Electron detection device and electron beam inspection device

Abstract

PURPOSE: To provide an electron detection device and an electron beam inspection device, and detect electrons emitted from a sample with high sensitivity and low noise.CONSTITUTION: An electron detection device includes an objective lens that reduces an emitted electron beam and applies it onto a sample, an electron beam scanning device that performs planar scanning of the electron beam applied onto the sample, means that accelerates the energy of electrons emitted from the sample by the electron beam applied by the electron beam scanning device, a first element having a light receiving surface of a first area which converts the accelerated electrons into light, a second element that collects the light converted from the electrons by the first element such that the area is smaller than the first area, and a third element having an area smaller than the first area which receives, detects, and converts the light collected by the second element into an electrical signal. The electron detection device is configured such that the light generated on the electron light receiving surface of the first element is collected using the second element and detected as an electrical signal by the third element, thereby detecting electrons with high sensitivity and low noise.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an electron detection device and an electron beam inspection device for detecting electrons emitted from a sample. [Background technology]

[0002] Semiconductor devices are shrinking every year according to Moore's Law, and cutting-edge devices are nearing the mass production stage. The minimum feature size of even the latest devices is less than 20 nm. In order to achieve this character size, exposure techniques capable of forming smaller patterns are required.

[0003] Conventionally, a laser beam with a wavelength of 193 nm has been used for exposure, but Since the dimensional limits have already been exceeded by a large margin, in recent years, lithography technology that utilizes EUV light with a wavelength of 13.5 nm has been actively pursued.

[0004] This technology is said to have originated in Japan, but companies such as ASML (registered trademark) have been researching and developing it for practical use for over 10 years. The optical system of the exposure tool itself is almost complete, but until a few years ago, EUV exposure tools were skipped for use in 20nm generation exposure because they were unable to obtain the light source power of 100W to 200W required for economical mass production.

[0005] Instead, so-called double or triple exposure techniques have been developed that can achieve even smaller feature sizes by repeating the exposure process multiple times using the existing 193 nm wavelength.

[0006] In principle, immersion lithography using a laser beam with a wavelength of 193 nm is repeated multiple times. This makes it possible to create patterns of arbitrarily small sizes. However, it is known that limitations of the repeated exposure process include the need for alignment accuracy on the order of nm with one digit higher precision than before, and large roughness resulting from the large molecular structure of chemically amplified resists optimized for 193 nm exposure. Currently, double patterning, which repeats the exposure process twice up to what is called the economic limit, is in practical use, and in relatively simple memory devices, it is used to achieve line & space of 20 nm to 16 nm.

[0007] Not only simple memory devices but also CPUs and logic devices need to reduce pattern sizes for function expansion and power consumption reduction. Logic devices use complex patterns without repetition, so creating logic devices by double or triple exposure requires quite complex patterns. To divide a pattern that can originally be realized by a single exposure into patterns on two photomasks so that they can be used in multiple exposures, very complex calculations are required. Depending on the pattern, the division calculation may diverge and the necessary result may not be obtained. For the purpose of avoiding these complexities, a lithography simplification technique called complementary lithography, mainly proposed by Intel (registered trademark), is being attempted to be used. This exposure method is characterized by using a pattern in which a complex logic circuit is reduced to a simple L&S pattern like that of a memory circuit. By doing so, since the pattern of a complex logic device is limited to only the L&S pattern that is most easily exposed and the process of cutting its lines, complex pattern division calculations are not required, the process is simple, and it is said that it can go up to about 8 nm in terms of calculation. For the purpose of avoiding these complexities, a lithography simplification technique called complementary lithography, mainly proposed by Intel (registered trademark), is being attempted to be used. This exposure method is characterized by using a pattern in which a complex logic circuit is reduced to a simple L&S pattern like that of a memory circuit. By doing so, since the pattern of a complex logic device is limited to only the L&S pattern that is most easily exposed and the process of cutting its lines, complex pattern division calculations are not required, the process is simple, and it is said that it can go up to about 8 nm in terms of calculation. For the purpose of avoiding these complexities, a lithography simplification technique called complementary lithography, mainly proposed by Intel (registered trademark), is being attempted to be used. This exposure method is characterized by using a pattern in which a complex logic circuit is reduced to a simple L&S pattern like that of a memory circuit. By doing so, since the pattern of a complex logic device is limited to only the L&S pattern that is most easily exposed and the process of cutting its lines, complex pattern division calculations are not required, the process is simple, and it is said that it can go up to about 8 nm in terms of calculation. For the purpose of avoiding these complexities, a lithography simplification technique called complementary lithography, mainly proposed by Intel (registered trademark), is being attempted to be used. This exposure method is characterized by using a pattern in which a complex logic circuit is reduced to a simple L&S pattern like that of a memory circuit. By doing so, since the pattern of a complex logic device is limited to only the L&S pattern that is most easily exposed and the process of cutting its lines, complex pattern division calculations are not required, the process is simple, and it is said that it can go up to about 8 nm in terms of calculation. For the purpose of avoiding these complexities, a lithography simplification technique called complementary lithography, mainly proposed by Intel (registered trademark), is being attempted to be used. This exposure method is characterized by using a pattern in which a complex logic circuit is reduced to a simple L&S pattern like that of a memory circuit. By doing so, since the pattern of a complex logic device is limited to only the L&S pattern that is most easily exposed and the process of cutting its lines, complex pattern division calculations are not required, the process is simple, and it is said that it can go up to about 8 nm in terms of calculation. For the purpose of avoiding these complexities, a lithography simplification technique called complementary lithography, mainly proposed by Intel (registered trademark), is being attempted to be used. This exposure method is characterized by using a pattern in which a complex logic circuit is reduced to a simple L&S pattern like that of a memory circuit. By doing so, since the pattern of a complex logic device is limited to only the L&S pattern that is most easily exposed and the process of cutting its lines, complex pattern division calculations are not required, the process is simple, and it is said that it can go up to about 8 nm in terms of calculation.

[0008] For the purpose of avoiding these complexities, a lithography simplification technique called complementary lithography, mainly proposed by Intel (registered trademark), is being attempted to be used. This exposure method is characterized by using a pattern in which a complex logic circuit is reduced to a simple L&S pattern like that of a memory circuit. By doing so, since the pattern of a complex logic device is limited to only the L&S pattern that is most easily exposed and the process of cutting its lines, complex pattern division calculations are not required, the process is simple, and it is said that it can go up to about 8 nm in terms of calculation. For the purpose of avoiding these complexities, a lithography simplification technique called complementary lithography, mainly proposed by Intel (registered trademark), is being attempted to be used. This exposure method is characterized by using a pattern in which a complex logic circuit is reduced to a simple L&S pattern like that of a memory circuit. By doing so, since the pattern of a complex logic device is limited to only the L&S pattern that is most easily exposed and the process of cutting its lines, complex pattern division calculations are not required, the process is simple, and it is said that it can go up to about 8 nm in terms of calculation.

[0009] When complex exposure methods such as double patterning and triple patterning are used as described above, the required photomasks increase each time split exposure is performed, and more mask inspections and precise measurements are required than before. For example, in double patterning, since two photomasks are sequentially overlapped and used, the absolute position accuracy and alignment accuracy of the lines formed on the mask between the two masks are more important than ever. Optical correction for exposing finer patterns also becomes more complex and precise. Masks using inverse lithography use even higher-order diffracted light, so the pattern size used for correction is even smaller. In mask production, defects of sizes that were previously negligible affect device yield, so it is necessary to reduce the defect density more than before, and even fine particles become the object of observation.

[0010] These trends are not limited to conventional 193nm exposure. In the future, when EUV is put into practical use, similar multi-patterning technologies will be used at even smaller sizes in the future, so the required inspection volume tends to increase exponentially.

[0011] Due to the above requirements, photomasks are always required to have even higher-resolution and faster defect inspection. This continues. A powerful method to meet these requirements is an ultra-high-speed electron beam inspection device.

Disclosure of the Invention

Problems to be Solved by the Invention

[0012] Conventional electron beam inspection devices are sufficient for resolution, but there is a problem that they are far slower than the required inspection speed.

[0013] The main reason is that the electron detection devices available in conventional electron beam inspection devices have problems with saturation characteristics, response speed, and lifespan where the dynamic range is narrow and linearity deteriorates. When the speed is increased, the image SNR deteriorates, so there is a problem that resolution and inspection speed cannot be achieved simultaneously.

