JP7374751B2 - Pad temperature adjustment device, pad temperature adjustment method, polishing device, and polishing system - Google Patents

Description

The present invention relates to a pad temperature adjustment device and a pad temperature adjustment method for adjusting the surface temperature of a polishing pad used for polishing a substrate such as a wafer. The present invention also relates to a polishing device incorporating a pad temperature adjustment device and a polishing system including at least one polishing device.

A CMP (Chemical Mechanical Polishing) apparatus is used in the process of polishing the surface of a substrate such as a wafer in the manufacture of semiconductor devices. The CMP apparatus has at least one polishing unit, which holds the wafer with a polishing head, rotates the wafer, and presses the wafer against a polishing pad on a rotating polishing table to polish the surface of the wafer. Polish. During polishing, a polishing liquid (slurry) is supplied to the polishing pad, and the surface of the wafer is flattened by the chemical action of the polishing liquid and the mechanical action of abrasive grains contained in the polishing liquid.

The polishing rate of a wafer depends not only on the polishing load of the wafer on the polishing pad, but also on the surface temperature of the polishing pad. This is because the chemical action of the polishing liquid on the wafer is temperature dependent. Therefore, in the manufacture of semiconductor devices, it is important to maintain the surface temperature of the polishing pad at an optimal value during wafer polishing in order to increase the wafer polishing rate and keep it constant.

Therefore, pad temperature adjustment devices have been conventionally used to adjust the surface temperature of polishing pads (see, for example, Patent Document 1 and Patent Document 2). In general, a pad temperature adjustment device includes a pad contact member that can come into contact with the surface (polishing surface) of a polishing pad, a liquid supply system that supplies temperature-controlled heating liquid and cooling liquid to the pad contact member, and a surface of the polishing pad. The polishing pad includes a pad temperature measuring device that measures temperature, and a control unit that controls the liquid supply system based on the surface temperature of the polishing pad measured by the pad temperature measuring device. The control unit controls the heating liquid and the cooling liquid based on the pad surface temperature measured by the pad temperature measuring device so that the surface temperature of the polishing pad reaches a predetermined target temperature and then maintains the surface temperature at the target temperature. Control the flow rate. For example, based on the difference between the target temperature and the surface temperature of the polishing pad, the control unit controls the operation amount of a flow rate regulating valve arranged in the heating liquid supply pipe of the liquid supply system and the flow rate arranged in the cooling liquid supply pipe. The flow rate of the heating liquid and cooling liquid supplied to the pad contacting member is controlled by PID controlling the operation amount of the regulating valve. Thereby, the surface temperature of the polishing pad can quickly reach the optimum value and can be maintained at this optimum value.

JP 2017-148933 Publication JP2018-027582A

The control unit of the CMP apparatus stores in advance PID parameters (ie, proportional gain P, integral gain I, and differential gain D) used in the PID control. When a CMP apparatus has a plurality of polishing units, the same PID parameter is used to adjust the surface temperature of the polishing pad in each polishing unit. Furthermore, a plurality of CMP apparatuses are often installed in a manufacturing factory for semiconductor devices and the like. Generally, the same PID parameters are stored in the control section of each CMP apparatus. That is, a plurality of CMP apparatuses adjust the surface temperature of the polishing pad of each polishing unit using a common PID parameter among these CMP apparatuses.

However, even though the same PID parameters are used, variations in temperature behavior occur due to machine differences between polishing units. As used herein, temperature behavior refers to changes in the surface temperature of the polishing pad over time from the time when the pad contacting member starts adjusting the surface temperature of the polishing pad to the time when the target temperature is reached.

FIG. 19 is a graph showing the temperature behavior in each of the four polishing units. In FIG. 19, the vertical axis represents the surface temperature of the polishing pad, and the horizontal axis represents time. In FIG. 19, time Ta is the time when the pad contact member contacts the polishing pad, that is, the time when the pad contact member starts adjusting the surface temperature of the polishing pad. In FIG. 19, time Tb is the time when the polishing head holding the wafer contacts the polishing pad, that is, the time when polishing the wafer is started. Note that in this specification, a curve showing temperature behavior as shown in FIG. 19 is referred to as a temperature behavior curve.

As shown in FIG. 19, the temperature behavior curves of each polishing unit are different from each other. As mentioned above, the wafer polishing rate also depends on the surface temperature of the polishing pad. Therefore, if the temperature behavior differs between each polishing unit, the polishing performance will also differ between each polishing unit. For example, in FIG. 19, temperature behavior curve R1 is located above temperature behavior curve R2. This means that the polishing performance (polishing rate) is different between the polishing unit having the temperature behavior curve R1 and the polishing unit having the temperature behavior curve R2. If polishing performance differs between each polishing unit, there is a possibility that the yield of products (ie, semiconductor devices) will be adversely affected.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a pad temperature adjustment device and a pad temperature adjustment method that can suppress variations in polishing performance between polishing units. Another object of the present invention is to provide a polishing device incorporating such a pad temperature adjustment device, and a polishing system including at least one polishing device.

In one aspect, there is provided a pad temperature adjustment device for causing the surface temperature of a polishing pad to reach a predetermined target temperature and thereafter maintaining the temperature at the target temperature, the device being capable of contacting the surface of the polishing pad, and configured to be able to a pad contact member in which a passage and a cooling passage are formed; a heating liquid supply pipe connected to the heating passage; a cooling liquid supply pipe connected to the cooling passage; A liquid supply system comprising a first flow control valve attached to the cooling liquid supply pipe and a second flow control valve attached to the cooling liquid supply pipe, and supplying temperature-adjusted heating liquid and cooling liquid to the pad contact member. , a pad temperature measuring device that measures the surface temperature of the polishing pad, and operation of the first flow control valve and the second flow control valve based on the difference between the measured value of the pad temperature measuring device and the target temperature. a control unit that performs PID control of the amount, and the control unit is configured to maintain a temperature behavior curve created based on the measured value of the pad temperature measuring device and the measurement time within a predetermined tolerance range. A storage unit storing a learned model constructed by machine learning and a data set including at least one temperature behavior parameter are input to the learned model, and a changed value of the PID parameter of the PID control is output. a processing device that performs calculations for the polishing pad, and the temperature behavior curve is defined as the temperature behavior curve of the polishing pad from the time when the pad contacting member starts adjusting the surface temperature of the polishing pad to the time when the target temperature is reached. is a curve showing a change over time in the surface temperature of the a data set consisting of a PID parameter for PID control and a film thickness parameter related to a film thickness of a substrate pressed against the polishing pad, and (c) a PID parameter for the PID control, a height of the polishing pad, and the polishing pad. A data set consisting of a polishing rate of a substrate polished at a pad temperature adjustment device is provided.

In one aspect, the learned model is constructed by deep learning using a neural network, and the control unit, when inputting a learning data set including the at least one temperature behavior parameter to the neural network, The weight parameters of the neural network are determined such that the neural network provides the PID parameters to be changed to maintain the temperature behavior curve within a predetermined tolerance range and the changed values of the PID parameters that are within the normal range. Adjustments are made to construct the learned model.
In one aspect, the control unit stores the temperature behavior curve created each time the substrate is polished by pressing the substrate against the polishing pad, and at least one temperature behavior parameter associated with the temperature behavior curve. The learning data set is created from the accumulated at least one temperature behavior parameter .
In one aspect, the temperature of the heating liquid, the rotation speed of a polishing head that presses a substrate against the polishing pad, the rotation speed of a polishing table to which the polishing pad is attached, dressing conditions of the polishing pad, polishing load of the polishing head, and the flow rate of the polishing liquid supplied to the polishing pad are further input into the learned model.
In one aspect, a pressing load of the pad contact member against the polishing pad, a temperature of a polishing liquid supplied to the polishing pad, an ambient temperature in a polishing unit in which the pad temperature adjustment device is arranged, and a supply pressure of the heating liquid. , and the supply pressure of the coolant are further input into the learned model.
In one aspect, the time at which the at least one temperature behavior parameter input to the trained model was obtained is further input to the trained model .

In one embodiment, while a pad contact member is in contact with the surface of a polishing pad, a heating liquid and a cooling liquid are respectively flowed into a heating channel and a cooling channel formed in the pad contact member, and a pad temperature measuring device is used to measure the temperature of the polishing pad. Attached to the heated liquid supply pipe connected to the heating flow path so as to measure the surface temperature of the polishing pad, make the surface temperature of the polishing pad reach a predetermined target temperature, and then maintain it at the target temperature. The operation amount of the first flow control valve attached to the coolant supply pipe connected to the cooling flow path and the second flow control valve attached to the coolant supply pipe connected to the cooling flow path are PID controlled, and the measured value of the pad temperature measuring device and its measurement are A trained model for maintaining a temperature behavior curve created based on a time point within a predetermined tolerance range is constructed by machine learning, and a dataset including at least one temperature behavior parameter is input to the trained model. and outputs a changed value of the PID parameter of the PID control, and the temperature behavior curve is determined from the time when the pad contacting member starts adjusting the surface temperature of the polishing pad to the time when the target temperature is reached. It is a curve showing a change in surface temperature of a polishing pad over time, and the data set includes (a) a data set consisting of a PID parameter of the PID control, a flow rate of the heating liquid, and a flow rate of the cooling liquid; ) a data set consisting of a PID parameter of the PID control and a film thickness parameter related to the film thickness of the substrate pressed against the polishing pad, and (c) a PID parameter of the PID control, the height of the polishing pad, and the A data set consisting of a polishing rate of a substrate polished with a polishing pad is provided.

In one aspect, at least one polishing unit includes a polishing table that supports a polishing pad, a polishing head that presses a substrate against the polishing pad, and a surface temperature of the polishing pad that reaches a predetermined target temperature; , a pad temperature adjustment device for maintaining the target temperature, the pad temperature adjustment device being able to contact the surface of the polishing pad, and having a heating channel and a cooling channel formed therein. a pad contact member, a heating liquid supply pipe connected to the heating flow path, a cooling liquid supply pipe connected to the cooling flow path, a first flow control valve attached to the heating liquid supply pipe, and the a second flow control valve attached to a cooling liquid supply pipe; a liquid supply system that supplies temperature-adjusted heating liquid and cooling liquid to the pad contact member; and a pad that measures the surface temperature of the polishing pad. a temperature measuring device; and a control unit that performs PID control on the operating amounts of the first flow control valve and the second flow control valve based on the difference between the measured value of the pad temperature measuring device and the target temperature. In order to maintain the temperature behavior curve created based on the measured value of the pad temperature measuring device and the measurement time point within a predetermined tolerance range, the control unit uses a learning constructed by machine learning to maintain the temperature behavior curve created based on the measured value of the pad temperature measuring device and the measurement time point within a predetermined tolerance range. a processing device that inputs a data set including at least one temperature behavior parameter into the learned model and executes an operation for outputting a changed value of the PID parameter of the PID control; , the temperature behavior curve represents a change in the surface temperature of the polishing pad over time from the time when the pad contacting member starts adjusting the surface temperature of the polishing pad to the time when the target temperature is reached. The data set is a data set consisting of (a) the PID parameter of the PID control, the flow rate of the heating liquid, and the flow rate of the cooling liquid, (b) the PID parameter of the PID control, and the polishing a data set consisting of film thickness parameters related to the film thickness of the substrate pressed against the pad; and (c) from the PID parameters of the PID control, the height of the polishing pad, and the polishing rate of the substrate polished by the polishing pad. A polishing apparatus is provided that is characterized by any one of the following data sets .

In one aspect, at least one polishing device, a relay device connected to the polishing device so that information can be transmitted and received, and a host control system connected to the relay device so that information can be transmitted and received, The polishing apparatus includes at least one polishing unit including a polishing table that supports a polishing pad, a polishing head that presses a substrate against the polishing pad, and a surface temperature of the polishing pad that reaches a predetermined target temperature; and a pad temperature adjustment device for maintaining the target temperature at the target temperature, the pad temperature adjustment device being able to contact the surface of the polishing pad and having a heating channel and a cooling channel formed therein. a heated liquid supply pipe connected to the heating flow path, a cooling liquid supply pipe connected to the cooling flow path, and a first flow control valve attached to the heating liquid supply pipe; a second flow control valve attached to the cooling liquid supply pipe; a liquid supply system that supplies temperature-adjusted heating liquid and cooling liquid to the pad contact member; and measuring the surface temperature of the polishing pad. a pad temperature measuring device; and a control unit that performs PID control on the operating amounts of the first flow control valve and the second flow control valve based on the difference between the measured value of the pad temperature measuring device and the target temperature. The host control unit of the host control system controls the machine to maintain a temperature behavior curve created based on the measured value of the pad temperature measuring device and the measurement time point within a predetermined tolerance range. A storage unit storing a learned model constructed by learning and a data set including at least one temperature behavior parameter are input to the learned model, and an operation for outputting a changed value of the PID parameter of the PID control. and a processing device for performing the above, wherein the temperature behavior curve is the surface temperature of the polishing pad from the time when the pad contacting member starts adjusting the surface temperature of the polishing pad to the time when the target temperature is reached. The data set includes (a) a data set consisting of the PID parameter of the PID control, the flow rate of the heating liquid, and the flow rate of the cooling liquid; (b) the data set of the PID parameter of the PID control; (c) a data set consisting of a PID parameter and a film thickness parameter related to the film thickness of the substrate pressed against the polishing pad; Provided is a polishing system characterized in that the data set is a data set of polishing rates of substrates .

In one aspect, the polishing apparatus includes at least one polishing device, a relay device connected to the polishing device so as to be able to transmit and receive information, and a host control system connected to the relay device so as to be able to transmit and receive information. includes at least one polishing unit including a polishing table that supports a polishing pad and a polishing head that presses a substrate against the polishing pad; a pad temperature adjustment device for maintaining a target temperature, the pad temperature adjustment device being able to contact the surface of the polishing pad and having a heating channel and a cooling channel formed therein; a member, a heating liquid supply pipe connected to the heating flow path, a cooling liquid supply pipe connected to the cooling flow path, a first flow control valve attached to the heating liquid supply pipe, and the cooling liquid. a second flow control valve attached to a supply pipe; a liquid supply system that supplies temperature-adjusted heating liquid and cooling liquid to the pad contacting member; and a pad temperature measurement that measures the surface temperature of the polishing pad. and a control unit that performs PID control of the operating amounts of the first flow control valve and the second flow control valve based on the difference between the measured value of the pad temperature measurement device and the target temperature. , the relay control unit of the relay device is constructed by machine learning in order to maintain a temperature behavior curve created based on the measured value of the pad temperature measuring device and the measurement time within a predetermined tolerance range. inputting into the trained model a storage unit storing a trained model, a combination of the measured value of the pad temperature measuring device and the measurement time point, and a data set including at least one temperature behavior parameter; , a processing device that executes an operation for outputting a changed value of a PID parameter of the PID control as needed while adjusting the surface temperature of the polishing pad , and the temperature behavior curve is determined based on the pad contact. A curve showing a change in the surface temperature of the polishing pad over time from the time when the member starts adjusting the surface temperature of the polishing pad to the time when the target temperature is reached, and the data set is (a) a data set consisting of the PID parameters of the PID control, the flow rate of the heating liquid, and the flow rate of the cooling liquid; (b) the PID parameters of the PID control and a film thickness related to the film thickness of the substrate pressed against the polishing pad; and (c) a data set consisting of a PID parameter of the PID control, a height of the polishing pad, and a polishing rate of a substrate polished with the polishing pad. A polishing system is provided.

According to the invention, by changing the PID parameters, the temperature behavior curve can be maintained within a predetermined tolerance range. As a result, variations in polishing performance between polishing units are suppressed, so product yield can be improved.

