JP2005114384A - Defect elemental analysis method and apparatus - Google Patents

Abstract

【Task】
In an analysis method for performing an elemental analysis of a defect by irradiating an electron beam onto a defect to obtain an X-ray spectrum, an acceleration voltage of the electron beam suitable for analyzing a subfilm defect is efficiently set.
[Solution]
The analysis device is equipped with a simulation execution unit, and as a simulation condition, the film material of the device in which the defect to be analyzed exists, the density of the film, the film thickness, the element that needs to be distinguished from the material of the film, and the analysis target Based on the conditions, the defect voltage is input and an X-ray spectrum is simulated to obtain an acceleration voltage suitable for analyzing the subfilm defects. The sample is irradiated with an electron beam at the calculated acceleration voltage. Thus, characteristic X-rays are detected.
[Selection] Figure 7

Description

  The present invention relates to an analysis apparatus and method for analyzing foreign matter and defects generated on a semiconductor wafer based on inspection information mainly in a semiconductor electronic circuit manufacturing process.

  The manufacture of semiconductor devices consists of a number of processes. The process can be broadly divided into a substrate process in which transistor elements are formed on a substrate and a wiring process in which wiring for connecting the elements is formed. Each of these steps is constituted by a combination of a plurality of processes such as a thin film formation process, a photolithography process, an etching process, an impurity introduction process, a heat treatment process, a planarization process, and a cleaning process. The number of such processing steps sometimes reaches several hundreds.

  If defects or foreign matter are generated on the semiconductor wafer due to inadequate or abnormal manufacturing conditions of the processing apparatus, the probability that a defect will occur in the product increases and the yield decreases. Therefore, inspections such as foreign matter inspection and appearance inspection are carried out for each main process, and whether or not the processing is normally performed is monitored. When an abnormality occurs, measures are taken for the corresponding device. At this time, since it is impossible to inspect all wafers for each processing process due to time and labor constraints, it is possible to inspect the sampled lots and wafers for each process consisting of several processes. To be implemented.

  In other words, as shown in FIG. 1, it is inspected whether the lot or wafer processed in the process group A composed of a plurality of processes is normal, that is, whether the occurrence state of defects and foreign matters is within the normal range. Is done. If everything is normal, the process proceeds to the process consisting of the next process group B. If not normal, a detailed inspection of the detected defects and foreign matter is performed, and the process group A is executed from the inspection result. Among them, a manufacturing apparatus such as a processing apparatus that causes such a defect or a foreign substance, that is, a problematic apparatus is identified, and measures are taken to prevent the defective apparatus from causing a defect or a foreign substance.

  The inspection device scans the wafer surface with a laser, detects the presence or absence of scattered light, or captures the shape of the pattern as an image and compares it with other images of the same pattern area, thereby detecting the position and number of singular points. You can get information about.

  Here, the “singular point” refers to a point that is output as a point where an abnormality has been found by the inspection of the inspection apparatus. The foreign object and the appearance abnormality are combined and hereinafter referred to as “defect”.

  In many cases, monitoring of an apparatus abnormality is performed using the number and density of defects detected by an inspection apparatus as a management index. If the number of defects exceeds a preset reference value, it is determined that an abnormality has occurred in the apparatus. As shown in FIG. 2, the defects are detected based on wafer map information using an optical microscope or a scanning electron microscope (Scanning Electron Microscope). , Hereafter referred to as SEM) and enlargement imaging, obtaining detailed information such as size, shape, texture, etc., and performing detailed inspections such as elemental analysis, cross-sectional observation, etc., to identify the device where the failure occurred and the content of the failure . Then, based on the result, measures are taken for devices and processes to prevent a decrease in yield.

  In recent years, review devices have been developed that have a function (Automatic Defect Review, hereinafter referred to as ADR) for automatically acquiring an enlarged image of foreign matter / defects based on inspection data from foreign matter inspection devices and appearance inspection devices. A method of automatically classifying acquired images (Automatic Defect Classification, hereinafter referred to as ADC) is also disclosed.

Here, when performing composition analysis on a defect, it is necessary to reliably irradiate the defect with an energy beam for analysis. The designation of the beam irradiation position is generally performed manually. However, since it takes time, it is desired to perform it automatically when the number of defects is large. In general, the time required for composition analysis is longer than the time required for review. Therefore, the composition analysis target may be narrowed down when the number of defects is large. Since this narrowing-down operation takes time when performed manually, it is desirable to perform it automatically when the number of defects is large.
For this composition analysis, Electron Probe Micro Analysis (hereinafter referred to as EPMA) is known.

  This is a method of irradiating an analysis object with an electron beam and detecting characteristic X-rays generated from the analysis object. Since the wavelength of the characteristic X-ray differs depending on the element, the element included in the analysis target can be identified by analyzing the wavelength of the obtained characteristic X-ray.

  Energy dispersion for obtaining composition information by irradiating a detected defect with an electron beam (hereinafter referred to as EB) and analyzing energy of characteristic X-rays emitted from the defect. A technique for performing analysis in an SEM equipped with an energy dispersive X-ray spectrometer (hereinafter referred to as EDX) is disclosed.

