System, method and apparatus for refining radio frequency transmission system models

• US 10,102,321 B2
• Filed: 11/10/2014
• Issued: 10/16/2018
• Est. Priority Date: 10/24/2014
• Status: Active Grant

• Current Assignee: Lam Research Corporation
• Original Assignee: Lam Research Corporation
• Inventors: Povolny, Henry S., Valcore, Jr., John C.
• Primary Examiner(s): Chad, Aniss
• Assistant Examiner(s): Gan, Chuen-Meei
• Application Number:
  US14/537,886
• Publication Number:
  US 20160117425A1
• Time in Patent Office: 1,436 Days
• CPC Class Codes: G06F 30/367 Design verification, e.g. using simulation, simulation program with integrated circuit emphasis [SPICE], direct methods or relaxation methods

Abstract

Systems and methods for defining a refined RF model of an RF transmission path includes identifying a selected initial term in the RF model having a least significant impact on a goodness of fit of the RF model with the selected initial term removed from the RF model. The initial term having the least significant impact on the goodness of fit is removed from the RF model. A selected non-initial term having a most significant impact on the goodness of fit of the RF model is identified and added to the RF model. The non-initial term being selected from a set of terms describing corresponding elements of the RF transmission path. The initial terms and non-initial terms having the most significant impact of goodness of fit are selected for a refined RF model.

Claims

1. A method of defining a chuck model for a radio frequency (RF) transmission path to calculate a wafer direct current (DC) bias at a chuck of a plasma chamber, comprising:

generating, by a computer, a cable model of an RF cable, a match model of an impedance matching circuit, a transmission model of an RF transmission line, and the chuck model, wherein the chuck model includes a series circuit model and a shunt circuit model, wherein the RF cable couples an RF generator to the impedance matching circuit and the RF transmission line couples the impedance matching circuit to the chuck;

identifying an initial term in the chuck model having a first level of impact on a goodness of fit of the chuck model, wherein the initial term is identified from a plurality of initial terms;

removing the initial term having the first level of impact on the goodness of fit from the chuck model, wherein the chuck model has a plurality of remaining ones of the plurality of initial terms after said removing of the initial term;

identifying a non-initial term having a second level of impact on the goodness of fit of the chuck model with the non-initial term added to the chuck model, wherein the non-initial term is not one of the plurality of initial terms and is identified from a plurality of non-initial terms, wherein the second level of impact is greater than the first level of impact;

selecting the remaining ones of the plurality of initial terms and the non-initial term having the second level of impact on the goodness of fit to update the chuck model;

measuring a voltage and current to generate a measured voltage and current, wherein said measuring is performed by a voltage and current probe within the RF generator, wherein the voltage and current probe is coupled to an output of the RF generator;

propagating the measured voltage and current via the cable model, the match model, and the transmission model to generate modeled voltage and current;

calculating the wafer DC bias at a top surface of the chuck of the plasma chamber by propagating the modeled voltage and current via the remaining ones of the plurality of initial terms and the non-initial term in the chuck model; and

controlling the plasma chamber based on the wafer DC bias to process a substrate placed within the plasma chamber.

2. The method of claim 1, wherein identifying the initial term in the chuck model having the first level of impact on the goodness of fit includes selecting the initial term in the chuck model, the plurality of initial terms being selected from a plurality of terms describing corresponding elements of the RF transmission path, the method further comprising:

determining a goodness of fit of the chuck model including the remaining initial terms; and

recording the goodness of fit of the chuck model including the remaining initial terms.

3. The method of claim 1, wherein identifying the non-initial term having the second level of impact on the goodness of fit includes selecting the non-initial term from the plurality of non-initial terms, the method further comprising:

determining a goodness of fit of the chuck model including the non-initial term and the remaining initial terms; and

recording the goodness of fit of the chuck model including the non-initial term and the remaining initial terms.

4. The method of claim 1, further comprising determining a third level of impact on the goodness of fit of the chuck model with the initial term and an additional initial term of the chuck model removed from the chuck model.