FIG. 1 is a schematic diagram showing one embodiment of a polishing unit arranged in a polishing apparatus. FIG. 2 is a top view of the temperature measurement area of the pad temperature measurement device. FIG. 3 is a side view of the temperature measurement area of the pad temperature measurement device. FIG. 4 is a horizontal cross-sectional view of one embodiment of a pad contact member. FIG. 5 is a schematic diagram showing one embodiment of the control section. FIG. 6 is a graph showing an example of a temperature behavior curve created by the control unit. FIG. 7 is a graph showing another example of the temperature behavior curve created by the control unit. FIG. 8 is a schematic diagram showing an example of multiple data sets. FIG. 9 is a diagram showing a normal distribution created based on the data set D2 at the measurement time point T2 shown in FIG. FIG. 10 is a schematic diagram showing an example of the configuration of an artificial intelligence installed in the control unit shown in FIG. 5. FIG. 11 is a flowchart for explaining a method for constructing the learned model shown in FIG. 10. FIG. 12 is a schematic diagram showing an example of the structure of a neural network. FIG. 13(a) is a schematic diagram showing an example of a polishing unit equipped with an eddy current type film thickness sensor, and FIG. 13(b) is a schematic cross-sectional view of the polishing unit shown in FIG. 13(a). FIG. 14 is a schematic diagram showing an example of a polishing unit equipped with a pad height sensor that measures the amount of change in the thickness of the polishing pad. FIG. 15 is a schematic diagram showing a modification of the configuration of the artificial intelligence shown in FIG. 10. FIG. 16 is a schematic diagram illustrating one embodiment of a polishing system including at least one polishing device. FIG. 17 is a schematic diagram illustrating another embodiment of a polishing system including at least one polishing device. FIG. 18 is a schematic diagram showing how the pad surface temperature is adjusted by a pad temperature adjustment member spaced upward from the surface of the polishing pad. FIG. 19 is a graph showing the temperature behavior in each of the four polishing units.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a schematic diagram showing one embodiment of a polishing unit arranged in a polishing apparatus. The polishing apparatus includes at least one polishing unit as shown in FIG. As shown in FIG. 1, the polishing unit includes a polishing head 1 that holds and rotates a wafer W, which is an example of a substrate, a polishing table 2 that supports a polishing pad 3, and a polishing liquid (e.g. A polishing liquid supply nozzle 4 that supplies a slurry) and a pad temperature adjustment device 5 that adjusts the surface temperature of the polishing pad 3 are provided. The surface (upper surface) of the polishing pad 3 constitutes a polishing surface on which the wafer W is polished.

The polishing head 1 is vertically movable and rotatable about its axis in the direction indicated by the arrow. The wafer W is held on the lower surface of the polishing head 1 by vacuum suction or the like. A motor (not shown) is connected to the polishing table 2, and is rotatable in the direction shown by the arrow. As shown in FIG. 1, polishing head 1 and polishing table 2 rotate in the same direction. The polishing pad 3 is attached to the upper surface of the polishing table 2.

The polishing unit further includes a dresser 20 for dressing the polishing pad 3 on the polishing table 2. The dresser 20 is configured to swing on the surface of the polishing pad 3 in the radial direction of the polishing pad 3. The lower surface of the dresser 20 constitutes a dressing surface made of a large number of abrasive grains such as diamond particles. The dresser 20 rotates while rocking on the polishing surface of the polishing pad 3, and dresses the surface of the polishing pad 3 by slightly scraping the polishing pad 3.

Polishing of the wafer W is performed as follows. The wafer W to be polished is held by the polishing head 1 and further rotated by the polishing head 1. On the other hand, the polishing pad 3 is rotated together with the polishing table 2. In this state, the polishing liquid is supplied to the surface of the polishing pad 3 from the polishing liquid supply nozzle 4, and the surface of the wafer W is further pressed against the surface of the polishing pad 3 (ie, the polishing surface) by the polishing head 1. The surface of the wafer W is polished by sliding contact with the polishing pad 3 in the presence of a polishing liquid. The surface of the wafer W is flattened by the chemical action of the polishing liquid and the mechanical action of the abrasive grains contained in the polishing liquid.

The pad temperature adjustment device 5 includes a pad contact member 11 that can contact the surface of the polishing pad 3 , a liquid supply system 30 that supplies temperature-adjusted heating liquid and cooling liquid to the pad contact member 11 , and a liquid supply system 30 that can contact the surface of the polishing pad 3 . The control unit 40 is provided to control at least the flow rates of the heating liquid and the cooling liquid so that the temperature reaches a predetermined target temperature and then maintains the temperature at the target temperature.

In this embodiment, the control unit 40 is configured to control the operation of the entire polishing apparatus including the pad temperature adjustment device 5. In the following description, an embodiment will be described in which the control unit 40 controls the operation of one polishing unit including the pad temperature adjustment device 5, but the present embodiment is not limited to this example. For example, when the polishing apparatus includes a plurality of polishing units, the control section 40 can individually control the operation of each polishing unit.

The liquid supply system 30 shown in FIG. 1 includes a heated liquid supply tank 31 as a heated liquid supply source that stores temperature-adjusted heated liquid, and a heated liquid supply pipe that connects the heated liquid supply tank 31 and the pad contact member 11. 32 and a heated liquid return pipe 33. One end of the heated liquid supply pipe 32 and the heated liquid return pipe 33 is connected to the heated liquid supply tank 31, and the other end is connected to the pad contact member 11.

The temperature-adjusted heating liquid is supplied from the heating liquid supply tank 31 to the pad contact member 11 through the heating liquid supply pipe 32, flows inside the pad contact member 11, and then returns from the pad contact member 11 through the heating liquid return pipe 33. It is returned to the supply tank 31. In this way, the heating liquid circulates between the heating liquid supply tank 31 and the pad contact member 11. In this embodiment, a heating source (for example, a heater) 48 is arranged in the heated liquid supply tank 31 . This heating source 48 heats the heating liquid stored in the heating liquid supply tank 31 to a predetermined temperature (set temperature).

A first on-off valve 41 and a first flow rate control valve 42 are attached to the heated liquid supply pipe 32 . The first flow control valve 42 is arranged between the pad contact member 11 and the first on-off valve 41. The first on-off valve 41 is a valve that does not have a flow rate adjustment function, whereas the first flow rate control valve 42 is a valve that has a flow rate adjustment function.

The liquid supply system 30 further includes a coolant supply pipe 51 and a coolant discharge pipe 52 connected to the pad contact member 11. The coolant supply pipe 51 is connected to a coolant supply source (for example, a cold water supply source) provided in a factory where the polishing apparatus is installed. The coolant is supplied to the pad contact member 11 through the coolant supply pipe 51 , flows through the pad contact member 11 , and is discharged from the pad contact member 11 through the coolant discharge pipe 52 . In one embodiment, the coolant flowing within the pad contacting member 11 may be returned to the coolant supply through the coolant drain tube 52.

A second on-off valve 55 and a second flow control valve 56 are attached to the coolant supply pipe 51 . The second flow control valve 56 is arranged between the pad contact member 11 and the second on-off valve 55. The second opening/closing valve 55 is a valve that does not have a flow rate adjustment function, whereas the second flow rate control valve 56 is a valve that has a flow rate adjustment function.

As the heating liquid supplied to the pad contact member 11, warm water is used. The hot water is heated to about 80° C. by the heating source 48 of the heating liquid supply tank 31, for example. If the surface temperature of the polishing pad 3 is to be raised more quickly, silicone oil may be used as the heating liquid. When silicone oil is used as the heating liquid, the silicone oil is heated to 100° C. or higher (for example, about 120° C.) by the heating source 48 of the heating liquid supply tank 31. As the cooling fluid supplied to the pad contact member 11, cold water or silicone oil is used. When using silicone oil as a cooling liquid, the polishing pad 3 can be quickly cooled by connecting a chiller as a cooling liquid supply source to the cooling liquid supply pipe 51 and cooling the silicone oil to below 0°C. can. Pure water can be used as cold water. A chiller may be used as a cooling liquid source to cool pure water to produce chilled water. In this case, the cold water flowing through the pad contact member 11 may be returned to the chiller through the coolant discharge pipe 52.

The heating liquid supply pipe 32 and the cooling liquid supply pipe 51 are completely independent pipes. Therefore, the heating liquid and the cooling liquid are simultaneously supplied to the pad contact member 11 without being mixed. The heating liquid return pipe 33 and the cooling liquid discharge pipe 52 are also completely independent piping. Therefore, the heating liquid is returned to the heating liquid supply tank 31 without being mixed with the cooling liquid, and the cooling liquid is discharged or returned to the cooling liquid supply without being mixed with the heating liquid.

The pad temperature adjustment system 5 further includes a pad temperature measuring device 39 that measures the surface temperature of the polishing pad 3 (hereinafter sometimes referred to as pad surface temperature), and the control unit 40 controls the pad temperature measuring device 39 to measure the surface temperature of the polishing pad 3 (hereinafter sometimes referred to as pad surface temperature). The first flow control valve 42 and the second flow control valve 56 are operated based on the measured pad surface temperature. The first on-off valve 41 and the second on-off valve 55 are normally open.

Pad temperature measuring device 39 is arranged above the surface of polishing pad 3 and is configured to measure the surface temperature of polishing pad 3 in a non-contact manner. The pad temperature measuring device 39 is connected to a control section 40 and further connected to a temperature display 45 via the control section 40. The pad temperature measuring device 39 may be an infrared radiation thermometer or a thermocouple thermometer that measures the surface temperature of the polishing pad 3, and measures the surface temperature of the polishing pad 3 to obtain the temperature distribution of the polishing pad 3. It may be a thermograph or a thermopile. In this embodiment, the pad temperature measuring device 39 is at least one temperature measuring device among an infrared radiation thermometer, a thermocouple thermometer, a thermograph, and a thermopile. If liquid (such as slurry) splashed by polishing the wafer W adheres to the pad temperature measuring device 39, the pad temperature measuring device 39 may not be able to accurately measure the surface temperature of the polishing pad 3. Therefore, pad temperature measuring device 39 is placed at a sufficiently high position from the surface of polishing pad 3.

FIG. 2 is a top view of the temperature measurement area of the pad temperature measurement device 39, and FIG. 3 is a side view of the temperature measurement area of the pad temperature measurement device 39. When the pad temperature measuring device 39 is a thermocouple thermometer, a thermograph, or a thermopile, the pad temperature measuring device 39 measures the surface of the polishing pad 3 in a region including the center CL of the polishing pad 3 and the outer circumference 3a of the polishing pad 3. It is configured to measure temperature (see dotted lines in Figures 2 and 3). When the pad temperature measuring device 39 is an infrared radiation thermometer, the pad temperature measuring device 39 measures the temperature of the polishing pad 3 in a part of the area between the center CL of the polishing pad 3 and the outer circumference 3a of the polishing pad 3. It is configured to measure the surface temperature (see the dashed line in FIGS. 2 and 3). When the pad temperature measuring device 39 is an infrared radiation thermometer, the polishing unit preferably has a position adjustment mechanism that can change the position of the pad temperature measuring device 39 in the radial direction of the polishing pad 3.

The pad temperature measuring device 39 measures the pad surface temperature in a non-contact manner and sends the measured value of the pad surface temperature to the control unit 40 . The pad temperature measuring device 39 measures the pad surface temperature at predetermined intervals (for example, every 100 ms). The control unit 40 controls the first flow control valve 42 and the second flow control valve based on the measured pad surface temperature so that the pad surface temperature reaches a preset target temperature and then maintains the pad surface temperature at the target temperature. Operate the flow control valve 56. The first flow control valve 42 and the second flow control valve 56 operate according to control signals from the control unit 40 and adjust the flow rate of the heating liquid and the flow rate of the cooling liquid supplied to the pad contact member 11. Heat exchange occurs between the heating liquid and cooling liquid flowing through the pad contact member 11 and the polishing pad 3, thereby changing the pad surface temperature.

Through such feedback control, the pad surface temperature reaches a predetermined target temperature, and thereafter is maintained at the target temperature. Feedback control is, for example, PID control. In this embodiment, the control unit 40 performs PID control on the operation amounts of the first flow control valve 42 and the second flow control valve 56 based on the difference between the surface temperature of the polishing pad 3 and a predetermined target temperature. In order to execute this PID control, PID parameters (that is, proportional gain P, integral gain I, and differential gain D) are input in advance to the control unit 40.

In other words, the operation amount of the first flow control valve 42 and the operation amount of the second flow control valve 56 are valve opening degrees. The amount of operation of the first flow rate control valve 42 is proportional to the flow rate of the heating liquid, and the amount of operation of the second flow rate control valve 56 is proportional to the flow rate of the cooling liquid. Preferably, the amount of operation of the first flow control valve 42 is directly proportional to the flow rate of the heating liquid, and the amount of operation of the second flow control valve 56 is directly proportional to the flow rate of the cooling liquid.

As the control unit 40, any control device can be used. For example, a dedicated computer or a general-purpose computer (eg, a personal computer) can be used as the control unit 40. In one embodiment, the control unit 40 may be a PLC (Programmable Logic Controller) or a server. Furthermore, the control unit 40 may include an FPGA (Field-Programmable Gate Array). The target temperature of the polishing pad 3 is determined depending on the type of wafer W or the polishing process, and the determined target temperature is input into the control unit 40 in advance.

Pad contact member 11 contacts the surface (ie, polishing surface) of polishing pad 3 in order to cause the pad surface temperature to reach a predetermined target temperature and then maintain it. In this specification, the mode in which the pad contact member 11 contacts the surface of the polishing pad 3 includes not only a mode in which the pad contact member 11 directly contacts the surface of the polishing pad 3, but also a mode in which the pad contact member 11 and the polishing pad 3 are in contact with each other. It also includes a mode in which the pad contact member 11 contacts the surface of the polishing pad 3 with a polishing liquid (slurry) present between the surface and the surface. In either embodiment, heat exchange is performed between the heating liquid and cooling liquid flowing through the pad contact member 11 and the polishing pad 3, thereby controlling the pad surface temperature.

As shown in FIG. 1, the pad temperature adjustment system 5 further includes a vertical movement mechanism (vertical movement mechanism) 71 that moves the pad contact member 11 perpendicularly to the surface of the polishing pad 3. The pad contact member 11 is held by a vertical movement mechanism 71. This vertical movement mechanism 71 is configured to be able to move the pad contact member 11 in the vertical direction with respect to the surface of the polishing pad 3. With such a configuration, the pad contact member 11 can directly contact the surface of the polishing pad 3. The vertical movement mechanism 71 is composed of a combination of a servo motor and a ball screw mechanism, or an air cylinder. According to such a vertical movement mechanism 71, the pad contact member 11 can be pressed against the surface of the polishing pad 3 with a predetermined pressing force. For example, the pressing force with which the pad contact member 11 is pressed against the surface of the polishing pad 3 is adjusted to a predetermined pressing force by controlling a pulse signal sent to a servo motor or the pressure of gas supplied to an air cylinder. .

The pad contact member 11 is configured to move perpendicularly to the surface of the polishing pad 3 to maintain a constant temperature of a region on the surface of the polishing pad 3 (a position in the radial direction of the polishing pad 3). There is. For example, the pad contact member 11 is arranged in a direction perpendicular to the surface of the polishing pad 3 so as to maintain a temperature of 55 degrees at a radial position of the polishing pad 3 at a distance of 100 mm from the center CL of the polishing pad 3. Move to. The user can arbitrarily determine (change) the surface temperature of the polishing pad 3 and the radial position of the polishing pad 3 controlled by the pad contact member 11. For example, the user may change the distance from the center CL of the polishing pad 3 from 100 mm to 200 mm as the radial position of the polishing pad 3, and change the surface temperature of the polishing pad 3 from 55 degrees to 70 degrees. You may. As a result, the pad contact member 11 is arranged above and below the surface of the polishing pad 3 so as to maintain the temperature at 70 degrees at a radial position of the polishing pad 3 at a distance of 200 mm from the center CL of the polishing pad 3. move in the direction.

Next, an example of the pad contact member 11 will be described with reference to FIG. 4. FIG. 4 is a horizontal cross-sectional view showing one embodiment of the pad contact member 11. As shown in FIG. As shown in FIG. 4, the pad contact member 11 has a heating channel 61 and a cooling channel 62 formed therein. The heating channel 61 and the cooling channel 62 extend adjacent to each other (alongside each other) and extend in a spiral shape. Further, the heating channel 61 and the cooling channel 62 have point-symmetrical shapes and have the same length.

As shown in FIG. 4, each of the heating channel 61 and the cooling channel 62 basically consists of a plurality of circular arc channels 64 having a constant curvature and a plurality of inclined channels 65 connecting these arc channels 64. It is configured. Two adjacent circular arc channels 64 are connected by each inclined channel 65 . According to such a configuration, the outermost periphery of each of the heating channel 61 and the cooling channel 62 can be arranged at the outermost periphery of the pad contact member 11. In other words, almost the entire pad contact surface composed of the lower surface of the pad contact member 11 is located below the heating channel 61 and the cooling channel 62, and the heating liquid and the cooling liquid quickly heat the surface of the polishing pad 3. and can be cooled.

