Plasma processing apparatus, plasma processing method and plasma processing analysis method

A plasma processing apparatus which performs plasma processing on a specimen with Advanced Process Control (APC) in use as control to suppress fluctuations in plasma processing by feedback control or feedforward control and is connected to an analysis unit configured to select a combination for the APC of plasma emission wavelength, time interval for the plasma emission wavelength, and a parameter for the plasma processing, and which includes an APC setup data storage portion which contains a stored changed-recipe parameter for plasma processing, the apparatus comprising:

• the analysis unit further configured to
• obtain a first regression line according to a regression equation representing correlation between a plasma emission intensity and a plasma processing result based on temporal change data of the plasma processing according to a fixed recipe with respect to a target etching result;
• obtain a second regression line according to the regression equation representing correlation between the plasma emission intensity and plasma processing result according to a changed recipe based on experimental data with respect to the target etching result;
• and
• select the combination for the APC of plasma emission wavelength, time interval for the plasma emission wavelength, and a parameter for the plasma processing, the selected combination for the APC of plasma emission wavelength, time interval for the plasma emission wavelength, and a parameter for the plasma processing making a difference between a gradient of the first regression line and a gradient of the second regression line smaller than a prescribed value and being selected based on a correlation of the first regression line,
• wherein the prescribed value corresponding with the selected combination is a value that causes the difference between the gradient of the first regression line and the gradient of the second regression line to be sufficiently small such that the plasma etching result can be estimated with respect to a target etching result using only the first regression line,
• wherein the changed recipe based on experimental data with respect to the target etching result is based on the changed-recipe parameter for plasma processing stored in the APC storage portion,
• wherein the changed-recipe parameter for plasma processing stored in the APC storage portion is a gas flow rate,
• wherein the difference between the gradient of the first regression line and the gradient of the second regression line is a stored degree of suitableness for the selected combination for the APC of plasma emission wavelength, the time interval for the plasma emission wavelength, and the parameter for the plasma processing,
• wherein the stored degree of suitableness for the selected combination for the APC is a weighted sum of a stored fixed-recipe residual value, a stored changed-recipe residual value, and a model difference value, and
• wherein the model difference value identifies a magnitude of a difference between a first relationship between an emission intensity monitor value and the etching result when a first combination of a first wavelength, a first time interval, a second wavelength, a second time interval, and the emission intensity monitor value calculated from fixed-recipe spectroscopic measurement data, and a second relationship between the emission intensity monitor value and the etching result when a second combination of the first wavelength, the first time interval, the second wavelength, the second time interval, and the emission intensity monitor value calculated from changed-recipe spectroscopic measurement data.

An analysis unit comprising:

• the analysis unit configured to select a combination of plasma emission wavelength, time interval for the plasma emission wavelength, and a parameter for plasma processing to suppress fluctuations under control in the plasma processing by feedback control or feedforward control, and which includes an APC setup data storage portion which contains a stored changed-recipe parameter for plasma processing,
• wherein a first regression line according to a regression equation representing correlation between a plasma emission intensity and a plasma processing result is obtained based on temporal change data of the plasma processing according to a fixed recipe with respect to a target etching result,
• wherein a second regression line according to the regression equation representing a correlation between the plasma emission intensity and plasma processing result is obtained according to a changed recipe based on experimental data with respect to the target etching result,
• wherein the combination of plasma emission wavelength, time interval for the plasma emission wavelength, and a parameter for plasma processing is selected, the selected combination for the APC of plasma emission wavelength, time interval for the plasma emission wavelength, and a parameter for the plasma processing making a difference between a gradient of the first regression line and a gradient of the second regression line smaller than a prescribed value and being selected based on a correlation of the first regression line, and
• wherein the prescribed value corresponding with the selected combination is a value that causes the difference between the gradient of the first regression line and the gradient of the second regression line to be sufficiently small such that the plasma etching result can be estimated with respect to a target etching result using only the first regression line,
• wherein the changed recipe based on experimental data with respect to the target etching result is based on the changed-recipe parameter for plasma processing stored in the APC storage portion,
• wherein the changed-recipe parameter for plasma processing stored in the APC storage portion is a gas flow rate,
• wherein the difference between the gradient of the first regression line and the gradient of the second regression line is a stored degree of suitableness for the selected combination for the APC of plasma emission wavelength, the time interval for the plasma emission wavelength, and the parameter for the plasma processing,
• wherein the stored degree of suitableness for the selected combination for the APC is a weighted sum of a stored fixed-recipe residual value, a stored changed-recipe residual value, and a model difference value, and
• wherein the model difference value identifies a magnitude of a difference between a first relationship between an emission intensity monitor value and the etching result when a first combination of a first wavelength, a first time interval, a second wavelength, a second time interval, and the emission intensity monitor value calculated from fixed-recipe spectroscopic measurement data, and a second relationship between the emission intensity monitor value and the etching result when a second combination of the first wavelength, the first time interval, the second wavelength, the second time interval, and the emission intensity monitor value calculated from changed-recipe spectroscopic measurement data.