[0014] To overcome this, the inventors have previously proposed an ultra-high-speed electron detection device that directly irradiates electrons onto an APD to detect the electrons (see Japanese Patent Application Laid-Open No. 2015-210998).

[0015] However, when attempting to detect a large amount of current with high sensitivity using a single APD, although the performance is much better than before, due to the dead time of the APD, counting misses occur, the input-output relationship does not become constant, and new problems arise such as deterioration of the SNR and linearity of the acquired image.

Means for Solving the Problems

[0016] The present invention aims to provide an ultra-high-speed electron detection device that enables high-speed inspection by an electron beam and a scanning electron beam inspection device incorporating the detection device.

[0017] In the ultra-high-speed electron detector that ultra-high-speed detects an electron beam emitted from a sample or reflected by a sample, secondary electrons emitted from the sample or electrons reflected by the sample are once applied to an ultra-high-speed scintillator plate having a response time on the order of ps to ns, which has a size necessary for signal electron detection, and converted into light. Then, the light generated by the scintillator is directly or reduced by a light guide or an optical lens and irradiated onto a small-area light detection element.

[0018] Therefore, the present invention provides, in an ultra-high-speed electron detector that ultra-high-speed detects an electron beam emitted from a sample or reflected by a sample or both electron beams, an energy normalization means for making an electron beam emitted from a sample and accelerated or an electron beam reflected by a sample or both electron beams into an electron beam of a constant energy, a first detector for converting the electron beam whose energy has been normalized by the energy normalization means into light, and a second detector for inputting, amplifying, and outputting an electrical signal the light converted by the first detector.

[0019] At this time, the energy stabilization means applies a predetermined voltage between the sample or a mesh provided near the surface of the sample and the first detector or a mesh provided near the surface of the first detector, so as to stabilize the energy of the electron beam.

[0020] In addition, the first detector is a scintillator that rapidly converts an electron beam into light.

[0021] In addition, the second detector is an MPPC element in which a plurality of avalanche diodes are arranged in parallel.

[0022] In addition, the second detector is a ring-shaped disk having a hole at the center through which the primary electron beam passes, and includes a plurality of independent detectors obtained by dividing the ring-shaped disk into a plurality of parts in the circumferential direction.

[0023] In addition, the voltage applied to the energy stabilization means is adjusted so that only the electrons emitted from the sample or the electrons reflected by the sample are detected and amplified by the first detector and the second detector.

[0024] In addition, it is configured as a scanning electron beam inspection device incorporating an ultra-fast electron detector.

Advantages of the Invention

[0025] In the present invention, since the second detector is composed of an MPPC element formed by arranging a large number of APDs in parallel electrically, the response speed is fast, and the dead time of the APD can be made negligibly small. Therefore, it is possible to avoid the counting omission phenomenon that occurred when irradiating a large number of electrons, which was a conventional problem.

[0026] In addition, since the electrons emitted or reflected from the sample are once converted into light, they do not interact with the electron beam, and thus can be freely transported to another place without loss. As a result, even if there are a plurality of scintillator light sources, etc., they can be efficiently detected independently.

[0027] In addition, an electronic detection device with a large input / output dynamic range can be obtained. Since a large number of APDs that operate independently when a large number of electrons or photons are incident each detect electrons or photons, electrons or photons can be detected without substantially generating a dead time. Electrons can be detected linearly from a small number of electrons to a large number of electrons.

[0028] In addition, since the light emitted by the scintillator can be reduced and irradiated onto an MPPC having a smaller area for detection, the dark noise due to the thermal noise proportional to the MPPC area can be reduced.

[0029] In addition, by accelerating while compressing or reducing the electron group that is generated on the sample surface and drifts while spreading, and irradiating a small scintillator, photons can be detected by a small scintillator and a small photon detection device, thereby reducing the thermal noise generated by the photon detection device. Since the scintillator can be made small, the cost can be reduced, and it becomes easy to arrange a plurality of them in the column.

[0030] In addition, a scintillator having a size or dimension necessary for detecting electrons can be used. Alternatively, since the signal electrons spread in all directions can be collected by an electron lens and irradiated onto the scintillator, the signal electrons generated on the sample surface can be detected almost completely. In addition, since the electric field used for secondary electron acceleration can be localized by using a conductive mesh and a shield tube with a large aperture ratio provided on the scintillator, it is possible not to adversely affect the beam orbit and aberration of the primary electrons.

[0031] In addition, since the life of the scintillator is very long, the life of the entire electron detection device is also very long, and it can be almost maintenance-free.

[0032] In addition, in a scanning electron beam inspection apparatus incorporating the ultra-high-speed electron detector of the present invention, an electron beam emitted or reflected from a sample (e.g., a mask or a semiconductor wafer) is detected at ultra-high speed, and dimensional measurement, defect detection, etc. can be performed in a short time, significantly improving throughput. Since there is almost no maintenance, the operating rate of the apparatus can be increased.

Example 1

[0033] FIG. 1 shows a configuration diagram of one embodiment of the present invention. In FIG. 1, while irradiating the sample 13 with the primary electron beam 31 and performing planar scanning, the generated secondary electrons and the reflected backscattered electrons are ultra-high-speed converted into light by the scintillator 9 which is the first detector, and the converted light is amplified by the MPPC 4 which is the second detector to output a signal. With this configuration, it becomes possible to detect signal electrons (secondary electrons, backscattered electrons, etc.) at ultra-high speed.

[0034] In FIG. 1, the electron gun 1 is a known one such as a thermionic cathode type, TFE, field emitter, photoexcitation type, etc. that generates the primary electron beam 31.

[0035] The electron gun control device 2 is a known one that supplies a high voltage, a bias voltage, a filament heating power supply in the case of a thermionic cathode type electron gun, etc. so that the electron gun 1 generates the primary electron beam 31.

[0036] The deflection electrode 3 is composed of a two-stage deflection electrode, the deflection electrode (upper) 3-1 and the deflection electrode (lower) 3-2. A predetermined deflection voltage is applied in pairs in the X direction and the Y direction to two-stage deflect the primary electron beam 31 and scan it in the X and Y directions above the sample 13. It is a known one.

[0037] The electron beam scanning control device 322 applies a predetermined scanning voltage to the deflection electrode 3 to narrow down the primary electron beam 31 onto the sample 13 and scan it in the X and Y directions.

[0038] The shielding tube 41 is for applying a negative voltage to prevent the secondary electrons emitted from the sample 13 from traveling upward along the axis and direct them toward the scintillator 9.

[0039] The blocking bias 5 is a negative bias voltage for blocking applied to the shielding tube 42 to prevent the secondary electrons emitted from the sample 13 from passing upward along the axis.

[0040] The bias circuit 61 applies a bias voltage to the MPPC 4 (to be described later with reference to FIG. 6 etc.).

[0041] The MPPC 4 is an abbreviation for Multi-Pixel Photon Counter, and is a photon counting device in which a Geiger-mode APD (avalanche diode) is multi-pixelized (to be described in detail with reference to FIG. 6 etc.).

[0042] The amplifier 6 amplifies the signal amplified by the MPPC 4.

[0043] The PC 7 is a personal computer that performs various processes and controls according to a program.

[0044] The display device 8 is a display for displaying images etc.

[0045] The scintillator 9 converts and amplifies electrons (such as secondary electrons and reflected electrons) into light at ultra-high speed (to be described later).

[0046] The mesh 10 applies an accelerating voltage for attracting secondary electrons etc. emitted from the sample 13.

[0047] The umbrella 11 forms an electric field such that secondary electrons etc. emitted from the sample 13 are accelerated by the voltage applied to the mesh 10 and collide efficiently with the scintillator 9. The umbrella 11 is made of a non-magnetic and conductive metal etc.

[0048] The objective lens 12 narrows down the primary electron beam 31 and irradiates the sample 13 with it.

[0049] The sample 13 is a sample (such as a photomask, a wafer, etc.) to be observed, inspected, and measured for length, and a bias voltage is applied to it by a sample bias circuit 14 as necessary.

[0050] The sample bias circuit 14 applies a bias voltage to the sample 13.

[0051] Next, according to FIGS. 2 to 5, a configuration for detecting and amplifying electrons (such as secondary electrons and reflected electrons) emitted and reflected from the sample 13 with high sensitivity and at ultra-high speed using the scintillator 9 and the MPPC 4 in FIG. 1 will be described in detail.

