extracts the electron beam from the electrons generated by the electron source 151, and the electron beam extracted by the extraction electrode 152. Beam accelerating electrode 153, deflection electrode 155 for deflecting accelerated electron beam 103 in the X and Y directions, objective lens 156 having a coil 1561 and a magnetic path 1562 for focusing electron beam 103 on the surface of sample 101, and electron beam 103 are irradiated The reflecting plate 154 that reflects the secondary electrons 105 generated from the sample 101, the secondary electron detector 106 that detects the secondary electrons reflected by the reflecting plate 154, and the 2 generated from the sample 101 irradiated with the electron beam 103. A secondary electron detector 106 that converts the secondary electrons 105 into an electrical signal and an X-ray detector 121 that detects X-rays generated from the sample 101 irradiated with the electron beam 103 are provided. Of beam 103 Energy (accelerating voltage) is adjusted by controlling the potential difference between the electron source 151 and the stage 102.

  The image generation / processing unit 160 converts an electrical signal output from the secondary electron detector 106 that has detected the secondary electrons 105 generated from the sample 101 into a digital image, and the image generation system 107 converts the electrical signal. An image processing system 110 for analyzing the coordinates 108 and the shape feature 109 of the sample 101 to be observed from the digital image, and a deflection system 111 for irradiating the coordinates 108 of the sample 101 to be observed with the electron beam 103. .

  The element search unit 170 converts the electric signal converted and output from the X-ray detector 121 that detects the X-rays emitted from the sample 101 irradiated with the electron beam 103 into the digital waveform 122 of the X-ray energy spectrum. The waveform generation system 123, the peak detection system 124 that extracts the wavelength of characteristic X-rays from the digital waveform converted by the waveform generation system 123, and the element by collating the characteristic X-ray extracted by the peak detection system 124 with the database 125 Is provided with an element search system 126 for estimating.

  The control / input / output unit 180 controls the voltage applied to the objective lens 156 of the electron optical system 150 and the potential setting difference between the electron source 151 and the stage 102, the column setting means 132, the display means 131, and the entire apparatus to obtain and A control unit / process 133 for processing the received data.

  The element search system 126 of the element search unit 170 stores a combination of characteristic X-ray wavelengths for each element in the periodic table in the database 125 and extracts characteristic X-rays extracted from the X-ray spectrum generated when the sample 101 is irradiated with the electron beam. And the database 125, the corresponding element and the peak position of the spectrum are displayed on the display means 131. The operator looks at the display means 131 and the element contained in the sample part that has emitted X-rays. -The composition can be estimated.

  The X-ray detector 121 includes a semiconductor detector (silicon detector, silicon drift detector) or a microcalorimeter. A waveform generation system 123 that converts an electric signal output from the X-ray detector 121 into a digital waveform of an X-ray energy spectrum is configured as an energy dispersion type or a wavelength dispersion type. Depending on the combination of the X-ray detector 121 and the waveform generation system 123, the counting speed and energy resolution of the X-rays detected by the X-ray detector 121 are different.

  When the X-ray detector 121 is configured with a semiconductor detector, counting can be performed at a higher speed than when the X-ray detector 121 is configured with a calorimeter. However, the energy when the energy generation type is used as the waveform generation system 123. Resolution degrades from about 10eV to about 100eV. Therefore, the characteristic X-ray energy, which makes it difficult for the peak detection system 124 to extract from the digital waveform of the X-ray energy spectrum, is about 3 kV or less when the X-ray detector 121 is composed of a semiconductor detector. When configured, it is about 1 kV or less.

  The energy of the characteristic X-ray is lower than the acceleration of the irradiating electron beam 103, and the acceleration of the irradiating electron beam 103 is 3 kV or more when the X-ray detector 121 is configured with a semiconductor detector, and is configured with a microcalorimeter Then, if the voltage is 1 kV or more, characteristic X-rays can be easily extracted. When the X-ray detector 121 is composed of a semiconductor detector, when the sample 101 is a silicon substrate, an element such as aluminum that is separated from the peak of the characteristic X-ray of the silicon substrate can be detected. is there. On the other hand, minute foreign matter containing tantalum or tungsten on a silicon substrate emits characteristic X-rays to 1.74 kV ± 0.05 kV when irradiated with an electron beam of 3 kV, but separates the overlap between elements of the characteristic X-ray peak. I can't. On the other hand, if the beam acceleration is changed to 15 kV by the column setting means 132, the characteristic X-ray peak is 8.15 kV for tantalum and 8.40 kV for tungsten, and there is a peak distance of 0.25 kV, and the overlap can be eliminated. .

