CN112445084B - Temperature control method and device of immersion lithography machine - Google Patents

Abstract

The present invention provides a temperature control method and device for an immersion lithography machine, including: exchanging heat between PCW and UPW of a lithography machine through a heat exchanger, and using a heater to heat the UPW, so as to increase the temperature of the UPW and control the temperature of the UPW. During rough adjustment, the trajectory of the UPW temperature, the flow control model of the PCW and the power control model of the heater are determined based on the detected three temperature values and the UPW temperature control target; the flow control model of the PCW and the heater are adaptively modified by fuzzy rules. The power control model is based on the power control model; the UPW after rough adjustment is passed through the remote transmission pipeline and passed into the lithography machine. After adjustment by the semiconductor refrigeration chip, it is divided into two channels of UPW injection, and the two channels of UPW injection are heated by two heaters respectively. ; When fine-adjusting the temperature of UPW, the adjustment amount of the cooling piece is determined by the power of the two heaters for two-way UPW injection, the flow rate of the two-way UPW injection and the temperature setting of the two-way UPW injection. The invention improves the temperature control precision of the lithography machine.

Description

A kind of temperature control method and device of immersion lithography machine

technical field

The invention belongs to the field of temperature control of immersion lithography machines, and more particularly relates to a temperature control method and device of immersion lithography machines.

Background technique

浸没液的温度变化将直接引起浸没液体的折射率和粘度的变化，从而导致曝光焦面偏移，引起数值孔径NA值的变化，进而使得光刻机分辨率降低和焦深的增大；另一方面，浸没液体温度变化将导致硅片和投影物镜温度变化，引起光学成像像差，最终将进一步降低浸没式光刻机的分辨率。因此，如何控制好浸液的温度并保持其稳定性是光刻机能正常工作的至关重要的因素。根据实际光刻机浸液流场温控需求，要求装置浸液温控精度达到+/-0.01℃。Immersion lithography machines are filled with immersion liquid between the last projection objective and the silicon wafer. According to the Rayleigh criterion

The temperature change of the immersion liquid will directly cause the change of the refractive index and viscosity of the immersion liquid, which will lead to the shift of the exposure focal plane and the change of the NA value of the numerical aperture, thereby reducing the resolution of the lithography machine and increasing the depth of focus; On the one hand, the temperature change of the immersion liquid will cause the temperature of the silicon wafer and the projection objective to change, causing optical imaging aberrations, which will ultimately further reduce the resolution of the immersion lithography machine. Therefore, how to control the temperature of the immersion liquid and maintain its stability is a crucial factor for the normal operation of the lithography machine. According to the actual temperature control requirements of the immersion liquid flow field of the lithography machine, the temperature control accuracy of the immersion liquid of the device is required to reach +/-0.01℃.

Chinese patent application 201020596742.2 describes an immersion liquid temperature control device for an immersion lithography machine. The thermoelectric cooling mechanism can ensure the temperature stability requirements of the immersion liquid flow field, and measure the temperature characteristics of the immersion liquid in real time. In the actual lithography machine, the immersion lithography machine has extremely high requirements on the immersion liquid. Generally, deionized and degassed ultrapure water is used, and the thermoelectric refrigeration mechanism is used to control the temperature of the immersion liquid, which is not conducive to the pollution control of the immersion liquid. Chinese patent application 201020596742.2 uses the heat exchange principle to control the temperature of the immersion liquid with a heat exchanger made of PFA and a flow servo valve, which reduces the pollution of the immersion liquid; and the patent 201020596742.2 mainly controls the temperature of the immersion liquid through TCU, The end immersion liquid lacks the ability of secondary temperature control and the flexibility of temperature control.

Chinese patent application 201010143659.4 describes a high-precision temperature control device and a parameter self-tuning method thereof. The provided device and method solve the matching problem of temperature control parameters and actual working conditions. The temperature control method of the device is one of the main temperature control methods of the previous generation lithography machine. This method has high requirements on the measurement of temperature control parameters, and the setting time is too long.

SUMMARY OF THE INVENTION

In view of the defects of the prior art, the purpose of the present invention is to provide a temperature control method and device for an immersion lithography machine, which aims to solve the problem that the existing lithography machine temperature control method and device lack secondary temperature control for the end immersion liquid. capacity, the lack of flexibility in temperature control, and the high requirements for the measurement of temperature control parameters, and the problem of too long settling time.

In order to achieve the above purpose, in a first aspect, the present invention provides a temperature control method for an immersion lithography machine, comprising the following steps:

The cooling water PCW of the plant is made to exchange heat with the ultrapure water UPW of the lithography machine through a heat exchanger. The temperature of the PCW is lower than the temperature of the UPW to cool the UPW, and the heater is used to heat the UPW to cool the UPW. Raise the temperature, adjust the UPW temperature roughly, control the flow rate of PCW and the power of the heater so that the UPW temperature can quickly reach the UPW temperature control target, and maintain it stably;

When the UPW temperature is roughly adjusted, the temperature before the heat exchange of the PCW, the temperature before the heat exchange of the UPW, and the temperature of the UPW after the heat exchange and heater control are detected, and the UPW temperature is controlled based on the detected three temperature values and the UPW temperature. The target determines the trajectory of the UPW coarse adjustment target temperature, the flow control model of the PCW and the power control model of the heater; and the temperature fluctuation before the heat exchange of the PCW and the temperature fluctuation before the heat exchange of the UPW are detected according to the fuzzy rules. Modify the flow control model of PCW and the power control model of heater;

The UPW after rough adjustment is passed through the remote transmission pipeline and passed into the interior of the lithography machine. After being adjusted by the semiconductor refrigeration chip, it is divided into two channels of horizontal UPW injection and vertical UPW injection, and the two channels of UPW injection are respectively injected through two heaters. After heating, the two channels of UPW injection finally meet in the main field area of the lithography machine; inside the lithography machine, the temperature of the UPW is finely adjusted by the semiconductor refrigeration chip and the heater of the two channels of UPW injection to compensate for the remote transmission tube. The environmental error introduced by the circuit and the thermal interference caused by the UPW fluid control components make the temperature of the UPW in the main field of the lithography machine stable at the UPW temperature control target; The set value of the two-way UPW injection is determined by the power of the two heaters for the two-way UPW injection, the flow rate of the two-way UPW injection and the temperature set value of the two-way UPW injection. The temperature set value of the two-way UPW injection is determined by the UPW The target and the two-way UPW injection flow rate are determined.

In an optional embodiment, the UPW temperature control target is set by the working requirements of the main field of the lithography machine.

In an optional embodiment, the rough adjustment of the UPW temperature specifically includes the following steps:

S1, to avoid sharp changes in the input of UPW temperature setting value and the output of UPW temperature control value during the rough adjustment control process, and use the relaxation factor α to soften the trajectory of the UPW rough adjustment target temperature:

y _r (k+i)=(1-α ⁱ )*SV+α ⁱ y(k)

Among them, α∈(0,1), SV is the UPW temperature control target, y _r (k+i) is the UPW temperature set value at the ith sampling period, and y(k) is the current UPW temperature control output;

由P个零输入响应

叠加零状态响应A_jΔU_M，j(k)得到：S2, the predicted value of P sampling periods in the prediction domain

Response by P zero inputs

Superimposing the zero-state response A _j ΔU _M,j (k) yields:

为预测模型得出的第k+j个输出值，

为第k+j个时刻的零状态值，Δu_j(k)为对应执行器在k时刻的控制量，A_j为对应执行器单位阶跃响应模型，ΔU_M，j(k)为对应执行器控制量的变化量，A为对应执行器单位阶跃响应模型，ΔU_M(k)为所有执行器控制量的变化量；j∈1，2；in,

is the k+jth output value derived from the prediction model,

is the zero state value at the k+jth time, Δu _j (k) is the control amount of the corresponding actuator at time k, A _j is the unit step response model of the corresponding actuator, ΔU _{M, j} (k) is the corresponding execution is the change of the control variable of the actuator, A is the step response model of the corresponding actuator unit, ΔU _M (k) is the change of the control variable of all the actuators; j∈1,2;

In the rough adjustment, the heat exchanger and the heater are set, and the two control variables act together on the temperature of the feedback point of the rough adjustment module. At this time, j=1, 2 respectively represents the two actuators of the heat exchanger servo valve and the heater. The actuator A _j matrix is the corresponding actuator unit step response model;

Under the action of the control quantity of M sampling periods whose control domain is equal to M, the predicted value of P sampling periods in the prediction domain is:

