Large particle monitoring with laser power control for defect inspection

Abstract

A semiconductor wafer is inspected using a primary laser beam and a secondary laser beam. The secondary laser beam precedes the primary laser beam and has a lower power than the primary laser beam. Particles on the semiconductor wafer having a size that meets a threshold are detected using the secondary laser beam. In response to the particle being detected, the power of the primary laser beam and the power of the secondary laser beam are reduced. The particle passes through the primary laser beam while the primary laser beam is at the reduced power. After the particle has passed through the primary laser beam while the primary laser beam is at the reduced power, the power of the primary laser beam and the power of the secondary laser beam are restored in a controlled manner that is slower than a single step.

Description

[Technical Field]

[0001] TECHNICAL FIELD This disclosure relates to laser power modulation, and more particularly to laser power modulation for large particle monitoring. [Background technology]

[0002] [Related Applications] This application claims priority to U.S. Provisional Patent Application No. 63 / 082,484, filed September 24, 2020, the entire contents of which are incorporated by reference for all purposes.

[0003] As semiconductor design rules tighten, defect sizes are correspondingly shrinking. Defect inspection tools, also known as particle inspection tools, or simply inspection tools, must be able to detect small particles in advanced design layers (e.g., those with the tightest design rules). For laser scattering-based inspection tools, the required laser power density will continue to increase, or they will be unable to detect increasingly smaller particles. Therefore, newer inspection tools have higher laser power densities and smaller focal spot sizes.

[0004] However, laser heating of particle defects can cause the particle to explode (i.e., particle ablation). Increasing the laser power density reduces the threshold size, i.e., the size above which particle ablation occurs. Particle ablation can cause a single large particle to shred into hundreds of smaller particles that fall into the surrounding area on the semiconductor wafer, causing particle contamination. Higher laser power densities can also damage films deposited on the wafer, and possibly the wafer itself (e.g., when high power densities are combined with long exposure times). These issues can be mitigated by dynamically adjusting the laser power during wafer inspection. Dynamic adjustment of laser power can be called laser power modulation (LPM). [Prior art documents] [Patent documents]

[0005] [Patent Document 1] U.S. Patent Application Publication No. 2018 / 0038803 [Patent Document 2] International Publication No. 2004 / 001382 Summary of the Invention [Problem to be solved by the invention]

[0006] Therefore, there is a need for an improved method and system for laser power modulation during wafer inspection. [Means for solving the problem]

[0007] In certain embodiments, a method inspects a semiconductor wafer using a primary laser beam and a secondary laser beam. The secondary laser beam precedes the primary laser beam and has a lower power than the primary laser beam. The method also uses the secondary laser beam to detect particles on the semiconductor wafer having a size that meets a threshold, and reduces the power of the primary laser beam and the secondary laser beam in response to the particle being detected. The particle passes through the primary laser beam with the primary laser beam at the reduced power. The method further includes restoring the power of the primary laser beam and the secondary laser beam in a controlled manner that is slower than a single step after the particle passes through the primary laser beam with the primary laser beam at the reduced power.

[0008] In certain embodiments, the system includes an optical system that delivers a primary laser beam and a secondary laser beam to a semiconductor wafer. The secondary laser beam precedes the primary laser beam and has a lower power than the primary laser beam. The system also includes a detector that uses the secondary laser beam to detect particles on the semiconductor wafer having a size that meets a threshold. The optical system includes a modulator that, in response to the particle being detected, reduces the power of the primary laser beam and the secondary laser beam when the particle passes through the primary laser beam, and restores the power of the primary laser beam and the secondary laser beam in a controlled manner that is slower than a single step after the particle has passed through the primary laser beam.

[0009] For a better understanding of the various implementations described, reference should be made to the following drawings in conjunction with the detailed description below, which may not be to scale. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram of components of a semiconductor wafer inspection system according to certain embodiments. [Figure 2] FIG. 10 illustrates the secondary beam spot and main beam spot positions relative to a particle, the main beam and secondary beam signals for detecting the particle, and step beam power reduction and recovery to avoid ablation of the particle, according to certain embodiments. [Figure 3] 10A-10C illustrate various ways in which the power of the primary beam and the secondary beams may be restored in a controlled manner that is slower than a single step according to certain embodiments. [Figure 4] 10 is a graph showing modeled minimum ablation size and secondary beam sensitivity versus laser power. [Figure 5] FIG. 1 illustrates a laser beam power modulation system having an acousto-optic modulator (AOM) according to certain embodiments. [Figure 6]1 is a flowchart of a method for modulating the power of a primary laser beam and a secondary laser beam to avoid large particle ablation while inspecting a semiconductor wafer according to certain embodiments. DETAILED DESCRIPTION OF THE INVENTION

[0011] Like reference characters refer to corresponding parts throughout the drawings and specification.

[0012] Reference will now be made in detail to various embodiments, examples of which are illustrated in the accompanying drawings. In the following detailed description, numerous specific details are set forth to provide a consistent understanding of the various embodiments being described. However, it will be apparent to those skilled in the art that the various embodiments described may be practiced without such specific details. In other instances, well-known methods, procedures, components, circuits, and networks are not described in detail so as not to unnecessarily obscure aspects of the embodiments.