  The element search system 126 stores a combination of characteristic X-ray wavelengths for each element of the periodic table in the database 125, and the characteristic X-ray wavelength extracted from the X-ray spectrum generated when the sample 101 is irradiated with the electron beam and the database 125. And the corresponding element list of each acceleration is stored in the database 125 as observation data. Further, for each defect ID, the shape feature amount and the element list for each acceleration can be searched from the observation data in the database 125.

  FIG. 1B is a schematic diagram of an overall configuration diagram of a modified example of the review apparatus 100 illustrated in FIG. 1A. The review apparatus 200 shown in FIG. 1B is different from the review apparatus 100 shown in FIG. 1A in the configuration of the deflection electrode 255 of the electro-optical column 250, and the objective lens 256 including the coil 2561 and the magnetic path 2562. . 1B, the same reference numerals as those in FIG. 1A have the same functions as those described in FIG. 1A. As shown in FIGS. 1A and 1B, even if the configurations of the electro-optic columns 104 and 204 are different, a system for realizing this embodiment can be configured.

  FIG. 2A shows a rough flow and FIG. 2B shows a detailed flow of the element analysis processing flow that enables detection of elements and concentrations of micro defects using the review apparatus 100 or 200 shown in FIGS. 1A and 1B. Shown in Here, although the observation target is a defect, the flowchart of the present invention can be applied to a plurality of observation points that specify coordinates.

  First, in the rough flow shown in FIG. 2A, first, a defect to be dealt with is extracted (S251), and then the extracted target defect is controlled to the first acceleration energy and irradiated with an electron beam to perform measurement. (S252) U. Next, this measurement result is evaluated, and it is determined whether or not it is necessary to perform measurement by irradiating an electron beam of different energy (S253). If it is determined in S253 that it is not necessary to perform measurement by irradiating an electron beam with different energy, the measurement is terminated. On the other hand, when it is determined that it is necessary to perform measurement by irradiating an electron beam with different energy, the measurement is performed by irradiating the sample with the electron beam controlled to the second acceleration energy (S254). The composition is classified by statistical analysis and the observation is completed.

Next, each step of the rough flow described in FIG. 2A will be described in detail using the flowchart of FIG. 2B.
First, in detailed steps corresponding to S251 in FIG. 2A, the semiconductor wafer as the sample 101 is loaded onto the stage 102 of the electro-optic column 104 (S201). Next, inspection data detected by the foreign substance inspection device and the appearance inspection device is read into the control unit 133 of the control / input / output unit 180 (S202), and a target defect is extracted from the inspection data (S203).

  Next, in the detailed steps corresponding to S251 in FIG. 2A, first, the control unit 133 of the control / input / output unit 180 uses the column setting unit 132 to perform observation with the first beam acceleration. Is set to the potential difference between (S204). Next, an image of the sample 101 to be scanned by irradiating the semiconductor wafer as the sample 101 with the electron beam 103 is obtained, and for the coordinate correction using the image of the reference pattern (mark) for alignment included in this image Then, the wafer alignment is performed by controlling the stage 102 or the deflection electrode 155 (S205). Next, the control of the stage 102 and the control of the electron beam 103 by the deflection electrode 155 are combined so that the defect to be observed enters the observation field of view of the electron optical system 150 of the scanning electron microscope (SEM) ( S206). Next, the sample 101 is imaged by scanning the electron beam and detecting secondary electrons reflected from the reflector 154 and incident on the secondary electron detector 106 among the secondary electrons 105 generated from the sample 101 ( (S207) The electric signal output from the secondary electron detector 106 is converted by the image generation system 107 to form a digital image (S208).

  Next, the image processing system 110 extracts the coordinates and shape feature amounts of the target defect from the digital image formed by the image generation system 107, and classifies the types of the target defects using the extracted coordinate and shape feature amount information. (S209). When forming a digital image, there is also a method of detecting a target defect coordinate by low-magnification imaging in advance and acquiring a zoom image of the target defect. The shape feature amount is a parameter used to classify the appearance of the target defect.