Japanese Unexamined Patent Publication No. 08-148111

Japanese Patent Laid-Open No. 10-27833 JP-A-6-82399 JP-A-6-325720 T. Kijima et.al: “Monte Carlo Calculations on the Passage of Electrons through Thin Films Irradiated by 300keV Electrons”, IEICE TRANS.ELECTRON, Vol.E78-C, No.5, pp.557-563 (1995) E. Acosta et.al: “Monte Carlo simulation of x-ray emission by kilovolt electron bombardment”, J. Appl. Phys. 83, pp. 6038-6049 (1998). Z.J.Ding et.al: “Monte Carlo simulation of xx-ray spectra in electron probe microanalysis: Comparison of continuum with experiment”, J. Appl. Phys. 76, pp. 7180-7187 (1994)

  Here, the foreign matter on the semiconductor wafer is not always on the outermost surface. For example, in the film formation process, when a foreign substance is generated during film formation, the surface of the foreign substance is covered with a material used for film formation. Further, as described above, since inspection is performed every several series of processes, the defect detected by the inspection may have occurred several processes before, not just the process immediately before the inspection. Here, when the film forming process is included in the several processes, the surface of the defect is covered with the material used for the film forming. In this way, a defect covered with a certain material is hereinafter referred to as “subfilm defect”.

  Here, as a feature of EPMA, continuous X-rays due to bremsstrahlung of electrons are generated in addition to characteristic X-rays from a region where electrons are diffused. Continuous x-rays do not have a specific peak. An element is detected when the intensity of the characteristic X-ray becomes stronger than the continuous X-ray. These characteristic X-rays and continuous X-rays are collectively referred to as “detected X-rays” hereinafter.

  As shown in FIG. 3, the behavior of electrons in a substance at the time of EB irradiation is such that when the acceleration voltage of EB irradiated at the time of analysis increases, the region where electrons diffuse becomes larger and the region where detection X-rays can be obtained increases. There is a relationship. As shown in FIG. 4, even when the acceleration voltage of EB is the same, the region where electrons diffuse depending on the substance is different, that is, the region where detection X-rays are obtained is different.

  Therefore, as shown in FIG. 5, when analyzing a subfilm defect, it is necessary to consider the acceleration voltage depending on the material of the film, the thickness of the film covered with the foreign matter, and the like. For example, as shown in FIG. 5A, when the region where the characteristic X-ray is obtained does not reach the subfilm defect, the obtained spectrum does not include element information of the subfilm defect. In addition, as shown in FIG. 5B, the region where the detected X-ray is obtained is larger than the size of the subfilm defect, and the continuous characteristic X-ray energy is higher than the characteristic X-ray intensity from the subfilm defect. If the intensity of the X-ray exceeds the level, the subfilm defect element cannot be detected.

  For these reasons, it is important to set the acceleration voltage in order to detect the element of the subfilm defect by EPMA. However, since the setting of the acceleration voltage in such a case has been determined empirically by the user, it takes time because trial and error are required.

  Here, a technique for detecting the composition of foreign matter under the film by changing the acceleration voltage is disclosed in Patent Document 3. Further, Patent Document 4 discloses a method for setting an acceleration voltage in consideration of an electron penetration depth. However, there is no disclosure of a method that explicitly indicates an acceleration voltage suitable for the analysis of the subfilm defect.

  An object of the present invention is to solve the above-mentioned problems and to provide means for efficiently setting a suitable acceleration voltage for subfilm defects.

  In order to achieve the above object, according to the present invention, there is provided a method of irradiating an object with an electron beam and performing elemental analysis of a defect of a device based on the obtained X-ray spectrum. The step of obtaining the X-ray spectrum by calculation and the step of determining the acceleration voltage of the electron beam to be irradiated based on the calculation are performed.

  In order to achieve the above object, the present invention provides an apparatus for elemental analysis of device defects, an electron source that irradiates the device with an electron beam, an electron optical system that focuses the electron beam on the device, and an emission. A detection system for detecting the emitted X-rays, a simulation condition acquisition unit, an acceleration voltage calculation unit, and a control device for irradiating an electron beam with the acceleration voltage.

  Further, in the present invention, in a method for comparing X-ray spectra obtained by irradiating a defect with an electron beam, an X-ray spectrum in a region having no defect having the same background as the defect is obtained as a reference spectrum, In the defect spectrum, a comparison region is set in the vicinity of a peak that does not exist in the reference spectrum, and a difference from another defect spectrum is calculated in the comparison region.

  As described above, according to the present invention, the material of the device film in which the defect to be analyzed exists, the density of the film, the film thickness, the element that needs to be distinguished from the material of the film, the defect to be analyzed Since the acceleration voltage capable of detecting the subfilm defect is calculated by simulation with the size of the input as an input, it is possible to easily set the acceleration voltage suitable for the elemental analysis of the subfilm defect.

  In addition, in determining the degree of similarity of the defect spectrum, since the degree of similarity is determined only in the region representing the feature of the element, high-precision similarity determination is realized regardless of the elements present in the background portion. be able to.