[0013] FIG. 1 is a diagram of components of a system 100 for inspecting a semiconductor wafer 116 according to certain embodiments. The system 100 uses a primary laser beam ("primary beam") 106 and a secondary laser beam ("secondary beam") 108 to inspect the semiconductor wafer 116. The secondary beam 108 has lower power than the primary beam 106 and may therefore be referred to as an LPM beam. Illumination optics within the system 100 include mirrors 110 and other optical components (not shown in FIG. 1 for simplicity) that deliver the primary beam 106 and the secondary beam 108 to the wafer 116. (Examples of other optical components within the illumination optics are shown in FIG. 5 and discussed below.) The primary beam 106 and the secondary beam 108 form respective spots 112 and 114 on the wafer 116. Because secondary beam 108 precedes (i.e., leads) main beam 106, a particle 118 on wafer 116 will pass through secondary beam 108 (i.e., in spot 114) and then through main beam 106 (i.e., in spot 112). Wafer 116 can be rotated during inspection (e.g., using a rotating chuck (not shown) on which wafer 116 is mounted), which in turn can cause particle 118 to pass through secondary beam 108 and then through main beam 106.

[0014] Light scattered from each spot 112 and 114 is collected by imaging objective 120 and directed to beam splitter 122. Beam splitter 122 directs the majority (e.g., 98%) of the beam power associated with main beam 106 and secondary beams 108 (those scattered from spots 112 and 114) to a main channel detector 124, which detects defects on the wafer 116. The light from secondary beam 108 directed to main channel detector 124 is offset from the light from main beam 106 directed to main channel detector 124. By appropriately positioning the apertures of main channel detectors 124, light from main beam 106 can be received by main channel detector 124 while rejecting light from secondary beam 108. The main channel detectors 124 detect defects on the wafer 116, including defects that are small enough not to be ablated by the main beam 106.

[0015] Beam splitter 122 redirects some of the beam power (e.g., 2%) of primary beam 106 and secondary beam 108 (scattered from spots 112 and 114) toward a photomultiplier tube (PMT) or PMT array 138 (or other detector). Lenses 126, 130, and 136 focus the light from secondary beam 108 redirected by beam splitter 122 onto PMT or PMT array 138. This light can pass into pupil plane 128 between lenses 126 and 130 and into wafer image plane 132 between lenses 130 and 136. A beam-blocking mask 134 (e.g., located between lenses 130 and 136) blocks the light from primary beam 106, preventing it from reaching PMT or PMT array 138.

[0016] Imaging objective 120, beam splitter 122, lenses 126, 130, and 136, and beam-blocking mask 134 at least partially comprise the collection optics for system 100. The collection optics may also include other optical elements, which are not shown for simplicity.

[0017] The power of the secondary beam 108 is sufficient to allow the PMT or PMT array 138 to detect particles large enough to pose a risk of ablation by the primary beam 106 (i.e., particles having a size that meets a threshold), such as particle 118, before they pass through the primary beam 106. Yet, the power of the secondary beam 108 is low enough to prevent ablation of large particles by the secondary beam 108 itself. The primary beam 106 has a higher power than the secondary beam 108 so that it can detect smaller defects that are unlikely to be ablated by the primary beam 106 and would not be detected using the secondary beam 108. Information from the PMT or PMT array 138 indicating the presence of large particles that pose a risk of ablation by the primary beam 106 can be used to reduce the power of the primary beam 106 until the large particles have passed through the primary beam 106, thereby avoiding ablation. For example, the PMT or PMT array 138 may assert a signal 140 in response to detecting a large particle (i.e., a particle having a size that meets a threshold) that poses a risk of ablation by the main beam 106. The power of the main beam 106 may be temporarily reduced in response to the assertion of the signal 140.

[0018] In certain embodiments, the illumination optics in the system 100 includes a diffractive optical element (DOE) 104 (or other beam splitter) that splits the third laser beam 102 into a main beam 106 and a secondary beam 108. As a result, the main beam 106 and the secondary beam 108 can have similar beam sizes and intensity profiles.

[0019] FIG. 2 shows the positions of spot 114 of secondary beam 108 and spot 112 of main beam 106 relative to particle 118. Spot 114 precedes (i.e., leads) spot 112 by a distance d. Distance d is, for example, in the range of 100-150 μm. Particle 118 first passes through spot 114 and then passes through spot 112 the distance d later. The passage of particle 118 through spot 114 produces a signal 200 (i.e., a secondary beam signal) indicating the presence of particle 118. Because secondary beam 108 has sufficient power to detect particle 118 but insufficient power to ablate particle 118, the passage of particle 118 through spot 114 does not result in ablation of particle 118. Without power attenuation of main beam 106, the passage of particle 118 through spot 112 produces a signal 202 (i.e., a main beam signal) indicating the presence of particle 118. Signal 202 is greater than signal 200 because the power of main beam 106 is higher than the power of secondary beam 108. The power of main beam 106 is sufficient not only to detect particle 118 but also to ablate it, which would cause additional contamination of wafer 116 (FIG. 1).