  Next, the deflection electrode 155 is controlled by the deflection system 111 based on the defect coordinate information extracted from the digital image, and the defect coordinate position on the sample is irradiated with the electron beam 103 (S210). Next, the X-ray emitted from the sample 101 irradiated with the electron beam 103 is detected by the X-ray detector 121 and converted into an electric signal (S211), and the electric signal is converted into an X-ray energy spectrum by the waveform generation system 123. (S212). Next, the wavelength of the characteristic X-ray is extracted from the digital waveform by the peak detection system 124, the element is estimated by collating the characteristic X-ray and the database by the element search system (S213), and the composition classification of the target defect is performed (S214). ).

  Next, it is checked whether composition classification has been completed for all target defects (S215). If the result of the check is NO, the position coordinate information of the next unclassified target defect is used to control the position of the stage 102 so that the next target defect falls within the observation field of view of the electron optical system 150. By repeating the flow from S204 to S214, the composition of all target defects is classified. In the control unit / process 133, a list of measurement data for each target defect constituted by a digital image and a digital waveform obtained by processing along the above flow is created and held. Further, for each target defect, a first feature list extracted from the digital feature by a peak detection system and an element search system based on a shape feature amount list of the target defect such as a defect size and a defect type analyzed by an image processing system. Create and hold.

  When the composition classification is completed for all target defects (in the case of YES in S215), the process proceeds to a step corresponding to S253 in FIG. 2A, and an electron source 151 for accelerating the electron beam 103 irradiating the sample 101 It is determined whether it is necessary to change the potential difference between the stages 102 and perform measurement again (S216).

  Here, in S213, the peak detection system 124 extracts the wavelength of the characteristic X-ray from the digital waveform of the X-ray energy spectrum. At the same time, the element search system 126 has an element that may overlap the characteristic X-ray peak. Extract candidates.

  Therefore, in S216, if the element search system 126 does not extract an element candidate that has the possibility of overlapping characteristic X-ray peaks, the determination result is NO and the composition classification step (S230) by statistical analysis described later is performed. move on.

  On the other hand, if the element search system 126 extracts an element candidate that has the possibility of overlapping characteristic X-ray peaks, the determination result is set to YES, and the element candidate that has the possibility of overlapping characteristic X-ray peaks. Is extracted (S217).

  Next, it is determined as YES in S216, and the defect including the element candidate extracted as the possibility of the overlap of the characteristic X-ray peak in S217 is set as the defect to be observed, and the extracted defect to be observed is as follows. Steps corresponding to S254 in FIG. 2A, that is, measurement is performed by irradiating the sample with the electron beam controlled to the second acceleration energy.

  The process corresponding to S254 is almost the same as the process described in S252. First, the target defect from which the element candidate is extracted is automatically extracted, and the sample is irradiated with the electron beam accelerated by the second acceleration energy. The wavelength of the characteristic X-ray extracted from the X-ray spectrum generated sometimes and the database are collated, the corresponding element is extracted, and the second element list is created. Further, when capturing the defect position and size data on the sample detected by the inspection apparatus, defects having a defect size smaller than the threshold are extracted to narrow down the observation target for the first acceleration, and further to the observation target for the second acceleration. It is also effective to add a defect larger than the threshold.

  The second acceleration energy of the electron beam 103 is not limited to the case where the second acceleration energy is set higher than the first acceleration energy, and may be set to a low acceleration energy. In any case, in the present invention, the composition classification of the target defect is performed using the electron beam 103 set to the second acceleration energy in the following flow.

  In the step corresponding to S254 in FIG. 2A, first, a target defect to be irradiated with the electron beam 103 set to the second acceleration energy is extracted from the inspection data (S217). Next, the potential difference between the electron source 151 and the stage 102 is set by the column setting means 132 (S218). Next, an image of the sample 101 to be scanned by irradiating the semiconductor wafer as the sample 101 with the electron beam 103 set to the second acceleration energy is obtained, and an image of a reference pattern (mark) for alignment included in this image is obtained. The wafer alignment is performed by controlling the stage 102 or the deflection electrode 155 to correct the coordinates (S219). Next, by combining the control of the stage 102 and the control of the electron beam 103 by the deflection electrode 155, the defect extracted as the observation object at the second acceleration voltage is detected by the electron optical system 150 of the scanning electron microscope (SEM). Enter the observation field (S220). Next, the sample 101 is imaged by scanning the electron beam and detecting secondary electrons reflected from the reflector 154 and incident on the secondary electron detector 106 among the secondary electrons 105 generated from the sample 101 ( (S221) The electric signal output from the secondary electron detector 106 is converted by the image generation system 107 to form a digital image (S222).

