Apparatus and method estimating breakdown voltage of silicon dioxide film using neural network model

An apparatus for estimating a breakdown voltage of a silicon dioxide film, the apparatus comprising:

• one or more semiconductor chips configured to:
• control a plurality of different first test voltages to be applied to each of first test dies selected from among test dies;
• generate breakdown voltage information associated with the first test dies based on respective current levels generated by each of the first test dies in response to the plurality of different first test voltages;
• generate a breakdown voltage estimation model by updating a parameter of a neural network model based on the breakdown voltage information;
• control a plurality of different second test voltages to be applied to each of second test dies selected from among the test dies and distinct from the first test dies, wherein the plurality of different second test voltages are selected from among the plurality of different first test voltages, and have levels lower than levels of breakdown voltages for the respective first test dies;
• detect respective current levels generated by each of the second test dies in response to the plurality of different second test voltages;
• and
• estimate breakdown voltages of the second test dies without destroying respective silicon dioxide films of the second test dies by detecting respective current levels generated by each of the second test dies in response to the plurality of different second test voltages.

An operating method for a system estimating a breakdown voltage of a silicon dioxide film, the operating method comprising:

• applying a plurality of different first test voltages to a first test die, and detecting respective current levels generated by the first test die in response to the plurality of different first test voltages;
• generating breakdown voltage information associated with the first test die based on the current levels associated with the plurality of different first test voltages;
• generating a breakdown voltage estimation model by updating a parameter of a neural network model based on the breakdown voltage information;
• applying to a second test die, second test voltages having a lower level than a predetermined breakdown voltage, wherein the second test voltages are selected from among the plurality of different first test voltages;
• reading respective current levels output from the second test die in response to the second test voltages;
• and
• estimating a breakdown voltage of the second test die without destroying a silicon dioxide film of the second test die by inputting the respective current levels to the breakdown voltage estimation model.