[0020] To avoid this ablation, the power 204 of the main beam 106 is set to a value P main , i.e., the default power of the main beam 106 when no particles of sufficient size to cause ablation are detected, to P main / AR, where AR is the attenuation ratio (i.e., the factor by which the power 204 is reduced). secondary Let P be the default power of the secondary beam 108. main P secondary It can be made equivalent to the ratio between (e.g., AR=P main / P secondary ) The power of the main beam 106 is P main From P secondary This attenuation to P prevents the particles 118 from being ablated. secondaryThe value of is chosen to maximize particle detection without causing particle ablation, and is therefore the desired power to which the main beam 106 power is reduced in the presence of large particles. In the example of FIG. 2, the power 204 of the main beam 106 is increased to P in a single step before the particle 118 reaches the spot 112. main From P main / AR, and the reduced power P main / AR, and after the particle 118 has traversed the spot 112, main / AR to P main That is, the power 204 is reduced from a first power level to a second power level, held at the reduced power level while the particle 118 traverses the spot 112, and then restored to the first power level. In certain embodiments, the power 204 of the main beam 106 is reduced to the reduced power P for a preset period (i.e., preset duration) sufficient for the particle 118 to traverse the spot 112. main / AR is maintained.

[0021] According to certain embodiments, since the main beam 106 and the secondary beams 108 are generated by splitting the third beam 102 (FIG. 1), the power of the main beam 106 can be reduced by appropriately attenuating the power of the third beam 102 (i.e., by modulating the third beam 102 to reduce its power). main From P main By attenuating the power of the third beam 102, the power of the secondary beam 108 is also reduced. That is, the power of the secondary beam 108 can be reduced by a factor of P main / AR and P secondary From P secondary / AR, and the power of the main beam 106 is reduced to P main At the same time, P secondary can be restored to.

[0022] One particle 118 is shown in Figure 2. The detection of this single particle 118 is called the primary trigger. In the post-trigger state caused by this primary trigger, the power of the main beam 106 is reduced to P main From P main / AR, and the power of the secondary beam 108 is reduced to P secondary From P secondary / AR. Multiple large particles may be located close to each other, so that during the post-trigger state, a second large particle may pass through spot 114 of secondary beam 108. The post-trigger state may end without the second large particle reaching spot 112 of main beam 106, but the power of main beam 106 will then be reduced to P main / AR is P main In this case, the second particle will be ablated because the primary beam 106 has already been restored to its original state. It is therefore desirable to be able to detect this second particle using the secondary beam 108 in the second state. Such detection is called a secondary trigger, and the post-trigger state is then extended so that the primary beam 106 is maintained at a reduced power P main / AR and the secondary beam 108 is kept in place. However, the reduced power P secondary Detecting this second particle is difficult because the laser power density is not sufficient to detect it. Compounding this problem is the fact that increasing the laser power density of an inspection tool makes even smaller particles detectable.

[0023] This problem can be at least partially alleviated by restoring the power of the main beam 106 and the power of the secondary beam 108 in a controlled manner that is slower than a single step. FIG. 3 illustrates various ways in which the power 300 of the main beam 106 can be restored in such a controlled manner, and the power of the secondary beam 108 can be restored accordingly, according to certain embodiments. In certain embodiments, the power 300 of the main beam 106 (and therefore the power of the secondary beam 108) is restored in a series of steps 302, 304, and 306. Each step 302, 304, and 306 corresponds to a different power level. While FIG. 3 shows three steps (i.e., with two intermediate steps 302 and 304), fewer steps (i.e., two steps) or more steps (e.g., 3-7 steps or 5-7 steps) may be used. The first step 302 has an attenuation ratio A1 less than AR, and the power of the main beam 106 in the first step 302 is P main / A1 (the power of the secondary beam 108 is P secondary The second step 304 has an attenuation ratio A2 less than A1, and the power of the main beam 106 in the second step 304 is P main / A2 (the power of the secondary beam 108 is P secondary In the third step 306 (or more generally the final step), the power of the main beam 106 is increased to P main (The power of the secondary beam 108 is P secondary (The power is fully recovered when the power is increased.) The value of AR is set to, for example, a range of 100 to 400. The value of A1 is set to, for example, a range of 20 to 50. The value of A2 is set to, for example, a range of 2 to 10. Power 300 is P main / AR (the power of the secondary beam 108 is P secondary / AR) has a preset duration, for example, in the range of 300-400 μs or 400-500 μs (e.g., depending on the spin speed of the wafer 116). The duration of each intermediate step 302 and 304 has a preset duration, for example, in the range of 300-400 μs.

[0024] In certain embodiments, the power 300 of the main beam 106 (and thus the power of the secondary beam 108) is recovered by ramping up the power of the main beam 106 and the secondary beam 108 according to a substantially smooth (e.g., smooth within the available resolution) curve, such as curve 308 or 312. In certain embodiments, the power 300 of the main beam 106 (and thus the power of the secondary beam 108) is recovered by ramping up the power of the main beam 106 and the secondary beam 108 according to a substantially linear (e.g., linear within the available resolution) slope 310.

