Match efficiency-variation compensation
AssigneeAdvanced Energy Industries, Inc.
InventorsDenis Shaw, Holger Philipp Kley, Suryan Emani
Life of this patent
Estimating the term…
Abstract
This disclosure describes systems, methods, and apparatus for identifying and compensating for variances between match networks. Each match network can be calibrated at the factory against a golden load and a reference load to turn a generalized model of the match network into an operation model. The operation model can be stored with the match network or shipped with the match network, and used by an operator to determine a correction factor during processing. This correction factor can be used to adjust a generator setpoint, frequency, or aspects of the match network.
Claims17
A system comprising:
- a generator;
- a match network u comprising:
- an input coupled to the generator;
- an output configured for electrical coupling to a plasma load within a plasma chamber;
- and
- variable reactance components with variable positions described by states, C1 to CK;
- an operation model, E_u(C,f), characterizing a deviation from nominal of the match network u over its operating range, the operation model, E_u(C,f), having parameters, p1 to pL, determined via calibration using one or more golden load fixtures, the states, C1 to CK, and one or more generator frequencies, f;
- a correction component configured to, during a processing recipe of the generator and the match network u:
- access an operating frequency, f′, of the generator;
- access the states, C1 to CK, after the match network u has been tuned for the plasma load;
- input the operating frequency, f′, as well as the states, C1 to CK, to the operation model, E_u(C,f), to generate a correction factor ε;
- and
- instruct the generator and/or match network u to adjust based on the correction factor ε.
A non-transitory, tangible computer readable storage medium, encoded with processor readable instructions to perform a method for identifying M constraints that can be used to form an operation model of a match network that accounts for variances from nominal, the method comprising, for M combinations of golden load fixtures and test generator frequencies f, generating a set of M constraints by:
- applying power from the test generator to a selected one of the golden load fixtures through a match network at a selected one of the frequencies f;
- tuning the match network to the selected one of the golden load fixtures and a reference load;
- measuring an input characteristic between the test generator and the match networks;
- measuring an output characteristic after the selected one of the golden load fixtures;
- using the input and output characteristics to extrapolate an intermediate characteristic between the match network and the selected one of the golden load fixtures;
- calculating one or more characteristics of the match network for a selected one of the M combinations based on the input characteristic, the intermediate characteristic, and the output characteristic.
A method comprising:
- for M combinations of golden load fixtures and test generator frequencies f, generating a set of M constraints by:
- applying power from the test generator to a selected one of the golden load fixtures through a match network at a selected one of the frequencies f;
- tuning the match network to the selected one of the golden load fixtures in series with a reference load;
- measuring an input characteristic between the test generator and the match network;
- measuring an output characteristic after the selected one of the golden load fixtures;
- using the input and output characteristics to extrapolate an intermediate characteristic between the match network and the selected one of the golden load fixtures;
- calculating one or more characteristics of the match network for a selected one of the M combinations based on the input characteristic, the intermediate characteristic, and the output characteristic;
- and
- using the one or more characteristics of the match network for the selected one of the M combinations to form one of the M constraints.
Dependencies are inferred from the claim text.