S3, after the prediction model is constructed, the control quantity is solved, and the objective function J is expressed as:

Among them, R and Q are the weight matrices considering the tracking effect and the actuator output, respectively, and W(k) is all the set values y _r (k+i) i=1, 2, ..., P in the prediction domain P; ΔU _M ^T (k) is the transpose of ΔU _M (k);

Use the gradient descent method to obtain the change of the control variable when the objective function is the smallest;

可求得：make

Available:

u _j (k)=u _j (k-1)+Δu _j (k) j=1, 2

Among them, Δu _j (k) is the variation of the control amount of the corresponding actuator at time k, d _j is a matrix composed of known A _j , R and Q, u _j (k) corresponds to the control amount of the actuator, and finally solved The control amount u _j (k) of each actuator at each moment;

S4, the continuous sampling correction of the prediction error is equivalent to the superposition of impulse effects; the perturbation of the model parameters comes from the temperature fluctuations of the input UPW and PCW, and fuzzy rules are used to use the filtered PCW temperature deviation Ep and UPW temperature deviation Eu as the fuzzy The input of the rules determines the model parameters, and the transfer function of the change of the heat exchanger servo valve flow rate on the output temperature is simplified to a first-order time-delay model;

Sliding filtering is performed on the sampled PCW and UPW, and the filtering adopts the average of nearly ten sampling values;

Determine the difference between the filtered Ep and Eu at the current sampling time and the PCW temperature difference Ep0 and UPW temperature difference Eu0 when the model parameter adjustment was triggered last time, and judge whether it exceeds the limit deviation Em; if it exceeds Em, update the model parameters according to the fuzzy rules, and Record the values of Ep and Eu this time and assign them to Ep0 and Eu0. Otherwise, no fuzzy rule adjustment is performed, and Ep0 and Eu0 are initialized to 0.

In an optional embodiment, when the difference between the roughly adjusted UPW temperature and the set value is less than the preset value Ec, the fine adjustment of the UPW is started;

The ambient temperature affects the temperature of the remote transmission line, and the semiconductor refrigeration chip controls the temperature fluctuation of the UPW caused by the increase or decrease of the ambient temperature. And the heating power of the servo valve to adjust the flow to determine;

The two UPW injection flow rates are Q ₁ and Q ₂ respectively; the relationship between the temperature rise and the flow rate caused by the heating of the servo valve is determined by prior experiments as f(Q), and the relationship between the temperature rise and flow rate corresponding to the rated power of the heater is known is H(Q), ε is the safety factor to ensure system redundancy;

The upper limit r _ymax of the UPW temperature setting value ry interval after the fine adjustment of the semiconductor refrigeration sheet is adjusted is: r _{ymax =} min(r _z1 -εf(Q ₁ ),r _z2 -εf(Q ₂ ));

The lower limit r _ymin must ensure that the output temperature under the rated power of the heater is not less than the set values r _z1 and r _z2 of the two-way UPW liquid injection temperature;

r _ymin =max(r _z1 -εH(Q ₁ ),r _z2 -εH(Q ₂ ))

r _y ∈[r _ymin ,r _ymax ];

Taking into account the thermal interference of the servo valve and the heater power range, the average value of the adjustable range is taken as the set value of the UPW temperature after the semiconductor refrigeration chip is adjusted;

The two-way UPW liquid injection temperature setting value needs to be decomposed by combining the actual liquid injection flow rate Q ₁ , Q ₂ and the UPW temperature control target SV:

r _z1 =T ₁ (SV, Q ₁ )

r _z2 = T ₂ (SV, Q ₂ )

Among them, r _z1 =T ₁ (SV, Q ₁ ), r _z2 =T ₂ (SV, Q ₂ ) are the set values r _z1 , r _z2 and SV of the two-way UPW injection temperature obtained through prior knowledge, respectively and the relational expression between each flow Q ₁ , Q ₂ ;

UPW fine-tuning controllers all use PID controllers to complete the tracking adjustment of the set value.

In an optional embodiment, the simplified first-order time-delay model of the transfer function of the effect of the flow change of the servo valve on the output temperature is:

为修正后的伺服阀控制阶跃响应矩阵，A₁为初始化的伺服阀控制的阶跃响应矩阵。Among them, K is the gain of the initialization state, δ is the multiple of the initial gain with the disturbance, T is the inertia time, τ is the pure delay, s is the Laplace operator, and G(s) is the change of the servo valve flow to the output Transfer function model of temperature effect,

is the corrected step response matrix of servo valve control, A ₁ is the step response matrix of initial servo valve control.

In a second aspect, the present invention provides a temperature control device for an immersion lithography machine, comprising: a coarse adjustment module and a fine adjustment module;

The rough adjustment module is placed in the external factory of the lithography machine, and includes four fluid interfaces, namely: PCW input interface in the factory cooling water PCW circulation loop, PCW output interface in the PCW circulation loop, and ultrapure water UPW circulation loop. The UPW input interface and the interface of the UPW remote transmission insulation pipeline; the external factory services of the lithography machine include: PCW circulation loop, UPW circulation loop and UPW remote transmission insulation pipeline; the UPW flowing out of the UPW circulation loop passes through the UPW remote The thermal insulation pipeline flows to the lithography machine;

The UPW recovered from the lithography machine flows into the UPW circulation loop, the temperature of the PCW in the PCW circulation loop is adjustable, and the PCW circulation loop and the UPW circulation loop conduct heat exchange through a heat exchanger; Heat exchange with the ultrapure water UPW of the lithography machine, the temperature of the PCW is lower than the temperature of the UPW to cool the UPW, and the heater is used to heat the UPW to heat the UPW, and the temperature of the UPW is roughly adjusted and controlled. The flow rate of the PCW and the power of the heater make the temperature of the UPW quickly reach the vicinity of the temperature control target of the UPW and maintain it stably; when the temperature of the UPW is roughly adjusted, the temperature before the heat exchange of the PCW, the temperature before the heat exchange of the UPW, and the temperature of the heat The temperature of the UPW after exchange and heater control is determined, based on the detected three temperature values and the UPW temperature control target, determining the trajectory of the UPW coarse adjustment target temperature, the flow control model of the PCW and the power control model of the heater; The temperature fluctuation before heat exchange of PCW and the temperature fluctuation before heat exchange of UPW are used to adaptively modify the flow control model of PCW and the power control model of heater by using fuzzy rules;

The fine adjustment module is placed inside the lithography machine and includes three fluid interfaces, namely: the interface for receiving the UPW remote transmission insulation pipeline of the UPW and the interface for the two-way UPW injected into the main field of the lithography machine; The UPW adjusted by the adjustment module passes through the remote transmission pipeline and passes into the interior of the lithography machine. After being adjusted by the semiconductor refrigeration chip, it is divided into two channels of horizontal UPW injection and vertical UPW injection, and the two channels of UPW injection are respectively injected through two heaters. After heating, the two channels of UPW injection finally meet in the main field area of the lithography machine; inside the lithography machine, the temperature of the UPW is finely adjusted by the semiconductor refrigeration chip and the heater of the two channels of UPW injection to compensate for the remote transmission tube. The environmental error introduced by the circuit and the thermal interference caused by the UPW fluid control components make the temperature of the UPW in the main field of the lithography machine stable at the UPW temperature control target; The set value of the two-way UPW injection is determined by the power of the two heaters for the two-way UPW injection, the flow rate of the two-way UPW injection and the temperature set value of the two-way UPW injection. The temperature set value of the two-way UPW injection is determined by the UPW The target and the two-way UPW injection flow rate are determined.

In an optional embodiment, the UPW temperature control target is set by the working requirements of the main field of the lithography machine.