[0025] Figure 4 is a graph 400 illustrating the principle behind the controlled power recovery of Figure 3. Graph 400 shows the relationship between modeled minimum ablation size 402 (i.e., the smallest particle size at which ablation will occur) and modeled secondary beam sensitivity 404 (i.e., the smallest particle size that can be detected using secondary beam 108) and laser power, in terms of particle size and laser power in arbitrary units ("au"). The x-axis gives laser power and has a logarithmic scale, while the y-axis gives particle size and has a linear scale. The controlled power recovery of Figure 3 utilizes the difference between secondary beam sensitivity 404 and minimum ablation size 402.

[0026] At the far right of graph 400 (i.e., point 410 on the x-axis), the power of main beam 106 is P main , the power of the secondary beam 108 is P secondary At the leftmost side of the graph 400 (i.e., x=0.005), the power of the main beam 106 is P main / AR, the power of the secondary beam 108 is P secondary At point 406 on the x-axis, the power of the main beam 106 is P main / A1, the power of the secondary beam 108 is P secondary At point 408 on the x-axis, the power of the main beam 106 is P main / A2, the power of the secondary beam 108 is P secondary / A2. Point 406 corresponds to step 302 (FIG. 3). Point 408 corresponds to step 304 (FIG. 3).

[0027] At point 410 (i.e., at the default, full power level), secondary beam sensitivity 404 equals 0.15, which is approximately equal to the minimum ablation size 402. On the other hand, at x=0.005 (i.e., at the reduced power level), secondary beam sensitivity 404 is only 0.4, which is significantly higher than the minimum ablation size at point 410. Particles having sizes in the range of approximately 0.15 to 0.4 that overshoot spot 114 at the reduced power level and then overshoot spot 112 at the full power level in a single step upon beam power restoration would be undetected and ablated using secondary beam 108. However, if beam power is only partially restored until step 302 (FIG. 3), which corresponds to point 406, secondary beam sensitivity 404 improves significantly (to just above 0.2), thereby improving the ability to detect ablation-sized particles using secondary beam 108 during step 302. This improvement increases the probability of successful secondary triggering and decreases the probability of particle ablation. Further partial restoration of beam power to step 304 (FIG. 3) further improves secondary beam sensitivity 404, further increasing the probability of successful secondary triggering and further decreasing the probability of particle ablation. At point 408, secondary beam 108 can be used to detect the smallest size particles that would be ablated at full power levels.

[0028] FIG. 5 illustrates a laser beam power modulation system 500 used to generate the primary beam 106 and the secondary beam 108 according to certain embodiments. The system 500 includes a laser 502 that generates a laser beam 504. In certain embodiments, the laser 502 is a green laser, and the laser beam is green (e.g., has a frequency in the range of 500-565 nm). For example, the laser 502 is a 532 nm mode-locked laser. In certain embodiments, the laser 502 includes a laser cavity (oscillator) and a second harmonic generator (SHG) that frequency-doubles the laser light emitted from the laser cavity. For example, the laser cavity may include a Nd:YVO4 crystal lasing at a fundamental wavelength of 1064 nm, and the SHG converts the 1064 nm light to 532 nm light to produce the laser beam 504.

[0029] Acousto-optic modulator (AOM) 506 receives laser beam 504 and splits laser beam 504 into a primary beam 528 (i.e., a zeroth order beam) and multiple diffracted beams 530 (e.g., ±1st order diffraction beams, ±2nd order diffraction beams, etc.) based on an input signal received from driver 514. AOM 506 performs this splitting by diffracting a portion of laser beam 504 into diffracted beams 530 based on an input signal received from driver 514. Thus, when AOM 506 is operating, the power of primary beam 528 is less than the power of laser beam 504. The input signal specifies the amount of diffraction performed by AOM 506, which in turn specifies the amount of power reduction in primary beam 528, which corresponds to the amount of splitting performed.

[0030] In certain embodiments, driver 514 is a radio frequency (RF) driver and the input signal is an RF signal. The power of primary beam 528 depends on the input signal, as does the power of multiple diffracted beams 530. For example, increasing the RF power of the input signal increases the diffraction performed by AOM 506, thereby decreasing the power of primary beam 528, and therefore diffracting more of laser beam 504 into multiple diffracted beams 530. Similarly, decreasing the RF power of the input signal decreases the diffraction performed by AOM 506, thereby increasing the power of primary beam 528, and therefore diffracting less laser beam 504 into multiple diffracted beams 530.

[0031] Slit 524 acts as a filter that transmits primary beam 528 and dumps (i.e., filters out) multiple diffracted beams 530. In certain embodiments, slit 524 includes a pair of mirrors with a gap between them that transmits primary beam 528 through the gap while reflecting multiple diffracted beams 530 to beam dump 526.