In the embodiment of the present invention described below, a device for observing defects and elemental analysis is described as a semiconductor wafer. However, the present invention is not limited to this, and other devices may be used.
Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  First, with reference to FIG. 6, a specific example of each apparatus and its connection configuration on the wafer production line will be described. In the figure, 20 is a data management server, 21 is a semiconductor manufacturing apparatus, 22 is an inspection apparatus, 23 is a review apparatus, 24 is an analysis apparatus, 25 is a review / analysis apparatus, and 26 is a network.

  In the figure, the production line has a configuration in which a semiconductor wafer production device 21, an inspection device 22, a review device 23, an analysis device 24, and a review / analysis device 25 are interconnected by a data management server 20 and a network 26. Yes.

  The manufacturing apparatus 21 is used for manufacturing a semiconductor wafer such as an exposure apparatus or an etching apparatus. The inspection device 22 inspects a defect position. For example, a light beam spot is scanned on a semiconductor wafer, and the defect position is specified from the degree of irregular reflection, or formed from two chips. An image of a certain pattern is acquired, these images are compared, a portion that is different between these images is regarded as a defect, and the defect position is detected.

  The review device 23 is for observing defects based on the inspection information of the inspection device 22. The review device 23 moves the stage on which the semiconductor wafer is mounted, and on the semiconductor wafer based on the defect position information output from the inspection device 22. Position the target defect and observe the defect. As the observation method, for example, an optical microscope or SEM is used.

  The analysis device 24 performs elemental analysis using, for example, EDX or Auger electron spectroscopy. Auger electron spectroscopy is a method that detects and analyzes Auger electrons emitted from an object when the object is irradiated with an electron beam, and is a well-known method. The review / analysis device 25 is a device that can perform defect observation and elemental analysis.

  Note that these inspection, observation, and analysis apparatuses are not necessarily separated, and may be combined, for example, such that inspection and review can be performed in the same apparatus.

  The data management server 20 manages data obtained by the inspection device 22, the review device 23, the analysis device 24, and the review / analysis device 25. The review device 23 and the analysis device 24 include the data management server 20 It is possible to acquire information such as the defect position coordinates output from the inspection apparatus 22 via.

  Although an example of connection is shown here, such a connection configuration may be used as long as data can be used between devices.

  Next, an embodiment of the present invention used for such a connection configuration will be described.

  FIG. 7 is a configuration diagram showing an embodiment of an elemental analyzer according to the present invention. WF is a semiconductor wafer, EB is an electron beam, 1 is an electron source, 2 is an electron optical system, 3 is an imaging device using a scanning electron microscope, 4 and 5 are detectors, 6 is a semiconductor X-ray detector, 7 is XY stage, 8 is monitor, 9 is computer, 10 is controller, 11 is monitor, 12 is computer, 13 is secondary electron, 14 is reflected electron, 15 is characteristic X-ray, 16 is computer, 17 is calculation A condition acquisition unit, 18 is an EDX condition calculation unit, and 19 is a monitor. This embodiment corresponds to the review / analysis device 21 in FIG.

  In the figure, an electron source 1, an electron optical system 2, detectors 4 and 5, a semiconductor X-ray detector 6 and an XY stage 7 constitute an SEM, which is mounted on the XY stage 7. The semiconductor wafer WF is used as the imaging device 3.

A semiconductor wafer WF to be subjected to review and elemental analysis is mounted on an XY stage 7. The XY stage 7 is controlled to move in the X and Y directions by the control device 10 based on a control signal from the computer 9. The imaging device 3 using the SEM enlarges and images the semiconductor wafer WF fixed to the XY stage 7. That is, the electron beam EB emitted from the electron source 1 is converged and scanned by the electron optical system 2 to irradiate the semiconductor wafer WF to be measured, and secondary electrons obtained from the semiconductor wafer WF by this irradiation. 13 and reflected electrons 14 are detected by detectors 4 and 5, respectively, and processed by computer 9 to generate an SEM image of semiconductor wafer WF.
This SEM image can be observed on the monitor 8. The semiconductor X-ray detector 6 detects the characteristic X-ray 15 emitted from the defect of the semiconductor wafer WF by irradiation with the electron beam EB, and converts it into an electric signal. This electrical signal is processed by the computer 12 and displayed on the monitor 11 as a spectrum. In the imaging device 3, by controlling the position of the XY stage 7, any position on the semiconductor wafer WF can be analyzed.

  In the above configuration, the user first calculates the material of the surface film, the film thickness, the density, the element assumed to be contained in the defect, that is, the element to be separated from the material of the surface film, and the element analysis in the calculation condition acquisition unit 17. Enter the minimum size of the target defect.

  The EDX condition calculation unit 18 calculates an acceleration voltage suitable for elemental analysis according to the input condition.

  The acceleration voltage value calculated by the EDX condition calculation unit 18 is sent to the computer 9. The computer 9 controls the power supply (not shown) of the electron optical system 2 to set the calculated acceleration voltage value. In this way, the electron optical system 2 irradiates the sample with an electron beam under the condition where the acceleration voltage is set, and the characteristic X-ray 15 emitted from the sample and the defect portion is detected by the semiconductor X-ray detector 6 for elemental analysis. Is done.