  Next, the image processing system 110 extracts the coordinates of the target defect from the digital image formed by the image generation system 107 (S223). When forming a digital image, there is also a method of detecting a target defect coordinate by low-magnification imaging in advance and acquiring a zoom image of the target defect.

  Next, the deflection electrode 155 is controlled by the deflection system 111 based on the defect coordinate information extracted from the digital image, and the observation target defect on the sample is irradiated with the electron beam 103 set to the second acceleration energy (S224). Next, X-rays emitted from the sample 101 irradiated with the electron beam 103 are detected by the X-ray detector 121 and converted into an electric signal (S225), and the electric signal is converted into an X-ray energy spectrum by the waveform generation system 123. (S226). Next, the wavelength of the characteristic X-ray is extracted from the digital waveform by the peak detection system 124, the element is estimated by collating the characteristic X-ray and the database by the element retrieval system (S227), and the composition classification of the target defect is performed (S228). ).

  Next, it is checked whether composition classification has been completed for all target defects (S229). If the result of the check is NO, the position coordinate information of the next unclassified target defect is used to control the position of the stage 102 so that the next target defect falls within the observation field of view of the electron optical system 150. By repeating the flow from S220 to S228, composition classification is performed for all target defects. In the control unit / process 133, a list of measurement data for each target defect constituted by a digital image and a digital waveform obtained by processing along the above flow is created and held. Furthermore, for each target defect, the second feature list extracted from the digital feature by the peak feature detection system and the element search system based on the shape feature quantity list of the target defect such as the defect size and defect type analyzed by the image processing system. Create and hold.

  When the composition classification is completed for all target defects (YES in S229), the process proceeds to the composition classification step (S230) by statistical analysis, and the composition classification is performed by statistical analysis of the target defects. A detailed flow of the composition classification step (S230) by statistical analysis will be described with reference to FIG. 2C.

  In the composition classification step (S230) by statistical analysis, first, an observation database in which a measurement data list, a shape feature amount list, a first element list, and a second element list are linked to each target defect is created ( S2301). The first element list and the second element list are read from this observation database for each target defect, the element is estimated in consideration of the detection limit of characteristic X-rays depending on the electron beam acceleration and the defect size, and the third element. A list is created (S2302). For target defects in which undetected elements that cannot be detected by increasing the electron beam acceleration exist between the first element list and the second element list, the defect size is determined from the shape feature amount list created in each flow of S252 and S254. Is read and the threshold value of the electron beam acceleration capable of detecting characteristic X-rays is calculated. If either the first acceleration or the second acceleration exceeds the threshold value, an undetected element is added to the third element list. (S2303). The third element list is associated with each target defect and added to the observation database (S2304). By creating the third element list by the above procedure, composition classification is performed by statistical analysis of the target defect.

  When using three or more multi-acceleration electron beams, create an element list corresponding to each acceleration and link the element list for each acceleration, the list of measurement data, and the shape feature list for each observation point. The attached observation database is formed. In this observation database, the combination of accelerations of the element list is different for each observation point.

  Read the element list for each observation point from the observation database created in S2304, estimate the elements in consideration of the characteristic X-ray detection limit depending on the electron beam acceleration energy (acceleration voltage) and defect size, etc. Is created (S2305). In particular, for target defects in the element list that have undetected elements that cannot be detected by increasing electron beam acceleration, the defect size is read from the shape feature list, and a threshold value for electron beam acceleration that can detect characteristic X-rays is calculated. If the acceleration exceeds the threshold, an undetected element is added to the estimated element list. The created estimated element list is associated with each target defect and added to the observation database (S2306).