In an optional embodiment, the rough adjustment module performs rough adjustment on the UPW temperature, which specifically includes the following steps:

S1, to avoid sharp changes in the input of UPW temperature setting value and the output of UPW temperature control value during the rough adjustment control process, and use the relaxation factor α to soften the trajectory of the UPW rough adjustment target temperature:

y _r (k+i)=(1-α ⁱ )*SV+α ⁱ y(k)

Among them, α∈(0,1), SV is the UPW temperature control target, y _r (k+i) is the UPW temperature set value at the ith sampling period, and y(k) is the current UPW temperature control output;

由P个零输入响应

叠加零状态响应A_jΔU_M，j(k)得到：S2, the predicted value of P sampling periods in the prediction domain

Response by P zero inputs

Superimposing the zero-state response A _j ΔU _M,j (k) yields:

为预测模型得出的第k+j个输出值，

为第k+j个时刻的零状态值，Δu_j(k)为对应执行器在k时刻的控制量，A_j为对应执行器单位阶跃响应模型，ΔU_M，j(k)为对应执行器控制量的变化量，A为对应执行器单位阶跃响应模型，ΔU_M(k)为所有执行器控制量的变化量；j∈1，2；in,

is the k+jth output value derived from the prediction model,

is the zero state value at the k+jth time, Δu _j (k) is the control amount of the corresponding actuator at time k, A _j is the unit step response model of the corresponding actuator, ΔU _{M, j} (k) is the corresponding execution is the change of the control variable of the actuator, A is the step response model of the corresponding actuator unit, ΔU _M (k) is the change of the control variable of all the actuators; j∈1,2;

In the rough adjustment, the heat exchanger and the heater are set, and the two control variables act together on the temperature of the feedback point of the rough adjustment module. At this time, j=1, 2 respectively represents the two actuators of the heat exchanger servo valve and the heater. The actuator A _j matrix is the corresponding actuator unit step response model;

Under the action of the control quantity of M sampling periods whose control domain is equal to M, the predicted value of P sampling periods in the prediction domain is:

S3, after the prediction model is constructed, the control quantity is solved, and the objective function J is expressed as:

Among them, R and Q are the weight matrices considering the tracking effect and the actuator output, respectively, and W(k) is all the set values y _r (k+i) i=1, 2, ..., P in the prediction domain P; ΔU _M ^T (k) is the transpose of ΔU _M (k);

Use the gradient descent method to obtain the change of the control variable when the objective function is the smallest;

可求得：make

Available:

u _j (k)=u _j (k-1)+Δu _j (k) j=1, 2

Among them, Δu _j (k) is the variation of the corresponding actuator control amount, d _j is a matrix composed of known A _j , R and Q, u _j (k) corresponds to the actuator control amount, and finally each moment is solved The control quantity u _j (k) of each actuator;

S4, the correction link includes the continuous sampling correction of the prediction error and the correction of the control model. The correction of the prediction error is to use the error of the predicted output and the actual output as the correction reference for the next prediction, which is equivalent to the superposition of the influence of the pulse; Correction of the control model Considering that the perturbation of the model parameters mainly comes from the temperature fluctuations of the input UPW and PCW, the fuzzy rules are used to determine the model parameters with the filtered PCW temperature deviation Ep and UPW temperature deviation Eu as the input of the fuzzy rules. The transfer function of the effect of servo valve flow change on output temperature is simplified to a first-order time-delay model;

Sliding filtering is performed on the sampled PCW and UPW, and the filtering adopts the average of nearly ten sampling values;

Determine the difference between the filtered Ep and Eu at the current sampling time and the PCW temperature difference Ep0 and UPW temperature difference Eu0 when the model parameter adjustment was triggered last time, and judge whether it exceeds the limit deviation Em; if it exceeds Em, update the model parameters according to the fuzzy rules, and Record the values of Ep and Eu this time and assign them to Ep0 and Eu0. Otherwise, no fuzzy rule adjustment is performed, and Ep0 and Eu0 are initialized to 0.

其中，r_z1＝T₁(SV，Q₁)、r_z2＝T₂(SV，Q₂)分别为通过先验知识得到的两路UPW注液温度的设定值r_z1、r_z2与SV和各路流量Q₁、Q₂之间的关系式；UPW精调节的控制器皆采用PID控制器完成对设定值的跟踪调节。In an optional embodiment, when the difference between the roughly adjusted UPW temperature and the set value is less than the preset value Ec, the fine adjustment module starts to finely adjust the UPW; the ambient temperature affects the remote transmission pipeline. Influenced by temperature, the semiconductor refrigeration chip in the fine adjustment module controls the UPW temperature fluctuation caused by the increase or decrease of the ambient temperature. The heating power of the servo valve is determined by the heating power of the servo valve; the two-way UPW injection flow rates are Q1 and Q2 respectively; the relationship between the temperature rise and the flow rate caused by the heating of the servo valve is determined by a priori experiment as f(Q), and it is known that the rated power of the heater corresponds to the temperature The relationship between liter and flow rate is H(Q), and ε is the safety factor to ensure system redundancy; the upper limit r _ymax of the UPW temperature setting value ry interval after fine-tuning semiconductor refrigeration chip adjustment is: r _{ymax =} min(r _z1 -εf(Q ₁ ),r _z2 -εf(Q ₂ )); the lower limit of r _ymin should ensure that the output temperature under the rated power of the heater is not less than the set value of the two-way UPW injection temperature r _z1 , r _z2 ;r _ymin =max(r _z1 -εH(Q ₁ ),r _z2 -εH(Q ₂ )); r _y ∈[r _ymin ,r _ymax ]; considering the thermal interference of the servo valve and the heater power range, it can be taken as The average value of the adjustment interval is used as the set value of the UPW temperature after adjustment of the semiconductor refrigeration chip; the set value of the two-way UPW liquid injection temperature needs to be decomposed by combining the actual liquid injection flow Q1, Q2 and the UPW temperature control target SV:

Among them, r _z1 =T ₁ (SV, Q ₁ ), r _z2 =T ₂ (SV, Q ₂ ) are the set values r _z1 , r _z2 and SV of the two-way UPW injection temperature obtained through prior knowledge, respectively The relational expression between Q ₁ and Q ₂ of each channel; UPW fine adjustment controllers all use PID controllers to complete the tracking adjustment of the set value.

In an optional embodiment, the simplified first-order time-delay model of the transfer function of the influence of the flow rate change of the servo valve on the output temperature in the rough adjustment module is:

为经模糊规则确定增益δ修正之后的伺服阀阶跃响应模型，A₁为初始伺服阀阶跃响应模型。Among them, K is the gain of the initialization state, δ is the multiple of the initial gain with the disturbance, T is the inertia time, τ is the pure delay, s is the Laplace operator, and G(s) is the change of the servo valve flow to the output Transfer function model of temperature effect,

is the step response model of the servo valve after the gain δ is corrected by the fuzzy rule, and A ₁ is the initial step response model of the servo valve.

In general, compared with the prior art, the above technical solutions conceived by the present invention have the following beneficial effects:

The invention provides a temperature control method and device for an immersion lithography machine. The observed UPW and PCW disturbance data are used as input, and are transformed into the gain change of a first-order inertial time-delay control model through fuzzy rules, so as to correct the prediction The model parameters of servo valve-heat exchanger and heater control in the control, realize dual-actuator control and overcome the influence of PCW and UPW temperature fluctuations on the target temperature of the coarse adjustment module, which can ensure temperature control in the case of small overshoot Accuracy and stability and robustness of the system.

Description of drawings

1 is a flowchart of a temperature control method for an immersion lithography machine provided by an embodiment of the present invention;

2 is a schematic structural diagram of the principle of a temperature control device for an immersion lithography machine provided by an embodiment of the present invention;

3 is a block diagram of a temperature control system for an immersion lithography machine provided by an embodiment of the present invention;

FIG. 4 is a membership function diagram of Eu for δ adaptive adjustment provided by an embodiment of the present invention;

Fig. 5 is the membership function diagram of the Ep of the δ adaptive adjustment provided by the embodiment of the present invention;

6 is a flow chart of temperature control of an immersion lithography machine provided by an embodiment of the present invention;

FIG. 7 is a flow chart of predictive control of model parameter adaptation provided by an embodiment of the present invention;

In all drawings, the same reference numerals are used to denote the same elements or structures, wherein: 100 is a coarse adjustment module, 200 is a fine adjustment module, 1 is a heat exchanger, 2 is a flow servo valve 1, and 3 is a heating Device 1, 4 are circulating pumps, 5 is temperature sensor 1, 6 is temperature sensor 2, 7 is microprocessor 1, 8 is microprocessor 2, 9 is temperature sensor, 10 is remote heat preservation pipeline, 11 is semiconductor Refrigeration plate, 12 is buffer tank, 13 is flow servo valve 2, 14 is flow servo valve 3, 15 is heater 2, 16 is heater 3, 17 is host computer, 18 is temperature sensor 3, 19 is temperature sensor 4 , 20 is the temperature sensor 5 .

Detailed ways

In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention.

To this end, the present invention provides an effective method for controlling the temperature of the immersion liquid, utilizing multiple heat exchangers and multi-stage servo flow control to achieve a steady-state accuracy of +/- 0.01°C.