  Note that the computers 9 and 12 may be connected via a network so that the processing content of the computer 12 is read by the computer 9 and displayed on the monitor 8. Further, the computer 12 may perform the processing in the computer 12, and the electrical signal output from the semiconductor X-ray detector 6 may be processed by the computer 9 to obtain spectral information.

  Further, the processing of the computer 16 including the processing of the calculation condition acquisition unit 17 and the EDX condition calculation unit 18 may be performed by the computer 9.

  Here, the acceleration voltage suitable for elemental analysis will be described.

  In the manufacturing process, inspection is performed for each series of processes as described above. Therefore, the defect coordinates detected in a certain inspection process are compared with the defect coordinates detected in the inspection process performed in the previous stage, and a series of defects are subtracted by subtracting the detected defects. Only defects newly generated in the process can be targeted. Here, as shown in FIG. 5, when the depth at which the detected X-ray is obtained does not reach the subfilm defect, the element information of the defect cannot be obtained. In addition, when the range of detection X-rays is larger than the size of the defect and the intensity of continuous X-rays at the energy of the characteristic X-ray exceeds the intensity of the characteristic X-ray from the defect, Elemental information cannot be obtained.

  Therefore, the first condition is that at least an in-film defect within the range of the film thickness to be analyzed is an object, and an acceleration voltage that can detect characteristic X-rays of elements contained in the under-film defect is a suitable acceleration voltage. And

  Further, in the detected X-ray spectrum, there is a case where the characteristic X-ray energy is close, so that it is difficult to distinguish. In such a case, if any of the elements to be separated has two or more characteristic X-ray energies, a suitable acceleration voltage can be used to detect elements that can be easily separated within the range where elemental analysis of the subfilm can be performed. Voltage. For example, when Si and W are separated, the characteristic X-ray energy of Si is 1.739 keV, whereas the characteristic X-ray energy (Mα) of W is 1.789 keV, and it is difficult to separate them. However, W also has characteristic X-ray energy at 8.394 keV called Lα1. Therefore, in a process that requires separation of Si and W, an acceleration voltage at which the characteristic X-rays are emitted is a suitable acceleration voltage. This is the second condition.

  Here, when setting the acceleration voltage that satisfies the second condition, it is assumed that the first condition is satisfied.

  Next, a first embodiment of the calculation condition acquisition method of the calculation condition acquisition unit 17 will be described.

  As shown in FIG. 4, how the region where the detection X-rays in the substance are obtained varies depending on the target material even at the same acceleration voltage. Such a situation can be approximately reproduced by computer simulation. That is, by giving the conditions necessary for the simulation, it is possible to obtain the range of the acceleration voltage that satisfies the above-mentioned preferable plan.

  Conditions necessary for the computer simulation include the material, density, and film thickness of the film. When there are a plurality of films, the condition is given by the number of films. In addition, in order to determine whether or not the element contained in the subfilm defect can be separated with respect to the material of the film, an element that may be contained in the defect, that is, an element to be separated is input. Further, in order to calculate the volume from which the characteristic X-rays from the subfilm defects are emitted, the minimum dimension of the defect nucleus to be detected is input.

  These inputs may be performed by a GUI displayed on the computer 16. An example of the GUI is shown in FIG. Input conditions can be registered for each combination of product type and process name. First, the product name and process name are input from an input device such as a keyboard, or selected from a pull-down menu. Then enter the material, density, and thickness. When there are a plurality of films, they can be input for each layer. The layer can be specified by entering a number or using the up and down buttons.

  As the material, a chemical formula may be directly input using an input device such as a keyboard, or a pre-registered material name may be selected from a pull-down menu. Alternatively, the periodic table of elements may be displayed by pressing a reference button, and the corresponding element may be input by selecting with an input device such as a mouse.

  Similarly, the density and thickness may be input directly by an input device such as a keyboard, or a pre-registered numerical value may be selected from a pull-down menu. Further, the density may be associated with the material name in advance, and the density corresponding to the input of the material name may be automatically input.

  The fractional element may be entered directly using an input device such as a keyboard, and the periodic table of the element is displayed by pressing the reference button, and the corresponding element is selected using an input device such as a mouse. You may input by.

  When input of all items is completed, these conditions are stored in association with the product name and process name by pressing a registration button. The stored data can be easily recalled by specifying the product name as the process name and pressing the recall button.

  A second embodiment of the calculation condition acquisition method of the calculation condition acquisition unit 17 will be described with reference to FIG.

  A processing condition database in which processing conditions are registered for each type and process of semiconductor wafer is created and connected to a network. The processing conditions refer to the film material, density, film thickness, tolerance, etc. in the case of a film forming process, and how much thickness is etched in the etch back process to determine the final film thickness. Refers to a condition such as In addition, information such as which process is inspected is added to this database in accordance with the processing procedure. In addition, a defect element database in which detected elements are registered for defects detected by dummy wafer inspection or the like in a single processing apparatus in each process is created and connected to a network. A computer including the calculation condition acquisition unit is also connected to the network.