  In S213, when the wavelength of the characteristic X-ray extracted from the X-ray spectrum generated when the sample is irradiated with the low acceleration beam is collated with the database, the peak detection system 124 considers the energy resolution and the wavelength of the characteristic X-ray. In the element search system 126, element candidates that may overlap characteristic X-ray peaks are extracted.

  In S217, the target defect from which such element candidates are extracted is selected, and the characteristic X extracted from the X-ray spectrum generated when the sample is irradiated with the high acceleration beam (second electron beam acceleration energy) in S227. The wavelength of the line is compared with the database, the corresponding elements are extracted, and an element list for each acceleration energy of the electron beam is created.

  Further, when the defect position and size data on the specimen detected by the inspection apparatus in S202 is taken in, the defect whose defect size is smaller than a preset threshold is extracted and observed as the first electron beam acceleration energy. The observation target is narrowed down, and a defect larger than a preset threshold value may be added as a defect observed after the second electron beam acceleration energy in S217.

  FIG. 3A shows that when a minute defect 302 and a relatively large defect 303 exist on a silicon substrate 301, each defect is irradiated with an electron beam 103 accelerated with a predetermined acceleration energy. It is the figure which showed typically the state in which the characteristic X-ray has generate | occur | produced.

  FIG. 3B shows the case where the target defect is silicon (Si), the size of the target defect is 7.5 nm, 15 nm, and 25 nm, and the detected characteristic X-ray energy is 1.74 kV. FIG. 3C shows a case where the target defect is aluminum (Al), the size of the target defect is 7.5 nm, 15 nm, and 25 nm, and the detected characteristic X-ray energy is 1.49 kV.

  Silicon, which is the material of the substrate, increases when the acceleration energy of the electron beam is increased, but aluminum, which is the signal of the target defect, exhibits a peak characteristic in which the acceleration energy of the electron beam is about 3 kV maximum. Therefore, it can be seen that, in order to detect a minute aluminum defect, the electron beam acceleration can be set efficiently to 3 kV.

  Fig. 4A shows the beam acceleration energy dependence when the electron beam irradiation material is made of titanium nitride microdefects. The target defect sizes are 7.5nm, 15nm, and 25nm, and the detected characteristic X-ray energy Is 0.39 kV for nitrogen. On the other hand, FIG. 4B is a diagram showing the beam acceleration energy dependency when the electron beam irradiation material is made of a fine defect of titanium nitride, and the characteristic X-ray to be detected when the size of the target defect is 7.5 nm, 15 nm, and 25 nm. The energy of is 4.5 kV for titanium. Nitrogen, a light element, decreases with increasing electron beam acceleration, while titanium, a heavy element, increases to 15 kV beam acceleration. Therefore, it is understood that fine defects of titanium nitride cannot be detected unless nitrogen is detected with an electron beam acceleration energy (acceleration voltage) of 3 kV or less and titanium is detected with 5 kV or more.

  In order to classify the composition of fine defects including light elements and heavy elements as components, it is necessary to measure characteristic X-rays by changing the acceleration energy of the electron beam into two or more types. Furthermore, even when a plurality of defects including light elements and heavy elements as components are automatically classified, it is necessary to measure characteristic X-rays by changing the acceleration energy of the electron beam into two or more types.

  FIG. 5 is a diagram showing the beam acceleration energy dependency of the detectable region for calculating the defect size threshold in the target defect extraction in S217 of the flowchart shown in FIG. 2B. Increasing the acceleration energy of the electron beam increases the threshold of defect size that can be detected. Therefore, it is possible to perform characteristic X-ray measurement for a large defect with an electron beam with high acceleration energy, and it is not necessary to perform characteristic X-ray measurement with an electron beam with low acceleration energy that easily causes erroneous detection elements.

  FIG. 6A and FIG. 6B are diagrams showing an example in which a 1.74 kV characteristic X-ray overlap process is performed in the composition classification step (S230) in the statistical analysis of the flowchart shown in FIG. 2B. 601 and 611 are examples in which the first acceleration energy of the electron beam 103 is 3 kV, and 602 and 612 are examples in which the second acceleration energy of the electron beam 103 is 15 kV.

  When extracting 1.74 kV characteristic X-rays using the X-ray spectrum generated when the target defect is irradiated with the electron beam 103 accelerated at 3 kV as the first acceleration energy, the element search system 126 considers the energy resolution. Then, elements such as 601 and 611 are extracted as element candidates having characteristic X-rays of 1.64 kV to 1.84 kV by overlap processing.