The heating liquid supply pipe 32 is connected to the inlet 61a of the heating channel 61, and the heating liquid return pipe 33 is connected to the outlet 61b of the heating channel 61. The coolant supply pipe 51 is connected to the inlet 62a of the cooling channel 62, and the coolant discharge pipe 52 is connected to the outlet 62b of the cooling channel 62. The inlets 61a and 62a of the heating channel 61 and the cooling channel 62 are located at the periphery of the pad contact member 11, and the outlets 61b and 62b of the heating channel 61 and the cooling channel 62 are located at the periphery of the pad contact member 11. Centrally located. Therefore, the heating liquid and the cooling liquid flow spirally from the periphery of the pad contacting member 11 toward the center. The heating channel 61 and the cooling channel 62 are completely separated, and the heating liquid and the cooling liquid are not mixed within the pad contact member 11.

As shown in FIG. 4, since the heating channel 61 and the cooling channel 62 are adjacent to each other, the heating channel 61 and the cooling channel 62 are arranged not only in the radial direction of the polishing pad 3 but also in the radial direction of the polishing pad 3. They are lined up along the circumference. Therefore, while the polishing table 2 and polishing pad 3 are rotating, the polishing pad 3 in contact with the pad contact member 11 exchanges heat with both the heating liquid and the cooling liquid.

As shown in FIG. 1, the pad temperature adjustment device 5 may include a heating liquid pump 47 attached to the heating liquid supply pipe 32. The heated liquid pump 47 is connected to the control unit 40, and the control unit 40 is configured to be able to control the operation of the heated liquid pump 47. For example, the control unit 40 can adjust the pressure of the heating liquid supplied to the pad contacting member 11 by controlling the rotational speed of the heating liquid pump 47. Furthermore, when the pad temperature adjustment device 5 includes the heating liquid pump 47, the first flow rate adjustment valve 42 attached to the heating liquid supply pipe 32 may be omitted. By controlling the rotational speed of the heating liquid pump 47 by the control unit 40, the flow rate of the heating liquid supplied to the pad contact member 11 can be adjusted. In this case, the control unit 40 stores in advance a relational expression or a data table indicating the relationship between the rotational speed of the heating liquid pump 47 and the flow rate of the heating liquid.

Although not shown, the pad temperature adjustment device 5 may include a coolant pump attached to the coolant supply pipe 51. In this case, the coolant pump is connected to the control unit 40, and the control unit 40 is configured to be able to control the operation of the coolant pump. For example, the control unit 40 can adjust the pressure of the coolant supplied to the pad contact member 11 by controlling the rotational speed of the coolant pump. Furthermore, when the pad temperature adjustment device 5 has a coolant pump, the second flow rate adjustment valve 56 attached to the coolant supply pipe 51 may be omitted. By controlling the rotational speed of the coolant pump by the control unit 40, the flow rate of the coolant supplied to the pad contact member 11 can be adjusted. In this case, the control unit 40 stores in advance a relational expression or a data table indicating the relationship between the rotational speed of the coolant pump and the flow rate of the coolant.

The polishing apparatus includes at least one polishing unit described above. When the polishing apparatus includes a plurality of polishing units, the control section 40 functions as a common control section among the plurality of polishing units. More specifically, the control unit 40 is configured to control the operation of each polishing unit and the operation of the pad temperature adjustment device 5 included in each polishing unit. Furthermore, a plurality of polishing apparatuses are often installed in semiconductor device factories and the like. That is, in such a factory, a plurality of polishing apparatuses each having a plurality of polishing units are installed.

FIG. 5 is a schematic diagram showing one embodiment of the control unit 40. As shown in FIG. The control unit 40 shown in FIG. 5 is a dedicated or general-purpose computer (for example, a personal computer). In one embodiment, the control unit 40 may be a PLC (Programmable Logic Controller) or a server. Furthermore, the control unit 40 may include an FPGA (Field-Programmable Gate Array). The control unit 40 shown in FIG. 5 includes a storage device 110 in which programs, data, etc. are stored, and a CPU (central processing unit) or GPU (graphic processing unit) that performs calculations according to the programs stored in the storage device 110. A processing device 120, an input device 130 for inputting data, programs, and various information into the storage device 110, and an output device 140 for outputting processing results and processed data are connected to a network such as the Internet. A communication device 150 is provided.

The storage device 110 includes a main storage device 111 that can be accessed by the processing device 120, and an auxiliary storage device 112 that stores data and programs. The main storage device 111 is, for example, a random access memory (RAM), and the auxiliary storage device 112 is a storage device such as a hard disk drive (HDD) or a solid state drive (SSD).

The input device 130 includes a keyboard and a mouse, and further includes a recording medium reading device 132 for reading data from a recording medium, and a recording medium port 134 to which the recording medium is connected. The recording medium is a computer-readable recording medium that is a non-transitory tangible object, such as an optical disk (e.g., CD-ROM, DVD-ROM) or a semiconductor memory (e.g., USB flash drive, memory card). be. Examples of the recording medium reading device 132 include optical drives such as CD-ROM drives and DVD-ROM drives, and card readers. An example of the recording medium port 134 is a USB port. The program and/or data stored on the recording medium is introduced into the computer via the input device 130 and stored in the auxiliary storage device 112 of the storage device 110. The output device 140 includes a display device 141 and a printing device 142.

As described with reference to FIG. 19, variations in temperature behavior occur due to machine differences between polishing units. In this embodiment, in order to suppress such variations in temperature behavior, the processing device 120 of the control unit 40 controls the polishing pad 3 to a predetermined temperature based on the measured value of the pad temperature measuring device 39 and the measurement time point. A temperature behavior curve is created until the target temperature is reached, and at least one of the temperature behavior parameters is changed so that the temperature behavior curve when polishing the next wafer W is maintained within a predetermined tolerance range. It is composed of Note that a temperature behavior curve is created every time a wafer W is polished, and the processing device 120 of the control unit 40 stores a plurality of temperature behavior curves in the storage device 110. The temperature behavior curve and the predetermined tolerance range will be described below with reference to FIGS. 6 and 7.

FIG. 6 is a graph showing an example of the temperature behavior curve, and FIG. 7 is a graph showing another example of the temperature behavior curve. FIG. 6 is a graph showing an example in which the temperature behavior curve falls within the tolerance range described later, and FIG. 7 is a graph showing an example in which the temperature behavior curve falls outside the tolerance range. 6 and 7, the vertical axis represents the surface temperature of the polishing pad 3 (pad surface temperature), and the horizontal axis represents time.

Note that, as shown in FIG. 7, while the pad temperature adjustment device 5 is converging the pad surface temperature to the target temperature, the pad surface temperature may temporarily exceed the target temperature. This phenomenon is called "overshoot." In this specification, "the surface temperature of the polishing pad 3 (pad surface temperature) has reached the target temperature" is a concept that includes this overshoot. More specifically, when an overshoot occurs, the point in time when the pad surface temperature once reaches the target temperature is not judged as the point in time when the pad surface temperature reaches the target temperature, and after overshoot, the pad surface temperature reaches the target temperature. The point at which the pad surface temperature converges is determined to be the point at which the pad surface temperature reaches the target temperature.

As described above, first, the control unit 40 gives a command to the vertical movement mechanism 71 to bring the pad contact member 11 into contact with the surface of the polishing pad 3 to start adjusting the pad surface temperature. As a result, the surface temperature of the polishing pad 3 begins to rise. Time Ta in FIGS. 6 and 7 is the time when the pad contact member 11 contacts the surface of the polishing pad 3, that is, the time when the pad contact member 11 starts adjusting the pad surface temperature. Next, the control unit 40 gives a command to the polishing head 1 to press the surface of the wafer W held by the polishing head 1 against the polishing pad 3 to start polishing the wafer (substrate) W. Time Tb in FIGS. 6 and 7 is the time when the wafer W contacts the surface (polishing surface) of the polishing pad 3, that is, the time when polishing of the wafer W is started. As shown in FIGS. 6 and 7, when the wafer W at room temperature comes into contact with the surface of the polishing pad 3 after the adjustment of the pad surface temperature has started, the pad surface temperature is temporarily lowered.

The control unit 40 controls at least the flow rates of the heating liquid and the cooling liquid flowing through the heating channel 61 and the cooling channel 62 of the pad contact member 11, respectively, so that the pad surface temperature reaches a predetermined target temperature. Time Tc in FIGS. 6 and 7 is the time when the pad surface temperature reaches a predetermined target temperature. Thereafter, the control unit 40 controls at least the flow rates of the heating liquid and the cooling liquid flowing through the heating channel 61 and the cooling channel 62 of the pad contacting member 11, respectively, so that the pad surface temperature maintains the target temperature.

The temperature behavior curve R is a curve showing the change in pad surface temperature over time (i.e., temperature behavior) from the time Ta when the pad contact member 11 starts adjusting the pad surface temperature to the time Tc when the target temperature is reached. be. The control unit 40 creates a temperature behavior curve R based on the measured value of the pad surface temperature sent from the pad temperature measuring device 39 and the measurement time point, and stores this temperature behavior curve R.

The pad temperature measuring device 39 measures the pad surface temperature at predetermined intervals (for example, every 100 ms), and the measured values of the pad surface temperature are sequentially sent to the control unit 40. Every time the pad temperature measuring device 39 sends the measured value of the pad surface temperature and the measurement time point, the control unit 40 sequentially plots the measured value on a graph as shown in FIG. At the time point Tc when the surface temperature reaches the target temperature, the plurality of plot points are connected with a smooth curve. Thereby, the control unit 40 can acquire the temperature behavior curve R.

In FIGS. 6 and 7, the tolerance range is an area sandwiched between an upper limit curve Ru that defines the upper limit of the tolerance range and a lower limit curve Rl that defines the lower limit of the tolerance range. This tolerance range is a range in which variations in the temperature behavior curve R can be tolerated. The control unit 40 determines that the temperature behavior curve R is outside the tolerance range when at least a portion of the temperature behavior curve R is outside the tolerance range (see FIG. 7). The tolerance range can be determined, for example, by experiments conducted by the manufacturer of the polishing device. This tolerance range is commonly used by each polishing unit.

Alternatively, the tolerance range may be created from a plurality of temperature behavior curves obtained when a predetermined number (for example, 100) of wafers W are polished. An example of a method for creating a tolerance range from a plurality of temperature behavior curves will be described below.

As shown in FIGS. 6 and 7, the pad temperature measuring device 39 measures the pad surface temperature at predetermined time intervals (that is, at each measurement time point T1, T2, T3, T4,...Tn), Measured values of the pad surface temperature are sent to the control unit 40 at predetermined time intervals. The first measurement time point T1 corresponds to the above-mentioned time point Ta, and the measurement time point Tn corresponds to the above-mentioned time point Tc. In this embodiment, the time Tb at which polishing of the wafer W is started is between the measurement time T1 (=Ta) and the measurement time Tn (=Tc). Note that in FIGS. 6 and 7, the intervals between adjacent measurement points are enlarged to avoid cluttering the drawings. The actual spacing between adjacent measurement time points is smaller than the spacing shown in FIGS. 6 and 7, for example 100 ms.

The control unit 40 stores a plurality of pad surface temperature measurements from time Ta (=T1) to time Tc (=Tn) in association with measurement time points T1, T2, T3, T4, . . . Tn. Furthermore, the control unit 40 executes this operation every time a predetermined number of wafers W are polished. As a result, the control unit 40 accumulates a plurality of data sets of pad surface temperature measurement values associated with the same measurement time points T1, T2, . . . Tn. FIG. 8 is a schematic diagram showing an example of multiple data sets. Note that when creating the multiple data sets shown in Figure 8, combinations of pad surface temperature measurements and their measurement points that constitute the temperature behavior curve (see Figure 7) outside the allowable range are excluded. is preferred. That is, it is preferable that the plurality of data sets shown in FIG. 8 are composed only of combinations of pad surface temperature measurement values that constitute a temperature behavior curve (see FIG. 6) that falls within the allowable range and the measurement time points.

Next, the control unit 40 creates a normal distribution for each data set. FIG. 9 is a diagram showing a normal distribution created based on the data set D2 at the measurement time point T2 shown in FIG. The control unit 40 extracts one plot point of the upper limit curve Ru based on a normal distribution as shown in FIG. For example, the value of the pad surface temperature corresponding to 2σ of the normal distribution shown in FIG. 9 is extracted as one plot point Pu2 (see FIG. 6) constituting the upper limit curve Ru. The control unit 40 repeats this operation with the data sets corresponding to all measurement time points T1, T2, . . . Tn, and obtains a plurality of plot points Pu1, Pu2, . By connecting these plot points Pu1, Pu2, . . . Pun with a smooth curve, an upper limit curve Ru can be obtained. Note that in FIG. 6, only the plot point Pu2 is drawn as a representative of the plot points forming the upper limit curve Ru.

Similarly, the value of the pad surface temperature corresponding to −2σ of the normal distribution shown in FIG. 9 is extracted as one plot point Pl2 (see FIG. 6) constituting the lower limit curve Rl. The control unit 40 repeats this operation with the data sets corresponding to all measurement time points T1, T2, . . . Tn, and obtains a plurality of plot points Pl1, Pl2, . By connecting these plot points with a smooth curve, a lower limit curve Rl is obtained. Note that in FIG. 6, only the plot point Pl2 is drawn as a representative of the plot points forming the lower limit curve Rl. The tolerance range thus obtained corresponds to a range of ±2σ of the average of a plurality of temperature behavior curves obtained when a predetermined number of wafers W are polished.

Another example of a method for creating a tolerance range from a plurality of temperature behavior curves obtained when a predetermined number (for example, 100 wafers) of wafers W is polished is to create a reference temperature behavior curve from among a plurality of temperature behavior curves. This is the method of choice. The reference temperature curve is, for example, the temperature behavior curve located most centrally when a plurality of temperature behavior curves are drawn on the same graph. The permissible range is defined by an upper limit curve Ru and a lower limit curve Rl that are vertically separated by a predetermined percentage from the reference temperature curve. More specifically, the upper limit curve Ru that defines the upper limit of the allowable range is obtained by multiplying the value of each pad surface temperature that constitutes the selected reference temperature behavior curve by a predetermined coefficient (+C%) that is a positive number. It is obtained by connecting the values obtained by using a smooth curve. The lower limit curve Rl that delimits the lower limit of the allowable range is a value obtained by multiplying the value of each pad surface temperature constituting the selected reference temperature behavior curve by a predetermined coefficient (-C%), which is a negative number. Obtained by connecting them with smooth curves. In this case, the allowable range is defined by an upper limit curve Ru and a lower limit curve Rl that are vertically separated by ±C% from the reference temperature curve. Generally, a value in the range of 5 to 20 is used as the value of C, which is the predetermined ratio.

As shown in FIG. 7, if the temperature behavior curve R exceeds a predetermined tolerance range, it may adversely affect the yield of the product (ie, semiconductor device). Therefore, it is preferable that the temperature behavior curve R always falls within a predetermined tolerance range. An example of a method for keeping the temperature behavior curve R within a predetermined tolerance range is that if the temperature behavior curve R created when polishing a wafer W exceeds a predetermined tolerance range, the temperature behavior curve R created when polishing a wafer W is A method of changing at least one of the temperature behavior parameters used. The temperature behavior parameters to be changed and the values of their changes can be determined, for example, by experiments carried out by the manufacturer of the polishing device.

However, in the present embodiment, in order to maintain the temperature behavior curve R within a predetermined tolerance range, the control unit 40 uses artificial intelligence (AI) to determine the temperature behavior parameter to be changed and its value. Predict change values.

Note that in this specification, the temperature behavior parameter is a general term for the parameters that can change the temperature behavior described above. Typical examples of temperature behavior parameters include the following:
1) Flow rate of the heating liquid 2) Flow rate of the cooling liquid 3) PID parameters for determining the operating amounts of the first flow control valve 42 and the second flow control valve 56 4) Rotational speed of the heating liquid pump 5) Cooling liquid pump 6) Supply pressure of heating liquid 7) Supply pressure of cooling liquid 8) Temperature of heating liquid 9) Temperature of cooling liquid 10) Set temperature of heating source 48 11) Temperature of polishing liquid 12) Flow rate of polishing liquid 13 ) The dropping position of the polishing liquid 14) The rotation speed of the polishing head 1 15) The rotation speed of the polishing table 2 16) The polishing load of the wafer W on the polishing pad 3 17) The pressing load of the pad contact member 11 on the polishing pad 3 18) Dressing conditions 19) Ambient temperature inside the polishing unit 20) Position of pad temperature measuring device 39 in the radial direction of polishing pad 3 21) Time

The time described in item 21) above means the time at which the control unit 40 acquires the temperature behavior parameters described in items 1) to 20) above, and the temperature behavior parameters described in items 1) to 20) above. is a value related to Here, in order to adjust the temperature of the polishing pad 3 by changing the temperature behavior parameter, it is necessary to monitor the temperature behavior parameter, which changes from moment to moment. This is a very important factor in performing temperature control. Therefore, in this specification, the temperature behavior parameter is defined to include "time."