  When the user inputs the product name and process name of the wafer to be analyzed, the calculation condition acquisition unit accesses the processing condition database, and the process sandwiched between the previous inspection in the corresponding product and process. A group is selected, and the material, density, and film thickness of the film in the corresponding process are acquired. When there are two or more generated films, the data is acquired for each layer.

  Further, the section selected by the previous processing condition database is selected from the defect information database, and the element of the defect generated in the section is acquired and set as the fractional element. The database does not have to manage all the information, but may manage only a part of the information. Further, the database may be included in a computer constituting the SEM, and any form may be used as long as information can be acquired from the computer constituting the SEM. In addition, as shown in FIG. 9, the calculation condition acquisition unit and the EDX condition calculation unit may be connected to the SEM via a network, and have a function in which the calculated EDX condition is reflected during the analysis of the SEM. Any configuration is acceptable.

  Next, an embodiment of the EDX condition calculation unit 18 will be described.

  As described above, the EDX condition calculation calculates an acceleration voltage suitable for analysis by performing a simulation from the condition acquired from the calculation condition acquisition unit. The simulation method is a Monte Carlo method that estimates the overall state by probabilistically determining the scattering process and energy loss process of each incident electron using random numbers and calculating this for many electrons. Can be used.

  According to the Monte Carlo method, it is possible to simulate an electron diffusion region when a material is irradiated with EB and a spectrum of X-rays generated in association therewith. The calculation method of the electron behavior and the X-ray spectrum by the Monte Carlo method is Non-Patent Document 1 for the electron scattering process, Non-Patent Document 2 for the calculation of the X-ray spectrum, and Non-Patent Document for the calculation of the spectrum of the continuous X-ray. It is disclosed in Document 3 and the like.

  In the simulation, the X-ray spectrum obtained under the conditions can be calculated by giving the conditions obtained from the calculation condition acquisition unit.

  FIG. 10 shows a schematic diagram of an X-ray spectrum obtained when EB is irradiated on the analysis target. As described above, the obtained spectrum is composed of continuous X-rays having no specific peak and characteristic X-rays having energy specific to the substance.

  When the peak intensity of characteristic X-ray is a and the intensity of continuous X-ray in the energy of the characteristic X-ray is b, in principle, the presence of an element corresponding to the characteristic X-ray can be detected when a / b> 1. become. Hereinafter, the a / b is described as “characteristic X-ray evaluation value”.

Here, there are elements having a plurality of characteristic X-ray energies. However, when high-energy characteristic X-rays are targeted, it is necessary to increase the accelerating voltage of EB to be irradiated, resulting in poor spatial resolution and sample damage. Therefore, priority is given to low energy characteristic X-rays.
Next, a first embodiment of a method for determining an acceleration voltage suitable for the above-described analysis by simulation will be described with reference to FIG.

  First, for a characteristic X-ray of an element to be sorted inputted in the calculation condition acquisition unit, that is, an element included in the defect, a range of acceleration voltage having a characteristic X-ray evaluation value exceeding 1 is calculated. That is, first, in STEP 100, a model is set based on the conditions obtained from the calculation condition acquisition unit. For example, the input material and film thickness are set as simulation conditions by the number of input films. And the foreign material of the input size is arrange | positioned in the lowest layer of the input film | membrane. The shape is a shape such as a sphere or a cube that is easy to model geometrically.

  Next, in STEP 101, the material of the foreign material is set to one of the elements set as the separation element.

  Next, in step 102, the acceleration voltage is set to an initial value. This may be the minimum acceleration voltage value of the analyzer, or may be registered in advance by another input means.

  Next, a simulation is performed at the acceleration voltage set in STEP 103 to calculate a spectrum.

  Next, in STEP 104, a characteristic X-ray evaluation value is calculated based on the characteristic X-ray energy of the element set in STEP 101 as a foreign material.

  Next, in STEP 105, it is determined whether the acceleration voltage used for the simulation is a final value for ending the simulation. This value may be the maximum acceleration voltage value of the analyzer, or may be registered in advance by another input means.

  If the acceleration voltage value is not the final value for ending the simulation, the acceleration voltage is increased at STEP 106 and the process is repeated from STEP 103 again. The increase amount of the acceleration voltage may be set in advance by the device, or may be registered in advance by another input means, but at least three types of acceleration voltages so that at least the tendency of change can be read. Set to simulate with.

  In this example, the procedure for simulating by increasing the acceleration voltage has been described. However, the effect is the same even if the simulation is performed by decreasing the acceleration voltage.

  If the accelerating voltage is the end value, it is determined in STEP 107 whether there is an element that has not yet been simulated as an element that is to be classified as a defect, and if there is, the defect element is selected as another classification target in STEP 108. Set to element and repeat from STEP102.

  When the simulation is completed for all elements to be sorted, the results are displayed in STEP109.

  An example of the result display is shown in FIG. On the display screen, conditions set by the condition acquisition unit of the simulation, such as a product name, a process name, and a minimum detection size, are displayed. The graph is a two-dimensional graph with the acceleration voltage on which the simulation is performed as the horizontal axis and the characteristic X-ray evaluation value at the acceleration voltage as the vertical axis.

