Method for discriminating wafer region and wafer region discrimination apparatus for performing the same
A method of discrimination a wafer region and a wafer region discrimination apparatus performing the same are disclosed. A wafer region discrimination apparatus includes a probe including an optical sensor and configured to obtain optical spectrum signals of which signal values vary based on a thickness characteristic of a wafer at a plurality of positions on the wafer and a processor configured to identify whether a current optical spectrum signal is obtained from a die region on the wafer in which a chip is disposed using a region discrimination model based on a neural network that uses the current optical spectrum signal obtained by the probe as an input.
1 . A wafer region discrimination apparatus comprising:
a probe comprising an optical sensor and configured to obtain optical spectrum signals of which signal values vary based on a thickness characteristic of a wafer at a plurality of positions on the wafer; and
a processor configured to identify whether a current optical spectrum signal obtained by the probe is obtained from a die region on the wafer in which a chip is disposed or from a non-die region on the wafer in which the chip is not disposed using a region discrimination model based on a neural network that uses the current optical spectrum signal obtained by the probe as an input, wherein the region discrimination model is trained using training data which include optical spectrum signals corresponding to the die region of the wafer and optical spectrum signals corresponding to the non-die region of the wafer.
2 . The wafer region discrimination apparatus of claim 1 , wherein the non-die region comprises at least one of an edge region in the wafer in which the chip is not disposed and a scribe line region between die regions.
3 . The wafer region discrimination apparatus of claim 1 , wherein the processor is configured to identify whether the current optical spectrum signal is obtained from a region corresponding to a wafer ring frame for fixing the wafer using the region discrimination model.
4 . The wafer region discrimination apparatus of claim 3 , wherein the processor is configured to identify whether the current optical spectrum signal is obtained from an air region outside the wafer ring frame using the region discrimination model.
5 . The wafer region discrimination apparatus of claim 1 , wherein the processor is configured to estimate a thickness of a thin film of the wafer based on the current optical spectrum signal when it is determined that the current optical spectrum signal is obtained from the die region.
6 . The wafer region discrimination apparatus of claim 5 , wherein the processor is configured to estimate the thickness of the thin film of the wafer using a wafer thickness estimation model based on a neural network that uses the current optical spectrum signal as an input.
7 . The wafer region discrimination apparatus of claim 1 , wherein the processor is configured to discriminate a type of a region corresponding to each of the positions of the wafer using the plurality of optical spectrum signals and the region discrimination model.
8 . The wafer region discrimination apparatus of claim 1 , wherein the optical spectrum signals obtained by the probe comprise signals obtained as light in a plurality of wavelength bands reflects off the wafer.
9 . The wafer region discrimination apparatus of claim 1 , wherein the region discrimination model is configured to discriminate a first signal region corresponding to a die region and a second signal region corresponding to a non-die region from a signal of a predetermined wavelength extracted from optical spectrum signals obtained along a movement trajectory of the probe.
10 . The wafer region discrimination apparatus of claim 1 , wherein the region discrimination model is a region discrimination model based on a convolutional neural network.
11 . A method of discriminating a wafer region, the method comprising:
obtaining optical spectrum signals of which signal values vary based on a thickness characteristic of a wafer at a plurality of positions on the wafer through a probe; and
identifying whether a current optical spectrum signal obtained by the probe is obtained from a die region on the wafer in which a chip is disposed or from a non-die region on the wafer in which the chip is not disposed using a region discrimination model based on a neural network that uses the current optical spectrum signal obtained by the probe as an input, wherein the region discrimination model is trained using training data which include optical spectrum signals corresponding to the die region of the wafer and optical spectrum signals corresponding to the non-die region of the wafer.
12 . The method of claim 11 , wherein the identifying of whether the current optical spectrum signal is obtained from the die region on the wafer in which the chip is disposed comprises identifying whether the current optical spectrum signal is obtained from a region corresponding to a wafer ring frame for fixing the wafer using the region discrimination model.
13 . The method of claim 11 , wherein the identifying of whether the current optical spectrum signal is obtained from the die region on the wafer in which the chip is disposed comprises identifying whether the current optical spectrum signal is obtained from an air region outside the wafer ring frame using the region discrimination model.
14 . The method of claim 11 , further comprising:
estimating a thickness of a thin film of the wafer based on the current optical spectrum signal when it is determined that the current optical spectrum signal is obtained from the die region.
15 . The method of claim 11 , wherein the region discrimination model is configured to discriminate a first signal region corresponding to a die region and a second signal region corresponding to a non-die region from a signal of a predetermined wavelength extracted from optical spectrum signals obtained along a movement trajectory of the probe.