By changing at least one of the temperature behavior parameters shown in 1) to 20) above, the temperature behavior (that is, the temperature behavior curve) can be changed. For example, when the flow rate and/or temperature of the heating liquid is increased or the flow rate and/or temperature of the cooling liquid is decreased, the surface temperature of the polishing pad 3 increases more quickly. When the PID parameter is changed, the operating amounts of the first flow control valve 42 and the second flow control valve 56 are changed, and as a result, the flow rate of the heating liquid and the flow rate of the cooling liquid are changed. In particular, changing the proportional gain P of the PID parameter has a large effect on changes in temperature behavior.

When a heated liquid pump 47 (see FIG. 1) is attached to the heated liquid supply pipe 32, the flow rate and supply pressure of the heated liquid can be changed by changing the rotational speed of the heated liquid pump 47. Similarly, if a coolant pump (not shown) is attached to the coolant supply pipe 51, the flow rate and supply pressure of the coolant can be changed by changing the rotational speed of the coolant pump. . Therefore, by changing the rotational speed of the heating liquid pump 47 and/or the rotational speed of the cooling liquid pump, the temperature behavior can be changed.

Further, the control unit 40 may change the set temperature of a heat source (for example, a heater) 48 arranged in the heated liquid supply tank 31. Thereby, the temperature of the heating liquid supplied to the pad contact member 11 is changed, so that the temperature behavior can be changed.

The polishing liquid (slurry) supplied onto the polishing pad 3 lowers the pad surface temperature raised by the pad contact member 11. Therefore, changing the temperature and/or flow rate of the polishing liquid changes the amount by which the pad surface temperature decreases, resulting in a change in temperature behavior. For the same reason, the temperature behavior can be changed by changing the dropping position of the polishing liquid.

When a wafer W held by a rotating polishing head 1 is pressed against a polishing pad 3 supported by a rotating polishing table 2, frictional heat is generated between the wafer W and the polishing pad 3. Surface temperature increases. The amount of this frictional heat changes depending on the rotational speed of the polishing head 1, the rotational speed of the polishing table 2, and the polishing load of the wafer W on the polishing pad 3. Therefore, the temperature behavior can be changed by changing the rotational speed of the polishing head 1, the rotational speed of the polishing table 2, and/or the polishing load.

Further, when the pad contact member 11 is pressed against the surface of the rotating polishing pad 3 by the vertical movement mechanism 11, frictional heat is generated between the pad contact member 11 and the polishing pad 3, and this frictional heat also causes the pad surface to Temperature rises. The amount of this frictional heat changes depending on the pressing load of the pad contact member 11 against the polishing pad 3. Therefore, by changing the pressing load with which the vertical movement mechanism 11 presses the pad contact member 11 against the surface of the polishing pad 3, the temperature behavior can be changed.

Dresser 20 dresses the surface of polishing pad 3 according to preset dressing conditions. The dressing conditions include, for example, the rotational speed of the dresser 20, the pressing load of the dresser 20 against the polishing pad 3, and the like. When the dressing conditions are changed, the roughness of the surface of the polishing pad 3 after being dressed by the dresser 20 changes. As a result, the amount of frictional heat generated between the wafer W being polished and the polishing pad 3 and the amount of frictional heat generated between the pad contact member 11 and the polishing pad 3 change, so the dressing conditions are changed. By changing it, the temperature behavior can be changed.

Since the ambient temperature within the polishing unit also influences the temperature behavior, the ambient temperature is also one of the temperature behavior parameters. For example, the slope of the temperature behavior curve in a polishing unit where the ambient temperature is 20°C is smaller than the slope of the temperature behavior curve in a polishing unit where the ambient temperature is 25°C.

As described above, when the pad temperature measuring device 39 is an infrared radiation thermometer, the position of the pad temperature measuring device 39 in the radial direction of the polishing pad 3 can be adjusted. When the position of the pad temperature measuring device 39 in the radial direction of the polishing pad 3 is changed, the measurement area of the pad temperature measuring device 39 on the polishing pad 3 changes. Therefore, the measured value of the pad temperature measuring device 39 changes before and after changing its position. Since the control unit 40 operates the first flow control valve 42 and the second flow control valve 56 based on the measured pad surface temperature, the control unit 40 controls the position of the pad temperature measuring device 39 in the radial direction of the polishing pad 3. By changing it, the temperature behavior can be changed.

FIG. 10 is a schematic diagram showing an example of the configuration of artificial intelligence installed in the control unit 40 shown in FIG. 5. As shown in FIG. In the artificial intelligence shown in FIG. 10, machine learning using a neural network or quantum computing is performed to construct a learned model for maintaining the temperature behavior curve R within a predetermined tolerance range. Generally, trained models are constructed to output prediction or diagnostic results for input data. For example, when at least one temperature behavior parameter is input into a trained model, the trained model determines the at least one temperature behavior parameter that should be changed in order to maintain the temperature behavior curve R within a predetermined tolerance range, and the value of the change. Predict and output.

FIG. 11 is a flowchart for explaining a method for constructing the learned model shown in FIG. 10. When constructing a trained model, first, data is collected (see step 1) and a collection of raw data is created (see step 2). The data collection performed in step 1 is extensive. For example, the collected data includes the measured values of various sensors placed in the polishing device, the materials of each component device placed in the polishing device, parameters input into the polishing device by the operator, etc. If the imaging device is placed in a polishing device, the image data acquired by the imaging device is also included. Furthermore, the collected data may include processed data obtained by processing sensor measurement values, image data, etc., and may also include a database storing various data of the polishing device and search data. A collection of these raw data is sometimes referred to as "big data."

Next, a training dataset necessary for constructing a trained model is created from the collection of raw data (see step 3). The learning data set is a data set necessary for constructing a learned model for maintaining the temperature behavior curve R within a predetermined tolerance range, and is also referred to as "teacher data." This learning data set is normal data, abnormal data, reference data, or mixed data. Mixed data means a data set in which normal data and abnormal data are mixed at a predetermined ratio. For example, the mixed data may be composed of 80% normal data and 20% abnormal data, or may be composed of 90% normal data and 10% abnormal data. Generally, mixed data includes more normal data than abnormal data, and the ratio of normal data to frequently used abnormal data is 8:1 or 9:1. In this way, the proportion of normal data is selected from the range of 70 to 100%, and the proportion of abnormal data is selected from the range of 0 to 30%, so that the sum of the proportion of normal data and the proportion of abnormal data is 100%. It is preferable to use mixed data created by selecting from a range.

The training data set includes, for example, at least one temperature behavior parameter. As described above, the temperature behavior parameters include the flow rate of the heating liquid, the flow rate of the cooling liquid, the PID parameter for determining the operation amount of the first flow control valve 42 and the second flow control valve 56, and the rotation speed of the heating liquid pump. , rotational speed of the cooling liquid pump, supply pressure of the heating liquid, supply pressure of the cooling liquid, temperature of the heating liquid, temperature of the cooling liquid, set temperature of the heating source 48, temperature of the polishing liquid, flow rate of the polishing liquid, The dropping position, the rotation speed of the polishing head 1, the rotation speed of the polishing table 2, the polishing load of the wafer W on the polishing pad 3, the pressing load of the pad contact member 11 on the polishing pad 3, the dressing conditions, the ambient temperature in the polishing unit, etc. , the position of the infrared radiation thermometer in the radial direction of the polishing pad 3, and the time. The learning data set may be stored in advance in the storage device 110 of the control unit 40 or may be provided to the control unit 40 via the communication device 150. Furthermore, the learning data set may include a plurality of temperature behavior curves stored in the storage device 110, and may further include the measured values of the pad temperature measuring device 39 used in creating each temperature behavior curve, It may also include a combination of the measurement time points.

Next, machine learning using a neural network or quantum computing is performed (see step 4) to construct a trained model for maintaining the temperature behavior curve R within a predetermined tolerance range (see step 5). Machine learning for building trained models includes learning using normal data as a training dataset, learning using abnormal data as a training dataset, learning using reference data as a training dataset, and learning using mixed data. Includes training used as a training dataset. Machine learning for constructing a trained model also includes learning different from the above learning. For example, a trained model may be constructed by performing learning without using a training dataset (that is, learning without "teacher data") or reinforcement learning.

As machine learning using a neural network or quantum computing, a deep learning method is suitable. The deep learning method is a learning method based on a neural network with multiple hidden layers (also called intermediate layers). In this specification, machine learning using a neural network including an input layer, two or more hidden layers, and an output layer is referred to as deep learning.

FIG. 12 is a schematic diagram showing an example of the structure of a neural network. The learned model is constructed by a deep learning method using a neural network as shown in FIG.

The neural network shown in FIG. 12 has an input layer 301, a plurality of (in the illustrated example, five) hidden layers 302, and an output layer 303. When normal data is used as a learning data set, the control unit 40 uses the normal data to adjust weight parameters configuring a neural network in order to construct a learned model. More specifically, the control unit 40 maintains the temperature behavior curve within a predetermined tolerance range when data including at least one temperature behavior parameter created for learning is input to the input layer 301 of the neural network. The weight parameters of the neural network are adjusted so that the temperature behavior parameter to be changed in order to change the temperature behavior parameter and data corresponding to the changed value are output from the output layer 303 of the neural network. For example, when data including at least one temperature behavior parameter is input to the input layer 301 of the neural network, the PID parameter that should be changed to maintain the temperature behavior curve within a predetermined tolerance, and the The weight parameters of the neural network are adjusted so that the modified values are output from the output layer 303 of the neural network.

The changed values of the PID parameters output from the output layer 303 are compared with a normal range, which is a collection of PID parameters when the temperature behavior curve was within a predetermined tolerance range. If the changed value of the PID parameter output from the output layer 303 deviates from the normal range, when the data containing at least one temperature behavior parameter created for learning is input again to the input layer 301 of the neural network, the output The weight parameters are automatically adjusted so that the changed value of the PID parameter output from the layer 303 falls within the normal range. In this manner, the trained model is constructed by repeatedly inputting at least one temperature behavior parameter into the input layer, comparing the output value from the output layer with a normal range, and adjusting the weight parameters.

Furthermore, it is preferable that the control unit 40 input verification data to the neural network and verify whether the data output from the neural network corresponds to data included in the normal range. The verification data may be created by extracting a part of the learning data set created in step 3 in advance. Alternatively, the entire learning data set created in step 3 may be used as verification data. In this case, the entire training data set created in step 3 is input again to the trained model, and weight parameter adjustment is repeated using the same training data set.

In one embodiment, the neural network may have an input layer 301' different from input layer 301. For example, data different from the temperature behavior parameter may be input to the input layer 301', or a temperature behavior parameter different from the temperature behavior parameter input to the input layer 301 may be input. Examples of data input to the input layer 301' include a plurality of temperature behavior curves stored in the storage device 110 and/or measured values of the pad temperature measuring device 39 used in creating each temperature behavior curve. , is a combination of the measurement points. Another example of data input to the input layer 301' is the usage time of consumables used in the polishing unit. Examples of consumables include the polishing pad 3, a retainer ring (not shown) that prevents the wafer W from flying out from the polishing head 1 during polishing, and a retainer ring (not shown) that is placed below the polishing head 1 and that protects the wafer W. It includes a membrane (not shown) for pressing the polishing pad 3 with a predetermined pressing force.

Still other examples of data input to the input layer 301' include temperature behavior parameters (for example, temperature of the heating liquid, rotational speed of the polishing head 1, polishing (rotational speed of table 2, dressing conditions, polishing load of polishing head 1, flow rate of polishing liquid, etc.). Still other examples of data input to the input layer 301' include temperature behavior parameters representing changes over time in the environment that affect temperature behavior (for example, the pressing load of the pad contacting member 11, the temperature of the polishing liquid, the polishing (atmospheric temperature within the unit, heating liquid supply pressure, cooling liquid supply pressure, etc.).

Neural networks can be used to generate data different from temperature behavior parameters, temperature behavior parameters that represent changes over time in state variables that influence temperature behavior, and/or temperature behavior parameters that represent changes over time in the environment that influence temperature behavior. By inputting , the output layer 303 of the neural network can output a more accurate predicted value (ie, the temperature behavior parameter to be changed and its changed value). For example, neural networks can provide more accurate temperature behavior parameters to be changed, taking into account changes over time in state quantities that influence temperature behavior, and/or changes over time in the environment that influences temperature behavior. The changed value of can be output from the output layer 303.

Furthermore, the neural network may have an output layer 303' different from the output layer 303. The output layer 303' outputs, for example, data different from the temperature behavior parameter to be changed and its change value. An example of the data output from the output layer 303' is an optimal temperature behavior curve representing the time course of the optimal pad surface temperature adjusted when polishing the wafer W, and/or a pad surface constituting the optimal temperature behavior curve. It is a combination of temperature and the time point at which it is measured. Other examples of data output from the output layer 303' are other temperature behavior parameters different from PID parameters. For example, the output layer 303' may output other temperature behavior parameters different from the PID parameters and their modified values to be modified in order to maintain the temperature behavior curve within a predetermined tolerance. Alternatively, when the PID parameter output from the output layer 303 is changed, the output layer 303' may output a predicted value of a change in a temperature behavior parameter other than the PID parameter.

The trained model constructed in this way is stored in the storage device 110 (see FIG. 5). The control unit 40 operates according to a program electrically stored in the storage device 110. That is, the processing device 120 of the control unit 40 inputs data including at least one temperature behavior parameter to the input layer 301 of the trained model, and changes the temperature behavior curve R to maintain it within a predetermined tolerance range. An operation is performed to output from the output layer 303 at least one temperature behavior parameter to be measured and its changed value.

As described above, when the PID parameter is changed, the operating amounts of the first flow control valve 42 and the second flow control valve 56 are changed, and as a result, the flow rate of the heating liquid and the flow rate of the cooling liquid are changed. Changes in the PID parameters (in particular, changes in the proportional gain P) have a direct and significant effect on changes in the temperature behavior (ie, the temperature behavior curve). Below, an example will be described in which a PID parameter and its changed value are outputted from the output layer 303 of the learned model as the temperature behavior parameter to be changed.

As a first example, the control unit 40 inputs the PID parameter, the flow rate of the heating liquid, and the flow rate of the cooling liquid to the input layer 301. The trained model determines, when the PID parameters, the flow rate of the heating liquid, and the flow rate of the cooling liquid are input into the input layer 301, the PID parameters to be changed in order to maintain the temperature behavior curve R within a predetermined tolerance range; The output layer 303 is configured to output the changed value from the output layer 303. The trained model may be constructed to output at least one of the PID parameters (eg, proportional gain P) and its modified value from the output layer 303.

In one embodiment, the control unit 40 inputs the time at which each data of the PID parameter, heating liquid flow rate, and cooling liquid flow rate (see item 21 above) inputted to the input layer 301 is input to the input layer 301'. You can also enter it. Alternatively, in addition to the PID parameter, the flow rate of the heating liquid, and the flow rate of the cooling liquid, the control unit 40 inputs into the input layer 301 the time at which each data of the PID parameter, the flow rate of the heating liquid, and the flow rate of the cooling liquid was acquired. You may.

The temperature behavior parameters described in items 1) to 20) above are acquired by various sensors arranged in the polishing unit. The times at which various sensors take measurements of temperature behavior parameters are different from each other. For example, the time at which a sensor that measures the flow rate of the heating liquid acquires the flow rate value of the heating liquid is different from the time at which the sensor that measures the flow rate of the cooling liquid acquires the flow rate value of the cooling liquid. Furthermore, the time from when the temperature behavior parameters acquired by each sensor are transmitted to the control section 40 until the control section 40 receives the temperature behavior parameters also differs from each other. For example, the time when the control unit 40 receives the flow rate value of the heating liquid transmitted from the sensor that measures the flow rate of the heating liquid is determined by the time when the control unit 40 receives the flow rate value of the heating liquid transmitted from the sensor that measures the flow rate of the cooling liquid. The received time is different. This is because the distance from each sensor to the control unit 40 is different, so the length of the cable extending from each sensor to the control unit 40 is different, and the equipment such as an amplifier placed in each sensor is different. It is.