  The range of the horizontal axis is the acceleration voltage initial value (shown as Vmin in the figure) and the final acceleration voltage value (shown as Vmax in the figure) described in FIG. The graph in the figure shows the results of the substances A and B that are the objects of separation. Here, the accelerating voltage that can be detected is in the range where the characteristic X-ray evaluation value> 1 in the simulation. However, since the intensity of the spectrum actually obtained has statistical fluctuations in the X-ray emission process, Considering the above, a certain value α greater than 1 is displayed as a detectable threshold value.

  The acceleration voltage values (indicated as VA and VB in the figure) that can be detected are also displayed on the graph. From this graph, the user can determine that if the acceleration voltage is equal to or higher than VB, both the subfilm defects composed of the substance A and the substance B can be analyzed. At the same time, an acceleration voltage range in which all substances targeted for separation can be detected is displayed as “detectable acceleration voltage”. If there is an element that cannot be detected regardless of how the acceleration voltage is set, this fact is displayed.

  The acceleration voltage obtained from this display is set in the apparatus as the acceleration voltage at the time of analysis, or set in a recipe, so that the elemental analysis of subfilm defects in the product / process is made suitable for analysis. Can be done.

  In addition, even if the display indicates that detection is impossible, it is possible to know in advance that no defect information is included in the spectrum when the analysis is performed, so that the element that causes the defect is not included. This is effective because it can avoid erroneous determination.

  In the graph display, a curve may be applied to the points plotted in the graph and displayed as a continuous curve. In this case, the continuous curve may display the acceleration voltage at which the characteristic X-ray evaluation value is α or more as the acceleration voltage value at which detection is possible.

  Further, only one of the graph and the “detectable acceleration voltage” may be displayed.

  Further, the value of the detectable acceleration voltage may be automatically registered in the recipe of the analyzer without performing the display.

  Here, α is determined in consideration of the statistical fluctuation of the X-ray emission process in the spectrum actually obtained as described above. The value may be given by the user as a constant or as a function of the continuous X-ray intensity value. An embodiment in the case of giving by function will be described below.

  In general, when the intensity of the X-ray is N, it is known that the standard deviation σ of the statistical fluctuation width is given by √N. Therefore, for example, a condition is set such that the peak intensity of characteristic X-rays in the simulation is 3σ or more of the intensity of continuous X-rays. In this case, the characteristic X-ray evaluation value α at which characteristic X-rays are detected can be given by (N + 3 * √N) / N, where N is the intensity of continuous X-rays. The actual value of N depends on the signal addition time at the time of analysis execution, but it is only necessary to take a correspondence between the continuous X-ray intensity at the preset signal addition time and the intensity calculated by the simulation. In this case, for example, when the intensity of continuous X-rays at a certain energy is calculated as 20, α is about 1.67.

  A second embodiment of the method for calculating the acceleration voltage range will be described below.

  In the first embodiment shown in FIG. 11, the simulation is performed for all the elements to be separated, and the installation value of the acceleration voltage at the time of analysis is determined by taking the logical product of the detectable acceleration voltages of the individual elements. In this case, if the number of elements to be separated increases, the simulation takes time.

  Therefore, among the elements to be separated, the highest acceleration voltage is required for detection in order to evaluate the element that is most difficult to detect in order to evaluate detectability and the minimum acceleration voltage required for element detection. An element is selected, and an acceleration voltage necessary for analysis is set by simulation of only the element.

An index for detecting elements that are difficult to detect is a mass absorption coefficient. This is a coefficient related to the rate at which the intensity decreases when X-rays of a certain energy pass through the substance, and has the relationship shown in the following equation (1).
I / I0 = exp (-[mass absorption coefficient] * ρt) (1)
Here, I is the measured X-ray intensity, I0 is the generated X-ray intensity, ρ is the density of the substance through which X-rays pass, and t is the transmission distance.

  The greater the mass absorption coefficient, the greater the rate at which X-rays are absorbed. That is, it is considered that the element having the largest mass absorption coefficient among the elements to be separated is difficult to detect. The mass absorption coefficient is expressed as a function value of the intensity of transmitted X-rays for each substance. Therefore, it is possible to improve the efficiency by comparing the mass absorption coefficient in the characteristic X-ray energy of the element constituting the film, selecting the element having the largest mass absorption coefficient, and performing simulation.

  In this case, a flow for performing the simulation is shown in FIG. Steps that are the same as in FIG. 11 are given the same numbers, and descriptions thereof are omitted.

  First, after STEP 100, in STEP 200, the mass absorption coefficient of the element of the film obtained from the calculation condition acquisition unit is referred from a separately stored mass absorption coefficient database.

  Usually, since this value is discretely given as a coefficient corresponding to a certain energy, a mass absorption coefficient corresponding to an arbitrary energy is calculated by interpolating from a mass absorption coefficient in the vicinity of the energy. This is calculated for each element given as a separation target.

  Next, in STEP 201, an element having the largest mass absorption coefficient among the mass absorption coefficients calculated in STEP 200 is selected as a simulation target. Moreover, the element with the highest excitation energy for radiating characteristic X-rays among the fractionated elements is selected as a simulation target in the same manner.

