Apparatus and methods for three dimensional reticle defect smart repair

1 . A method comprising:

emitting a first electron beam toward an object on a semiconductor substrate, the first electron beam emitted at a first accelerate voltage;

emitting a second electron beam toward the object on the semiconductor substrate, the second electron beam emitted at a second accelerate voltage, the second accelerate voltage being higher than the first accelerate voltage;

obtaining one or more images of the object on the semiconductor substrate based on the first accelerate voltage and the second accelerate voltage and a plurality of detectors positioned at respective locations, the one or more images of the object including one or more reticle defect images, wherein each reticle defect image is an image depicting at least one reticle defect of the object, and wherein each reticle defect image includes morphology information and composition information of the object and the at least one reticle defect;

analyzing the one or more images of the object on the semiconductor substrate with one or more artificial intelligence assisted object detection algorithms of a control module, wherein the control module includes a training module of training set data, an analysis model, memory resources, processing resources, and communication resources, and analyzing the one or more images of the object on the semiconductor substrate including: identifying a reticle defect type based on the one or more reticle defect images of the object;

performing a defect diagnosis by various parameters with the one or more artificial intelligence assisted object detection algorithms of the control module including:

generating a predicted reticle defect data, wherein the predicted reticle defect data includes a predicted thickness, a predicted composition, a predicted structure, a predicted shape, a predicted size, a predicted roughness, or a predicted length, width, and depth; and

comparing the predicted reticle defect data with a target reticle defects data, wherein the target reticle defects data includes parameters of a defect-free reticle;

performing a pre-defect repair diagnosis with the control module including determining whether a reticle cleaning process is performed or a reticle repair process is performed;

when the reticle repair process is determined to be performed in performing the pre-defect repair diagnosis, the reticle repair process is performed including:

determining a type of gas for repairing the reticle defect type by the control module;

adjusting process conditions by the control module for the reticle repair process based on the results of analyzing the one or more images of the object on the semiconductor substrate;

repairing the at least one reticle defect of the object using the adjusted process conditions, wherein the adjusted process conditions are determined to be predictive, and repairing the at least one reticle defect includes selectively supplying the type of gas, wherein the type of gas is at least one of a deposition gas or an etching gas, and wherein, when the type of gas is determined to be the etching gas, the etching gas is diluted in real-time based on at least one of the morphology information and composition information to avoid over-etching of underlying structures underneath the reticle defect;

when the reticle cleaning process is determined to be performed in performing the pre-defect repair diagnosis, the reticle cleaning process is performed including:

determining a type of cleaning chemical for cleaning the reticle defect type by the control module; and

performing a cleaning recipe on the reticle with the type of cleaning chemical.

2 . The method of claim 1 , wherein obtaining one or more images of an object based on a plurality of detectors positioned at respective locations includes:

receiving a first signal reflected from the object at a first detector among the plurality of detectors positioned at a first location;

receiving a second signal reflected from the object at a second detector among the plurality of detectors positioned at a second location; and

receiving a third signal reflected from the object at a third detector among the plurality of detectors positioned at a third location,

wherein the third location is opposite the object.

3 . The method of claim 2 , wherein the first, second, and third signals include signals from secondary electrons radiated from the object upon signals emitted from a signal source opposite of the object contacts the object.

4 . The method of claim 3 , wherein the first, second, and third signals include morphology information of the object.

5 . The method of claim 3 , wherein the first detector is laterally positioned from the object and the second detector is vertically positioned from the object,

wherein the first location of the first detector is closer to the object than the second location of the second detector.

6 . The method of claim 3 , wherein receiving the first signal reflected from the object at the first detector among the plurality of detectors positioned at the first location includes:

receiving the first signal at the first detector at a first angle from a plane in which the object is on, the first angle being between about 0 degrees and about 45 degrees.