In order for the output layer 303 of the neural network to more accurately output at least one temperature behavior parameter to be changed in order to maintain the temperature behavior curve R within a predetermined tolerance range and its change value, the control unit 40 It is preferable to match the times of each temperature behavior parameter obtained by. However, as described above, it is difficult to match the times at which the control unit 40 acquires each temperature behavior parameter. Therefore, time is additionally input to the input layer 301 (or input layer 301') of the neural network. The neural network into which time is input calculates each predicted value that matches the measurement time of multiple temperature behavior parameters from the temporal change of multiple temperature behavior parameters input other than time, and based on this predicted value. , outputs the PID parameter to be changed and its changed value from the output layer 303. Thereby, a more accurate output value can be obtained in consideration of the time difference between each temperature behavior parameter acquired by the control unit 40.

In one embodiment, the control unit 40 includes a combination of the measurement value of the pad temperature measurement device 39 and the measurement time point, and/or the temperature behavior created based on the measurement value of the pad temperature measurement device 39 and the measurement time point. The curve may be further input to the input layer 301' of the trained model. In this case as well, the learned model transmits the PID parameter (or at least one of the PID parameters) to be changed in order to maintain the temperature behavior curve R within a predetermined tolerance range and the change value from the output layer 303. Output.

The control unit 40 changes the PID parameters according to the output from the learned model, and polishes the next wafer W while adjusting the surface temperature of the polishing pad 3 according to the changed PID parameters. The combined data set of PID parameters, heating fluid flow rate, and cooling fluid flow rate is the combination of parameters that most influences the change in temperature behavior (i.e., temperature behavior curve). Therefore, by inputting the data set of the first example to the input layer 301, it can be expected that the trained model will output the most appropriate PID parameter and its modified value. As mentioned above, the trained model may further output from its output layer 303' an optimal temperature behavior curve and/or a combination of pad surface temperatures and measurement time points that constitute the optimal temperature behavior curve.

In one embodiment, the trained model includes other temperature behavior parameters different from the PID parameters and their modified values that are to be modified to maintain the temperature behavior curve within a predetermined tolerance range from the output layer 303'. may be output. In this case, the control unit 40 updates the PID parameter output from the output layer 303 to the changed value, and simultaneously updates other temperature behavior parameters output from the output layer 303' to the changed value. In this case, it is possible to more effectively prevent the temperature behavior curve from deviating from the permissible range.

As a second example, the control unit 40 inputs the temperature of the heating liquid, the rotation speed of the polishing head 1, the rotation speed of the polishing table 2, the dressing conditions, the polishing load of the polishing head 1, and the flow rate of the polishing liquid to the input layer 301. Enter. The learned model calculates the temperature when the temperature of the heating liquid, the rotation speed of the polishing head 1, the rotation speed of the polishing table 2, the dressing conditions, the polishing load of the polishing head 1, and the flow rate of the polishing liquid are input to the input layer 301. The output layer 303 is configured to output the PID parameter (or at least one of the PID parameters) to be changed in order to maintain the behavior curve R within a predetermined tolerance range and the changed value thereof.

The second example data set is a representative combination of temperature behavior parameters representing changes over time in state quantities that influence temperature behavior. For example, the temperature of the heating liquid is a state quantity that changes moment by moment while the pad temperature adjustment device 5 adjusts the surface temperature of the polishing pad 3, and the change in the temperature of the heating liquid over time is determined by the temperature behavior. influence changes in Therefore, by inputting the data set of the second example into the input layer 301, the output layer 303 determines the PID parameters to be changed and their The changed value can be output.

Similar to the first example, the control unit 40 may input time to the input layer 301' (or input layer 301) of the learned model. Further, the control unit 40 learns a combination of the measured value of the pad temperature measuring device 39 and the measurement time point, and/or a temperature behavior curve created based on the measured value of the pad temperature measuring device 39 and the measurement time point. It may also be input to the input layer 301' of the completed model. Furthermore, the trained model outputs from its output layer 303' other temperature behavior parameters different from the PID parameters and their modified values to be modified in order to maintain the temperature behavior curve within a predetermined tolerance. Alternatively, the optimal temperature behavior curve and/or the combination of the pad surface temperature constituting the optimal temperature behavior curve and the measurement time point may be output.

The control unit 40 changes the PID parameters according to the output from the learned model, and polishes the next wafer W while adjusting the surface temperature of the polishing pad 3 according to the changed PID parameters. When adjusting the pad surface temperature by constructing a learned model into which the data set of the second example is input, there is no need to modify each component of the polishing apparatus. That is, the data set of the second example is composed only of parameters that are constantly monitored even in existing polishing equipment. Therefore, by simply installing this learned model in the control unit 40, the temperature behavior curve R can be maintained within a predetermined tolerance range without modifying the existing polishing apparatus. As a result, a polishing apparatus with reduced variation in polishing performance can be provided at a low cost.

In one embodiment, the first example dataset may be input to the input layer 301 (or input layer 301') in addition to the second example dataset. As mentioned above, the data set of the second example is a representative combination of temperature behavior parameters representing changes over time in state quantities that influence temperature behavior. Therefore, by inputting the combination of the data set of the first example and the data set of the second example to the input layer 301, the output layer 303 takes into account changes over time in the state quantities that affect temperature behavior. Then, the PID parameter to be changed and its changed value are output. As a result, the neural network can output the PID parameters to be changed and their changed values with increased accuracy.

As a third example, the control unit 40 inputs the pressing load of the pad contacting member 11, the temperature of the polishing liquid, the ambient temperature in the polishing unit, the supply pressure of the heating liquid, and the supply pressure of the cooling liquid to the input layer 301. do. The learned model generates a temperature behavior curve when the pressing load of the pad contact member 11, the temperature of the polishing liquid, the ambient temperature in the polishing unit, the supply pressure of the heating liquid, and the supply pressure of the cooling liquid are input to the input layer 301. The output layer 303 is configured to output the PID parameter (or at least one of the PID parameters) to be changed in order to maintain R within a predetermined tolerance range and the changed value thereof.

The third example dataset is a representative combination of temperature behavior parameters representing changes in the environment over time that affect temperature behavior. Therefore, by inputting the third example data set to the input layer 301, the output layer 303 determines the PID parameters to be changed and their changes based on changes in the environment over time that affect temperature behavior. The value can be output.

Similar to the first example, the control unit 40 may input time to the input layer 301' (or input layer 301) of the learned model. Further, the control unit 40 learns a combination of the measured value of the pad temperature measuring device 39 and the measurement time point, and/or a temperature behavior curve created based on the measured value of the pad temperature measuring device 39 and the measurement time point. It may also be input to the input layer 301' of the completed model. Furthermore, the trained model outputs from its output layer 303' other temperature behavior parameters different from the PID parameters and their modified values to be modified in order to maintain the temperature behavior curve within a predetermined tolerance. Alternatively, the optimal temperature behavior curve and/or the combination of the pad surface temperature constituting the optimal temperature behavior curve and the measurement time point may be output.

The control unit 40 changes the PID parameters according to the output from the learned model, and polishes the next wafer W while adjusting the surface temperature of the polishing pad 3 according to the changed PID parameters. In order to adjust the pad surface temperature by constructing a trained model into which the data set of the third example is input, modification of the polishing device is required, but this trained model is installed in the control unit 40. This makes it possible to maintain the temperature behavior curve R within a predetermined tolerance range.

In one embodiment, the first example dataset may be input to the input layer 301 (or input layer 301') in addition to the third example dataset. As mentioned above, the third example data set is a representative combination of temperature behavior parameters representing changes in the environment over time that affect temperature behavior. Therefore, by inputting the combination of the first example data set and the third example data set to the input layer 301, the output layer 303 takes into account changes in the environment over time that affect temperature behavior. Then, the PID parameters to be changed and their changed values are output. As a result, the neural network can output the PID parameters to be changed and their changed values with increased accuracy.

Furthermore, in addition to the third example data set, the first example data set and the second example data set may be input to the input layer 301 (or input layer 301'). By inputting the combination of the first example data set, the second example data set, and the third example data set to the input layer 301, the output layer 303 outputs state quantities that affect temperature behavior. The PID parameters to be changed and their changed values are output in consideration of changes over time and changes over time in the environment. As a result, the neural network can output the PID parameters to be changed and their changed values with increased accuracy.

As a fourth example, the control unit 40 inputs the combination of the measured value of the pad temperature measuring device 39 and the measurement time point only to the input layer 301'. The learned model determines the PID that should be changed in order to maintain the temperature behavior curve R within a predetermined tolerance range when a combination of the measured value of the pad temperature measuring device 39 and its measurement time point is input to the input layer 301'. It is constructed to output a parameter (or at least one of the PID parameters) and its modified value from the output layer 303. Similar to the first example, the trained model may output from its output layer 303' an optimal temperature behavior curve and/or a combination of pad surface temperatures and their measurement points that constitute the optimal temperature behavior curve. .

The control unit 40 changes the PID parameters according to the output from the learned model, and polishes the next wafer W while adjusting the surface temperature of the polishing pad 3 according to the changed PID parameters. When adjusting the pad surface temperature by constructing a learned model into which the data set of the fourth example is input, there is no need to modify each component of the polishing apparatus, as in the second example. Therefore, simply by installing this trained model in the control unit 40, the temperature behavior curve R can be maintained within a predetermined tolerance range.

In the embodiment described above, at least one film thickness parameter related to the thickness of the film of the wafer W polished by the polishing unit may be further input into the input layer 301 (or input layer 301'). In other words, the control unit 40 may input at least one temperature behavior parameter and at least one film thickness parameter to the input layer 301. In this case, the control unit 40 may further input at least one data set among the data sets shown in the first to fourth examples to the input layer 301 (or input layer 301'). In this specification, the film thickness parameter is a general term for index values related to the thickness of a film formed on the surface of a wafer (substrate) W. As explained below, the film thickness parameter includes, for example, a film thickness signal acquired by a film thickness sensor, and a film thickness value obtained by calculating (or converting) the film thickness signal. .

Conventionally, in a polishing process for polishing a wafer W, a film thickness sensor is used to measure the thickness of the film formed on the surface of the wafer W in order to detect the polishing end point, which is the point at which a desired target film thickness is reached. is being detected. For example, when the film of the wafer W to be polished by the polishing unit is a conductive film, the thickness of the conductive film is detected using an eddy current film thickness sensor. The eddy current type film thickness sensor applies a high-frequency alternating current to a coil to induce eddy current in the conductive film of the wafer W, and detects the thickness of the conductive film from changes in impedance caused by the magnetic field of this eddy current. It is composed of Any sensor can be used as the film thickness sensor as long as it can detect the thickness of the film formed on the surface of the wafer W, but in the following, an eddy current film thickness sensor, which is an example of a film thickness sensor, will be used. An example of detecting the thickness of a conductive film on a wafer W using the method will be described.

FIG. 13(a) is a schematic diagram showing an example of a polishing unit equipped with an eddy current type film thickness sensor, and FIG. 13(b) is a schematic cross-sectional view of the polishing unit shown in FIG. 13(a). In FIGS. 13A and 13B, components of the polishing unit other than the polishing head 1, polishing table 2, polishing pad 3, and eddy current film thickness sensor 7 are not illustrated. FIG. 13(b) shows the eddy current film thickness sensor passing below the wafer. The configuration of the polishing unit of this embodiment, which is not particularly described, is the same as the configuration of the polishing unit of the embodiment described above, and therefore, the redundant description thereof will be omitted.

As shown in FIGS. 13(a) and 13(b), the eddy current type film thickness sensor 7 is embedded in the polishing table 2, and rotates around the central axis of the polishing table 2 as the polishing table 2 rotates. revolve. The central axis of the polishing table 2 extends vertically through the center CL of the polishing pad 3 shown in FIGS. 2 and 3. The eddy current film thickness sensor 7 scans the surface of the wafer W every time the polishing table 2 rotates, and acquires a film thickness signal at at least one measurement point on the wafer W. The eddy current film thickness sensor 7 is connected to the control section 40 , and at least one film thickness signal acquired by the eddy current film thickness sensor 7 is sent to the control section 40 . This film thickness signal changes according to changes in the thickness of the conductive film on the wafer W. The film thickness signal is one of the film thickness parameters related to the thickness of the film of the wafer W to be polished. The control unit 40 can monitor the polishing progress of the wafer W based on the film thickness signal. For example, the control unit 40 can determine the polishing end point when the film thickness signal reaches a predetermined threshold.

In one embodiment, the control unit 40 may be configured to be able to obtain the film thickness value of the conductive film by calculating the received film thickness signal. The actual film thickness value of the conductive film, which changes according to the progress of polishing the wafer W, is also one of the film thickness parameters. The control unit 40 can monitor the polishing progress of the wafer W based on the film thickness value. For example, the control unit 40 can determine the point in time when the film thickness value of the conductive film obtained by calculation reaches a desired target film thickness as the polishing end point.

In this embodiment, each time the eddy current film thickness sensor 7 acquires a film thickness signal of the wafer W, the control unit 40 sends at least one temperature behavior parameter and at least one film thickness parameter to the input layer 301. input. When the eddy current type film thickness sensor 7 acquires a plurality of film thickness signals at a plurality of measurement points, the control unit 40 may input all the film thickness signals to the input layer 301, or may input all the film thickness signals to the input layer 301, or A representative value of the signal may be input to the input layer 301. The representative value of the film thickness signal is, for example, one of the average value, maximum value, and minimum value of all the film thickness signals. Alternatively, the control unit 40 may input some film thickness signals selected from all the film thickness signals to the input layer 301. For example, the control unit 40 inputs the average value, maximum value, and minimum value of all film thickness signals to the input layer 301.

When the film thickness value of the conductive film is acquired, the control unit 40 inputs the film thickness value obtained by calculation to the input layer 301 (or , input layer 301'). The film thickness value is input to the input layer 301 (or input layer 301') every time the eddy current film thickness sensor 7 acquires a film thickness signal of the wafer W. When the eddy current type film thickness sensor 7 acquires a plurality of film thickness signals at a plurality of measurement points, the control unit 40 inputs all the film thickness values obtained from all the film thickness signals to the input layer 301. Alternatively, a representative film thickness value obtained from representative values of a plurality of film thickness signals may be input to the input layer 301. The representative film thickness value is, for example, a film thickness value obtained from one of the average value, maximum value, and minimum value of all film thickness signals. Alternatively, the control unit 40 may input selected film thickness values obtained from some film thickness signals selected from all film thickness signals to the input layer 301. For example, the control unit 40 inputs into the input layer 301 three selected film thickness values obtained from the average value, maximum value, and minimum value of all film thickness signals.

As described above, the trained model outputs the PID parameter to be changed and its changed value from its output layer 303. In this embodiment, the trained model further predicts the polishing end point of the wafer W every time the eddy current film thickness sensor 7 acquires a film thickness signal of the wafer W, and outputs the predicted polishing end point. It is constructed to output from layer 303 (or output layer 303'). In one embodiment, the learned model is used from the present time to the polishing end point each time the eddy current film thickness sensor 7 acquires a film thickness signal of the wafer W, instead of or in addition to the predicted polishing end point. The polishing completion time may be predicted and the predicted polishing completion time may be output from the output layer 303 (or output layer 303'). This predicted polishing end point and/or polishing completion time is based on polishing with the pad surface temperature adjusted according to the optimal PID parameters while taking into account the thickness of the film being polished from time to time (i.e., changing). This is the polishing end point and/or polishing completion time of the wafer W. Therefore, the thickness of the film on the wafer W to be polished can be made to match the target film thickness more accurately.

In the embodiment described above, the amount of change in the thickness of the polishing pad 3 may be further input into the input layer 301 (or input layer 301'). In other words, the control unit 40 may input at least one temperature behavior parameter and the amount of change in the thickness of the polishing pad 3 to the input layer 301. In this case, the control unit 40 may further input at least one data set among the data sets shown in the first to fourth examples to the input layer 301 (or input layer 301').

The amount of change in the thickness of the polishing pad 3 is a parameter that affects the polishing rate of the wafer W. That is, when the thickness of the polishing pad 3 changes, the polishing rate of the wafer W also changes. Therefore, the amount of change in the thickness of the polishing pad 3 is one of the parameters representing changes in the polishing environment over time. Therefore, in this embodiment, the pad surface temperature is adjusted in order to keep the polishing rate at an optimal state. Specifically, the learned model calculates the optimal value when the amount of change in the thickness of the polishing pad 3 is input into its input layer 301 (or input layer 301') in addition to at least one temperature behavior parameter. It is constructed to predict the temperature behavior curve necessary to maintain the polishing rate, and output from the output layer 303 the PID parameters to be changed and their changed values in order to achieve this predicted temperature behavior curve. . An example of a polishing unit that can measure the amount of change in the thickness of the polishing pad 3 will be described below.