[0032] In certain embodiments, one or more mirrors direct the primary beam 528 and the plurality of diffracted beams 530 toward the slit 524. The one or more mirrors can include a first mirror (M1) 518 and a second mirror (M2) 520. The first mirror 518 reflects the primary beam 528 and the plurality of diffracted beams 530 and directs them toward the second mirror 520. The second mirror 520 reflects the primary beam 528 and the plurality of diffracted beams 530 and directs them toward the slit 524. The first mirror 518 and the second mirror 520 are positioned between the AOM 506 and the slit 524 along the optical paths of the primary beam 528 and the plurality of diffracted beams 530. The first mirror 518 is positioned between the AOM 506 and the second mirror 520 along the optical paths of the primary beam 528 and the plurality of diffracted beams 530.

[0033] In certain embodiments, the system 500 further includes an attenuator 508 that attenuates the primary beam 528. The attenuator 508 may also attenuate the plurality of diffracted beams 530. For example, the attenuator 508 may direct some of the light from the primary beam 528 and the plurality of diffracted beams 530 to a beam dump 516. The attenuator 508 may be positioned between the AOM 506 and one or more mirrors that direct the primary beam 528 and the plurality of diffracted beams 530 toward the slit 524. For example, the attenuator 508 may be positioned between the AOM 506 and the first mirror 518.

[0034] In certain embodiments, sensor 522 (e.g., a green sensor) receives light from primary beam 528 that leaks through an individual mirror of the one or more mirrors. Because primary beam 528 is not completely reflected by the individual mirror, some of the light from primary beam 528 leaks through (i.e., is transmitted through) the individual mirror. For example, the individual mirror may be a second mirror 520, through which light from primary beam 528 leaks and is provided to sensor 522. Sensor 522 may include an aperture that filters out light from multiple diffracted beams 530 that leaks through second mirror 520, allowing sensor 522 to sense only light from primary beam 528. The power of the light leaked through the individual mirror and sensed by sensor 522 is related to (e.g., a percentage of) the power of primary beam 528. By sensing this light, sensor 522 generates feedback indicative of the power of primary beam 528 and provides that feedback to controller 510 .

[0035] A controller 510 (e.g., a microcontroller) generates a control signal 512 based in part on feedback from a sensor 522. In certain embodiments, in the process of generating the control signal 512, the controller 510 filters out noise (e.g., kilohertz-scale noise) from the feedback. The controller 510 provides the control signal 512 to a driver 514, which generates an input signal based on the control signal 512 and provides it to the AOM 506. For example, the control signal 512 specifies the RF power of the input signal. In certain embodiments, the control signal 512 is a DC voltage (V DC ), which is fed to the input of driver 514.

[0036] The primary beam 528 transmitted by the slit 524 is provided to a harmonic generator 540, which converts a first portion of the primary beam 528 into a converted beam 542, i.e., a harmonic of the primary beam 528. Because the conversion efficiency of the harmonic generator 540 is less than 100 percent, the harmonic generator 540 also transmits a second portion 544 of the primary beam 528. The harmonic generator 540 includes a nonlinear optical crystal (e.g., a cesium lithium borate (CLBO) crystal). In certain embodiments, the harmonic generator 540 is a fourth-order harmonic generator, so that the frequency of the converted beam 542 can be four times the fundamental frequency of the laser beam emitted by the laser cavity in the laser 502 and twice the frequency of the primary beam 528. In certain embodiments, the primary beam 528 is green and the converted beam 542 is ultraviolet (UV). For example, the wavelength of the primary beam 528 is set to 532 nm and the wavelength of the converted beam 542 is set to 266 nm.

[0037] In certain embodiments, a beam reduction telescope (BRT) 532 converges and focuses the primary beam 528 onto a harmonic generator 540. The beam reduction telescope 532 is located on the path of the primary beam 528 between the slit 524 and the harmonic generator 540. The beam reduction telescope 532 can change the shape of the beam spot of the primary beam 528 from circular to elliptical. The beam reduction telescope 532 can include a spherical lens 534 and a pair of cylindrical lenses 536 and 538, where the spherical lens 534 can be located closest to the slit 524.

[0038] Because converted beam 542 is a harmonic of primary beam 528, it has a different frequency than primary beam 528 and therefore a different frequency than second portion 544 of primary beam 528. Frequency separator 546 receives converted beam 542 and second portion 544 of the primary beam, transmits converted beam 542, and dumps second portion 544 of the primary beam.

[0039] In certain embodiments, sensor 560 (e.g., a UV sensor) receives light from converted beam 542 that leaks through frequency separator 546. The power of the light that leaks through frequency separator 546 and is sensed by sensor 560 is related to (e.g., a percentage of) the power of converted beam 542. By sensing this light, sensor 560 generates feedback indicative of the power of converted beam 542 and provides the feedback to controller 510. Controller 510 can generate control signal 512 based in part on the feedback from sensor 560. For example, controller 510 can generate control signal 512 using feedback from both sensor 560 and sensor 522, or can generate control signal 512 using feedback from sensor 560 but without feedback from sensor 522. Alternatively, controller 510 can generate control signal 512 using feedback from sensor 522 but without feedback from sensor 560. In certain embodiments, in the process of generating its control signal 512, controller 510 filters out noise (eg, kilohertz-scale noise) from the feedback from sensor 560 and / or sensor 522.