  On the other hand, comparing the wavelength of the characteristic X-ray extracted from the X-ray spectrum generated when the target defect is irradiated with the electron beam 103 accelerated at 15 kV as the second acceleration energy and the database, the element is shown as 602 in FIG. 6A. There are cases where it can be specified, and there are cases where it cannot be specified as 612 in FIG. 6B. When the element cannot be specified as in 612 in FIG. 6B, the element can be estimated in consideration of a defect size limit that allows detection of characteristic X-rays with an electron beam accelerated at 15 kV as the second acceleration energy.

  FIG. 6C shows a combination of conditional branches for element estimation. When the target defect is irradiated with the electron beam 103 accelerated at 3 kV as the first acceleration energy, an element candidate having a characteristic X-ray near the X-ray spectrum generated from the target defect is extracted (621). Next, when the target defect is irradiated with the electron beam 103 accelerated at 15 kV as the second acceleration energy, an element candidate having characteristic X-rays in the vicinity of the X-ray spectrum generated from the target defect is extracted (622) And (623) when not extracted. If an element candidate having characteristic X-rays in the vicinity of the X-ray spectrum generated from the target defect is extracted when the target defect is irradiated with the electron beam 103 accelerated at 15 kV (622), the element can be estimated. (624). On the other hand, if an element candidate having characteristic X-rays in the vicinity of the X-ray spectrum generated from the target defect is not extracted when the target defect is irradiated with the electron beam 103 accelerated at 15 kV (623), the element is estimated. There is a case where the element can be estimated (626) and a case where the element cannot be estimated (626).

  FIG. 7 is a diagram showing the configuration of the observation database 700 formed by the flowchart shown in FIG. 2C.

  An observation database is created by repeating the flow shown in FIG. 2B. A digital image 702 is obtained by the flow shown in FIG. 2C, and a shape feature quantity list of target defects such as the defect size 704 and defect type 701 analyzed by the image processing system The digital waveform 703, the first element list 705 and the second element list 706 extracted by the peak detection system and the element search system are stored in the observation database 700 for each defect ID 710.

  From this observation database 700, the first element list 705 and the second element list 706 of each defect ID 710 are read, and the detection limit of characteristic X-rays depending on the electron beam acceleration energy and the defect size shown in FIG. The third element list 707 is created by estimating the elements by the characteristic X-ray overlap process shown in FIGS. 6A to 6C. The third element list 707 is stored in the category of each defect ID 710 in the observation database 700.

  Particularly, a defect 710 (any one of ID1, ID2,...) In which an undetected element that cannot be detected when the acceleration energy of the electron beam is increased is present between the first element list 705 and the second element list 706 is present. Then, the defect size 704 is read from the shape feature quantity list (not shown) of the observation database 700, and the threshold value of the acceleration energy of the electron beam capable of detecting characteristic X-rays is calculated based on the graph shown in FIG. If either the first acceleration or the second acceleration exceeds the threshold, an undetected element is added to the third element list 707.

  By creating the third element list 707 by the above procedure, composition classification is performed by statistical analysis of target defects.

  When using an electron beam with three or more types of acceleration energy, an element list corresponding to 705 and 706 corresponding to each acceleration energy is created, and for each observation point, an element list for each acceleration energy, a list of measurement data, and a shape The feature quantity list is stored in the category of each defect ID 710 in the observation database 700. In this observation database 700, the combination of element list accelerations may be different for each defect ID 710.

  An element list 705 to 707 is read from the observation database 700 for each defect ID 710, and the element list is estimated by taking into account the characteristic X-ray detection limit depending on the acceleration energy of the electron beam and the defect size 704, etc. The estimated element list thus created is stored in the category of each defect ID 710 in the observation database 700. In particular, a defect ID 710 in which an undetected element that cannot be detected when the acceleration energy of the electron beam is increased is present in the element list reads out a defect size 704 that is a list of shape features in the observation database 700 based on the graph shown in FIG. Then, a threshold value of acceleration energy of the electron beam capable of detecting characteristic X-rays is calculated, and when the acceleration energy exceeds the calculated threshold value, an undetected element is added to an estimated element list (not shown).