Specifically, FIG. 1 is a flowchart of a temperature control method for an immersion lithography machine provided by an embodiment of the present invention; as shown in FIG. 1 , the following steps are included:

S101, through a heat exchanger, the factory cooling water PCW is made to exchange heat with the ultrapure water UPW of the lithography machine, and the temperature of the PCW is lower than the temperature of the UPW, so as to cool the UPW, and use a heater to heat the UPW, so as to reduce the temperature of the UPW. Raise the temperature of the UPW, adjust the temperature of the UPW roughly, and control the flow of the PCW and the power of the heater so that the temperature of the UPW can quickly reach the UPW temperature control target and maintain it stably;

S102, when roughly adjusting the temperature of the UPW, detect the temperature before the PCW performs heat exchange, the temperature before the UPW performs the heat exchange, and the temperature of the UPW after the heat exchange and heater control, based on the detected three temperature values and the UPW The temperature control target determines the trajectory of the UPW coarse adjustment target temperature, the flow control model of the PCW, and the power control model of the heater; and based on the detected temperature fluctuations before the PCW performs heat exchange and UPW before heat exchange, using fuzzy rules Adaptively correct the flow control model of the PCW and the power control model of the heater;

S103, pass the coarsely adjusted UPW into the lithography machine through the remote transmission pipeline, and then be divided into two channels of horizontal UPW injection and vertical UPW injection after adjustment by the semiconductor refrigeration chip, and the two channels of UPW are respectively injected by the two heaters. The liquid injection is heated, and finally the two UPW liquid injections meet in the main field area of the lithography machine; inside the lithography machine, the temperature of the UPW is finely adjusted by the semiconductor cooling chip and the two-way UPW liquid injection heater to compensate for the distance. The environmental error introduced by the transmission line and the thermal interference caused by the UPW flow control components make the temperature of the main field UPW of the lithography machine stable at the UPW temperature control target; The set value of the UPW temperature is determined by the power of the two heaters for the two-way UPW injection, the flow rate of the two-way UPW injection and the temperature set value of the two-way UPW injection. The temperature set value of the two-way UPW injection is determined by the The UPW target and the two-way UPW injection flow rate are determined.

For the specific control method, refer to the detailed introduction in the following embodiments:

In a specific embodiment, as shown in FIG. 2 , the present invention provides an immersion liquid temperature control device of an immersion lithography machine, so that the immersion liquid input into the immersion unit can realize decoupling under different flow rates and different input temperature conditions. Fixed value, independent control of grading and sub-module control. The wide-area temperature adjustment is realized in the coarse adjustment module of SUBFAB, and the specificity of the two-way injection is independently controlled in the fine adjustment module inside the FAB lithography machine to achieve high temperature stability control. The device can achieve temperature control accuracy within ±0.0025°C under steady state conditions.

Due to the limited space of the immersion lithography machine, the device of the present invention is divided into two parts: a coarse adjustment module 100 and a fine adjustment module 200 . The coarse adjustment module is placed in the SUBFAB, as shown in Figure 2. The fine adjustment module is placed inside the lithography machine. As shown in Figure 2, the coarse adjustment module has four fluid interfaces, including the input PCW IN of the factory cooling water PCW, and the output interface PCW OUT. The refrigerant PCW used for temperature control enters and exits from the pair of inlets and outlets. The input temperature range is in 10℃～18℃, the ultrapure water UPW input interface UPW IN of the object whose temperature is controlled, the input temperature is 20℃～24℃, and the ultrapure water UPW remote transmission heat preservation between the two modules leading to the fine adjustment module 200 Line 10. The fine adjustment module 200 includes three fluid interfaces, receiving the UPW inlet of the coarse adjustment module, and the two-way injection UPW OUT1, UPW OUT2 for injecting into the submerged head.

The coarse adjustment module 100 is located in the SUBFAB (unclean area), and the module includes: adjusting the flow servo valve 2 to control the PCW flow leading to the heat exchanger 1, and using the heat exchanger 1 to control the temperature of pure water; The power is used to provide the heating capability to control the temperature of the ultrapure water, and the circulating pump 4 forms a loop for the controlled UPW, which further improves the temperature uniformity in the pipeline and improves the temperature stability. The temperature sensor 5 is used to collect the UPW temperature output by the coarse adjustment module, which is used for trajectory softening and feedback correction in this stage of temperature control. The model parameters are optimized. MCU7 is used as the microprocessor of the coarse adjustment module to collect the temperature signals from the temperature sensors 5/6 and 9 with the help of the built-in A/D module, and to solve the adjustment amount of the servo valve and the heater through the controller. The built-in D/A module is transmitted to the servo valve and heater, which plays the role of temperature regulation.

The fine adjustment module 200 is located inside the lithography machine in the FAB environment, and is used for the last temperature control to provide the immersion head with ultrapure water that conforms to the temperature gradient and stability of the main field. The semiconductor refrigeration sheet (TEC) 11 is used to control the temperature rise or fall of the UPW transmitted through the remote heat preservation pipeline 10 from the coarse adjustment module. The buffer tank 12 is used to suppress the high-frequency disturbance of the UPW and further improve the temperature stability. The flow servo valve 13/14 is used to control the flow of the two-way injection to meet the flow stability requirements of the submerged head. The heaters 15/16 are used to adjust the temperature of the two-way injection. The temperature sensor 18 is used for the first-stage temperature control of the module, and provides feedback signals for feedback control. Temperature sensors 19/20 respectively provide feedback signals for two-way injection temperature control. The microprocessor MCU8 is used to collect the temperature sensor signal, and through the two-stage PID controller, the power of the corresponding TEC and the power of the heater are calculated through the D/A module to convert the analog-digital signal. The device provides a visualization device, the upper computer 17 can communicate with the MCU7/8 of the two modules through the TCP/IP protocol, and the upper computer 17 can provide functions such as visual operation.

The temperature control device of the immersion lithography machine adopts hierarchical control, and independent processors are used to control the modules independently of each other, which is conducive to fault diagnosis and provides a visual operation interface. The coarse adjustment module utilizes the coordinated control of the heater and the servo valve to simultaneously realize the heating and cooling control of the temperature of the immersion liquid UPW. The TEC module in the fine adjustment module can realize heating and cooling control at the same time, so that the temperature of the UPW that is affected by the environment through the remote pipeline can be controlled, and it can adapt to the compact space inside the lithography machine. Temperature fluctuations are hardware filtered to improve stability. The last two-stage liquid injection device adopts the actuator temperature control combined with flow control to take into account the flow parameters, so that the average temperature and stability of the final injection into the immersion head are closer to the final needs.

The temperature control device uses a heat exchanger to control the cooling temperature of the UPW immersion liquid. Once the temperature overshoot occurs, the temperature can only be increased by the UPW itself. And in order to adapt to the input UPW, the temperature is higher or lower than the set value, and to improve the temperature control range of the system, so the servo valve is used to control the PCW flow through the heat exchanger heat exchange series heater to directly heat, while providing temperature rise. and the ability to control cooling. Different from the multi-closed-loop control or hierarchical control of multi-control variables, the system is based on predictive control. The hardware only uses a temperature sensor at the end of the coarse adjustment module as the feedback point, and the algorithm only uses the feedback variable as the input of the controller. The cost and complexity of the system is reduced, while the control of the coarse temperature adjustment module with improved predictive control enables coordinated control of dual actuators. The main flow field of the immersion head puts forward extremely stringent requirements on the temperature accuracy and stability of the temperature sensor 5. This temperature control device needs to ensure the predictive control model in the case of external disturbances such as the UPW and PCW input temperature fluctuations collected by the temperature sensor 6/9. accuracy and temperature control stability.

Specifically, the present invention provides an immersion temperature control method, which converts the observed external disturbance data into model parameters of predictive control through fuzzy rules, so as to realize dual-actuator control and overcome the rough adjustment caused by PCW and UPW temperature fluctuations. The influence of the target temperature of the module can ensure the accuracy and stability of temperature control and the robustness of the system in the case of small overshoot.

The fine adjustment module provided by the embodiment of the present invention is a two-stage temperature control method. The two-stage temperature control method is to realize the temperature at the final control end temperature sensors 19 and 20 in two stages: the primary temperature control adopts PI control to ensure the rapid response capability of the control system; the secondary temperature control adopts stable PI control to ensure control. accuracy of the system. In order to allow the actuator of the control system to have a wide adjustable space, the set value and the first-stage temperature of the last two-way injection temperature control are determined in combination with the prior knowledge of the injection flow, the heating power of the servo valve and the temperature change. control setpoint.