[0040] In addition to receiving feedback from sensors 560 and / or 522, controller 510 receives signal 140 from PMT or PMT array 138 (FIG. 1) (or from other detectors). Signal 140 can be used by controller 510 to generate control signal 512 in addition to or instead of feedback from sensors 560 and / or 522. For example, controller 510 can generate control signal 512 in response to assertion of signal 140 that modulates primary beam 528 in a controlled manner, as shown in FIG.

[0041] In certain embodiments, frequency separator 546 includes a Perambroca prism 550, a third mirror (M3) 552, and a fourth mirror (M4) 556. Third mirror 552 is positioned in the beam path of converted beam 542 between Perambroca prism 550 and fourth mirror 556. Perambroca prism 550 spatially separates converted beam 542 from second portion 544 of the primary beam. The front surface of Perambroca prism 550 also reflects a portion of second portion 544 of the primary beam toward beam dump 548. Third mirror 552 is a dichroic mirror, which reflects converted beam 542 while sending second portion 544 of the primary beam toward beam dump 554, thereby dumping second portion 544 of the primary beam.

[0042] By reflecting converted beam 542, third mirror 552 directs converted beam 542 to fourth mirror 556, which reflects converted beam 542 and transmits converted beam 542 out of frequency separator 546. Third mirror 552 can also reflect some (e.g., less than 5%) of second portion 544 of the primary beam along a spatially separated path to fourth mirror 556. Fourth mirror 556 can also be dichroic, which reflects some (e.g., less than 5%) of the light from the second portion of the primary beam and directs it to beam dump 558. That is, frequency separator 546 essentially dumps second portion 544 of the primary beam (e.g., green light not converted to UV by harmonic generator 540) while transmitting converted beam 542 (e.g., UV light generated by harmonic generator 540).

[0043] The converted beam 542 transmitted by frequency separator 546 is fed to DOE 104 (or other beam splitter), which splits the converted beam 542 into primary beam 106 and secondary beam 108. Converted beam 542 is an example of third beam 102 (FIG. 1).

[0044] In certain embodiments, controller 510 may include a processor circuit and may include embedded non-volatile memory that stores instructions (e.g., one or more programs) executed by the processor circuit. For example, controller 510 may be a microcontroller with embedded flash memory or other non-volatile memory. Alternatively, the controller may be coupled to a separate non-volatile memory that stores instructions executed by the processor circuit. The non-volatile memory, whether embedded within or coupled to controller 510, is an example of a non-transitory computer-readable storage medium. Execution of the instructions by the processing circuitry enables controller 510 to perform the functions described herein in connection with controller 510 (e.g., generating control signal 512). In certain embodiments, a filter (e.g., an analog filter, such as an RC filter) is provided within controller 510 and selectively coupled to driver 514 in response to assertion of signal 140 (e.g., after a predetermined period of time has elapsed since assertion of signal 140), such that the power of primary beam 528 (and therefore the power of converted beam 542) can be ramped up according to a substantially smooth curve (e.g., curve 308 or 312 in FIG. 3).

[0045] In certain embodiments, another type of modulator is used to modulate primary beam 528 instead of AOM 506. For example, a Pockels cell can be used.

[0046] FIG. 6 is a flowchart of a method 600 for avoiding large particle ablation by modulating the power of a primary laser beam (e.g., primary beam 106 in FIGS. 1 and 5) and a secondary laser beam (e.g., secondary beam 108 in FIGS. 1 and 5) while inspecting a semiconductor wafer (e.g., wafer 116 in FIG. 1), according to certain embodiments.

[0047] In certain embodiments of the method 600, a third laser beam (e.g., third beam 102 in FIG. 1 or converted beam 542 in FIG. 5) is generated (602). For example, a harmonic generator (e.g., harmonic generator 540 in FIG. 5) is used to convert a portion of a fourth laser beam (e.g., primary beam 528 in FIG. 5) into the third laser beam (604), thereby making the third laser beam a harmonic of the fourth laser beam. The third laser beam is split (606) into a primary laser beam and a secondary laser beam. For example, the third laser beam is provided to a diffractive optical element (e.g., DOE 104 in FIGS. 1 and 5) (608), which splits the third laser beam into a primary laser beam and a secondary laser beam.

[0048] The semiconductor wafer is inspected 610 using a primary laser beam and a secondary laser beam, which precedes the primary laser beam (e.g., by a distance d in FIG. 3) and has a lower power than the primary laser beam.

[0049] Using the secondary laser beam, particles (e.g., particles 118 in Figures 1, 3, and 5) on the semiconductor wafer that have a size that meets a threshold (i.e., their particle size is above or equal to the threshold) are detected (612).