[0052] FIG. 2 shows an explanatory diagram of the main part of the present invention (Part 1). FIG. 2(a) shows an example in which the mesh 10 is arranged on the entire front surface (reflected electron detection part 331 and + secondary electron detection part 321) of the detector (first detector + second detector), and FIG. 2(b) shows an example in which the mesh 10 is arranged on the secondary electron detection part 321 on the front surface of the detector (first detector + second detector). The other configurations are the same.

[0053] In FIG. 2(a), the detector (first detector + second detector) 34 is a detector having a structure in which the scintillator 9, which is the first detector in FIG. 1, and the MPPC 4, which is the second detector, are stacked (see FIGS. 6 to 10 and their descriptions). It has a reflected electron detection part 331 on the inside and a secondary electron detection part 321 on the outside. Since high-energy reflected electrons gather at the center and low-energy secondary electrons gather at the peripheral part, the reflected electron detection part 331 is provided at the central part, and the secondary electron detection part 321 is provided outside it, so that the reflected electrons and the secondary electrons can be separated and detected respectively.

[0054] Mesh 10 applies a positive voltage to accelerate the low-energy secondary electrons emitted from sample 13. Here, an example is shown where it is arranged in front of the entire surfaces of both the reflected electron detection unit 331 and the secondary electron detection unit 321 of detector 34. In this case, the high-energy reflected electrons reflected from sample 13 form a Gaussian distribution in the upward direction of the axis, and many of them are detected by the reflected electron detection unit 331 shown closer to the axis. On the other hand, the low-energy secondary electrons emitted from sample 13 are accelerated in the direction of mesh 10 to which a high voltage is applied, collide with the secondary electron detection unit 321 shown in the figure, and also slightly with the inner reflected electron detection unit 331, and are amplified and detected. However, overall, the secondary electron component in the secondary electron detection unit 321 is larger, and as a result, it can be regarded as detecting secondary electrons.

[0055] Also, in Fig. 2(b), since mesh 10 is arranged only in the front part of the secondary electron detection unit 321 and not in the front part of the reflected electron detection unit 331, all the secondary electrons can be amplified and detected by accelerating and attracting all the secondary electrons that are accelerated and attracted to the secondary electron detection unit 321.

[0056] As described above, by arranging mesh 10 in front of detector (first detector + second detector) 34 and applying a positive voltage, low-energy secondary electrons and the like emitted from sample 13 are accelerated and attracted and amplified and detected by the outer secondary electron detection unit 321. On the other hand, high-energy reflected electrons and the like reflected by sample 13 can be detected and amplified by the inner reflected electron detection unit 331.

[0057] At this time, when the primary electron beam 31 is narrowed by the objective lens 12 and irradiated onto the sample 13, the energy of the secondary electrons emitted is 0.1 eV to 1. several eV. Since this is accelerated by 10 kV applied to the mesh 10, the rate of variation of the secondary electrons when they collide with the scintillator 9, which is the first detector constituting the detector 34, is (0.1 eV to 1. several eV) / 10,000, which is approximately ±0.01%. That is, the secondary electrons emitted from the sample 13 are accelerated by 10 kV applied to the mesh 10, and the variation in the energy of the secondary electrons when they collide with the scintillator 9 is adjusted to be a constant energy of approximately ±0.01% or less. As a result, the energy of the secondary electrons emitted from the sample 13 is made constant (here, constant at 10 kV ±0.01%) and input to the scintillator 9, and it becomes possible to convert it into the amount (number) of light that exactly corresponds to the number of secondary electrons. Then, after converting to light, it is possible to amplify with, for example, the MPPC 4, which is the second detector, and detect a signal corresponding to the number of the secondary electrons.

[0058] Also, since the reflected electrons reflected by the sample 13 have almost the same energy as, for example, 10 kV of the primary electrons, when input to the scintillator 9, which is the first detector, the variation is approximately 0.01% or less, similar to the case of the secondary electrons described above, and it becomes possible to output a signal corresponding to the number of reflected electrons from the sample 13 from the detector 34. This will be described in detail sequentially below.

[0059] Figure 3 shows an explanatory diagram of the main part of the present invention (part 2). In Figure 3, an energy filter 101 is arranged between the sample 13 and the detector (first detector + second detector) 34 to separate the high-energy reflected electrons 33 and the low-energy secondary electrons 32, and the reflected electron detection unit 331 and the secondary electron detection unit 321 are separated and amplified and detected respectively.

[0060] In FIG. 3, EXB101 is an energy filter that applies both an electric field and a magnetic field to separate electrons (backscattered electrons, secondary electrons) by varying the deflection angle as shown according to the difference in their energies, and independently detect each of them. By providing EXB101, electrons with low energy (such as secondary electrons) can be amplified and detected by the outer secondary electron detector 321, and electrons with high energy (such as backscattered electrons) can be amplified and detected by the inner backscattered electron detector 331.

[0061] As described above, by arranging the energy filter 101 between the sample 13 and the detector (the first detector + the second detector) 34, electrons with low energy (such as secondary electrons) can be amplified and detected by the outer secondary electron detector 321, and electrons with high energy (such as backscattered electrons) can be amplified and detected by the inner backscattered electron detector 331.

[0062] FIG. 4 shows an explanatory diagram of the main part of the present invention (part 3). In FIG. 4, the secondary electron detector 321 in FIG. 2(b) is lowered downward to get as close as possible to the sample 13, and it is devised so that secondary electrons emitted from the sample 13 can be efficiently collected. That is, the aperture is enlarged as seen from the sample 13, and a mesh 10 is provided on the front surface to apply a high voltage (for example, 10 KV), accelerating and attracting the secondary electrons emitted from the sample 13, and enabling efficient amplification and detection.

[0063] As described above, by bringing the secondary electron detector 321 closer to the sample 13 and enlarging the aperture, applying a high voltage (for example, 10 KV) to the mesh 10 arranged on the front surface of the secondary electron detector 321 to efficiently accelerate and collect the secondary electrons emitted from the sample 13, and enabling amplification and detection.

[0064] FIG. 5 shows an explanatory diagram of the main part of the present invention (part 4). In FIG. 5, the potential of the mesh 10 is controlled to detect the difference.

[0065] Fig. 5(a) shows an example without bias (normal state with +10 kV applied to mesh 10). In the case of Fig. 5(a), as shown in the figure, the detector 34 can detect both reflected electrons and secondary electrons.

[0066] Fig. 5(b) shows a state with bias. In the case of Fig. 5(b), for example, a voltage (negative voltage of several V to several tens of V) is applied to mesh 10 such that secondary electrons are reflected and cannot be detected. Therefore, secondary electrons cannot be detected, and highly energetic reflected electrons can be amplified and detected.

[0067] As described above, if an arbitrary voltage is applied to mesh 10, only electrons with energy equal to or higher than the applied voltage can collide with detector 34 and be amplified and detected. By taking the difference between the two, the number or quantity of either one of the electrons can be accurately detected.

[0068] Fig. 6 shows an example of the detector of the present invention. Fig. 6 shows a schematic configuration of the MPPC4, which is the second detector constituting detector 34. Fig. 6(a) shows a photo of the MPPC (perspective view), Fig. 6(b) shows a schematic diagram, and Fig. 6(c) shows an example of an equivalent circuit of the MPPC.

[0069] In Fig. 6(a), the primary electron passage hole is a hole through which the primary electron beam passes. The primary electron beam passing through the hole is narrowed by the objective lens 12 and scanned in a plane while irradiating the sample 13, emitting secondary electrons and reflecting reflected electrons.

[0070] Fig. 6(b) is a schematic view (the scintillator 9 is omitted) of the MPPC4 and the mesh 10 disposed in front of it as seen from below. There is a primary electron passage hole at the center, and a mesh 10 is disposed in front of the MPPC4. A positive high voltage (e.g., 10 kV) is applied to the mesh 10. After accelerating the secondary electrons emitted from the sample 13 to a certain energy, they are made to collide with a scintillator 9 (not shown) disposed between the MPPC4 and the mesh 13 to be converted into light. This light is incident on the illustrated MPPC4 for amplification and a signal is output.