FIG. 14 is a schematic diagram showing an example of a polishing unit equipped with a pad height sensor that measures the amount of change in the thickness of the polishing pad. In FIG. 14, illustration of the pad temperature adjustment device 5 is omitted. The configuration of the polishing unit according to this embodiment, which is not particularly described, is the same as the configuration of the polishing unit according to the embodiments described above, and therefore, the redundant explanation will be omitted.

In the polishing unit shown in FIG. 14, the dresser 20 is connected to a dresser shaft 24, and an air cylinder 25 is provided at the upper end of the dresser shaft 24. The dresser shaft 24 is rotatably supported by the dresser arm 21. Further, the dresser shaft 24 and the dresser 20 are movable up and down relative to the dresser arm 21. The air cylinder 25 is a device that applies a dressing load to the polishing pad 3 (that is, a pressing load of the dresser 20 against the polishing pad 3) to the dresser 20. The dressing load can be adjusted by the air pressure supplied to the air cylinder 25. The air cylinder 25 presses the dresser 20 against the surface (ie, polishing surface) of the polishing pad 3 with a predetermined load via the dresser shaft 24 .

The dresser arm 21 is driven by a motor (not shown) and is configured to swing about the support shaft 19 . This dresser arm 21 allows the dresser 20 to swing in the radial direction of the polishing pad 3 while contacting the polishing pad 3. The dresser shaft 24 is rotated by a motor (not shown) installed in the dresser arm 21, and the rotation of the dresser shaft 24 causes the dresser 20 to rotate around its axis.

A pad height sensor (surface height measuring device) 27 that measures the height of the surface of the polishing pad 3 is fixed to the dresser arm 21 . Further, a sensor target 28 is fixed to the dresser shaft 24 so as to face the pad height sensor 27 .

When the air cylinder 25 is driven, the dresser 20, the dresser shaft 24, and the sensor target 28 move up and down together. On the other hand, the vertical positions of the dresser arm 21 and the pad height sensor 27 are fixed. The pad height sensor 27 measures the relative position of the dresser 20 in the vertical direction with respect to the dresser arm 21 when the dresser 20 is in contact with the surface (i.e., the polishing surface) of the polishing pad 3. indirectly measure the height of the surface. Sensor target 28 is coupled to dresser 20 so that pad height sensor 27 can measure the height of the surface of polishing pad 3 during conditioning of polishing pad 3. The pad height sensor 27 can be any type of sensor, such as a linear scale sensor, a laser sensor, an ultrasonic sensor, or an eddy current sensor. The pad height sensor 27 is connected to the control section 40 , and the measured value of the thickness of the polishing pad 3 obtained by the pad height sensor 27 is sent to the control section 40 .

The amount of change in the thickness of the polishing pad 3 is determined as follows. First, the air cylinder 25 is driven to bring the dresser 20 into contact with the unworn surface of the polishing pad 3. In this state, the pad height sensor 27 measures the initial position of the dresser 20 (the initial surface height of the polishing pad 3), and the control unit 40 acquires the measured value of the initial position of the dresser 20. After the polishing process for one or more wafers W is completed, the dresser 20 is brought into contact with the surface of the polishing pad 3 again, and in this state, the pad height sensor 27 measures the position of the dresser 20 again. do. The control unit 40 obtains a measured value of the position of the dresser 20 (that is, the surface height of the worn polishing pad 3). Since the dresser 20 is displaced downward as the polishing pad 3 is worn, the control unit 40 calculates the difference between the measured value of the initial surface height of the polishing pad 3 and the measured value of the surface height of the worn polishing pad 3. , the amount of change in the thickness of the polishing pad 3 can be determined.

Usually, dressing of the polishing pad 3 is performed every time one wafer W is polished. Dressing is performed before or after polishing the wafer W, or during polishing of the wafer W. To calculate the amount of change in the thickness of the polishing pad 3, the measured value of the pad height sensor 27 obtained during any dressing is used.

The dresser 20 swings on the polishing pad 3 in its radial direction by swinging the dresser arm 21 . The measured value of the surface height of the polishing pad 3 is sent from the pad height sensor 27 to the control unit 40, where the average of the measured values of the surface height of the polishing pad 3 during dressing is determined. Note that for one dressing operation, the dresser 20 reciprocates over the polishing pad 3 once or multiple times.

In this embodiment, each time the control unit 40 acquires the amount of change in the thickness of the polishing pad 3, in addition to at least one temperature behavior parameter, the control unit 40 determines the amount of change in the thickness of the polishing pad 3 from the input layer 301 (or , input layer 301'). The trained model predicts the temperature behavior curve required to maintain the optimum polishing rate, and outputs from the output layer 303 the PID parameters to be changed and their change values to achieve this predicted temperature behavior curve. Output. In this way, the output layer 303 takes into account the amount of change in the thickness of the polishing pad 3 that affects the polishing rate, and determines the PID parameters that should be changed in order to maintain the optimum polishing rate and their change values. Output. The control unit 40 changes the PID parameters according to the output from the learned model, and polishes the next wafer W while adjusting the surface temperature of the polishing pad 3 according to the changed PID parameters.

In one embodiment, the polishing rate may also be input into input layer 301 (or input layer 301'). The control unit 40 can obtain the actual polishing rate by dividing the thickness of the polished film by the polishing time (that is, the time from the start of polishing to the end point of polishing). By further inputting the actual polishing rate to the input layer 301 (or input layer 301'), the neural network can output the PID parameters to be changed and their changed values with higher accuracy. can.

In one embodiment, when the input layer 301 is input with at least one temperature behavior parameter and the amount of change in thickness (and polishing rate) of the polishing pad 3, the output layer 303 (or output layer 303') is The trained model may be constructed to further output dressing conditions. The dressing conditions include, for example, the rotation speed of the dresser 20, the pressing load of the dresser 20 against the polishing pad 3, the swing speed of the dresser 20, and the like. By dressing the polishing pad 3 using the dressing conditions output from the output layer 303, the surface condition of the polishing pad 3 can be maintained at or close to the surface condition for maintaining the optimum polishing rate.

The control unit 40 operates according to a program electrically stored in the storage device 110. That is, the control unit 40 controls the heating liquid and cooling so that the pad surface temperature reaches a predetermined target temperature based on the pad surface temperature measured by the pad temperature measuring device 39, and then maintains the pad surface temperature at the target temperature. PID control is performed by controlling at least the flow rate of the liquid, creating a temperature behavior curve until the pad surface temperature reaches the target temperature, and inputting at least one temperature behavior parameter into a trained model constructed by machine learning. A calculation is executed to output the changed value of the PID parameter, and when polishing the next wafer W, a step of controlling the flow rates of the heating liquid and the cooling liquid using the changed PID parameter is executed.

A program for causing the control section 40 to execute the above steps is recorded on a computer-readable recording medium that is a non-temporary tangible object, and is provided to the control section 40 via the recording medium. Alternatively, the program may be provided to the control unit 40 via the communication device 150 via a communication network such as the Internet. The program provided to the control unit 40 is installed in the storage device 110 by the processing device 120. Furthermore, when updating an old program to a new program (for example, upgrading the program), the new program is provided to the control unit 40 via the communication device 150 or the input device 130. The processing device 120 installs the provided new program into the storage device 110 and updates the old program. The processing device 120 may install the new program on the storage device 110 after uninstalling the old program from the storage device 110.

In this way, even if the pad surface temperature is adjusted while changing the PID parameters using artificial intelligence, the temperature behavior range R may deviate from the predetermined allowable range, as shown in FIG. When the temperature behavior curve R deviates from a predetermined tolerance range, the control unit 40 preferably outputs an alarm. An operator of the polishing apparatus who receives the warning can detect that an abnormality in temperature behavior has occurred in the polishing unit to which the warning was output. Furthermore, the operator can check the state of the wafer W in which abnormal temperature behavior has occurred.

After outputting the alarm, or simultaneously with outputting the alarm, the control unit 40 may stop the operation of the polishing unit to which the alarm has been output. Thereby, the operator of the polishing apparatus can collect the wafer W in which abnormal temperature behavior has occurred and check the state of the wafer W and the state of the polishing unit. In one embodiment, after stopping the operation of the polishing unit, the control unit 40 executes the step of placing the wafer W being polished in the polishing unit to which the alarm has been output in a wafer shelter part (not shown). It's okay. The wafer shelter part is preferably provided within the polishing unit. Thereby, the operator of the polishing apparatus can easily and reliably recover the wafer W in which abnormal temperature behavior has occurred.

A wafer cassette containing a predetermined number (for example, 25) of wafers W may be erroneously transported to the polishing apparatus. In this embodiment, the control unit 40 can determine whether the wafer cassette is being conveyed incorrectly. More specifically, when a wafer cassette is erroneously transported to a polishing apparatus, the polishing recipe used to polish the wafer W housed in this wafer cassette is different from the polishing recipe that should originally be used. If the wafer W is polished with a polishing recipe different from the polishing recipe originally used, it is assumed that the temperature behavior curve will deviate significantly from the allowable range. In this embodiment, the operator can collect the wafer W in which abnormal temperature behavior has occurred, and therefore can easily confirm whether or not the wafer cassette has been erroneously transported. As a result, it is possible to prevent all of the wafers W accommodated in the wafer cassette from being discarded.

FIG. 15 is a schematic diagram showing another example of the configuration of the artificial intelligence shown in FIG. 10. The artificial intelligence shown in FIG. 15 has a feature of automatically updating a trained model. More specifically, when it is determined that the changed value of the PID parameter output from the output layer 303 of the trained model is within the normal range, the control unit 40 controls the PID parameter and its changed value. It is stored in the storage device 110 as an additional learning data set, and the learned model is automatically updated through machine learning (deep learning) based on the learning data set and the additional learning data set. Thereby, it is possible to improve the accuracy of the PID parameter (at least one temperature behavior parameter) output from the learned model and its changed value.

Machine learning (deep learning) can learn not only the combination of the PID parameter (at least one temperature behavior parameter) to be changed and its change value, but also various elements. Therefore, the learned model constructed by machine learning can be used for diagnosing or predicting the state and/or abnormality of the pad temperature adjustment device 5. Furthermore, the trained model constructed by machine learning can be used to diagnose or predict the state and/or abnormality of the polishing unit. Below, for convenience of explanation, the trained model according to the embodiment described above may be referred to as "trained model 1." Furthermore, the trained model described below may be referred to as "trained model 2."

Trained model 2 is constructed using the neural network described with reference to FIG. This learned model 2 inputs into the input layer 301 a temperature behavior curve R created based on the measured value of the pad surface temperature device 39 and the measurement time point, and a plurality of temperature behavior curves stored in the storage device 110. It is constructed to predict the lifespan of the polishing pad 3 and/or the lifespan of the dresser 20 upon input, and output the results from the output layer 303. Alternatively, the learned model 2 may input the temperature behavior curve R created based on the measured value of the pad surface temperature device 39 and the measurement time point, and a plurality of temperature behavior curves stored in the storage device 110 to the input layer 301. The system may be configured to diagnose abnormalities in the polishing head 1 , polishing table 3 , and polishing liquid supply nozzle 4 when the input is input, and output the results from the output layer 303 .

When the wafer W is polished by pressing the wafer W against the polishing pad 3, the polishing pad 3 deteriorates and the amount of frictional heat generated between the polishing pad 3 and the wafer W decreases. Therefore, as the polishing pad 3 deteriorates, the slope of the temperature behavior curve gradually decreases. Alternatively, as the polishing pad 3 is ground by the dresser 20, the depth of the grooves (not shown) provided on the surface of the polishing pad 3 becomes smaller, and eventually the grooves disappear. Due to such a phenomenon, the amount of frictional heat generated between the polishing pad 3 and the wafer W may decrease or increase, and the slope of the temperature behavior curve changes. Therefore, the learned model 2 includes a temperature behavior curve R created based on the measured values of the pad surface temperature device 39 inputted to the input layer 301 and the measurement time points, and a plurality of temperature behavior curves R stored in the storage device 110. The temperature behavior curves are compared, the life of the polishing pad 3 is predicted from the amount of decrease in the slope of the temperature behavior curves, and the prediction result is output from the output layer 303.

Furthermore, when the surface of the polishing pad 3 is dressed using the dresser 20, the dressing surface of the dresser 20 deteriorates, and the roughness of the surface of the polishing pad 3 after dressing becomes smaller. Alternatively, if the dressing surface deteriorates, the surface of the polishing pad 3 may not be appropriately dressed. Therefore, as the dresser 20 deteriorates, the amount of frictional heat generated between the polishing pad 3 and the wafer W decreases or increases, and the slope of the temperature behavior curve gradually decreases or increases. Therefore, the learned model 2 includes a temperature behavior curve R created based on the measured values of the pad surface temperature device 39 inputted to the input layer 301 and the measurement time points, and a plurality of temperature behavior curves R stored in the storage device 110. The temperature behavior curve R is compared with a plurality of temperature behavior curves, the lifespan of the dresser 20 is predicted from the amount of decrease or increase in the slope of the temperature behavior curve, and the prediction result is output from the output layer 303.

In order to prevent the polishing liquid (slurry) from adhering to the bottom surface of the pad contact member 11, a coating film may be attached to the bottom surface of the pad contact member 11. In this case, the trained model 2 inputs a temperature behavior curve R created based on the measured value of the pad surface temperature meter 39 and the measurement time point, and a plurality of temperature behavior curves R stored in the storage device 110. It may be constructed to diagnose the wear state of the coating film when input to the layer 301 and output the result from the output layer 303.

The material of the coating film is, for example, Teflon (registered trademark), which has a relatively high heat insulation effect. Since the pad contact member 11 is pressed against the surface of the polishing pad 3 with a predetermined pressing load, the coating film is worn every time the pad surface temperature is adjusted. As the wear of the coating film progresses, the slope of the temperature behavior curve increases. Therefore, the learned model 2 includes a temperature behavior curve R created based on the measured values of the pad surface temperature device 39 inputted to the input layer 301 and the measurement time points, and a plurality of temperature behavior curves R stored in the storage device 110. The temperature behavior curves are compared, the life of the coating film is predicted from the amount of increase in the slope of the temperature behavior curves, and the prediction result is output from the output layer 303.

The temperature behavior curve also changes due to a change in the amount of frictional heat generated between the wafer W held by the polishing head 1 and the surface of the polishing pad 3. For example, if the polishing load of the wafer W on the polishing pad 3 deviates from a desired value, the temperature behavior curve changes. The temperature behavior curve also changes when the supply amount and/or temperature of the polishing liquid deviates from a desired value, or the drop position of the polishing liquid deviates from a desired position. Furthermore, the temperature behavior curve also changes when the rotational speed of the polishing head 1 and/or the rotational speed of the polishing table 3 deviates from desired values. Therefore, the trained model 2 includes a temperature behavior curve R created based on the measured value of the pad surface temperature device 39 inputted to the input layer 301 and the measurement time point, and a plurality of temperature behavior curves R stored in the storage device 110. By comparing the temperature behavior curves, abnormalities in the polishing head 1, polishing table 3, and polishing liquid supply nozzle 4 can be diagnosed.

The control unit 40 operates according to a program electrically stored in the storage device 110. That is, the control unit 40 controls the heating liquid and cooling so that the pad surface temperature reaches a predetermined target temperature based on the pad surface temperature measured by the pad temperature measuring device 39, and then maintains the pad surface temperature at the target temperature. Control at least the flow rate of the liquid, create a temperature behavior curve R until the pad surface temperature reaches the target temperature, and combine the created temperature behavior curve R and a plurality of temperature behavior curves stored in the storage device 110. A step of diagnosing the lifespan of the polishing pad 3 and/or the lifespan of the dresser 20 is executed by inputting the learned model constructed by machine learning. Instead of the step of diagnosing the lifespan of the polishing pad 3 and/or the lifespan of the dresser 20, the control unit 40 may execute a step of diagnosing wear of the coating film, or may perform a step of diagnosing wear of the coating film, or detect an abnormality in the polishing head 1, or the polishing table 3. A step of diagnosing an abnormality in the polishing liquid supply nozzle 4 and an abnormality in the polishing liquid supply nozzle 4 may be performed.