[0050] In response to the particle being detected, the power of the primary laser beam and the power of the secondary laser beam are reduced (614). For example, the power of the primary laser beam and the power of the secondary laser beam are reduced in a single step as shown in FIG. 3. The particle is detected when the primary laser beam reaches a reduced power (e.g., reduced power level P main / AR) and passes through the main laser beam.

[0051] In certain embodiments, the power of the third laser beam is reduced (616) to reduce the power of the primary laser beam and the secondary laser beam. For example, reducing the power of the fourth laser beam (618) reduces the power of the third laser beam, and thus the power of the primary laser beam and the secondary laser beam. A signal (e.g., signal 140 in FIGS. 1 and 5) may be asserted in response to a particle being detected. The power of the fourth laser beam may be reduced in response to the assertion of the signal. If the power of the fourth laser beam is modulated using an AOM (e.g., AOM 506 in FIG. 5) in the path of the fourth laser beam, the power of the fourth laser beam may be reduced by increasing the RF power supplied to the AOM (e.g., by driver 514 controlled by controller 510 in FIG. 5).

[0052] After the particle passes through the primary laser beam while the primary laser beam is at the reduced power, the power of the primary laser beam and the power of the secondary laser beam are restored (620) in a controlled manner that is slower than a single step. In certain embodiments, the power of the primary laser beam and the power of the secondary laser beam are restored (622) in a series of steps (e.g., steps 302, 304, and 306 in FIG. 3). For example, the series of steps can include 3 to 7 steps (e.g., 5 to 7 steps). In other certain embodiments, the power of the primary laser beam and the power of the secondary laser beam are ramped up (624) according to a substantially smooth curve (e.g., curves 308 or 312 in FIG. 3) or according to a substantially linear slope (e.g., substantially linear slope 310 in FIG. 3).

[0053] In certain embodiments, the power of the third laser beam is restored 626 in a controlled manner to restore the power of the primary laser beam and the power of the secondary laser beam. For example, the third laser beam can be restored in a series of steps (e.g., 3-7 steps) (e.g., 5-7 steps). In another example, the power of the third laser beam is ramped up according to a substantially smooth curve or a substantially linear slope.

[0054] In certain embodiments, the power of the fourth laser beam is restored in a controlled manner (628) to restore the power of the third laser beam. For example, the fourth laser beam can be restored in a series of steps (e.g., 3-7 steps), (e.g., 5-7 steps), (e.g., at least 3 steps, or at least 5 steps). In another example, the power of the fourth laser beam can be ramped up according to a substantially smooth curve or a substantially linear slope. According to certain embodiments, the power of the fourth laser beam can be restored by reducing the RF power supplied to the AOM (i.e., reducing the RF power relative to the increased RF power used to reduce the power of the fourth laser beam). For example, reducing the RF power supplied to the AOM can restore it to its level prior to reducing the power of the fourth laser beam in step 618.

[0055] Although method 600 includes a number of operations that appear to occur in a particular order, method 600 may include more or fewer operations. Two or more operations may be combined into a single operation. The performance of two or more operations may overlap. For example, test operation 610 may be performed continuously while other operations of method 600 are performed.

[0056] The above description is for purposes of explanation and has been written with reference to specific embodiments. However, the illustrative discussion above is not intended to be exhaustive or to limit the claims to the precise form disclosed. Many modifications and variations are possible in light of the above teachings. The embodiments have been selected to illustrate the principles underlying the claims and their practical application, and to enable those skilled in the art to use the embodiments with various modifications appropriate to the specific use envisioned.

Claims

1. 1. A method comprising: Inspecting a semiconductor wafer using a primary laser beam and a secondary laser beam, wherein the secondary laser beam precedes the primary laser beam and has a lower power than the primary laser beam; using the secondary laser beam to detect particles on the semiconductor wafer having a size that meets a threshold; reducing the power of the primary laser beam and the secondary laser beam in response to the particle being detected, so that the particle passes through the primary laser beam while the primary laser beam is at the reduced power; and after the particle passes through the primary laser beam with the primary laser beam at the reduced power, restoring the power of the primary laser beam and the power of the secondary laser beam in a plurality of steps in a predetermined controlled manner; method.

2. 2. The method of claim 1, wherein the plurality of steps comprises 3 to 7 steps.

3. 2. The method of claim 1, wherein the plurality of steps includes at least three steps.

4. 10. The method of claim 1, wherein the power of the primary laser beam and the power of the secondary laser beam are ramped up according to a substantially smooth curve when the power of the primary laser beam and the power of the secondary laser beam are restored in the controlled manner.

5. 10. The method of claim 1, wherein the power of the primary laser beam and the power of the secondary laser beam are ramped up according to a substantially linear slope when restoring the power of the primary laser beam and the power of the secondary laser beam in the controlled manner.

6. 2. The method of claim 1, further comprising splitting a third laser beam into the main laser beam and the secondary laser beam; reducing the power of the third laser beam when reducing the power of the main laser beam and the power of the secondary laser beam; and When the power of the main laser beam and the power of the secondary laser beam are restored in the controlled manner, the power of the third laser beam is restored in the controlled manner. method.