[0071] Figure 6(c) shows an equivalent circuit example of the MPPC. The MPPC 4 is configured by connecting in parallel a plurality of D1 (avalanche diodes) arranged in parallel as shown in the figure, and further connecting R1 (quenching resistors) in series to each of them in parallel. When light is input to one D1 (avalanche diode), the D1 in the Geiger mode discharges and a current flows. When the current flows, the current is limited by R1 (quenching resistor) to be below the discharge voltage, and the discharge stops, and one pulse is output. When two photons are input simultaneously, they are in parallel, and a pulse having a peak approximately twice as large is output. Similarly, when n photons are input simultaneously, an n-fold pulse is output.

[0072] The bias power supply 43 is a power supply that applies a bias voltage for applying to D1 to maintain it in the Geiger mode state.

[0073] The current detection device 44 detects the pulse current generated by D1 and converts it into a voltage signal.

[0074] By using the MPPC 4 having the above configuration, the secondary electrons and reflected electrons reflected from the sample 13 are converted into light by the scintillator 9 with a constant energy, and then amplified and detected by the MPPC 4 at ultra-high speed and with high precision. Next, the characteristics of the MPPC in Figure 6 will be described.

[0075] As a whole, the device with a size of several millimeters square is the MPPC 4, which is formed by arranging a large number of small APDs (D1) with a size of several microns square in parallel, from several tens to tens of thousands or more. This MPPC 4 can be purchased from Hamamatsu Photonics (registered trademark), etc. Smaller devices may also be fabricated and used. Compared with the conventional PMT (photomultiplier tube), it has a volume of less than one ten-thousandth. It is a plate-shaped solid element made of silicon, and since it does not have glass or vacuum-sealed parts like a PMT, it is very robust. The weight of a single device is on the order of grams and is very light.

[0076] Each APD (D1) that makes up MPPC4 is electrically connected in parallel via a resistor for avalanche amplification control, called quenching resistor R1, as shown in Fig. 6(c). One end of the parallel circuit is connected to a bias power supply 43 for controlling the avalanche phenomenon of about 50V to 100V, and the other end is connected to a current detection device 44 of a virtual ground input. Since there are slight variations in characteristics among the APDs (D1) that make up the array of MPPC4, the bias voltage applied to the entire array is experimentally determined so that all APD (D1) elements forming the array are in the Geiger mode. When one photon is input to each APD in the Geiger mode, avalanche amplification occurs where the amplification factor reaches even one million times. Of course, it may be set to a voltage with other amplification factors.

[0077] A quenching resistor R1 with a fixed value for current limiting is connected to each APD, and a constant voltage is applied to the terminals. Therefore, when avalanche amplification occurs, the electrical resistance of the APD becomes negligibly small and acts just like a switch, so a constant current flows through the quenching resistor R1. Since the APDs are connected in parallel, the sum of the avalanche currents generated in each APD is output to the current detection device 44 connected to the output terminal.

[0078] Since a large number of APDs are spatially spread out in MPPC4, the probability that an incident photon enters the same APD is extremely small. Therefore, for example, if one photon is input to MPPD4, one unit of current is generated, and if N photons are input simultaneously, N units of current are generated. From this property, it is possible to know from the output current how many photons or electrons are incident on MPPC4 simultaneously. Since the output signal is an analog signal proportional to the number of incident photons, it is converted into a digital signal using an AD conversion device and taken into a PC for use in imaging. Since the current output by each APD element depends on the value of the quenching resistor R1 and is not strictly constant, normalization processing may be performed on a computer so that it becomes strictly constant.

[0079] Also, as a signal addition method, after once converting whether each APD is in an avalanche state into a digital signal of 0 or 1 inside the silicon chip, a digital sum-of-products operation is performed to calculate the number of AAPDs in the avalanche state, and a digital MPPC of a method that uses this as the total output of the entire MPPC4 can also be used. In this case, since the output is already converted into a digital signal in advance, there is no need to convert it into a digital signal with an AD converter. Also, the signal addition result is very accurate. In such an integrated IC, image processing can also be performed inside the chip.

[0080] FIG. 7 shows a detector example (part 2) of the present invention. FIG. 7 shows a schematic configuration when the scintillator 9, which is the first detector and the second detector constituting the detector 34, and the MPPC4 are incorporated into the apparatus of FIG. 1.

[0081] FIG. 7(a-1) shows a top view of the detector 34, and FIG. 7(a-2) shows a cross-sectional view, and these show a configuration example when there is no mesh 10.

[0082] In FIG. 7(a-1), the MPPC4 has a hole for a primary electron beam opened at the center, and the primary electron beam passes through this hole portion from top to bottom, is narrowed by the objective lens 12 of FIG. 1, and irradiates above the sample 13 while performing planar scanning.

[0083] FIG. 7(a-2) shows a cross-sectional view. Secondary electrons emitted from the sample 13 are accelerated and attracted by a positive voltage (for example, 5 KV here) applied to the scintillator 9, collide with the scintillator 9, and convert the secondary electrons into light. The converted light enters the MPPC4, is amplified, and outputs a signal. At this time, a negative voltage is applied to the shield tube 41 to suppress the secondary electrons from traveling upward on the axis, make them travel in the direction of the positive voltage applied to the scintillator 9, and improve the collection efficiency of the secondary electrons.

[0084] In addition, in FIGS. 7(b-1) and (b-2), the mesh 10 is provided between the sample 13 shown in the figure and the scintillator 9, and a positive acceleration voltage (5 KV in the figure) is applied between the mesh 10 and the scintillator 9 (in FIGS. 7(a-1) and (a-2), a positive voltage is applied between the sample 13 and the scintillator 9), and a positive voltage (for example, several V to several tens of V) is also applied between the sample 13 and the mesh 10 to efficiently collect the secondary electrons emitted from the sample 13. Others are the same as in FIGS. 7(a-1) and (a-2), so the description is omitted.

[0085] Here, the configuration of FIG. 7 will be described in detail.

[0086] In FIG. 7, the primary electron beam 31 generated by the electron gun 1 in FIG. 1 and accelerated is reduced to a desired beam spot size by the objective lens 12 arranged directly above the sample 13 and then scanned on the surface of the sample 13. Signal electrons such as secondary electrons and reflected electrons generated on the surface of the sample 13 are detected by a detector composed of a scintillator 9 and an MPPC 4. Since the signal electrons rise while spreading in a conical shape in the vertical direction from the point on the surface of the sample 13 irradiated with the primary electron beam 31, the signal electrons are likely to be distributed axially symmetrically with respect to the primary electron beam axis. In order to efficiently detect the signal electrons, it is necessary to provide a hole (hole for primary electron beam) for the primary electrons to pass through at the center of the detector and arrange it in the vicinity of the primary electron beam axis.

[0087] Therefore, as shown in (a-1) and (b-1) of FIG. 7, a scintillator 9 is provided directly on the light-receiving surface of the MPPC 4, and the surface of the scintillator 9 is coated with a thin aluminum conductive film. The main purpose of the aluminum thin film is to prevent incident electrons from charging on the surface of the scintillator 9, but it also functions as an antireflection film so that the light generated by the scintillator 9 does not leak outside the MPPC 4. The scintillator 9 on the MPPC 4 may be attached with an adhesive considering the refractive index, or the scintillator 9 may be directly deposited by evaporation, CVD, sputtering, or the like. As the scintillator 9, an inorganic scintillator, a direct-transition semiconductor scintillator, a ceramic material, a plastic scintillator containing heavy atoms, a halide scintillator using the luminescence of inner-shell transitions, a liquid, or the like can be used.

[0088] Also, there is a hole (hole for primary electron beam) through which the primary electron beam passes, and a shield tube 41 is provided so that the electric and magnetic fields in the peripheral part do not affect the primary electron beam. The primary electron beam passes through the shield tube 41. The shield tube 41 is made of a non-magnetic material, has a diameter of several millimeters, a width, and a thickness of 1 mm or less. There is inevitably a gap between the detector and the shield tube 41. Since electrons hitting the gap are not detected, a conductive umbrella-shaped member is used to extend the electric field outward and repel the electrons so that the electrons do not enter the gap and go toward the detector.