The block diagram of the control system of the coarse adjustment module and the fine adjustment module is shown in Figure 3, where:

The model parameter adaptive predictive control in the coarse adjustment module is used to control the output temperature of the coarse adjustment module with the prediction control as the algorithm framework in order to realize the dual actuator control of the heater and the servo valve. As shown in control block diagram 3, it includes the following steps:

X001: Soften track

y _r (k+i)=(1-α ⁱ )*SV+α ⁱ y(k)

Avoid sharp changes in the set value input and control value output during the control process, and use the relaxation factor α to soften the trajectory of the UPW coarse adjustment target temperature, α ∈ (0,1), the temperature control system of the lithography machine pays attention to stability The larger the α, the better the stability and the slower the adjustment. Take α=0.8; SV is the final temperature target value, y _r (k+i) is the temperature setting value in the ith sampling period, y(k) is the current output temperature value.

After the target trajectory is determined, an improved dynamic matrix control method is used to track and control the target trajectory. The parameters of the improved part of the dynamic matrix will be adaptively modified with the observation of temperature disturbance.

X002 prediction model:

The heat exchanger and heater models in the current process model can be identified by the step response response and can be stabilized after the step signal is input for N sampling periods, so N is used as the modeling time of the dynamic matrix control.

可以由P个零输入响应

叠加零输入响应A_jΔU_M,j(k)得到：So the predicted value of the prediction domain P sampling periods

can be responded by P zero inputs

Superimposing the zero input response A _j ΔU _M,j (k) yields:

Combined with the system in this paper, two actuators, a cooling heat exchanger and a heater, are set in the temperature coarse adjustment module. The two control variables work together with the temperature of the feedback point of the coarse adjustment module. At this time, j=1 and 2 represent the servo valve and the heater respectively. The A _j matrix of the two actuators is the unit step response model of the actuator, and ΔU _M,j is the variation of the actuator control amount.

Therefore, under the action of the control amount of M sampling periods whose control domain is equal to M, the obtained predicted value of P sampling periods in the prediction domain can be obtained as follows.

X003 scrolling optimization:

After the prediction model is built, the control variables are solved. Considering the tracking effect of the target temperature value, it is necessary to minimize the control deviation in the prediction domain P and consider that the actuator variation amplitude should not be too large to cause impact. The objective function J can be expressed as:

where Q and R are the weight matrices considering the tracking effect and actuator fluctuations, respectively, and W(k) is all the set values in the prediction domain P. The gradient descent method is used to obtain the change of the control variable when the objective function is the smallest.

可求得：make

Available:

D=L(A ^T QA+R) ^-1 A ^T Q

θ=[1,0,...,0]

u _j (k)=u _j (k-1)+Δu _j (k) j=1, 2

Among them, D represents the set of d _j , L represents the transformation matrix composed of θ, θ represents the control quantity of the sampling period to be solved, and finally the control quantity u _j (k) of each actuator at each moment is solved. .

X004 Error correction

The feedback correction link of predictive control often compensates the prediction error directly, that is, the prediction error obtained after sampling at this moment is regarded as the impulse excitation to the prediction model, and the prediction value of P prediction cycles after this moment is corrected; Correction is different from compensating and correcting prediction errors. Optimizing the model can better reflect the compensation of prediction errors from the mechanism analysis. This paper states that the system adopts the method of compensating the prediction error and superimposing the parameter correction of the prediction model to deal with the prediction error and model perturbation.

X004_A Prediction Error Correction

为k时刻预测的k+1时刻的输出值。The continuous sampling correction of the prediction error is equivalent to the superposition of the influence of the pulse. The specific formula is composed as: e(k+1) is the prediction error at the next moment, y(k+1) is the output value at the moment k+1,

is the predicted output value at time k+1 at time k.

为：So the output predicted value of P sampling periods after correction

for:

The error correction matrix h is the influence matrix of the zero input response on the future P modeling cycles, which can be obtained by transforming the unit step response A _j above. After the prediction model correction and updating A _j , the error correction matrix h also follows Change. further:

可由

乘系数矩阵S得到：Therefore, after the correction, the zero input response of P sampling periods of the entire prediction domain at the next iteration is performed.

by

Multiply the coefficient matrix S to get:

X004_B Prediction Model Correction

Different semiconductor factories have different business conditions, and the fluctuations of the supplied UPW and PCW are different. The existing conditions are that the UPW temperature fluctuates between 20°C and 24°C and the PCW temperature fluctuates between 10°C and 18°C. The ideal condition is that the UPW is stable. Around 22°C, PCW is stable around 14°C. The analysis of the system shows that the perturbation of the model parameters comes from the temperature fluctuations of the input UPW and PCW. The analysis of the heat exchanger model shows that the temperature of the UPW and PCW affects the gain link in the transfer function. The step response model and the transfer function are consistent in the time domain characteristics. The adaptive parameter δ is used to characterize the influence of fluctuations in the UPW temperature between 20 °C and 24 °C and the PCW temperature between 10 °C and 18 °C on the step response model of the servo valve + heat exchanger model. The effect is monotonic but nonlinear. The higher the PCW temperature, the smaller the gain, the higher the UPW temperature, the larger the gain, and there are other small flow fluctuations such as input PCW and UPW interference, so it is impossible to accurately determine the model according to the linear correspondence in actual control. parameter. Therefore, the X003 fuzzy rule is adopted, and the PCW temperature deviation Ep and the UPW temperature deviation Eu obtained by filtering are used as the input of the fuzzy rule, and the model parameters are corrected for these two main disturbances. The transfer function that controls the flow of the servo valve through the heat exchanger for the effect on the output temperature can be simplified to a first-order time-delay model, where K is the gain in the initialization state, δ is the multiple of the initial gain changing with the disturbance, T is the inertia time, τ is the pure delay. As follows:

Taking the UPW input temperature of 22°C and PCW of 14°C as the initialized reference value, Ep(k) and Eu(k) are the difference between the currently sampled UPW temperature, the PCW temperature and the initialized UPW temperature value of 22°C, and the PCW initialization temperature of 14°C respectively. value, perform sliding filtering on the sampled PCW and UPW, and the filtering adopts the average of nearly ten sampled values:

Initialize Ep0 and Eu0 to 0. Determine the difference between the filtered Ep and Eu at the current sampling time and the Ep0 and Eu0 when the model parameter adjustment was triggered last time, and judge whether it exceeds the limit deviation Em; if it exceeds Em, update the model parameters according to the following fuzzy rules, and record the The value of the next Ep and Eu is assigned to Ep0 and Eu0, otherwise the fuzzy rule adjustment will not be performed.

According to the prior experimental test, if the temperature of PCW and UPW exceeds Em=0.5℃ after filtering, the output temperature fluctuation will fluctuate greatly, so whether Ep and Eu are greater than Em is used as the judgment of fuzzy rule tuning model parameters. Threshold, if Ep>Em or Eu>Em, the model parameters are adjusted to adapt to model changes.

如果

则δ偏大；如果

则δ偏小。同样的，Ep为PCW的当前采样温度值和初始值14℃的差值，对Ep进行了限幅Ep_max＝4℃，Ep_min＝-4℃，波动区间

对于

且

和

且

δ具有确定的大小状态，即前者最大PB，后折最小PS，并由先验实验整定了有限个工况下的模型参数，确定输入条件下符合实际模型的δ值。但是，对于已知参数之外的中间状态，δ的状态具有不确定性。所以采用模糊逻辑可以解决这一问题。The following is the use of fuzzy logic to adaptively adjust the size of δ. The current state of δ is divided into five states: maximum, too large, suitable, too small, and minimum. The corresponding δ are PB, PM, PS, Z0, NS, NM, NB, respectively, and the corresponding δ values are 5, 2.5 , 1.6, 1, 0.6, 0.4, 0.2. Among them, Eu is the difference between the current sampling temperature value of UPW and the initial value of 22°C, and the Eu is limited by Eu _max = 2°C, Eu _min = -2°C, and the fluctuation range is

if

then δ is too large; if

Then δ is small. Similarly, Ep is the difference between the current sampling temperature value of PCW and the initial value of 14°C, and Ep is limited by Ep _max = 4°C, Ep _min = -4°C, and the fluctuation range is

for

and

and

and

δ has a certain size state, that is, the former is the largest PB, and the back is the smallest PS. The model parameters under a limited number of working conditions are set by a priori experiments, and the δ value that conforms to the actual model under the input conditions is determined. However, for intermediate states outside the known parameters, the state of δ has uncertainty. So using fuzzy logic can solve this problem.