7. 7. The method of claim 6, wherein splitting the third laser beam comprises providing the third laser beam to a diffractive optical element, which splits the third laser beam into the main laser beam and the secondary laser beam.

8. 7. The method of claim 6, further comprising generating the third laser beam by converting a portion of a fourth laser beam into the third laser beam using a harmonic generator, thereby making the third laser beam a harmonic of the fourth laser beam.

9. 9. The method of claim 8, When reducing the power of the third laser beam, reducing the power of the fourth laser beam; and When the power of the third laser beam is restored in the controlled manner, the power of the fourth laser beam is restored in the controlled manner. method.

10. 10. The method of claim 9, wherein restoring the power of the fourth laser beam in the controlled manner comprises restoring the power of the fourth laser beam in multiple steps.

11. 11. The method of claim 10, wherein the plurality of steps consists of 3 to 7 steps.

12. 11. The method of claim 10, wherein the plurality of steps includes at least three steps.

13. 10. The method of claim 9, wherein the power of the fourth laser beam is ramped up according to a substantially smooth curve when the power of the fourth laser beam is restored in the controlled manner.

14. 10. The method of claim 9, wherein the power of the fourth laser beam is ramped up according to a substantially linear slope when the power of the fourth laser beam is restored in the controlled manner.

15. 10. The method of claim 9, further comprising validating a signal in response to the particle being detected; increasing radio frequency (RF) power supplied to an acousto-optic modulator (AOM) on the path of the fourth laser beam when the power of the fourth laser beam is reduced; and reducing the RF power supplied to the AOM when restoring the power of the fourth laser beam in the controlled manner; method.

16. 1. A system comprising: an optical system for delivering a primary laser beam and a secondary laser beam to the semiconductor wafer, the secondary laser beam leading the primary laser beam and having a lower power than the primary laser beam; a detector that uses the secondary laser beam to detect particles on the semiconductor wafer having a size that meets a threshold; the optical system comprising a modulator, the modulator being configured to: reducing the power of the primary laser beam and the power of the secondary laser beam when the particle passes through the primary laser beam; and After the particle has passed through the primary laser beam, the power of the primary laser beam and the power of the secondary laser beam are restored in a plurality of steps in a predetermined controlled manner. system.

17. 17. The system of claim 16, wherein the plurality of steps comprises 3 to 7 steps.

18. 17. The system of claim 16, wherein the plurality of steps includes at least three steps.

19. 17. The system of claim 16, wherein the modulator restores the power of the primary laser beam and the power of the secondary laser beam by ramping up the power of the primary laser beam and the power of the secondary laser beam according to a substantially smooth curve.

20. 17. The system of claim 16, wherein the modulator restores the power of the primary laser beam and the power of the secondary laser beam according to a substantially linear ramp line.

21. 17. The system of claim 16, the detector enabling a signal in response to the particle being detected; The system further comprises a controller that receives the signal, In response to the signal being valid, the controller reduces the power of the primary laser beam and the power of the secondary laser beam when the particle passes through the primary laser beam; and the controller restoring the power of the primary laser beam and the power of the secondary laser beam in the controlled manner after the particle has passed through the primary laser beam. system.

22. 22. The system of claim 21, further comprising a beam splitter that splits the third laser beam into the main laser beam and the secondary laser beam; In response to the signal being valid, the controller reduces the power of the third laser beam when the particle passes within the primary laser beam; and the controller restoring the power of the third laser beam in the controlled manner after the particle has passed through the primary laser beam. system.

23. 23. The system of claim 22, wherein the beam splitter comprises a diffractive optical element.

24. 23. The system of claim 22, further comprising a harmonic generator for converting a portion of a fourth laser beam into the third laser beam, thereby making the third laser beam a harmonic of the fourth laser beam; In response to the signal being valid, the controller reduces the power of the fourth laser beam when the particle passes within the primary laser beam; and the controller restoring the power of the fourth laser beam in the controlled manner after the particle has passed through the main laser beam. system.

25. 25. The system of claim 24, the modulator is an acousto-optic modulator (AOM); a radio frequency (RF) driver coupled between the controller and the AOM to provide RF power to the AOM; In response to the signal being valid, the controller causes the RF driver to increase the RF power supplied to the AOM when the particle passes through the primary laser beam, thereby decreasing the power of the fourth laser beam when the particle passes through the primary laser beam; and the controller causes the RF driver to reduce the RF power supplied to the AOM after the particle has passed through the main laser beam, thereby restoring power of the fourth laser beam in the controlled manner after the particle has passed through the main laser beam. system.

26. 26. The system of claim 25, wherein the controller restores the power of the fourth laser beam in multiple steps.

27. 27. The system of claim 26, wherein the plurality of steps comprises 3 to 7 steps.

28. 27. The system of claim 26, wherein the plurality of steps includes at least three steps.

29. 26. The system of claim 25, wherein the controller restores the power of the fourth laser beam according to a substantially smooth curve.

30. 26. The system of claim 25, wherein the controller restores the power of the fourth laser beam according to a substantially linear ramp line.

Images