[0089] Also, when performing electron acceleration to cause the scintillator 9 to emit light in front of the detector, as shown in FIGS. 7(b-1) and (b-2), a conductive mesh 10 with a high aperture ratio (preferably 90% or more) is arranged at a position several millimeters away from the scintillator 9, and a voltage of about 1 to 10 kV is applied between the mesh 10 and the conductive thin film provided on the scintillator surface. The mesh 10 It is also desirable to apply a slightly positive potential of about 10 V to the sample 13 so that the signal electrons generated on the surface of the sample 13 can reach the mesh 10 itself. It is desirable to keep this space in a vacuum state higher than 10^-3 Pascal so as not to cause discharge.

[0090] FIG. 8 shows a detector example (No. 3) of the present invention. The schematic configuration in the case of making the primary electron beam ring-shaped and scanning the sample 13 in a plane while focusing it into a thin point with the objective lens 12 is shown. Since the others are the same as those in FIG. 7, the description thereof is omitted.

[0091] (a-1) of FIG. 8 shows a top view of the detector 34, and (a-2) of FIG. 8 shows a cross-sectional view, which shows a configuration example in the case where there is no mesh 10.

[0092] In (a-1) of FIG. 8, the MPPC 4 is formed with a ring-shaped hole for the primary electron beam in a ring shape, and the ring-shaped primary electrons pass through this hole portion from top to bottom, are focused into a point by the objective lens 12, and scan the sample 13 while irradiating it.

[0093] (a-2) of FIG. 8 shows a cross-sectional view. The secondary electrons emitted from the sample 13 are accelerated and attracted by the positive voltage (for example, 5 KV) applied to the scintillator 9 and collide with the scintillator 9 to convert the secondary electrons into light. The converted light is incident on the MPPC 4 and amplified to output a signal. At this time, a negative voltage is applied to the ring-shaped shield tube 41 to prevent the secondary electrons from traveling upward on the axis, and the secondary electrons travel in the direction of the positive voltage applied to the scintillator 9, improving the collection efficiency of the secondary electrons. At this time, it becomes possible to detect secondary electrons and reflected electrons at the central portion of the detector composed of the scintillator 9 and the MPPC 4, and it becomes possible to increase the detection efficiency.

[0094] Also, in FIGS. 8(b-1) and (b-2), a mesh 10 is provided between the sample 13 shown in the figure and the scintillator 9, and a positive acceleration voltage (5 KV in the figure) is applied between the mesh 10 and the scintillator 9 (in FIGS. 8(a-1) and (a-2), a positive voltage is applied between the sample 13 and the scintillator 9), and a positive voltage (e.g., several V to several tens of V) is also applied between the sample 13 and the mesh 10 to efficiently collect the secondary electrons emitted from the sample 13. The rest is the same as in FIGS. 8(a-1) and (a-2), so the description is omitted. At this time, it becomes possible to detect secondary electrons and reflected electrons at the central part of the detector composed of the scintillator 9 and the MPPC 4, and it becomes possible to improve the detection efficiency.

[0095] Here, FIG. 8 is characterized by using a hollow electron beam (hollow beam) as the primary electron beam. The hollow electron beam is a ring-shaped beam and may be used to prevent repulsion between electrons when using a high-current electron beam. It can be focused on a single point on the surface of the sample 13 by the objective lens 12 in the same way as a normal beam. As shown in FIGS. 8(a-1) and (b-1), a scintillator 9 is also arranged at the center, and a ring-shaped shield tube 41 through which primary electrons pass is provided at its peripheral part. If necessary, a scintillator 9 can also be arranged outside the ring-shaped shield tube 41. By doing so, even when the electrons (such as secondary electrons and reflected electrons) generated on the surface of the sample 13 rise completely vertically, they can be accurately detected by the scintillator 9 existing directly above. The signal electrons flying to the peripheral part of the primary electron beam axis are detected by the scintillator 9 arranged outside the ring-shaped shield tube 41.

[0096] FIG. 9 shows a detector example (No. 4) of the present invention. FIG. 9 shows a schematic configuration when the scintillator 9 and the MPPC 4, which are the first detector and the second detector constituting the detector 34, are divided into, for example, four parts in the circumferential direction and incorporated into the apparatus of FIG. 1.

[0097] (a-1) of FIG. 9 shows a top view of the detector 34, and (a-2) of FIG. 9 shows a cross-sectional view, which shows a configuration example when the mesh 10 is not present.

[0098] In (a-1) of FIG. 9, the MPPC 4 has a hole for the primary electron beam opened at the center, and the surrounding portion is divided into four in the circumferential direction. Each portion is blocked by the partition 431 so that electrons and light do not enter the adjacent portions. The primary electron beam passes through the hole portion at the center from top to bottom, is narrowed by the objective lens 12 in FIG. 1, and scans the plane while irradiating above the sample 13.

[0099] (a-2) of FIG. 7 shows a cross-sectional view. The secondary electrons emitted from the sample 13 are accelerated and attracted by the positive voltage (e.g., 5 KV) applied to the four-divided scintillator 9 and collide with the scintillator 9, converting the secondary electrons into light. The converted light enters the four-divided MPPC 4 and is amplified to output a signal. At this time, a negative voltage is applied to the shield tube 41 to suppress the secondary electrons from traveling upward on the axis, and the secondary electrons travel in the direction of the positive voltage applied to the four-divided scintillator 9, improving the collection efficiency of the secondary electrons.

[0100] Also, in (b-1) and (b-2) of FIG. 9, the mesh 10 is provided between the illustrated sample 13 and the scintillator 9, a positive acceleration voltage (5 KV in the illustration) is applied between the sample 13 and the four-divided scintillator 9, and a positive voltage (e.g., 50 V) is also applied between the four-divided scintillator 9 and the mesh 10 to efficiently collect the secondary electrons emitted from the sample 13. Since the others are the same as (a-1) and (a-2) of FIG. 7, the description is omitted.

[0101] Here, the configuration of FIG. 9 will be described in detail. FIG. 9 shows the case where a 4CH (4-segment) detector is used. The 4CH detector is used to obtain information regarding the direction in which electrons ejected from the surface of sample 13 fly out. By adding or subtracting the signals of each detected CH, necessary components can be extracted to know the direction in which the electrons are flying out. 3D information can be obtained. Since the sensitivity output by each detector is not necessarily uniform, calibration is performed by multiplying by an appropriate coefficient so that the output sensitivity of each channel becomes the same before use. Utilizing this information makes it possible to emphasize the edge information of the surface structure of sample 13 and to obtain 3D information.

[0102] Also, as shown in FIG. 9, a shield tube 41 is provided at the location where the primary electron beam passes, and detectors that are electrically independent are arranged to be axially symmetric with respect to the axis of the primary electron beam. The heights are also made the same. The arrangement of the detectors may be not only planar but also three-dimensionally arranged in the height direction along the primary electron beam axis.

[0103] Also, partition-shaped electrodes are extended in all directions from the shield tube 41 through which the primary electron beam passes to surround the scintillator 9, and the detection range of the signal electrons flying to the detector can be separated. A bias voltage that generates a repulsive force against the signal electrons is applied to the shield tube 41 and the partition 431. Electrons coming from sample 13 are bent toward the scintillator 9 by this repulsive force and fly. The signal electrons are detected by their respective detectors, and an electrical signal is sent to the amplification device.

[0104] FIG. 10 shows a detector example (No. 5) of the present invention. FIG. 10 shows a schematic configuration when the first detector and the second detector, which are the scintillator 9 and the MPPC 4 that constitute the detector 34, are divided into, for example, 4 segments in the circumferential direction, and light is focused by the light guide 45 (FIGS. 10(a-1) and (a-2)) and the lens 46 (FIGS. 10(b-1) and (b-2)) to use a smaller MPPC 4 and this is incorporated into the apparatus of FIG. 1.

[0105] (a-1) of FIG. 10 shows a top view of the detector 34, and (a-2) of FIG. 10 shows a cross-sectional view, both of which illustrate a configuration example when the mesh 48 is absent.

[0106] In (a-1) of FIG. 10, the MPPC 4 has a hole for the primary electron beam opened at the center, and the surrounding portion is divided into four in the circumferential direction. Each portion is blocked by the baffle 47 so that electrons and light do not enter adjacent portions. The primary electron beam passes through the hole at the center from top to bottom, is narrowed by the objective lens 12 in FIG. 1, and irradiates the sample 13 while performing planar scanning.