7 . The method of claim 3 , wherein the second signal reflected from the object at the second detector among the plurality of detectors positioned at a second location includes:

receiving the second signal at the second detector at a second angle from a plane in which the object is on, the second angle being between about 45 degrees and about 75 degrees.

8 . The method of claim 2 , wherein obtaining one or more images of the object based on the plurality of detectors positioned at the respective locations includes:

receiving a fourth signal reflected from the object at a fourth detector among the plurality of detectors positioned at a fourth location,

wherein the fourth signal includes composition information of the object, and

wherein the fourth detector and the third detector are adjacently arranged, and the fourth location of the fourth detector is opposite the object.

9 . The method of claim 8 , wherein the fourth signal includes a signal from backscattered electrons radiated from the object upon respective signals emitted from a signal source opposite of the object contacts the object.

10 . The method of claim 1 , wherein the predicted reticle defect data further includes predicted morphology information and predicted composition information of the object are obtained based on the one or more artificial intelligence assisted object detection algorithms.

11 . The method of claim 10 , wherein the one or more artificial intelligence assisted object detection algorithms include a YOLO (You Only Look Once) model.

12 . The method of claim 1 , wherein the first accelerate voltage of the electron beam is 1 keV and the second accelerate voltage of the electron beam is 3 keV.

13 . A method comprising:

obtaining one or more images of an object on a semiconductor substrate based on a plurality of detectors positioned at respective locations,

wherein the one or more images of the object includes one or more reticle defect images, and each reticle defect image is an image depicting at least one reticle defect of the object, and the one or more reticle defect images include morphology information and composition information of the object and the at least one reticle defect, and

wherein obtaining one or more images of the object based on the plurality of detectors positioned at respective locations further includes:

receiving a first signal reflected from the object and a backscattered electron signal at a first detector positioned opposite of the object;

analyzing the one or more images of the object on the semiconductor substrate with one or more artificial intelligence assisted object detection algorithms of a control module, wherein the control module includes a training module of training set data an analysis model, memory resources, processing resources, and communication resources, and analyzing the one or more images of the object on the semiconductor substrate including: identifying a reticle defect type based on the one or more reticle defect images of the object;

performing a defect diagnosis by various parameters with the one or more artificial intelligence assisted object detection algorithms of the control module including:

generating a predicted reticle defect data, wherein the predicted reticle defect data includes a predicted thickness, a predicted composition, a predicted structure, a predicted shape, a predicted size, a predicted roughness, or a predicted length, width, and depth;

comparing the predicted reticle defect data with a target reticle defects data, wherein the target reticle defects data includes parameters of a defect-free reticle; and

performing a pre-defect repair diagnosis with the control module including determining whether a reticle cleaning process is performed or a reticle repair process is performed;

when the reticle repair process is determined to be performed in performing the pre-defect repair diagnosis, the reticle repair process is performed including:

determining a type of gas for repairing the reticle defect type by the control module;

adjusting process conditions by the control module for a reticle repair process based on the results of analyzing the one or more images of the object on the semiconductor substrate;

repairing the reticle defect of the object using the adjusted process conditions, wherein the adjusted process conditions are determined to be predictive, and repairing the at least one reticle defect includes selectively supplying the type of gas, wherein the type of gas is at least one of a deposition gas or an etching gas, and wherein, when the type of gas is determined to be the etching gas, the etching gas is diluted in real-time based on at least one of the morphology information and composition information of the at least one reticle defect;

when the reticle cleaning process is determined to be performed in performing the pre-defect repair diagnosis, the reticle cleaning process is performed including:

determining a type of cleaning chemical for cleaning the reticle defect type by the control module; and

performing a cleaning recipe on the reticle with the type of cleaning chemical, and wherein the composition information of the object and the at least one reticle defect being obtained from the backscattered electron signal received by the first detector.

14 . The method of claim 13 , wherein receiving the signal reflected from the object includes morphology information of the object and the at least one reticle defect.