The membership functions used by Eu and Ep are shown in Figure 4 and Figure 5. For the corresponding uncertain state, the fuzzy rules shown in Table 1 are adopted, where δ is the transfer function gain adjustment factor of the flow servo valve-heat exchanger. For example, as shown by the dotted lines in Figure 4 and Figure 5, when the temperatures of the currently sampled UPW and PCW are 0.5°C and 1.5°C higher than the initial values, respectively, the corresponding Eu50% probability is PS and the 50% probability is Z; corresponding to Ep is 75% PS and 25% PB. According to the linear interpolation method and the fuzzy rules in Table 1, the final calculation method of δ is:

δ=50%×75%PS:PS+50%×25%PS:PB+50%×75%Z:PS+50%×25%Z:PB=0.475.

Table 1 Eu and Ep correspond to δ fuzzy rule table

Ep\Eu NB NS Z PS PB NB PBδ=5 PMδ=2.5 PMδ=2.5 PSδ=1.6 Zδ=1 NS PMδ=2.5 PMδ=2.5 PSδ=1.6 Zδ=1 NSδ=0.6 Z PMδ=2.5 PSδ=1.6 Zδ=1 NSδ=0.6 NMδ=0.4 PS PSδ=1.6 Zδ=1 NSδ=0.6 NMδ=0.4 NMδ=0.4 PB Zδ=1 NSδ=0.6 NMδ=0.4 NMδ=0.4 NBδ=0.2

Through the above obtained δ, the current servo valve prediction model is set as:

Bring the updated step response matrix A into the prediction model and enter the next loop.

Setpoint optimization in the fine tuning module:

When the difference between the UPW temperature controlled by the coarse adjustment module and the set value is less than Ec, the fine adjustment module adjustment is started. From the coarse adjustment module to the fine adjustment module, a remote transmission pipeline is required, and the coarse adjustment module is in the subFAB area, and the fine adjustment module is inside the lithography machine in the Fab area. The influence of ambient temperature on the temperature of the pipeline is obvious, and there are two polar effects of high temperature and low temperature. Therefore, a semiconductor refrigeration chip is added to the first stage of the fine-tuning module, and its size is small to adapt to the internal area of the lithography machine. The space constraint is no longer dependent on the compressor or other externally introduced refrigerants and other actuators that take up a large space for cooling, and can be heated and cooled through the two-way adjustment of the current, which can reduce the impact caused by the increase or decrease of the ambient temperature. UPW temperature fluctuations are controlled. The set value is determined by the set value of the two-way injection at the end, the power of the actuator and the heating power of the servo valve that adjusts the flow. The liquid injection flow rate Q ₁ , Q ₂ ; the relationship between the temperature rise and the flow rate caused by the heating of the servo valve is determined by a priori experiment as f(Q), and the relationship between the temperature rise and flow rate corresponding to the rated power of the known heater is H(Q) . ε is a safety factor of 1.5 to ensure system redundancy.

Therefore, the upper limit r _ymax of the UPW temperature setting value ry interval adjusted by the semiconductor refrigeration chip controlled by the first stage of the fine adjustment module is: r _ymax =min(r _{z1 -εf} (Q ₁ ),r _z2 -εf(Q ₂ ));

The lower limit should at least ensure that the output temperature under the rated power of the heater is not less than the set value r _z1 r _z2 ;

r _ymin =max(r _z1 -εH(Q ₁ ),r _z2 -εH(Q ₂ ))

r _y ∈[r _ymin ,r _ymax ];

Taking into account the thermal interference of the servo valve and the heater power range, there is a large redundancy to adjust when there is a large disturbance, and the average value of the adjustable range is taken as the set value of the first-level control of the fine-tuning module

The two-way injection includes horizontal injection and vertical injection. The flow and temperature-controlled liquids flow through the submerged head cavity and reach the main flow field in different pipeline environments. However, in the final intersection of the main flow field area, the SV design needs to be met. Set value requirements, and ensure that the temperature gradient and stability are within a certain range. Due to the different flow rates of the two channels of liquid injection, the heat load caused by the same power environment will reflect different changes in temperature, which means that the two-channel set values need to be combined with the actual liquid injection flow rates Q ₁ , Q ₂ and the final The set value SV of the main flow field temperature is decomposed and obtained.

r _z1 =T ₁ (SV, Q ₁ )

r _z2 = T ₂ (SV, Q ₂ )

The controllers in the fine-tuning module all use the PID controller to complete the tracking adjustment of the set value.

In a specific embodiment, the present invention provides a temperature control process of a lithography machine, as shown in FIG. 6 , including the following steps:

Step S101: Initialize the coarse adjustment controller, the initialization parameters include the step response matrix A _j of the servo valve and the heater model and the impulse response matrix h for error correction, the set value SV, the set value softening trajectory parameter α, modeling Time domain N, prediction domain P, control domain M, parameter adaptive adjustment threshold Ep0, Eu0 and gain adjustment factor δ.

以及精调节模块中第一二级控制器的PID参数。Step S102: Initialize the set value SV, the given liquid injection flow rate Q1, Q2; the redundant safety factor ε, the heating temperature rise of the servo valve, the correlation function f(Q)H(Q) of the rated power temperature rise of the heater and the flow rate , the set value of the main field and the flow rate of liquid injection, as well as the correlation function between the set value of each branch and the two

And the PID parameters of the first-level controller in the fine-tuning module.

Step S200 : performing predictive control of model parameter adaptation on the coarse adjustment module, and the control flow is shown in FIG. 7 . For the flow of FIG. 7, reference may be made to the descriptions in the foregoing method embodiments. It should be noted that the P in the prediction P step in FIG. 7 is a concept with the prediction domain P drawn in the previous control block diagram 3 . In FIG. 7 , the U that minimizes the objective function J is obtained by gradient descent, which refers to the change in the control variable when the objective function is minimized by using the gradient descent method as pointed out in the rolling optimization in step X003. Step S103: Determine whether the temperature output by the coarse adjustment module is within the target error |SV-Tx|<ec, if so, go to Step S104, otherwise, go to Step S200 again.

Step S104: Obtain the first-stage output Ty of the fine-tuning module.

Step S105: Acquire the second-stage outputs Tz1 and Tz2 of the fine-tuning module.

Step S106: R _z1 =T ₁ (SV, Q ₁ ) R _z2 =T ₂ (SV, Q ₂ ) according to the set value of the main field, the flow rate of liquid injection, and the set value of each branch and the correlation function between them Determine the set value R _z1 R _z2 of the two channels.

Step S107: According to the given liquid injection flow rate Q1, Q2; the redundant safety factor ε, the heating temperature rise of the servo valve, the correlation function f(Q)H(Q) of the rated power temperature rise of the heater and the flow rate, and the obtained The two-way set value R _z1 R _z2 . The set value R _y is determined according to the established relational expression.

Step S108: Calculate the adjustment amount of the semiconductor refrigeration sheet TEC according to the aforementioned primary temperature control setting value R _y and the first-level temperature control PI parameter.

Step S109: Calculate the adjustment amount of the two-way heater according to the aforementioned primary temperature control set value R _z1 R _z2 and the PI parameter of the second temperature control.

Step S110: It is judged whether the two-way liquid injection temperature control satisfies |Tz-Rz|<0.01°C, if yes, go to step S111, otherwise, restart S200.

Step S111 : Determine whether the two-channel liquid injection temperature control satisfies |Tz-Rz|<0.0025°C, if yes, go to step S112 , otherwise, restart S104 .

Step S112: whether to end the algorithm operation, otherwise restart S200, and if yes, end.

The invention provides a temperature control method and device for an immersion lithography machine, which adopts a coarse adjustment and a fine adjustment module for temperature control, and sets two-level adjustment in the fine adjustment module. In the coarse adjustment module, the dual actuators of the heater and the servo valve are adjusted through the predictive control of model parameters adaptive to realize the control of heating and cooling. The improved predictive control adopted by this device can take into account the judgment conditions of model parameter optimization, and the fuzzy rules used in parameter optimization can be more adaptively adjusted to an appropriate model according to the change of disturbance, and the parameters can solve the problem of mismatch of traditional predictive control models. . In the two-level control in the fine-tuning module, combined with different injection flow rates, different heating power during servo valve adjustment, and different immersion head environments, the setting values of the two-channel injection temperatures are different. Use prior knowledge to solve the two-level control It solves the problem of mutual coordinated control and avoids the problem that the set value of the previous stage is too low or too high, so that the two-way control of the latter stage cannot reach the set value or exceed the set value.