[0107] (a-2) of FIG. 10 shows a cross-sectional view. The secondary electrons emitted from the sample 13 are accelerated and attracted by a positive voltage (e.g., 5 KV) applied to the four-divided scintillator 9 and collide with the scintillator 9, converting the secondary electrons into light. The converted light enters the four-divided MPPC 4 and is amplified to output a signal. At this time, a negative voltage is applied to the shield tube 41 to suppress the secondary electrons from traveling upward on the axis, and the secondary electrons travel in the direction of the positive voltage applied to the four-divided scintillator 9, improving the collection efficiency of the secondary electrons.

[0108] Also, in (b-1) and (b-2) of FIG. 10, a mesh 48 is provided between the illustrated sample 13 and the scintillator 9, a positive acceleration voltage (50 V in the illustration) is applied between the sample 13 and the four-divided scintillator 9, and a positive voltage (e.g., 5 KV) is also applied between the four-divided scintillator 9 and the mesh 10 to efficiently collect the secondary electrons emitted from the sample 13. Also, in (b-2) of FIG. 10, instead of the light guide 45 in (a-2) of FIG. 10, a lens 46 is used to guide the light generated by the scintillator 9 to the small MPPC 4. Since the rest is the same as in (a-1) and (a-2) of FIG. 10, the description is omitted.

[0109] Here, the configuration of FIG. 10 will be described in detail. FIG. 10 shows an example in which the light generated by the scintillator 9 by electron beam irradiation is reduced and irradiated onto the MPPC 4. Different from a single APD, the MPPC 4 has an excellent feature that even if the overall area of the MPPC 4 is increased, the capacitance of each APD does not increase, so the deterioration of the electron detection speed is not observed. However, when the area of the MPPC 4 becomes large, noise called dark noise, which is caused by thermal noise, increases in proportion to the area. Since this noise is proportional to the absolute temperature, it can be reduced by cooling. However, if a cooling device is provided inside the vacuum device, the device becomes extremely complex, the volume increases, the cost also increases, and the maintenance becomes very difficult, so the advantages of the MPPC 4 may disappear. On the other hand, when the area of the MPPC 4 is reduced, the electron detection efficiency decreases and the image SNR deteriorates.

[0110] In the present invention, paying attention to this point, we have succeeded in keeping the electron detection efficiency high while making the area of the MPPC 4 as small as possible. First, the area of the scintillator 9 that converts electrons generated on the surface of the sample 13 into light is made as large as necessary and sufficient. Specifically, by using experiments or electron trajectory simulations, the location where the signal electrons fly is determined, and the scintillator 9 having a size necessary and sufficient is arranged at that location.

[0111] Thereby, the detection efficiency of the signal electrons generated on the surface of the sample 13 is increased. On the other hand, the light generated by the scintillator 9 is irradiated onto the MPPC 4 after being converged or its cross-sectional area is reduced using optical elements (after being reduced by the light guide 45 in (a-2) of FIG. 10, the lens 46 in (b-2) of FIG. 10, etc.). By doing so, the high electron detection efficiency is ensured by the large scintillator 9, and the small dark noise is realized by using the MPPC 4 as small as possible.

[0112] Here, for the MPPC4 to be used, it is desirable that the dimensions of each APD constituting the APD array be as small as possible (less than several microns), the number of elements be tens of thousands or more, and the area of the MPPC4 be 1 square mm or less. In order to reduce the light generated by the scintillator 9, as shown in (a-2) of FIG. 10, there are methods such as using the light guide 45 and methods such as using the optical lens 46 shown in (b-2) of FIG. 10. Furthermore, by using a mirror, a Fresnel lens, a holographic lens, or a diffraction element, it is also possible to obtain the effect of shortening the dimension in the longitudinal direction. When using a light guide or the like in contact with the photon detection device, it is desirable to match the refractive index of the contacting member so that unnecessary reflection does not occur and signal loss does not occur, and to newly insert an antireflection film. Furthermore, in order to select the wavelength components incident on the MPPC4, an optical filter for wavelength selection may be inserted in between.

[0113] FIG. 11 shows a configuration diagram (Part 1) of another embodiment of the present invention. This FIG. 11 shows an example in which the electrons emitted from the sample 13 are accelerated to cause the scintillator 9 to emit light, the area of which is reduced by the light guide 91 and incident on the MPPC4 having an area smaller than that of the scintillator 9. Since other configurations are the same as those in FIG. 1, the description thereof is omitted.

[0114] In FIG. 11, the light guide 91 is made of a single material such as glass or acrylic or a bundle such as an optical fiber bundle.

[0115] The MPPC4 is a device in which a large number of APDs with sides of several microns are arranged in an array. When a bias voltage is applied to the APD, initially, the amplification factor is the same as that shown by an ordinary photodiode. However, when the threshold voltage of about 50V is exceeded, the amplification factor suddenly increases, and finally, it reaches an operating mode called the Geiger mode, in which when one electron (or photon) is incident, amplification of nearly one million times occurs. In the Geiger mode, the MPPC4 has the sensitivity to detect one electron incident from the outside, but since the same amplification occurs even when one electron is generated due to internal thermal fluctuations, it is impossible to distinguish between the electron (photon) that entered only once and the electron generated by heat. Therefore, these become the cause of noise. The generation of thermoelectrons is proportional to the area of the APD. That is, the larger the area of the APD, the higher the frequency of generation of thermoelectrons and the increase in noise.

[0116] On the other hand, since the signal electrons emitted from sample 13 are scattered and emitted over a wide range, in order to efficiently detect the signal electrons, it is advantageous for the detector to have as large an area as possible.

[0117] As described above, the two requirements are contradictory to each other. In the present invention, in order to solve this contradiction, the area of the scintillator 9 for detecting electrons is made as large as possible, and by reducing the light generated by the scintillator 9, it is input to the MPPC4 having as small an area as possible, thereby finding a method that achieves both high detection efficiency and low noise. In FIG. 11, an example is shown in which a light guide 91 with different area ratios on the incident side and the emission side is used, the larger area side is connected to the scintillator 9, and the smaller area side is connected to the MPPC4 which is a photon detection device.

[0118] The light generated by the scintillator 9 having a large area is reduced by the light guide 91 and guided to the MPPC4 which is a photon detection device having a small area and converted into an electric current. By doing so, it is possible to achieve both high-efficiency detection and low dark noise.

[0119] Figure 12 shows a configuration diagram (part 2) of another embodiment of the present invention. This Figure 12 shows an example in which electrons emitted from sample 13 are accelerated to cause scintillator 9 to emit light, and the light is reduced in area by lens 51 and incident on MPPC 4, which is smaller than the area of scintillator 9. Since other configurations are the same as those in Figure 1, the description thereof is omitted.

[0120] In Figure 12, lens 51 is a lens that reduces light. Although a single-lens configuration is shown in the figure, a plurality of lenses may be used. For example, a lens system that emits a reduced parallel beam when a parallel beam is incident may be used. Signal electrons generated by irradiating sample 13 with a primary electron beam are accelerated and attracted by a positive voltage (e.g., 10 KV) applied to mesh 10 or the like and collide with scintillator 9. The colliding electrons emit a large amount of light. The emitted light is collected by lens 51 so that the cross-sectional area becomes 1 / 10 or less, and the area is reduced, and then is incident on MPPC 4, which is a photon detection device. In response to the incident light, MPPC 4 amplifies it and outputs a detection current.

[0121] By adopting the above configuration, since the area of MPPC 4 is 1 / 10 or less of the area of scintillator 9, it is possible to reduce the dark noise to 1 / 10 or less as compared with the case where MPPC 4 having the same area as scintillator 9 is used.

[0122] Figure 13 shows a configuration diagram (part 3) of another embodiment of the present invention. Figure 13 is an example of a configuration in which an electron lens 53 is used to reduce signal electrons (secondary electrons, reflected electrons, etc.) and input them to a detector (scintillator 9 + MPPC 4). Since other configurations are the same as those in Figures 11 and 12, the description thereof is omitted.

[0123] In FIG. 13, the electron lens 53 is an electron lens that reduces signal electrons (such as secondary electrons emitted from the sample 13 and reflected backscattered electrons) from the sample 13 to an area smaller than the aperture of the electron lens 531, and is an electrostatic lens or a magnetic lens. Here, before the electrons enter the scintillator 9, signal electrons (such as secondary electrons) that are generated and floating on the surface of the sample 13 are first reduced in cross-sectional area and energy-accelerated using an electron lens 531 such as an electrostatic lens or a magnetic lens having a first aperture, and then irradiated onto the scintillator 9 or a portion thereof having an area smaller than the first aperture. The signal electrons are detected by the MPPC 4, which is a photon detection device having a size smaller than the first aperture.