  Thereafter, processing corresponding to the STEP numbers described in FIG. 11 is sequentially performed.

  By doing so, the number of simulations can be reduced.

  A third embodiment of the method for calculating the acceleration voltage range will be described below.

  As described above, in the detected X-ray spectrum, if the energy of the characteristic X-ray is close, it may be difficult to identify the element. In such a case, it is desirable to increase the acceleration voltage within a range in which another element to be separated can be detected, and to set the acceleration voltage so as to obtain an energy peak that is not close to the defect. The simulation flow in that case is shown in FIG. Steps that are the same as in FIG. 13 are given the same numbers, and descriptions thereof are omitted.

  First, after STEP 100, in STEP 300, an element whose characteristic X-ray energy is close is searched for in the element to be separated from the film element obtained from the calculation condition acquisition unit. As the threshold value for determining proximity, an appropriate value may be registered in advance according to the resolution of the detector, or the user may register using another input means.

  Next, in STEP 301, it is determined whether or not there is an element whose characteristic X-ray energy is close. If not, STEP 302 is skipped and STEP 200 is executed.

  If there is an element with close characteristic X-ray energy, select an element with multiple characteristic X-ray energies from among the plurality of elements determined to have close characteristic X-ray energy in STEP 302 and close to others Another characteristic X-ray energy that is not to be set is set as an object to be considered for subsequent processing. At this time, if an element having a plurality of characteristic X-ray energies does not exist, this is displayed to the user and the processing is skipped.

  Step 200 and subsequent steps are the same as those in FIG. 13, but the element selected in STEP 302 calculates the mass absorption coefficient and the characteristic X-ray evaluation value for the characteristic X-ray energy set in the STEP.

By doing so, the spectrum of the element can be distinguished from other spectra.
Further, in this flow, STEP 200 and subsequent steps may be replaced with STEP 101 and subsequent steps in FIG. 11, and a simulation of all elements may be performed to set a more preferable acceleration voltage.

  Here, in the product wafer that has passed through the film formation process and the dummy process wafer in the processing apparatus for the film formation process, the generated defect exists under the film. However, elemental analysis can be easily performed.

  In the dummy wafer of the processing apparatus, information effective for estimating the cause of the defect may be obtained by performing the elemental analysis of the defect. For example, when Fe, Cr, and Ni, which are main components of stainless steel, are detected simultaneously from defects, it can be determined that there is a high possibility that the material is generated from the processing apparatus using stainless steel.

  On the other hand, there is a case where information effective for estimating a defect generation apparatus can be obtained by obtaining defect element information in a product wafer. For example, it can be determined that there is a high possibility that it has been generated from a processing apparatus using stainless steel as an apparatus member, such as the stainless steel. Further, when an element used in the reaction gas such as F is detected, it can be determined that there is a high possibility that the element is generated as a reaction product from an apparatus using the gas containing the element.

  Thus, in order to narrow down the apparatus in which the problem has occurred based on the contained elements, as shown in FIG. 15, it is necessary to compare the spectrum of the defect generated from each processing apparatus with the spectrum obtained by the defect analysis of the product. Necessary.

  This is shown in FIG. In FIG. 15, the spectra obtained by processing A and processing B show the analysis results of defects obtained by starting a dummy wafer on the processing apparatus. That is, these defects are defects generated in the apparatus for performing processing A and the apparatus for performing processing B, respectively. The spectrum obtained by the inspection shows the spectrum of defects on the product wafer. Moreover, the spectrum of the reference part shown in the figure represents the spectrum acquired in the same background as the defect part and the part where no defect exists.

  The letters “a”, “b”, “c”, and “d” appended to the spectrum peaks in the figure indicate that the respective peaks are characteristic X-ray peaks corresponding to the substances a, b, c, and d. Further, when the spectrum of the defect portion and the spectrum of the reference portion are compared, a defect containing the substance a in the apparatus for processing A, a defect containing the substance d in the apparatus for processing B, and a defect containing the substance a in the product wafer Can be estimated.

  Here, when the spectrum of the defective part of the product wafer is compared with the spectrum of the dummy wafer defective part of each processing step, the peak of the element other than the element contained in the defect, that is, the peak of the spectrum of the reference part is superimposed. A reasonable result cannot be obtained. For example, in the spectrum of the product wafer, when the spectrum of the defect portion is compared with the spectrum of the reference portion, it can be estimated that the defect portion contains the substance a, and it can be estimated that the defect is attached by the processing A. However, since the peaks of the substances b and c generated in the film forming process are high, there is a high possibility that it is determined that the spectrum is similar to the spectrum of the defect portion of the processing B.

  Thus, when comparing the spectrum obtained from the defective part of the product in this way with the spectrum obtained from the defective part of the dummy wafer of the apparatus, it is not obtained in the reference spectrum in the product wafer as shown in FIG. A spectrum comparison region is set in the vicinity of the element peak, and in this region, a comparison is made with a spectrum obtained from a defect of a dummy wafer of each processing apparatus.