A program for causing the control section 40 to execute the above steps is recorded on a computer-readable recording medium that is a non-temporary tangible object, and is provided to the control section 40 via the recording medium. Alternatively, the program may be provided to the control unit 40 via the communication device 150 via a communication network such as the Internet. The program provided to the control unit 40 is installed in the storage device 110 by the processing device 120. Furthermore, when updating an old program to a new program (for example, upgrading the program), the new program is provided to the control unit 40 via the communication device 150 or the input device 130. The processing device 120 installs the provided new program into the storage device 110 and updates the old program. The processing device 120 may install the new program on the storage device 110 after uninstalling the old program from the storage device 110.

FIG. 16 is a schematic diagram illustrating one embodiment of a polishing system including at least one polishing device. The polishing system shown in FIG. 16 includes a plurality of polishing apparatuses according to the embodiment described above, a plurality of relay apparatuses 500 connected to each polishing apparatus, and a host control system 600 connected to the plurality of relay apparatuses 500. Be prepared. Relay device 500 is a gateway such as a router, and includes a relay control unit 510, a relay communication device 515, and a relay storage device 512. The host control system 600 includes a host control unit 610, a host communication device 615, and a host storage device 612.

The communication device 150 (see FIG. 5) of the control unit 40 of the polishing apparatus is connected to the relay communication device 515 of the relay device 500 so as to be able to send and receive information by wireless communication (for example, high-speed WiFi (registered trademark)) or wired communication. There is. The relay communication device 515 of the relay device 500 is connected to the host communication device 615 of the host control system 600 so as to be able to transmit and receive information by wireless communication (for example, high-speed WiFi (registered trademark)) or wired communication. In this embodiment, each polishing apparatus is connected to a network (for example, the Internet) via a host control system 600 and a relay device 500.

Furthermore, the host control system 600 may be located within a factory where at least one polishing device is installed, or may be located outside a factory where at least one polishing device is installed. If the host control system 600 is located in a factory in which at least one polishing device is installed, the host control system 600 may be a host computer located in the factory, or may be a host computer located in the factory. It may be a built cloud computing system or a fog computing system. If the host control system 600 is located outside the factory where the at least one polishing device is installed, the host control system 600 is preferably a cloud computing system or a fog computing system. In this case, the host control system 600 is preferably connected to a plurality of factories in which at least one polishing device is installed.

In the example shown in FIG. 16, the relay device 500 is placed outside the polishing device. However, the invention is not limited to this example. For example, the relay device 500 may be placed inside a polishing device.

In the embodiment shown in FIG. 16, the host control unit 610 of the host control system 600 determines the PID parameters (at least one temperature behavior parameter) to be changed in order to bring the temperature behavior curve R into a predetermined tolerance range, and the change thereof. The value is determined by artificial intelligence (AI). The host storage device 612 of the host control unit 610 stores in advance the trained model 1 described with reference to FIGS. 10 to 15. Note that the host control unit 610 includes a processing device (not shown) corresponding to the processing device 120 shown in FIG. The processing device of the host control unit 610 reads out the learned model stored in the host storage device 612, inputs at least one temperature behavior parameter to the learned model, and outputs the PID parameter and its changed value. Perform the calculation for.

In one embodiment, the processing unit of the host controller 610 may input time to the input layer 301' (or input layer 301) of the trained model 1. Furthermore, the processing device of the host control unit 610 generates a temperature that is created based on a combination of the measured value of the pad temperature measuring device 39 and the measurement time point, and/or the measured value of the pad temperature measuring device 39 and the measurement time point. The behavior curve R may be further input to the input layer 301' of the learned model 1. Furthermore, the trained model 1 outputs other temperature behavior parameters different from the PID parameters and their changed values, which should be changed in order to maintain the temperature behavior curve within a predetermined tolerance range, from its output layer 303'. Alternatively, the combination of the optimum temperature behavior curve and/or the pad surface temperature constituting the optimum temperature behavior curve and the measurement time point thereof may be output.

In this embodiment, the control unit 40 of each polishing apparatus receives data including a combination of the measured value of the pad surface temperature meter 39 and its measurement time point, and at least one temperature behavior parameter input to the learned model 1. , is transmitted to the host control system 600 via the relay device 500. Having received this data, the host control unit 610 of the host control system 600 creates a temperature behavior curve R from the measured value of the pad surface temperature meter 39 and the measurement time point, and stores at least one temperature behavior parameter in the host storage. The PID parameters to be input to the input layer 301 of the learned model 1 stored in 612 and changed in order to maintain the temperature behavior curve within a predetermined tolerance range when polishing the next wafer W, and their change values. Execute the calculation to output .

The PID parameters output from the output layer 303 of the trained model 1 and their changed values are sent to the polishing device via the relay device 500. The control unit 40 of the polishing apparatus adjusts the surface temperature of the polishing pad 3 according to the received PID parameter and its changed value. When the output layer 303' outputs other temperature behavior parameters different from the PID parameters and their changed values, which should be changed in order to maintain the temperature behavior curve within a predetermined tolerance range, the control unit 40 updates other temperature behavior parameters to their changed values.

A first example of data that the polishing apparatus sends to the host control system 600 for input into the trained model 1 is the above-mentioned PID parameter, heating liquid flow rate, and cooling liquid flow rate. A second example of data that the polishing apparatus sends to the host control system 600 is the temperature of the heating liquid, the rotation speed of the polishing head 1, the rotation speed of the polishing table 2, dressing conditions, the polishing load of the polishing head 1, and the polishing liquid. The flow rate is A third example of data that the polishing apparatus sends to the host control system 600 is the temperature of the heating liquid, the rotation speed of the polishing head 1, the rotation speed of the polishing table 2, dressing conditions, the polishing load of the polishing head 1, and the polishing liquid. The flow rate is The data that the polishing apparatus sends to the host control system 600 to be input to the trained model 1 may be a combination of the first to third examples. Furthermore, the time described in item 21) above may be transmitted to the host control system 600 and input into the learned model 1. In this case, the trained model 1 calculates each predicted value by matching the measurement time of the multiple temperature behavior parameters from the temporal change of multiple temperature behavior parameters input other than time, and based on this predicted value. , outputs the PID parameter to be changed and its changed value from the output layer 303. Thereby, a more accurate output value can be obtained in consideration of the time difference between each temperature behavior parameter acquired by the control unit 40. Furthermore, in any of the examples, in order to input into the learned model 1, the polishing apparatus uses a combination of the measured value of the pad surface temperature device 39 and its measurement time point, and/or the measured value of the pad surface temperature device 39. The temperature behavior curve R created based on the measurement time point and the temperature behavior curve R may be further transmitted to the host control system 600.

Another example of data that the polishing apparatus sends to the host control system 600 for input into the trained model 1 is at least one film thickness parameter related to the film thickness of the wafer W. As described above, the film thickness parameter includes the film thickness signal acquired by the film thickness sensor and the actual film thickness value obtained by calculating (or converting) the film thickness signal. Still other examples of data that the polishing apparatus sends to the host control system 600 to be input into the learned model 1 are the amount of change in the thickness of the polishing pad 3 and the polishing rate.

When it is determined that the changed value of the PID parameter output from the output layer 303 of trained model 1 is within the normal range, the host control unit 612 of the host control system 600 changes the PID parameter and its changed value. is stored in the host storage device 612 as an additional training dataset, and the trained model 1 is automatically updated through machine learning (deep learning) based on the training dataset and the additional training dataset. . A huge amount of data including PID parameters to be changed and their changed values is sent to the host control system 600 from a plurality of polishing apparatuses. Accuracy with changed values can be improved in a short period of time.

The host storage device 612 of the host control system 600 may store the above-described learned model 2 in addition to the learned model 1. In this case, the host control system 600 may predict the lifespan of the polishing pad 3 and/or the lifespan of the dresser 20, or detect abnormalities in the polishing head 1, polishing table 3, and polishing liquid supply nozzle 4. May be diagnosed. If a coating film is provided on the bottom surface of the pad contact member 11, the learned model stored in the host storage device 612 of the host control system 600 may diagnose the wear state of the coating film.

In one embodiment, the above-described learned model 1 may be stored in the storage unit 110 of the control unit 40 of the polishing apparatus, and the above-described learned model 2 may be stored in the host storage device 612 of the host control system 600. In this case, the control unit 40 of the polishing apparatus calculates the PID parameter to be changed and its change value in order to maintain the temperature behavior curve R within a predetermined tolerance range, and the host control unit 612 of the host control system 600 , predict the life of the polishing pad 3 and/or the life of the dresser 20, or diagnose abnormalities in the polishing head 1, polishing table 3, and polishing liquid supply nozzle 4.

Alternatively, the learned model 2 described above may be stored in the storage unit 110 of the control unit 40 of the polishing apparatus, and the learned model 1 described above may be stored in the host storage device 612 of the host control system 600. In this case, the control unit 40 of the polishing apparatus predicts the lifespan of the polishing pad 3 and/or the lifespan of the dresser 20, or diagnoses abnormalities in the polishing head 1, polishing table 3, and polishing liquid supply nozzle 4. However, the host control unit 612 of the host control system 600 calculates the PID parameter to be changed and its change value in order to maintain the temperature behavior curve R within a predetermined tolerance range.

FIG. 17 is a schematic diagram illustrating another embodiment of a polishing system including at least one polishing device. The configuration of this embodiment, which is not particularly described, is the same as that of the embodiment shown in FIG. 16, so the redundant explanation will be omitted.

In the embodiment shown in FIG. 17, the relay control unit 510 of the relay device 500 determines the PID parameter (at least one temperature behavior parameter) to be changed in order to bring the temperature behavior curve R into a predetermined tolerance range, and the change value thereof. will be determined by artificial intelligence (AI). In this case, the polishing system is constructed as an edge computing system in which the relay device 500 is placed near the polishing device. The learned models described with reference to FIGS. 10 to 15 are stored in advance in the relay storage device 512 of the relay device 500. Note that for convenience of explanation, the trained model stored in the relay storage device 512 will be referred to as "trained model 3."

Relay control unit 510 includes a processing device (not shown) corresponding to processing device 120 shown in FIG. Like the learned model 1, the learned model 3 stored in the relay storage device 512 is a learned model constructed by machine learning in order to maintain the temperature behavior curve R within a predetermined tolerance range. In the trained model 3, in addition to at least one temperature behavior parameter, each time the pad temperature measuring device 39 measures the pad surface temperature, the measured value and measurement time are input into the input layer 301. The learned model 3 is constructed so that the PID parameters and their changed values are outputted from the output layer 303 at any time so that the measured value of the pad temperature measuring device 39 does not fall outside the allowable range. . That is, the trained model 3 inputs the combination of the measured value of the pad surface temperature meter 39 sent while adjusting the pad surface temperature and the measurement time point into the input layer 301 as needed, and adjusts the temperature. It is constructed to predict in real time whether or not the behavior curve R exceeds a predetermined tolerance range, and output the PID parameters to be changed and their changed values from the output layer 303 at any time.

The processing device of the relay control unit 510 reads the trained model 3 stored in the relay storage device 512, inputs at least one temperature behavior parameter to the input layer 301, and while adjusting the pad surface temperature. Combinations of sent measurement values from the pad surface temperature meter 39 and their measurement times are input into the input layer 301 as needed. Further, the processing device of the relay control unit 510 predicts in real time whether or not the temperature behavior curve R exceeds a predetermined tolerance range, and outputs the PID parameter to be changed and its changed value from the output layer 303. Perform calculations in real time. The calculated PID parameters and their changed values are outputted from the output layer 303 at any time while the pad surface temperature is being adjusted. The relay control unit 510 of the relay device 500 transmits the PID parameter output from the output layer 303 and its changed value to the control unit 40 of the polishing device.

The control unit 40 of the polishing apparatus, which has received the PID parameter to be changed and its changed value, adjusts the pad surface temperature while changing the PID parameter to the received changed value as needed.

In this embodiment, the relay processing device 500 predicts the PID parameters to be changed and their changed values at any time during polishing of the wafer W, and immediately sends the predicted PID parameters and their changed values to the polishing apparatus. . Therefore, it is possible to effectively prevent the temperature behavior curve R from deviating from a predetermined tolerance range.

In the polishing system according to the present embodiment, the relay control unit 510 of the relay device 500 diagnoses at least one temperature behavior parameter that is changed to bring the temperature behavior curve R into a predetermined tolerance range and the changed value. can be processed at high speed and output to a polishing device. On the other hand, information that does not need to be processed at high speed (for example, status information of each polishing unit) is transmitted from the polishing apparatus to the host control system 600 via the relay device 500. As a result, the relay control unit 510 of the relay device 500 does not need to perform unnecessary information processing, and can quickly process the determination of the PID parameter to be changed and its change value.

In one embodiment, the above-mentioned "time" may be additionally input to the input layer 301 (or input layer 301') of the trained model 3. In this case, the learned model 3 calculates each predicted value by matching the measurement time of the multiple temperature behavior parameters from the temporal change of multiple temperature behavior parameters input other than time, and based on this predicted value. , the PID parameters to be changed and their changed values are outputted from the output layer 303 as needed. This allows the output layer 303 to output a more accurate output value that takes into account the time difference between each temperature behavior parameter. As a result, it is possible to more effectively prevent the temperature behavior curve R from deviating from the predetermined tolerance range.

Further, the output layer 303' may output other temperature behavior parameters different from the PID parameters and their modified values to be changed in order to maintain the temperature behavior curve within a predetermined tolerance range. In this case, the control unit 40 adjusts the pad surface temperature while updating the PID parameter and other temperature behavior parameters to the changed values as needed, so that it is more effective to prevent the temperature behavior curve R from falling outside the predetermined tolerance range. can be prevented.

In one embodiment, the trained model 2 described above may be stored in the host storage device 612 of the host control system 600. Alternatively, the learned model 3 described above may be stored in the storage unit 110 of the control unit 40 of the polishing apparatus.

In the embodiment described above, the trained model 1 or the trained model 3 outputs the PID parameter to be changed and its changed value from its output layer 303, but the present invention is not limited to this example. For example, the trained model 1 or the trained model 3 may output a PID parameter to be changed and a program for calculating the changed value from its output layer 303. In this case, the control unit 40 of the polishing device, the relay control unit 510 of the relay device 500, or the host control unit 610 of the host control device 600 performs a process of calculating the change value of the PID parameter according to the program output from the output layer 303. Execute. Alternatively, the host control unit 610 of the host control system 600 or the relay control unit 510 of the relay device 500 executes the process of calculating the change value of the temperature behavior parameter according to the program output from the output layer 303, and the result is It may also be sent to a polishing device.

Alternatively, the trained model 1 or the trained model 3 may output the PID parameter to be changed and the correction coefficient for calculating the value of the changed PID parameter from its output layer 303. In this case, the control unit 40 of the polishing device, the relay control unit 510 of the relay device 500, or the host control unit 610 of the host control device 600 multiplies the current PID parameter by the correction coefficient output from the output layer 303. As a result, the changed PID parameter value can be obtained. Alternatively, the host control unit 610 of the host control system 600 or the relay control unit 510 of the relay device 500 multiplies the current PID parameters by the correction coefficient output from the output layer 303, thereby changing the PID parameters after the change. The value may be obtained and sent to the polishing device.

In the embodiment described above, the pad contact member (pad temperature control member) 11 that causes the pad surface temperature to reach a predetermined target temperature and then maintains the pad surface temperature at the target temperature contacts the surface of the polishing pad 3 (i.e., the polishing surface). are in contact. However, as shown in FIG. 18, pad contact member 11 may be spaced upward from the surface of polishing pad 3. In this case, the member designated by the reference numeral 11 in FIG. 18 functions as a pad temperature control member that adjusts the pad surface temperature in a non-contact manner, and therefore will be referred to as a "pad temperature control member" hereinafter.

Pad temperature control member 11 includes a pad heating source 11a that heats the surface of polishing pad 3 in a non-contact manner. In one embodiment, the pad temperature control member 11 may be the pad heating source 11a itself. Examples of the pad heating source 11a include a heater (especially an infrared heater) or a lamp (especially an infrared lamp) that emits radiant heat toward the surface of the polishing pad 3. When the pad heating source 11a is a heater or a lamp, the "temperature of the heating liquid" among the temperature behavior parameters described above is replaced with "temperature of the pad heating source."

Another example of the pad heating source 11a is a heated fluid injection device that injects heated fluid such as hot air, hot water, and superheated steam onto the surface of the polishing pad 3. When the pad heating source 11a is a heating fluid injection device, the heating fluid is supplied to the pad heating source 11a via a supply line (not shown). Further, among the temperature behavior parameters described above, "flow rate of heating fluid" is replaced with "injection amount of heating fluid", and "temperature of heating fluid" is replaced with "temperature of heating fluid".