  8A and 8B are diagrams showing the results of classifying defects based on the observation database 700 shown in FIG. For the third element list or the estimated element list in the observation database, FIG. 8A shows an example classified by defect type for each shape feature amount, and FIG. 8B shows an example classified by defect size for each defect composition.

  The function of classifying defects based on the observation database 700 can easily examine and classify the elements of fine defects with defect coordinates detected by the inspection device. This is a very important analytical tool for the prevention of occurrence.

  FIG. 9 is an example of a display screen 900 for an operator to register an arbitrary operation sequence in the apparatus in order to realize the processing shown in the flowcharts of FIGS. 2B and 2C.

  The inspection data selection icon 901 can display and select inspection data or register a file name. The target defect filter icon 902 is an icon for selecting a filter in order to extract a target defect from inspection data. The target defect selection icon 903 is an icon that displays the target defect candidate ID, inspection feature amount, and the like so that the operator can extract the target defect. An alignment recipe icon 904 is an icon for registering a sequence for automatically aligning a wafer by selecting alignment coordinates and a reference pattern. The defect imaging recipe icon 905 is an icon for setting a method and conditions for imaging a target defect by electron beam scanning. The X-ray measurement recipe icon 906 is an icon for setting an operation sequence of the X-ray detector and the waveform generation system.

  The first acceleration setting icon 911 is an icon for setting an electron beam acceleration for the first X-ray measurement or selecting an automatic multiple acceleration setting function. The second acceleration setting icon 912 is an icon for setting the electron beam acceleration of the second X-ray measurement. There may be a case where icons relating to a plurality of acceleration settings such as a third acceleration setting icon (not shown) are added. The overlap processing icon 913 is an icon for performing settings related to processing for estimating and adding an element whose characteristic X-ray peak overlaps the first element list. The automatic defect narrowing icon 914 is an icon for performing a setting relating to a process of extracting defects including elements overlapping in the first element list and narrowing down the target defect of the second X-ray measurement. The estimated element list icon 915 is an icon for making settings related to processing when creating the third element list or the estimated element list using the observation database. The composition classification recipe icon 916 is an icon for setting a method for classifying defects using the observation database 700.

  The effective icon 920 is an icon that is executed under the conditions set by each of the icons described above. In the case of FIG. 9, the result of classifying the third feature list or the estimated element list in the observation database with the defect type in the shape feature quantity list (corresponding to FIG. 8A) with the composition classification recipe icon 916 is displayed in the display area 930. The displayed state is shown.

  According to the present embodiment, based on the coordinate information of the fine defect on the semiconductor wafer detected by another inspection apparatus, the fine defect is observed by the review apparatus, and the element and the concentration of the fine defect can be simplified. It becomes possible to investigate. As a result, it becomes possible to identify the cause of the occurrence of defects in the manufacturing process of the semiconductor device in a relatively short time, and it is possible to take measures to prevent the occurrence of defects at an early stage.

  DESCRIPTION OF SYMBOLS 100 ... Review apparatus 101 ... Sample 102 ... Stage 103 ... Electron beam 104 ... Electro-optical column 105 ... Secondary electron 106 ... Secondary electron detector 107 ... Image Generation system 110 ... Image processing system 111 ... Deflection control system 121 ... X-ray detector 123 ... Waveform generation system 124 ... Peak detection system 125 ... Database 126 ... Element search system 131: Display means 132 ... Column control means 133 ... Control / processing unit 150 ... Electro-optical system 155 ... Deflection electrode 156 ... Objective lens 160 ... Image generation / processing unit 170 ... Element search part 180 ... Control / input / output part.

Claims (6)