Those skilled in the art can easily understand that the above are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention, etc., All should be included within the protection scope of the present invention.

Claims (10)

1. A temperature control method of an immersion lithography machine is characterized by comprising the following steps:

the method comprises the following steps of exchanging heat between plant cooling water PCW and ultrapure water UPW of a photoetching machine through a heat exchanger, wherein the temperature of the PCW is lower than that of the UPW so as to cool the UPW, heating the UPW by using a heater so as to heat the UPW, roughly adjusting the temperature of the UPW, and controlling the flow of the PCW and the power of the heater so as to enable the temperature of the UPW to quickly reach the vicinity of a UPW temperature control target and stably maintain the temperature of the UPW;

when the UPW temperature is coarsely regulated, detecting the temperature of the PCW before heat exchange, the temperature of the UPW before heat exchange and the temperature of the UPW after heat exchange and heater control, and determining the trace of the UPW coarsely regulated target temperature, the flow control model of the PCW and the power control model of the heater based on the three detected temperature values and the UPW temperature control target; according to the detected temperature fluctuation before heat exchange of the PCW and the temperature fluctuation before heat exchange of the UPW, a flow control model of the PCW and a power control model of the heater are corrected in a self-adaptive mode by adopting a fuzzy rule; the flow control model for adaptively correcting the PCW by adopting the fuzzy rule specifically comprises the following steps: determining PCW temperature fluctuation and UPW temperature fluctuation, taking the two temperature fluctuations as input of a fuzzy rule, and outputting parameters of a flow control model of the PCW according to a preset fuzzy rule so that the PCW temperature fluctuation and the UPW temperature fluctuation are considered by the flow control model of the PCW; the parameter of the PCW flow control model is a multiple delta of the initial gain changing along with disturbance;

controlling the step response moment for a corrected servo valveArray, A₁A step response matrix for the initialized servo valve control; the servo valve control step response matrix corresponds to a flow control model of the PCW;

the roughly adjusted UPW is introduced into a photoetching machine through a remote transmission pipeline, is adjusted by a semiconductor refrigerating sheet and then is divided into two paths of horizontal UPW injection liquid and vertical UPW injection liquid, the two paths of UPW injection liquid are respectively heated by two heaters, and finally the two paths of UPW injection liquid are intersected in a main flow field area of the photoetching machine; the temperature of the UPW is finely adjusted in the photoetching machine through a semiconductor refrigerating sheet and two UPW liquid injection heaters, and environmental errors introduced by a remote transmission pipeline and thermal interference brought by a UPW flow control component are compensated, so that the temperature of a main flow field UPW of the photoetching machine is stabilized at a UPW temperature control target; when the UPW temperature is finely adjusted, the set value of the UPW temperature after the semiconductor refrigerating sheet is adjusted is determined by the power of two heaters for two UPW liquid injection, the flow rate of the two UPW liquid injection and the set value of the temperature of the two UPW liquid injection, and the set value of the temperature of the two UPW liquid injection is determined by the UPW temperature control target and the flow rate of the two UPW liquid injection; the method specifically comprises the following steps: determining the upper limit and the lower limit of a UPW temperature set value regulated by a semiconductor chilling plate based on the flow of the two UPW liquid injections, the relationship between the temperature rise and the flow caused by the heating of the servo valve determined by a prior experiment, the relationship between the temperature rise and the flow corresponding to the rated power of a heater and the set value of the temperature of the two UPW liquid injections, and determining the UPW temperature set value based on the upper limit and the lower limit of the UPW temperature set value; and the set values of the two UPW injection temperatures are respectively determined by the actual injection flow of the two UPW injection temperatures and the prior relational expression of the UPW temperature control target.

2. The temperature control method of an immersion lithography machine according to claim 1, wherein the UPW temperature control target is set by operational requirements of a main flow field of the lithography machine.

3. The temperature control method of an immersion lithography machine according to claim 1, wherein the rough adjustment of the UPW temperature specifically comprises the steps of:

s1, avoiding the rapid change of UPW temperature set value input and UPW temperature control quantity output in the course of rough adjustment control, using the relaxation factor alpha to soften the trace of UPW rough adjustment target temperature:

y_r(k+i)＝(1-αⁱ)*SV+αⁱy(k)

wherein, alpha belongs to (0,1), SV is UPW temperature control target, y_r(k + i) is a UPW temperature set value in the ith sampling period, and y (k) is the output of the current UPW temperature control quantity;

s2, predicting values of P sampling periods in prediction domain

Responding by P zero inputs

Superposition of zero State response A_jΔU_M，j(k) Obtaining:

wherein,

for the k + j output value derived by the prediction model,

is a zero state value, Δ u, at the k + j time_j(k) Corresponding to the control quantity of the actuator at time k, A_jΔ U for unit step response model of corresponding actuator_M，j(k) Is the variation of the control quantity of the corresponding actuator, A is the unit step response model of the corresponding actuator, delta U_M(k) The variation of all actuator control quantities; j is an element of 1, 2;

in the rough regulation, two actuators of a heat exchanger and a heater are arranged, two control quantities jointly act on the temperature of a feedback point of a rough regulation module, and j is 1 and 2 respectively represent a servo valve of the heat exchanger and two actuators A of the heater_jThe matrix is a unit step response model of the corresponding actuator;

under the action of the control quantity of M sampling periods of the control domain equal to M, the predicted values of P sampling periods of the prediction domain are as follows:

s3, solving the control quantity after the prediction model is built, wherein an objective function J is expressed as:

wherein, R and Q are weight matrixes considering the tracking effect and the output quantity of the actuator respectively, and W (k) is all set values y in the prediction domain P_r(k+i)i＝1，2，…，P；ΔU_M ^T(k) Is Delta U_M(k) Transposing;

obtaining the control quantity change quantity when the target function is minimum by using a gradient descent method;

order to

The following can be obtained:

u_j(k)＝u_j(k-1)+Δu_j(k)j＝1，2

wherein，Δu_j(k) D is the amount of change of the control quantity of the corresponding actuator at the time k_jIs prepared from known A_jR and Q form a matrix, u_j(k) Corresponding to the control quantity of the actuator, finally solving the control quantity u of each actuator at each moment_j(k)；

S4, correcting continuous sampling of the prediction error is equivalent to superposition of pulse influence; the perturbation of the model parameters is derived from temperature fluctuation of input UPW and PCW, fuzzy rules are adopted, PCW temperature deviation Ep obtained through filtering and UPW temperature deviation Eu are used as input of the fuzzy rules to determine the model parameters, and a transfer function of the influence of the flow change of the servo valve of the heat exchanger on output temperature is simplified into a first-order time lag model;

performing sliding filtering on the sampled PCW and UPW, wherein the filtering adopts the average of sampling values for nearly ten times;

judging whether Ep and Eu after filtering at the current sampling moment are deviated from the PCW temperature difference Ep0 and the UPW temperature difference Eu0 when the parameters of the previous trigger model are adjusted, and judging whether the deviation exceeds a limited deviation Em; and if the model parameters exceed Em, updating the model parameters according to fuzzy rules, recording the values of Ep and Eu at the time, and assigning the values to Ep0 and Eu0, otherwise, not performing fuzzy rule adjustment, and initializing Ep0 and Eu0 to be 0.