  In the comparison method, the absolute value of the difference between the two waveforms in the set area is summed, a value normalized by the size of the area is calculated, and it is determined that the smaller the value, the more similar. This value is hereinafter referred to as similarity. In calculating the similarity, the absolute difference value of the differential value of the waveform may be used, or the absolute difference value of the second-order differential value of the waveform may be used.

  At this time, it can be determined that there is a high possibility that the defect has occurred in the processing apparatus having the defect having the smallest similarity.

  Also, as shown in FIGS. 17A and 17B, with reference to the spectrum of defects detected on the dummy wafer of the processing apparatus, a spectrum comparison region is provided in the vicinity of an element peak not included in the reference spectrum. It may be set and compared with the spectrum of product defects.

  Also in this case, it can be determined that the defect is highly likely to occur in a processing apparatus having a defect with the smallest similarity.

It is the flowchart explaining the outline | summary of the semiconductor manufacturing process. It is a figure which shows an example of the information obtained from an inspection apparatus or a review / analysis apparatus. It is sectional drawing of the sample explaining the difference of the electron diffusion area by an acceleration voltage. It is sectional drawing of the sample explaining the difference of the electron diffusion area by the material of a sample. It is sectional drawing of the sample which showed an example of the electron diffusion area | region in a subfilm defect analysis. It is the block diagram which showed an example of the apparatus structure in a production line. It is a block diagram which shows schematic structure of the review and analysis apparatus in this invention. It is a front view of the input screen of the simulation conditions in this invention. It is the block diagram which showed an example of the review and analysis apparatus structure in this invention. It is a graph which shows an example of an X-ray spectrum. It is the flowchart which showed one Example of the simulation procedure of this invention. It is a front view of the screen which shows the simulation result of this invention. It is the flowchart which showed one Example of the simulation procedure of this invention. It is the flowchart which showed one Example of the simulation procedure of this invention. It is a graph which shows the relationship between the acceleration energy of the electron beam of a defect part and a reference part for every process, and the spectrum intensity which show the concept of problem process candidate extraction by spectrum comparison of this invention. It is a graph which shows the relationship between the acceleration energy of the electron beam of a defect part and a reference part, and spectrum intensity which shows the Example of the spectrum comparison method by the comparison area setting of this invention. (A) And (b) is a graph which shows the relationship between the acceleration energy of an electron beam and spectrum intensity which showed the Example of the spectrum comparison method by the comparison area setting of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Electron source 2 ... Electron optical system 3 ... Electron microscope imaging device 4 ... Secondary electron detector 5 ... Backscattered electron detector 6 ... Semiconductor X-ray detector 7 ... XY stage 8 ... Monitor 9 ... Computer 10 ... Control device 11 ... Monitor 12 ... Computer 16 ... Computer 17 ... Calculation condition acquisition part 18 ... EDX condition calculation part 19 ... Monitor 20 ... Data management server 21 ... Manufacturing equipment 22 ... Inspection equipment 23 ... Review equipment 24 ... Analysis equipment 25 ... Review / Analysis equipment 26 ... Network

Claims (6)

A method for performing elemental analysis of a device defect by an X-ray spectrum obtained by irradiating an object with an electron beam,
Entering simulation conditions;
Obtaining an X-ray spectrum by calculation based on the input conditions;
And a step of determining an accelerating voltage of an electron beam to be irradiated based on the X spectrum obtained by the calculation.   2. The simulation condition includes at least one of a material, density, film thickness, an element that needs to be distinguished from the material of the film, and a defect size, which constitute a device. Elemental analysis method for defects.   The method further includes a step of displaying the simulation result on a screen as a graph indicating a relationship between an acceleration voltage, a characteristic X-ray intensity of the element of interest, and a ratio of continuous X-ray intensity with the characteristic X-ray energy. The elemental analysis method for defects according to claim 1.   2. The defect elemental analysis method according to claim 1, further comprising a step of displaying, on the screen, a range of acceleration voltage capable of detecting the element of interest as a result of the simulation. An electron source that emits an electron beam;
Electron optical system means for converging an electron beam emitted from the electron source and irradiating the sample;
Detection system means for detecting X-rays emitted from the sample by irradiation of the electron beam;
An input means for inputting simulation conditions;
Accelerating voltage calculating means for calculating the accelerating voltage of the electron beam using simulation conditions input from the input means;
Control means for controlling the electron optical system means to irradiate the electron beam with the acceleration voltage calculated by the acceleration voltage calculation means;
A signal obtained by detecting the X-ray generated from the sample irradiated with the electron beam by the electron optical system means with the detection system means in a state in which the acceleration voltage is controlled by the control means is processed into the sample. An analysis means for extracting information on contained elements;
An element analysis device for defects, comprising output means for outputting a result of analysis by the analysis means. A method for comparing X-ray spectra obtained by irradiating defects with an electron beam,
An X-ray spectrum in a non-existent region having the same background as the defect is obtained as a reference spectrum;
Set a comparison region in the vicinity of a peak that does not exist in the reference spectrum in the spectrum of the defect,
A spectrum comparison method comprising calculating a difference from another defect spectrum in the comparison region.

Images