15 . The method of claim 13 , wherein determining the type of gas for repairing the reticle defect type includes:

determining the type of gas for repairing the reticle defect type based on the morphology information and the composition information.

16 . The method of claim 13 , wherein obtaining one or more images of an object on a semiconductor substrate based on a plurality of detectors positioned at respective locations includes:

obtaining one or more plan view images.

17 . The method of claim 13 , wherein receiving the first signal reflected from the object includes receiving secondary electrons radiated from the object upon a light signal emitted from a signal source opposite of the object.

18 . The method of claim 13 , wherein obtaining one or more images of an object based on a plurality of detectors positioned at respective locations includes:

receiving a second signal reflected from the object at a second detector positioned laterally from the object at a first inclination angle;

receiving a third signal reflected from the object at a third detector positioned laterally from the object at a second inclination angle, the second inclination angle being greater than the first inclination angle; and

receiving a fourth signal reflected from the object at a fourth detector positioned opposite of the object.

19 . A method comprising:

obtaining one or more images of an object on a semiconductor substrate based on a plurality of detectors positioned at respective locations, the one or more images of the object including one or more reticle defect images, wherein each reticle defect image is an image depicting at least one reticle defect of the object, wherein the one or more reticle defect images include morphology information and composition information of the object and the at least one reticle defect, and the obtaining one or more images of the object based on the plurality of detectors includes:

receiving a first signal reflected from the object at a first detector;

receiving a second signal reflected from the object at a second detector;

receiving a third signal reflected from the object at a third detector; and

receiving a fourth signal reflected from the object and a backscattered electron signal at a fourth detector;

analyzing the one or more images of the object on the semiconductor substrate with one or more artificial intelligence assisted object detection algorithms of a control module, wherein the control module includes a training module of training set data, an analysis model, memory resources, processing resources, and communication resources, and analyzing the one or more images of the object on the semiconductor substrate including: identifying a reticle defect type based on the one or more reticle defect images of the object;

performing a defect diagnosis by various parameters with the one or more artificial intelligence assisted object detection algorithms of the control module including:

generating a predicted reticle defect data, wherein the predicted reticle defect data includes a predicted thickness, a predicted composition, a predicted structure, a predicted shape, a predicted size, a predicted roughness, or a predicted length, width, and depth;

comparing the predicted reticle defect data with a target reticle defects data, wherein the target reticle defects data includes parameters of a defect-free reticle;

performing a pre-defect repair diagnosis with the control module including determining whether a reticle cleaning process is performed or a reticle repair process is performed;

when the reticle repair process is determined to be performed in performing the pre-defect repair diagnosis, the reticle repair process is performed including:

determining a type of material for repairing the reticle defect type by a control module, wherein the control module includes training module of training set data, analysis model, memory resources, processing resources, and communication resources;

adjusting process conditions by the control module for a reticle repair process based on the results of analyzing the one or more images of the object on the semiconductor substrate; and

repairing the at least one reticle defect of the object using the adjusted process conditions, wherein the adjusted process conditions are determined to be predictive, and repairing the at least one reticle defect includes selectively supplying the type of gas, wherein the type of gas is at least one of a deposition gas or an etching gas, and wherein, when the type of gas is determined to be the etching gas, the etching gas is diluted in real-time based on at least one of the morphology information and composition information of the at least one reticle defect;

when the reticle cleaning process is determined to be performed in performing the pre-defect repair diagnosis, the reticle cleaning process is performed including:

determining a type of cleaning chemical for cleaning the reticle defect type by the control module; and

performing a cleaning recipe on the reticle with the type of cleaning chemical, and

wherein the composition information obtained from the backscattered electron signal, and the backscattered electron signal accounting for at least 40 percent of the morphology and composition information obtained for the object and the at least one reticle defect.

20 . The method of claim 19 , wherein

adjusting process conditions includes a concentration of the etching gas, an amount of the deposition gas, duration, or temperature in response to the real-time.