A stage part that can be moved in a plane by placing a sample in which a defect is detected in advance by another inspection apparatus,
An electron optical system that focuses and irradiates an electron beam on a sample placed on the stage; and
A secondary electron detector the electron beam is focused by electron optics to detect the secondary electrons generated from the sample irradiated,
An image generation / processing unit that receives a signal obtained by detecting the secondary electrons by the secondary electron detector, generates a secondary electron image of the sample, and processes the generated secondary electron image;
And X-ray detector for detecting X-rays generated from the defects on the specimen in which the electron beam is focused by the electron optical system is irradiated,
An element search unit which searches elements of the defect that caused the X-ray using information on characteristic X-ray signals the obtained by detecting the X-rays generated from the defects in the X-ray detector,
An electron beam application apparatus comprising a control unit for controlling the stage unit, the electron optical system, the image generation / processing unit, and the element search unit,
The electron optical system includes a control electrode for controlling the acceleration energy of the electron beam irradiated onto the sample,
The control unit controls the element search unit and the control electrode, and detects an X-ray detected by irradiating the defect on the sample with an electron beam controlled to the first acceleration energy by the control electrode. Among the defects irradiated with the electron beam controlled to the first acceleration energy by processing a signal in the element search unit, the defect having a characteristic X-ray peak wavelength overlapping between elements is controlled. electron beam apparatus which is characterized in that by irradiating the second controlled electron beam acceleration energy at the electrode is extracted as a defect detecting X-rays. The electron beam application apparatus according to claim 1, wherein the image generation / processing unit extracts the defect from a secondary electron image of the sample, and the element search unit extracts the defect by the image generation / processing unit. said control electrode by said first controlled electron beam acceleration energy detected X-rays generated by irradiating the information of the composition of the defect of the defect, and the second acceleration energy by the control electrode Extracting information on the composition of the defect in which X-rays generated by irradiating the controlled electron beam are detected, and the control unit extracts the type and composition of the defect extracted from the secondary electron image of the sample or the sample An electron beam application apparatus that outputs information on the size and composition of defects extracted from a secondary electron image . The electron beam application apparatus according to claim 1, wherein the image generation / processing unit is a secondary power supply of the sample.
Extracting the defect from the child image, the element search unit, X-rays generated by irradiating the electron beam is controlled to the first acceleration energy of the defect extracted by the image generating and processing section detects the first information of the composition of the defect, and extracts the second information of the composition of defects X-rays generated by irradiating the electron beam is controlled to the second acceleration energy by the control electrode is detected And classifying the defect composition extracted from the secondary electron image of the sample by performing statistical analysis using the first information of the defect composition and the second information of the defect composition. Features electron beam application equipment. Place the sample in which the defect was detected in advance by another inspection device on the stage part that can move in the plane,
A sample placed on the stage is irradiated with a focused electron beam by an electron optical system,
Secondary electrons generated from the sample irradiated with the focused electron beam are detected by a secondary electron detector;
Receiving a signal from the secondary electron detector that has detected the secondary electrons, generating a secondary electron image of the sample, processing the generated secondary electron image to detect the defect,
An X-ray detector detects X-rays generated from the defect on the sample irradiated with the focused electron beam,
An electron beam application apparatus for searching for an element of the defect that generated the X-ray using information on a peak wavelength of the characteristic X-ray from a signal obtained by detecting the X-ray generated from the defect by the X-ray detector A defect classification method using
Processing the X-ray detection signal obtained by detecting the X-rays generated from the defect by irradiating the defect on the sample by controlling the focused electron beam to a first acceleration energy; Among the defects irradiated with the electron beam controlled to the first acceleration energy, the defect having the characteristic X-ray peak wavelength overlapping between the elements is irradiated with the electron beam controlled to the second acceleration energy. defect classification method using the electron beam apparatus according to claim Rukoto issuing extract as a defect for detecting X-rays Te. The defect classification method using the electron beam application apparatus according to claim 4, wherein the defect is extracted from a secondary electron image of the sample, and the electron controlled to the first acceleration energy among the extracted defect. Information on the composition of defects in which X-rays generated by irradiation with a beam are detected, and information on the composition of defects in which X-rays generated by irradiation with an electron beam controlled to the second acceleration energy are detected. A defect classification method using an electron beam application apparatus, which extracts and outputs information on the size and composition of defects extracted from a secondary electron image of the sample . The defect classification method using the electron beam application apparatus according to claim 4, wherein the defect is extracted from a secondary electron image of the sample, and the electron controlled to the first acceleration energy among the extracted defects. First information on a composition of a defect in which X-rays generated by irradiation with a beam are detected, and composition of a defect in which X-rays generated by irradiation with an electron beam controlled to the second acceleration energy are detected The defect extracted from the secondary electron image of the sample through statistical analysis using the first information of the defect composition and the second information of the defect composition The defect classification method using the electron beam application apparatus characterized by classifying the composition of

Images