Pad temperature control member 11 may further include a pad cooling source 11b that cools the surface of polishing pad 3 in a non-contact manner. In FIG. 18, the pad cooling source 11b is drawn with a virtual line (dotted line). An example of the pad cooling source 11b is a cooling fluid ejecting device that ejects cooling fluid such as cold air and cold water onto the surface of the polishing pad 3. When the pad cooling source 11b is a cooling fluid injection device, the cooling fluid is supplied to the pad cooling source 11b via a supply line (not shown). Further, among the temperature behavior parameters described above, "flow rate of cooling fluid" is replaced with "injection amount of cooling fluid", and "temperature of cooling fluid" is replaced with "temperature of cooling fluid".

Another example of the pad cooling source 11b is a coolant injection device that injects a coolant such as dry ice onto the surface of the polishing pad 3. When the pad cooling source 11b is a coolant injection device, the coolant is supplied to the pad cooling source 11b via a supply line (not shown). Further, among the temperature behavior parameters described above, "flow rate of coolant" is replaced with "injection amount of coolant", and "temperature of coolant" is replaced with "temperature of coolant".

The embodiments described above have been described to enable those skilled in the art to carry out the invention. Various modifications of the above embodiments can be naturally made by those skilled in the art, and the technical idea of the present invention can be applied to other embodiments. Therefore, the invention is not limited to the described embodiments, but is to be construed in the broadest scope according to the spirit defined by the claims.

1 Polishing head 2 Polishing table 3 Polishing pad 4 Polishing liquid supply nozzle 5 Pad temperature adjustment device 11 Pad contact member (pad temperature adjustment member)
30 Liquid supply system 31 Heated liquid supply tank 32 Heated liquid supply pipe 33 Heated liquid return pipe 39 Pad temperature measuring device 40 Control part 41 First opening/closing valve 42 First flow rate control valve 47 Heated liquid pump 48 Heating source 51 Cooling liquid supply pipe 52 Coolant discharge pipe 55 Second opening/closing valve 56 Second flow rate control valve 61 Heating channel 62 Cooling channel 110 Storage device 120 Processing device 150 Communication device 301 Input layer 302 Hidden layer 303 Output layer 500 Relay device 510 Relay control section 512 Relay storage device 600 Host processing system 610 Host control unit 612 Host storage device

Claims (20)

A pad temperature adjustment device for causing the surface temperature of a polishing pad to reach a predetermined target temperature and then maintaining it at the target temperature,
a pad contact member capable of contacting the surface of the polishing pad and having heating channels and cooling channels formed therein;
A heating liquid supply pipe connected to the heating flow path, a cooling liquid supply pipe connected to the cooling flow path, a first flow control valve attached to the heating liquid supply pipe, and a cooling liquid supply pipe connected to the cooling liquid supply pipe. a second flow control valve attached thereto, a liquid supply system for supplying temperature regulated heating liquid and cooling liquid to the pad contacting member;
a pad temperature measuring device that measures the surface temperature of the polishing pad;
a control unit that performs PID control of the operation amounts of the first flow control valve and the second flow control valve based on the difference between the measured value of the pad temperature measuring device and the target temperature,
The control unit includes:
a storage unit storing a trained model constructed by machine learning in order to maintain a temperature behavior curve created based on the measured value of the pad temperature measuring device and the measurement time point within a predetermined tolerance range; and,
a processing device that inputs a data set including at least one temperature behavior parameter into the learned model and executes an operation for outputting a changed value of the PID parameter of the PID control ,
The temperature behavior curve is a curve showing a change in the surface temperature of the polishing pad over time from the time when the pad contacting member starts adjusting the surface temperature of the polishing pad to the time when the target temperature is reached. ,
The data set is
(a) a data set consisting of the PID parameters of the PID control, the flow rate of the heating liquid, and the flow rate of the cooling liquid;
(b) a data set consisting of a PID parameter of the PID control and a film thickness parameter related to the film thickness of the substrate pressed against the polishing pad;
(c) a data set consisting of the PID parameters of the PID control, the height of the polishing pad, and the polishing rate of the substrate polished with the polishing pad;
A pad temperature adjustment device characterized by being any one of the above . The trained model is constructed by deep learning using a neural network,
The control unit includes:
When a training data set including the at least one temperature behavior parameter is input to the neural network, the neural network determines a PID parameter to be changed in order to maintain the temperature behavior curve within a predetermined tolerance range. The pad temperature adjustment device according to claim 1, wherein the learned model is constructed by adjusting weight parameters of the neural network so as to obtain a changed value of the PID parameter that is within a normal range. . The control unit includes:
accumulating the temperature behavior curve created each time the substrate is polished by pressing the substrate against the polishing pad, and at least one temperature behavior parameter associated with the temperature behavior curve;
The pad temperature adjustment device according to claim 2, wherein the learning data set is created from the accumulated at least one temperature behavior parameter. The temperature of the heating liquid, the rotation speed of the polishing head that presses the substrate against the polishing pad, the rotation speed of the polishing table to which the polishing pad is attached, the dressing conditions of the polishing pad, the polishing load of the polishing head, and the polishing pad. The pad temperature adjusting device according to claim 1 , wherein a flow rate of polishing liquid supplied to the pad temperature adjusting device is further input to the learned model. The pressing load of the pad contact member against the polishing pad, the temperature of the polishing liquid supplied to the polishing pad, the ambient temperature within the polishing unit in which the pad temperature adjustment device is arranged, the supply pressure of the heating liquid, and the cooling The pad temperature adjustment device according to claim 1 , wherein a liquid supply pressure is further input into the learned model. The pad temperature adjustment device according to claim 1 , wherein a time at which the at least one temperature behavior parameter inputted into the trained model was acquired is further inputted into the trained model. While the pad contact member is in contact with the surface of the polishing pad, a heating liquid and a cooling liquid are respectively flowed into a heating channel and a cooling channel formed in the pad contact member,
Measuring the surface temperature of the polishing pad using a pad temperature measuring device ,
a first flow control valve attached to a heated liquid supply pipe connected to the heating flow path so as to cause the surface temperature of the polishing pad to reach a predetermined target temperature and then maintain it at the target temperature; PID controlling the operation amount of a second flow control valve attached to a coolant supply pipe connected to the cooling flow path;
Constructing a trained model by machine learning to maintain a temperature behavior curve created based on the measured value of the pad temperature measuring device and the measurement time within a predetermined tolerance range,
inputting a data set including at least one temperature behavior parameter into the learned model and outputting a changed value of the PID parameter of the PID control ;
The temperature behavior curve is a curve showing a change in the surface temperature of the polishing pad over time from the time when the pad contacting member starts adjusting the surface temperature of the polishing pad to the time when the target temperature is reached. ,
The data set is
(a) a data set consisting of the PID parameters of the PID control, the flow rate of the heating liquid, and the flow rate of the cooling liquid;
(b) a data set consisting of a PID parameter of the PID control and a film thickness parameter related to the film thickness of the substrate pressed against the polishing pad;
(c) a data set consisting of the PID parameters of the PID control, the height of the polishing pad, and the polishing rate of the substrate polished with the polishing pad;
A pad temperature adjustment method characterized by : The step of constructing the trained model is a step performed by deep learning using a neural network,
The deep learning comprises: when inputting a training data set including the at least one temperature behavior parameter to the neural network, changes are made from the neural network to maintain the temperature behavior curve within a predetermined tolerance range; 8. The learned model is constructed by adjusting the weight parameters of the neural network so as to obtain the correct PID parameters and the changed values of the PID parameters that are within a normal range. How to adjust the pad temperature. further comprising the step of accumulating the temperature behavior curve created each time the substrate is polished by pressing the substrate against the polishing pad, and at least one temperature behavior parameter associated with the temperature behavior curve;
9. The pad temperature adjustment method according to claim 8 , wherein the learning data set is created from the accumulated at least one temperature behavior parameter. The temperature of the heating liquid, the rotation speed of the polishing head that presses the substrate against the polishing pad, the rotation speed of the polishing table to which the polishing pad is attached, the dressing conditions of the polishing pad, the polishing load of the polishing head, and the polishing pad. 8. The pad temperature adjustment method according to claim 7 , further comprising inputting the flow rate of the polishing liquid supplied to the learned model. The pressing load of the pad contact member against the polishing pad, the temperature of the polishing liquid supplied to the polishing pad, the ambient temperature in the polishing unit in which the polishing pad is placed, the supply pressure of the heating liquid, and the temperature of the cooling liquid. 8. The pad temperature adjustment method according to claim 7 , wherein supply pressure is further input into the learned model. 8. The pad temperature adjustment method according to claim 7 , wherein a time at which the at least one temperature behavior parameter inputted into the trained model was acquired is further inputted into the trained model. at least one polishing unit comprising a polishing table that supports a polishing pad, and a polishing head that presses a substrate against the polishing pad;
A pad temperature adjustment device for causing the surface temperature of the polishing pad to reach a predetermined target temperature and then maintaining it at the target temperature,
The pad temperature adjustment device includes:
a pad contact member capable of contacting the surface of the polishing pad and having heating channels and cooling channels formed therein;
A heating liquid supply pipe connected to the heating flow path, a cooling liquid supply pipe connected to the cooling flow path, a first flow control valve attached to the heating liquid supply pipe, and a cooling liquid supply pipe connected to the cooling liquid supply pipe. a second flow control valve attached thereto, a liquid supply system for supplying temperature regulated heating liquid and cooling liquid to the pad contacting member;
a pad temperature measuring device that measures the surface temperature of the polishing pad;
a control unit that performs PID control of the operation amounts of the first flow control valve and the second flow control valve based on the difference between the measured value of the pad temperature measuring device and the target temperature,
The control unit includes:
a storage unit storing a trained model constructed by machine learning in order to maintain a temperature behavior curve created based on the measured value of the pad temperature measuring device and the measurement time point within a predetermined tolerance range; and,
a processing device that inputs a data set including at least one temperature behavior parameter into the learned model and executes an operation for outputting a changed value of the PID parameter of the PID control ,
The temperature behavior curve is a curve showing a change in the surface temperature of the polishing pad over time from the time when the pad contacting member starts adjusting the surface temperature of the polishing pad to the time when the target temperature is reached. ,
The data set is
(a) a data set consisting of the PID parameters of the PID control, the flow rate of the heating liquid, and the flow rate of the cooling liquid;
(b) a data set consisting of a PID parameter of the PID control and a film thickness parameter related to the film thickness of the substrate pressed against the polishing pad;
(c) a data set consisting of the PID parameters of the PID control, the height of the polishing pad, and the polishing rate of the substrate polished with the polishing pad;
A polishing device characterized by being any one of the above . The trained model is constructed by deep learning using a neural network,
The control unit includes:
When a training data set including the at least one temperature behavior parameter is input to the neural network, the neural network determines a PID parameter to be changed in order to maintain the temperature behavior curve within a predetermined tolerance range. 14. The polishing apparatus according to claim 13 , wherein the learned model is constructed by adjusting weight parameters of the neural network so as to obtain a changed value of the PID parameter that is within a normal range. The control unit includes:
accumulating the temperature behavior curve created each time the substrate is polished by pressing the substrate against the polishing pad, and at least one temperature behavior parameter associated with the temperature behavior curve;
The polishing apparatus according to claim 14 , wherein the learning data set is created from the accumulated at least one temperature behavior parameter. The temperature of the heating liquid, the rotation speed of the polishing head that presses the substrate against the polishing pad, the rotation speed of the polishing table to which the polishing pad is attached, the dressing conditions of the polishing pad, the polishing load of the polishing head, and the polishing pad. 14. The polishing apparatus according to claim 13 , wherein a flow rate of the polishing liquid supplied to the polishing apparatus is further input to the learned model. The pressing load of the pad contact member against the polishing pad, the temperature of the polishing liquid supplied to the polishing pad, the ambient temperature within the polishing unit in which the pad temperature adjustment device is arranged, the supply pressure of the heating liquid, and the cooling 14. The polishing apparatus according to claim 13 , wherein a liquid supply pressure is further input to the learned model. 14. The polishing apparatus according to claim 13 , wherein a time at which the at least one temperature behavior parameter inputted into the trained model was acquired is further inputted into the trained model. at least one polishing device;
a relay device connected to the polishing device so as to be able to transmit and receive information;
a host control system connected to the relay device so as to be able to transmit and receive information;
The polishing device includes:
at least one polishing unit comprising a polishing table that supports a polishing pad, and a polishing head that presses a substrate against the polishing pad;
A pad temperature adjustment device for causing the surface temperature of the polishing pad to reach a predetermined target temperature and then maintaining it at the target temperature,
The pad temperature adjustment device includes:
a pad contact member capable of contacting the surface of the polishing pad and having heating channels and cooling channels formed therein;
A heating liquid supply pipe connected to the heating flow path, a cooling liquid supply pipe connected to the cooling flow path, a first flow control valve attached to the heating liquid supply pipe, and a cooling liquid supply pipe connected to the cooling liquid supply pipe. a second flow control valve attached thereto, a liquid supply system for supplying temperature regulated heating liquid and cooling liquid to the pad contacting member;
a pad temperature measuring device that measures the surface temperature of the polishing pad;
a control unit that performs PID control of the operation amounts of the first flow control valve and the second flow control valve based on the difference between the measured value of the pad temperature measuring device and the target temperature,
The host control unit of the host control system includes:
a storage unit storing a trained model constructed by machine learning in order to maintain a temperature behavior curve created based on the measured value of the pad temperature measuring device and the measurement time point within a predetermined tolerance range; and,
a processing device that inputs a data set including at least one temperature behavior parameter into the learned model and executes an operation for outputting a changed value of the PID parameter of the PID control ,
The temperature behavior curve is a curve showing a change in the surface temperature of the polishing pad over time from the time when the pad contacting member starts adjusting the surface temperature of the polishing pad to the time when the target temperature is reached. ,
The data set is
(a) a data set consisting of the PID parameters of the PID control, the flow rate of the heating liquid, and the flow rate of the cooling liquid;
(b) a data set consisting of a PID parameter of the PID control and a film thickness parameter related to the film thickness of the substrate pressed against the polishing pad;
(c) a data set consisting of the PID parameters of the PID control, the height of the polishing pad, and the polishing rate of the substrate polished with the polishing pad;
A polishing system characterized by being any of the following . at least one polishing device;
a relay device connected to the polishing device so as to be able to transmit and receive information;
a host control system connected to the relay device so as to be able to transmit and receive information;
The polishing device includes:
at least one polishing unit comprising a polishing table that supports a polishing pad, and a polishing head that presses a substrate against the polishing pad;
A pad temperature adjustment device for causing the surface temperature of the polishing pad to reach a predetermined target temperature and then maintaining it at the target temperature,
The pad temperature adjustment device includes:
a pad contact member capable of contacting the surface of the polishing pad and having heating channels and cooling channels formed therein;
A heating liquid supply pipe connected to the heating flow path, a cooling liquid supply pipe connected to the cooling flow path, a first flow control valve attached to the heating liquid supply pipe, and a cooling liquid supply pipe connected to the cooling liquid supply pipe. a second flow control valve attached thereto, a liquid supply system for supplying temperature regulated heating liquid and cooling liquid to the pad contacting member;
a pad temperature measuring device that measures the surface temperature of the polishing pad;
a control unit that performs PID control of the operation amounts of the first flow control valve and the second flow control valve based on the difference between the measured value of the pad temperature measuring device and the target temperature,
The relay control unit of the relay device includes:
a storage unit storing a trained model constructed by machine learning in order to maintain a temperature behavior curve created based on the measured value of the pad temperature measuring device and the measurement time point within a predetermined tolerance range; and,
A combination of the measurement value of the pad temperature measuring device, the measurement time point thereof, and a data set including at least one temperature behavior parameter is input into the trained model to adjust the surface temperature of the polishing pad. and a processing device that executes an operation for outputting a changed value of the PID parameter of the PID control at any time ,
The temperature behavior curve is a curve showing a change in the surface temperature of the polishing pad over time from the time when the pad contacting member starts adjusting the surface temperature of the polishing pad to the time when the target temperature is reached. ,
The data set is
(a) a data set consisting of the PID parameters of the PID control, the flow rate of the heating liquid, and the flow rate of the cooling liquid;
(b) a data set consisting of a PID parameter of the PID control and a film thickness parameter related to the film thickness of the substrate pressed against the polishing pad;
(c) a data set consisting of the PID parameters of the PID control, the height of the polishing pad, and the polishing rate of the substrate polished with the polishing pad;
A polishing system characterized by being any of the following .

Images