A plasma processing method using a plasma processing apparatus which performs plasma processing on a specimen with Advanced Process Control (APC) in use as control to suppress fluctuations in plasma processing by feedback control or feedforward control and comprises an analysis unit to find a combination for the APC of plasma emission wavelength, time interval for the plasma emission wavelength, and a parameter for the plasma processing, the method comprising the steps of:

• storing a changed-recipe parameter for plasma processing;
• obtaining a first regression line according to a regression equation representing correlation between a plasma emission intensity and a plasma processing result based on temporal change data of the plasma processing according to a fixed recipe with respect to a target etching result;
• obtaining a second regression line according to a regression equation representing correlation between the plasma emission intensity and plasma processing result according to a changed recipe based on experimental data with respect to the target etching result;
• selecting the combination for the APC of plasma emission wavelength, time interval for the plasma emission wavelength, and a parameter for the plasma processing, the selected combination for the APC of plasma emission wavelength, time interval for the plasma emission wavelength, and a parameter for the plasma processing making a difference between a gradient of the first regression line and a gradient of the second regression line smaller than a prescribed value and being selected based on a correlation of the first regression line,
• wherein the prescribed value corresponding with the selected combination is a value that causes the difference between the gradient of the first regression line and the gradient of the second regression line to be sufficiently small such that the plasma etching result can be estimated with respect to a target etching result using only the first regression line,
• wherein the changed recipe based on experimental data with respect to the target etching result is based on a stored changed-recipe parameter for plasma processing,
• wherein the stored changed-recipe parameter for plasma processing is a gas flow rate,
• wherein the difference between the gradient of the first regression line and the gradient of the second regression line is a stored degree of suitableness for the selected combination for the APC of plasma emission wavelength, the time interval for the plasma emission wavelength, and the parameter for the plasma processing,
• wherein the stored degree of suitableness for the selected combination for the APC is a weighted sum of a stored fixed-recipe residual value, a stored changed-recipe residual value, and a model difference value, and
• wherein the model difference value identifies a magnitude of a difference between a first relationship between an emission intensity monitor value and the etching result when a first combination of a first wavelength, a first time interval, a second wavelength, a second time interval, and the emission intensity monitor value calculated from fixed-recipe spectroscopic measurement data, and a second relationship between the emission intensity monitor value and the etching result when a second combination of the first wavelength, the first time interval, the second wavelength, the second time interval, and the emission intensity monitor value calculated from changed-recipe spectroscopic measurement data.

An analysis method comprising:

• finding a combination of plasma emission wavelength, time interval for the plasma emission wavelength, and a parameter for plasma processing to suppress fluctuations under control in the plasma processing by feedback control or feedforward control,
• wherein a first regression line according to a regression equation representing a correlation between a plasma emission intensity and a plasma processing result is obtained based on temporal change data of the plasma processing according to a fixed recipe with respect to a target etching result,
• wherein a second regression line according to the regression equation representing correlation between the plasma emission intensity and plasma processing result is obtained according to a changed recipe based on experimental data with respect to the target etching result,
• wherein the combination of plasma emission wavelength, time interval for the plasma emission wavelength, and a parameter for plasma processing is selected, the selected combination for the Advanced Process Control (APC) of plasma emission wavelength, time interval for the plasma emission wavelength, and a parameter for the plasma processing making a difference between a gradient of the first regression line and the gradient of the second regression line smaller than a prescribed value, and a difference between an intercept of the first regression line and an intercept of the second regression line is smaller than a prescribed value and being selected based on a correlation of the first regression line,
• wherein the prescribed value corresponding with the selected combination is a value that causes the difference between the gradient of the first regression line and the gradient of the second regression line to be sufficiently small, and the difference between the intercept of the first regression line and the intercept of the second regression line is also sufficiently small, to allow the plasma etching result to be estimated with respect to a target etching result using only the first regression line,
• wherein the changed recipe based on experimental data with respect to the target etching result is based on a stored changed-recipe parameter for plasma processing,
• wherein the stored changed-recipe parameter for plasma processing is a gas flow rate,
• wherein the difference between the gradient of the first regression line and the gradient of the second regression line is a stored degree of suitableness for the selected combination for the APC of plasma emission wavelength, the time interval for the plasma emission wavelength, and the parameter for the plasma processing, and
• wherein the stored degree of suitableness for the selected combination for the APC is a weighted sum of a stored fixed-recipe residual value, a stored changed-recipe residual value, and a model difference value, and
• wherein the model difference value identifies a magnitude of a difference between a first relationship between an emission intensity monitor value and the etching result when a first combination of a first wavelength, a first time interval, a second wavelength, a second time interval, and the emission intensity monitor value calculated from fixed-recipe spectroscopic measurement data, and a second relationship between the emission intensity monitor value and the etching result when a second combination of the first wavelength, the first time interval, the second wavelength, the second time interval, and the emission intensity monitor value calculated from changed-recipe spectroscopic measurement data.