5. The method of claim 4, wherein determining the third level of impact on the goodness of fit of the chuck model with the initial term and the additional initial term of the chuck model removed from the chuck model includes:

selecting the additional initial term in the chuck model;

removing the additional initial term from the chuck model;

determining a goodness of fit of the chuck model including any remaining ones of the plurality of initial terms after said removing the additional initial term; and

recording the goodness of fit of the chuck model having the any remaining ones of the plurality of initial terms after said removing the additional initial term.

6. The method of claim 1, further including determining a third level of impact on the goodness of fit of the chuck model with the non-initial term and an additional non-initial additional term selected from the plurality of non-initial terms.

7. The method of claim 6, further comprising:

selecting the additional non-initial term from the plurality of non-initial terms;

adding the additional non-initial term to the chuck model;

determining a goodness of fit of the chuck model including the non-initial term, the additional non-initial term, and any remaining ones of the plurality of initial terms; and

recording the goodness of fit of the chuck model including the non-initial term, the additional non-initial term, and the any remaining ones of the plurality of initial terms.

8. The method of claim 1, wherein the non-initial term is a value of a resistor or an inductor or a capacitor of the shunt circuit model and is added to the chuck model after said removing of the initial term, wherein each of the plurality of initial terms is a value of the resistor or the inductor or the capacitor of the shunt circuit model and is present within the chuck model before said removing the initial term.

9. The method of claim 1, wherein the non-initial term is a value of a resistor or an inductor or a capacitor of the series circuit model and is added to the chuck model after said removing of the initial term, wherein each of the plurality of initial terms is a value of the resistor or the inductor or the capacitor of the series circuit model and is present within the chuck model before said removing the initial term.

10. A method of defining a chuck model for a radio frequency (RF) transmission path to calculate a wafer direct current (DC) bias at a chuck of a plasma chamber, comprising:

generating, by a computer, a cable model of an RF cable, a match model of an impedance matching circuit, a transmission model of an RF transmission line, and the chuck model, wherein the chuck model includes a series circuit model and a shunt circuit model, wherein the RF cable couples an RF generator to the impedance matching circuit and the RF transmission line couples the impedance matching circuit to the chuck;

selecting a first of a plurality of initial terms of the chuck model and a second of the plurality of initial terms in the chuck model;

removing the first and second initial terms from the chuck model;

determining a goodness of fit of the chuck model including remaining ones of the plurality of initial terms after said removing the first and second initial terms;

recording the goodness of fit of the chuck model including the remaining ones of the plurality of initial terms;

selecting a first additional term from a plurality of non-initial terms;

adding the first additional term to the chuck model;

determining a goodness of fit of the chuck model including the first additional term and the remaining ones of the plurality of initial terms;

recording the goodness of fit of the chuck model including the additional term and the remaining ones of the plurality of initial terms to update the chuck model;

measuring a voltage and current to generate a measured voltage and current, wherein said measuring is performed by a voltage and current probe within the RF generator, wherein the voltage and current probe is coupled to an output of the RF generator;

propagating the measured voltage and current via the cable model, the match model, and the transmission model to generate modeled voltage and current;

calculating the wafer DC bias at a top surface of the chuck of the plasma chamber by propagating the modeled voltage and current via the additional term and the remaining ones of the plurality of initial terms in the chuck model; and

controlling the plasma chamber based on the wafer DC bias to process a substrate placed within the plasma chamber.

11. The method of claim 10, wherein the first additional term is a value of a resistor or an inductor or a capacitor of the shunt circuit model.

12. The method of claim 10, wherein the first additional term is a value of a resistor or an inductor or a capacitor of the series circuit model.

13. The method of claim 10, wherein the match model is coupled to the cable model, the transmission model is coupled to the match model, and the chuck model is coupled to the transmission model.