[0124] With the above configuration, both the scintillator 9 and the MPPC 4 can be made small. Therefore, not only can the dark noise be reduced, but also the cost of the scintillator 9 can be reduced because a small size of the scintillator 9 is sufficient. In addition, since the size of the MPPC 4 can be reduced, there is more freedom in arranging it inside the lens barrel. Since there is no need to pass through lenses or light guides, the signal loss can be reduced, contributing to high-quality image formation. Since the detector can be made very small in this way, when a large number of electron beam devices are arranged in one electron beam apparatus and a large number of electron beams are simultaneously irradiated onto the surface of the sample 13, the signal electron groups generated can be measured simultaneously.

[0125] FIG. 14 shows another detector explanatory diagram (Part 1) of the present invention. FIG. 14(a) shows the overall configuration diagram, and FIG. 14(b) shows an example of an electron beam aperture.

[0126] In FIG. 14(a), the irradiation lens 51 makes the primary electron beam generated by the electron gun 1 parallel.

[0127] The electron beam aperture 52 is for extracting a plurality of predetermined thin portions from the parallel primary electron beam to form a plurality of primary electron beams. For example, as shown in Fig. 14(b), it has a plurality of circular apertures. Such an aperture may use a blanking type aperture that can be electrically opened and closed.

[0128] The support part 53 holds a detector composed of the MPPC 54 and the scintillator 53.

[0129] The first reduction lens 56 is the first lens that reduces the plurality of primary electron beams that have passed through the electron beam aperture 52.

[0130] The second reduction lens 57 is for further reducing the plurality of primary electron beams reduced by the first reduction lens 56, making them thinner and narrower, and performing planar scanning while irradiating the sample 13.

[0131] Next, the operation of the configuration in Fig. 14 will be described in detail.

[0132] (1) In Fig. 14, after the primary electron beam 31 generated by the electron gun 1 is once converted into a parallel beam by the illumination lens 51, it passes through the electron beam aperture 52 and is formed into a plurality of primary electron beams. The number of primary electron beams simultaneously irradiating the surface of the sample 13 increases in proportion to the number of primary electron beams. If the primary electron beam is passed through the center of the electron beam aperture 52, it becomes difficult to arrange the secondary electron detection device so that the orbits of the primary electrons and secondary electrons are separated. Therefore, it is desirable to provide a plurality of holes in the peripheral part.

[0133] (2) The formed plurality of primary electron beams pass through the support part 53 provided with holes so that the primary electron beams can pass through, pass through the first reduction lens 56 and the second reduction lens 57, and after the beam diameter is compressed, they are irradiated almost perpendicularly to the surface of the sample 13.

[0134] (3) The irradiated electrons generate secondary electrons on the surface of the sample 13. The generated secondary electrons are accelerated by a bias voltage (e.g., 10 KV) applied between the sample 13 and the conductive film provided on the surface of the scintillator 55, pass through the second reduction lens and the first reduction lens, and then enter the scintillator 55. The bias voltage is adjusted so that the height of the scintillator becomes the focal position of the secondary electrons. That is, it is adjusted so that an image of the primary electron spot is formed on the scintillator surface.

[0135] (4) The detector composed of the scintillator 55 and the MPPC 54 performs secondary electron trajectory calculation and is installed at the location where the secondary electrons return. A large number of detectors may be arranged in an array in advance, and addresses and the like may be provided for each element so that they can be electrically selected and used from a computer or the like as needed. With this function, it is possible to cope with the trajectory change of secondary electrons accompanying the change of the bias voltage.

[0136] As described above, the secondary electrons generated by irradiating the sample 13 with a plurality of divided primary electron beams are simultaneously detected by their respective detectors (scintillator 55, MPPC 54) in parallel, so that it is possible to acquire images at high speed.

[0137] FIG. 15 shows another detector explanatory diagram (Part 2) of the present invention. While FIG. 14 is composed of the support portion 53, the MPPC 54, and the scintillator 55, FIG. 15 is composed of the MPPC 54, the transparent support portion 531, and the scintillator 55. Since other configurations are the same as those in FIG. 14, the description thereof is omitted.

[0138] In FIG. 15, the transparent support portion 531 is characterized in that the scintillator 55 for detection is supported by a transparent member. As the transparent member used for the transparent support portion 532, various optical elements such as lenses and holograms are used. By using these optical elements, it becomes possible to reduce, enlarge, or move the position of the light generated by the scintillator 55. That is, the electron detector can be arranged anywhere.

[0139] FIG. 16 shows another detector explanatory diagram (part 3) of the present invention. FIG. 16 is characterized in that a mesh 58 is provided on the sample 13, and secondary electrons generated on the surface of the sample 13 are accelerated to a desired acceleration energy before entering the second reduction lens 57, thereby improving the detection efficiency of the secondary electrons.

[0140] In FIG. 16, the mesh 58 is a mesh provided on the sample 13, which accelerates secondary electrons emitted from the sample 13 (for example, 5 KV). This mesh 56 is provided on the sample 13, and an acceleration voltage (for example, 5 KV) is applied thereto to accelerate the secondary electrons. The accelerated secondary electrons pass through the second reduction lens 57 and the first reduction lens 56, are further accelerated, and finally collide with the scintillator 55. By accelerating immediately after the generation of secondary electrons, it is possible to avoid the generated secondary electrons from being scattered and dispersed, and the detection efficiency can be improved.

[0141] FIG. 17 shows another detector explanatory diagram (part 4) of the present invention. FIG. 17 further provides a mesh 59 on the scintillator 55 with respect to FIG. 16, so that the energy of secondary electrons passing through the second reduction lens 57 and the first reduction lens 56 is in a constant state, and they travel along the designed trajectory (travel along the designed trajectory without being affected by the mechanical shape of the second reduction lens 57 and the first reduction lens 56, etc.).

[0142] In FIG. 17, the mesh 59 is a mesh provided in front of the scintillator 55.

[0143] The mesh 58 is a mesh provided in front of the sample 13. Here, the meshes 59 and 58 are maintained at the same potential, so that the surrounding geometric structure does not affect the secondary electrons passing through the second reduction lens 57 and the first reduction lens 56, and they pass through the designed trajectory. Here, for example, 5 KV (the voltage for accelerating secondary electrons) is applied to the mesh 58 as shown in the figure, and 10 KV (the voltage for accelerating secondary electrons passing through the mesh 59) is applied to the mesh 59.

[0144] With the above configuration, when a plurality of primary electron beams are irradiated onto the sample 13 and scanned in parallel in a plane, secondary electrons and reflected electrons corresponding to the plurality of primary electron beams are respectively emitted and reflected from the sample 13. These emitted and reflected secondary electrons and reflected electrons are respectively accelerated upward in parallel by the mesh 58, pass through the second condenser lens 57 and the first reduction lens 56 at a constant energy, pass through the mesh 59, and are respectively accelerated in parallel by a positive voltage (for example, 10 KV) applied between the mesh 59 and the scintillator 55, and are respectively converted into light in parallel by the scintillator 55. These converted lights are respectively amplified and detected by the MPPC 54 in parallel, and signals can be output in parallel.

[0145] At this time, electrons (such as reflected electrons) accelerated to high energy take orbits inside, and electrons (such as secondary electrons) accelerated to low energy take orbits outside. Therefore, they are respectively detected and output in parallel by a plurality of divided detectors (scintillator 55, MPPC 54).

[0146] FIG. 18 shows an explanatory diagram of the detector (part 5) of the present invention. FIG. 18 is characterized in that a resolution higher than the size of the irradiated primary electron beam can be achieved and the throughput can be improved. This will be described in detail below.

[0147] (1) In FIG. 18, after primary electrons are emitted from the electron gun 1, they are accelerated to a desired energy to generate a hollow electron beam (see (b) in FIG. 18), which is converged to a point by the objective lens 121 and irradiated onto the sample 13. The acceleration voltage used is about 1 KV to 50 KV.

[0148] (2) The hollow electron beam (hologram beam) can be generated by uniformly irradiating a parallel electron beam onto an electron beam aperture 52 having concentric circular cutouts with an irradiation lens 51. The concentric circular holes may be formed by using a large number of small holes in a concentric circular pattern, or a square shape may be used, or they may be surrounded doubly. These apertures may also be in the form of blanking apertures that can electrically open and close a large array of apertures. A concentric circular aperture may be formed by selectively opening and closing a part of the array. Further, a grating lens or a thin film lens that forms a concave lens may be used to reduce aberration.

[0149] The electron beam aperture 52 is made of a non-magnetic material such as titanium, MO, or W, and it is desirable to heat it to 100 °C or higher to prevent contamination. It is desirable to have conductivity for antistatic purposes.

[0150] (3) The generated hollow electron beam is focused to a point by the objective lens 121. The focused primary electron beam is scanned in the XY plane by the deflection electrode 122. The deflection electrode 122 can be arranged at necessary locations such as above or below the objective lens 121 (in Fig. 18, the electrostatic electrode 122 is arranged below the objective lens 121).

[0151] (4) The focused primary electron beam irradiates the sample 13 to generate secondary electrons. The generated secondary electrons have very low energy and are emitted in all directions, so a point image cannot be formed by the objective lens 121 as it is. In this embodiment, a positive acceleration voltage is applied to the mesh 58 directly above the sample 13, and almost all of the generated secondary electrons are accelerated to have a desired energy. This acceleration voltage relatively reduces the energy variation of the secondary electron group, resulting in a secondary electron beam having a single uniform energy. Therefore, the higher this energy, the better the energy uniformity.

[0152] (5) The secondary electrons accelerated to the desired energy pass through the objective lens 121, and finally collide with the scintillator 55. After being decelerated by the mesh 59 so that the light emission becomes optimal, they collide with the scintillator and emit light. The emitted light is detected in each part divided into a plurality of MPPCs 4 directly or via a mirror (a high-sensitivity CCD or CMOS device may also be used). (6) As shown in FIG. 18, the light spot to be detected is an enlarged image of secondary electrons generated by the primary electron beam narrowed down to a desired size by the objective lens 121. Also, this point image is dynamically XY-scanned corresponding to the deflection of the primary electron beam. The enlarged image of the light spot has information on the secondary electron generation amount distribution at the irradiation point of the primary electron beam. When image analysis is performed using the scanning period of the primary electron beam to extract the detailed luminance distribution of the light spot from the image, changes in the sample surface in a region smaller than the spot size can be extracted. That is, the resolution of the primary electron beam optical system can be increased. This is just like drawing a picture by bundling the cores of 10 pencils. If the luminance distribution measurement is performed and separated into 10, the same resolution images written by each pencil can be obtained.

[0153] (7) For example, when secondary electrons generated from a spot size of a 10 nm primary electron beam are magnified 10,000 times using an electron optical system and an optical system, they become points of about 100 microns. Since the minimum pixel size of the current two-dimensional light detection device is about 1 micron, the luminance distribution of these points can be sufficiently separated and detected by the two-dimensional light detection device. Continuously extracting this point image by a computer, dividing the luminance of each part so as to correspond to each pixel, and performing image processing such as rotation correction and luminance correction at high speed to form an appropriate image, and rearranging the pixels in the correct order so that one image can be reconstructed, an image with a resolution smaller than the beam size can be obtained. Since these processes do not necessarily have to be performed in real time, the necessary processing may be performed by a computer for image processing after the image is acquired.

[0154] (8) For example, if a light point is divided into four parts and each part is made to correspond to a respective pixel, a resolution substantially equivalent to that obtained by irradiating with a beam size of 2.5 nm can be achieved, and throughput four times as high can be obtained. Normally, increasing the resolution causes a decrease in throughput, but the characteristic of this method is that both the resolution and the throughput can be increased.

[0155] (9) To obtain a higher resolution, it is very important that the scintillator composition is uniform and fine. When using particulate scintillators, it becomes possible to obtain a high resolution by making the particle size on the order of nanometers. In the case of plastic scintillators, etc., they are uniform at the nanolevel. Since a resolution on the order of nanometers can be achieved with a fluorescence microscope, if they can be separated to that extent, one point can be separated into a plurality of beam components.

[0156] (10) The above method has the same effect as collecting a plurality of beams in a multi-beam format in the extremely near vicinity. On the other hand, contrary to the commonly known multi-beam method, since only the central part of the lens axis is utilized, lens distortion is small, aberration correction is not required, and a multi-beam inspection apparatus can be easily realized. Since this effect can be obtained regardless of whether the beam spot size is increased or decreased, it becomes possible to realize an inspection apparatus with high resolution and ultra-high speed or low resolution and super-ultra-high speed. Conversely, if the beam size is made small from the beginning and the resolution of the light detection device is increased, a resolution of less than 1 nm, which has been impossible to achieve conventionally, and high throughput can be realized. In particular, when realizing high resolution, in the case of a hollow beam, since electron beam spreading due to electrostatic repulsion hardly occurs, it becomes easy to narrow down the electron beam.

[0157] Different from conventional multi-beam inspection apparatuses, since the detected secondary electrons are pulled up in the vertical direction and electrons are directly detected without using a beam separator, aberration is small and higher resolution and efficiency can be realized.

[0158] Furthermore, in the above method, since the beams are separated by image analysis, the direction in which the beams are arranged, which is a problem in a multi-beam inspection apparatus, does not pose a problem. Therefore, it can be easily applied to an inspection apparatus that performs high-speed inspection by continuously moving the XY stage at a constant speed.

Brief Description of the Drawings

[0159]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Explanation of Symbols

[0160] 1: Electron gun 2: Electron gun control device 3: Deflection electrode 3-1: Deflection electrode (upper) 3-2: Deflection electrode (lower) 31: Primary electron beam 32: Secondary electrons 321: Secondary electron detector 322: Electron beam scanning control device 33: Backscattered electrons 331: Backscattered electron detector 34: Detector (first detector + second detector) 4, 54: MPPC 41: Shield tube 411: Quenching resistor 42: Avalanche photodiode 421: Support glass 43: Bias power supply 431, 47: Diaphragm 44: Current detection device 441: Support part 45, 91: Light guide 46: Lens 5: Bias for passage prevention 51: Irradiation lens 52: Electron beam aperture 521: Hollow beam 522: Reflecting mirror 523: Optical magnification lens 524: Optical two-dimensional detector 53: Support part 531: Electron lens 532: Transparent support part 56: First reduction lens 57: Second reduction lens 6: Amplifier 61: Bias circuit 7: PC 8: Display device 9, 55: Scintillator 10, 48, 93, 58, 59: Mesh 101: Energy filter (EXB) 11: Umbrella 12, 121: Objective lens 122: Deflection electrode 13: Sample 131: Lens control circuit 14: Sample bias circuit

Claims

1. In an electron detection device for detecting electrons emitted from a sample, an electron gun that emits an electron beam, an objective lens that reduces the emitted electron beam and irradiates the sample, an electron beam scanning device that scans the electron beam irradiated on the sample in a plane, means for accelerating the energy of electrons emitted from the sample by the electron beam irradiated by the electron beam scanning device, a shield tube to which a negative voltage is applied through which the electron beam passes, and a first element having a light receiving surface with a first area, which is disposed around the shield tube and converts the accelerated electrons into light, a second element that condenses the light converted from electrons by the first element so as to be smaller than the first area, and a third element having an area smaller than the first area that receives, detects, and converts the light condensed by the second element into an electrical signal, characterized in that the light generated on the electron light receiving surface of the first element is condensed using the second element in direct contact with the surface of the third element that receives, detects, and converts the light into an electrical signal, and the third element detects it as an electrical signal, and the electrons are detected with high sensitivity, low noise, and by suppressing the upward travel of the electrons on the axis to improve efficiency.

2. The electron detection device according to claim 1, wherein the third element is an APD or an MPPPC.

3. The electron detection device according to claim 1 or claim 2, wherein the means for accelerating the energy of the electrons is a bias voltage applied to the sample.

4. The electron detection device according to any one of claims 1 to 3, wherein the means for accelerating the energy of the electrons is a bias voltage applied to a mesh provided between the sample and the first element.

5. The electron detection device according to any one of claims 1 to 4, further comprising an energy filter for separating the accelerated electrons.

6. The electron detection device according to claim 5, wherein the energy filter is a mesh.

7. The electron detection device according to claim 5, wherein the energy filter is an EXB.

8. An electron beam inspection device, characterized by incorporating any one of claims 1 to 7.

Images