14. A method of defining a chuck model for a radio frequency (RF) transmission path to calculate a wafer direct current (DC) bias at a chuck of a plasma chamber, comprising:

generating, by a computer, a cable model of an RF cable, a match model of an impedance matching circuit, a transmission model of an RF transmission line, and the chuck model, wherein the chuck model includes a series circuit model and a shunt circuit model, wherein the RF cable couples an RF generator to the impedance matching circuit and the RF transmission line couples the impedance matching circuit to the chuck;

designating a number of permanent terms and designating a number of random terms in the chuck model;

determining a goodness of fit of the chuck model with one of the permanent terms removed from the chuck model;

determining a goodness of fit of the chuck model with the one of the permanent terms of the chuck model removed from the chuck model and replaced with an additional term selected from a plurality of terms to update the chuck model;

measuring a voltage and current to generate a measured voltage and current, wherein said measuring is performed by a voltage and current probe within the RF generator, wherein the voltage and current probe is coupled to an output of the RF generator;

propagating the measured voltage and current via the cable model, the match model, and the transmission model to generate modeled voltage and current;

calculating the wafer DC bias at the chuck of the plasma chamber by propagating the modeled voltage and current via remaining ones of the permanent terms, the additional term, and the random terms in the chuck model; and

controlling the plasma chamber based on the wafer DC bias to process a substrate placed within the plasma chamber.

15. The method of claim 14, further comprising:

selecting the one of the permanent terms in the chuck model;

removing the one of the permanent terms from the chuck model; and

recording the goodness of fit of the chuck model including the remaining ones of the permanent terms and the random terms.

16. The method of claim 14, further comprising:

selecting the additional term from the plurality of terms;

adding the additional term to the chuck model; and

recording the goodness of fit of the chuck model including the additional term, the random terms, and the remaining ones of the permanent terms.

17. The method of claim 14, further comprising:

increasing a total order of the random terms to an increased order of random terms;

determining a goodness of fit of the chuck model with the increased order random terms;

recording the goodness of fit of the chuck model with the increased order random terms;

decreasing the total order of the random terms to a decreased order of random terms;

determining a goodness of fit of the chuck model with the decreased order random terms; and

recording the goodness of fit of the chuck model with the decreased order random terms.

18. The method of claim 14, wherein the number of the permanent terms and the number of the random terms is are added to output a total number of terms of the chuck model, the method further comprising:

increasing the total number of the terms to an increased number of terms;

determining a goodness of fit of the chuck model with the increased number of terms;

recording the goodness of fit of the chuck model with the increased number of terms;

decreasing the total number of terms to a decreased number of terms;

determining a goodness of fit of the chuck model with the decreased number of terms; and

recording the goodness of fit of the chuck model with the decreased number of terms.

19. A method for defining and optimizing a chuck model for a radio frequency (RF) path to calculate a wafer direct current (DC) bias at a chuck of a plasma chamber, comprising:

generating, by a computer, a cable model of an RF cable, a match model of an impedance matching circuit, a transmission model of an RF transmission line, and the chuck model, wherein the RF cable couples an RF generator to the impedance matching circuit and the RF transmission line couples the impedance matching circuit to the chuck;

changing a number of terms in the chuck model until a parameter level is achieved to optimize the chuck model to generate an optimized chuck model;

measuring a voltage and current to generate a measured voltage and current, wherein said measuring is performed by a voltage and current probe within the RF generator, wherein the voltage and current probe is coupled to an output of the RF generator;

propagating the measured voltage and current via the cable model, the match model, and the transmission model to generate model data at an output of the transmission model;

applying the model data to the optimized chuck model to calculate the wafer DC bias at the chuck; and

controlling the plasma chamber based on the wafer DC bias to process a substrate placed within the plasma chamber.

20. The method of claim 19, wherein the chuck model includes a shunt circuit model and a series circuit model, wherein said changing the number of terms includes deleting at least one of the terms, or adding at least one term to the terms, or a combination thereof, wherein the parameter level is a goodness of fit.