4. The temperature control method of an immersion lithography machine according to claim 1, wherein when the difference between the coarsely adjusted UPW temperature and the set value is less than a preset value Ec, the fine adjustment of the UPW is started;

the temperature of the remote transmission pipeline is influenced by the ambient temperature, the UPW temperature fluctuation caused by the rise or the fall of the ambient temperature is controlled by the semiconductor refrigerating sheet, and the set value is determined by the set value and the heater power of two paths of UPW liquid injection at the tail end and the heating power of a servo valve for adjusting the flow;

two paths of UPW liquid injection flow rates are respectively Q₁、Q₂(ii) a Determining the relationship between temperature rise and flow caused by heating of the servo valve as f (Q) by a priori experiment, knowing that the relationship between the temperature rise and the flow corresponding to the rated power of the heater is H (Q), wherein epsilon is a safety coefficient for ensuring the redundancy of the system;

fine-adjustment semiconductor refrigerating sheetAdjusted UPW temperature set value r_yUpper limit of interval r_ymaxComprises the following steps: r is_ymax＝min(r_z1-εf(Q₁)，r_z2-εf(Q₂))；

Lower limit r_yminEnsuring that the temperature output under the rated power of the heater is not less than the set value r of the temperatures of two UPW injection_z1，r_z2；

r_ymin＝max(r_z1-εH(Q₁)，r_z2-εH(Q₂))

r_y∈[r_ymin，r_ymax]；

Considering both the thermal interference of the servo valve and the power range of the heater, and taking the average value of the adjustable interval as the set value of the UPW temperature after the semiconductor chilling plate is adjusted;

two-way UPW liquid injection temperature set value needs to be combined with actual liquid injection flow Q₁、Q₂And decomposing the UPW temperature control target SV to obtain:

r_z1＝T₁(SV，Q₁)

r_z2＝T₂(SV，Q₂)

wherein r is_z1＝T₁(SV，Q₁)、r_z2＝T₂(SV，Q₂) Respectively obtaining set values r of two UPW injection temperatures obtained through priori knowledge_z1、r_z2And SV and each path of flow Q₁、Q₂The relation between;

and the controllers for UPW fine adjustment all adopt PID controllers to complete the tracking adjustment of the set value.

5. A method according to claim 3, wherein the first order time lag model for transfer function simplification of the effect of servo valve flow changes on output temperature is:

k is the gain of the initialization state, T is the inertia duration, tau is the pure time delay, s is the Laplace operator, and G(s) is a transfer function model of the influence of the flow change of the servo valve on the output temperature.

6. A temperature control device of an immersion lithography machine, comprising: a coarse adjustment module and a fine adjustment module;

the coarse adjustment module is arranged in an external plant of the photoetching machine, and comprises four fluid interfaces which are respectively as follows: a PCW input interface in the plant cooling water PCW circulation loop, a PCW output interface in the PCW circulation loop, a UPW input interface in the ultrapure water UPW circulation loop and an interface of a UPW remote transmission heat preservation pipeline; the external factory service of the photoetching machine comprises: the device comprises a PCW circulation loop, a UPW circulation loop and a UPW remote transmission heat preservation pipeline; the UPW flowing out of the UPW circulation loop flows to the photoetching machine through a UPW remote transmission heat preservation pipeline;

UPW recovered from the photoetching machine flows into the UPW circulation loop, the temperature of PCW in the PCW circulation loop is adjustable, and the PCW circulation loop and the UPW circulation loop exchange heat through a heat exchanger; the method comprises the following steps of exchanging heat between plant cooling water PCW and ultrapure water UPW of a photoetching machine through a heat exchanger, wherein the temperature of the PCW is lower than that of the UPW so as to cool the UPW, heating the UPW by using a heater so as to heat the UPW, roughly adjusting the temperature of the UPW, and controlling the flow of the PCW and the power of the heater so as to enable the temperature of the UPW to quickly reach the vicinity of a UPW temperature control target and stably maintain the temperature of the UPW; when the UPW temperature is coarsely regulated, detecting the temperature of the PCW before heat exchange, the temperature of the UPW before heat exchange and the temperature of the UPW after heat exchange and heater control, and determining the trace of the UPW coarsely regulated target temperature, the flow control model of the PCW and the power control model of the heater based on the three detected temperature values and the UPW temperature control target; according to the detected temperature fluctuation before heat exchange of the PCW and the temperature fluctuation before heat exchange of the UPW, a flow control model of the PCW and a power control model of the heater are corrected in a self-adaptive mode by adopting a fuzzy rule; the flow control model for adaptively correcting the PCW by adopting the fuzzy rule specifically comprises the following steps: determining PCW temperature fluctuation and UPW temperature fluctuation, using the two temperature fluctuations as input of fuzzy rule,outputting parameters of a flow control model of the PCW according to a preset fuzzy rule, so that the flow control model of the PCW considers PCW temperature fluctuation and UPW temperature fluctuation; the parameter of the PCW flow control model is a multiple delta of the initial gain changing along with disturbance;

for the modified servo valve control step response matrix, A₁A step response matrix for the initialized servo valve control; the servo valve control step response matrix corresponds to a flow control model of the PCW;

the fine adjustment module is arranged inside the photoetching machine, comprises three fluid interfaces, and is respectively as follows: receiving a UPW interface of the UPW remote transmission heat preservation pipeline and two UPW interfaces injected into a main flow field of the photoetching machine; the UPW adjusted by the coarse adjustment module is introduced into the photoetching machine through a remote transmission pipeline, is adjusted by a semiconductor refrigerating sheet and then is divided into two paths of horizontal UPW liquid injection and vertical UPW liquid injection, the two paths of UPW liquid injection are heated by two heaters respectively, and finally the two paths of UPW liquid injection are converged in a main flow field area of the photoetching machine; the temperature of the UPW is finely adjusted in the photoetching machine through a semiconductor refrigerating sheet and two UPW liquid injection heaters, and environmental errors introduced by a remote transmission pipeline and thermal interference brought by a UPW flow control component are compensated, so that the temperature of a main flow field UPW of the photoetching machine is stabilized at a UPW temperature control target; when the UPW temperature is finely adjusted, the set value of the UPW temperature after the semiconductor refrigerating sheet is adjusted is determined by the power of two heaters for two UPW liquid injection, the flow rate of the two UPW liquid injection and the set value of the temperature of the two UPW liquid injection, and the set value of the temperature of the two UPW liquid injection is determined by the UPW temperature control target and the flow rate of the two UPW liquid injection; the method specifically comprises the following steps: determining the upper limit and the lower limit of a UPW temperature set value regulated by a semiconductor chilling plate based on the flow of the two UPW liquid injections, the relationship between the temperature rise and the flow caused by the heating of the servo valve determined by a prior experiment, the relationship between the temperature rise and the flow corresponding to the rated power of a heater and the set value of the temperature of the two UPW liquid injections, and determining the UPW temperature set value based on the upper limit and the lower limit of the UPW temperature set value; and the set values of the two UPW injection temperatures are respectively determined by the actual injection flow of the two UPW injection temperatures and the prior relational expression of the UPW temperature control target.

7. The temperature control apparatus of an immersion lithography machine according to claim 6, wherein the UPW temperature control target is set by operational requirements of a main flow field of the lithography machine.

8. The temperature control apparatus of an immersion lithography machine according to claim 6, wherein the coarse adjustment module performs coarse adjustment of the UPW temperature, specifically comprising the steps of:

s1, avoiding the rapid change of UPW temperature set value input and UPW temperature control quantity output in the course of rough adjustment control, using the relaxation factor alpha to soften the trace of UPW rough adjustment target temperature:

y_r(k+i)＝(1-αⁱ)*SV+αⁱy(k)

wherein, alpha belongs to (0,1), SV is UPW temperature control target, y_r(k + i) is a UPW temperature set value in the ith sampling period, and y (k) is the output of the current UPW temperature control quantity;

s2, predicting values of P sampling periods in prediction domain

Responding by P zero inputs

Superposition of zero State response A_jΔU_M，j(k) Obtaining:

wherein,

for the k + j output value derived by the prediction model,

is a zero state value, Δ u, at the k + j time_j(k) Corresponding to the control quantity of the actuator at time k, A_jΔ U for unit step response model of corresponding actuator_M，j(k) Is the variation of the control quantity of the corresponding actuator, A is the unit step response model of the corresponding actuator, delta U_M(k) The variation of all actuator control quantities; j is an element of 1, 2;

in the rough regulation, two actuators of a heat exchanger and a heater are arranged, two control quantities jointly act on the temperature of a feedback point of a rough regulation module, and j is 1 and 2 respectively represent a servo valve of the heat exchanger and two actuators A of the heater_jThe matrix is a unit step response model of the corresponding actuator;

under the action of the control quantity of M sampling periods with the control domain equal to M, the predicted values of P sampling periods in the prediction domain are as follows:

s3, solving the control quantity after the prediction model is built, wherein an objective function J is expressed as:

wherein, R and Q are weight matrixes considering the tracking effect and the output quantity of the actuator respectively, and W (k) is a prediction domainAll set values y within P_r(k+i)i＝1，2，…，P；ΔU_M ^T(k) Is Delta U_M(k) Transposing;

obtaining the control quantity change quantity when the target function is minimum by using a gradient descent method;

order to

The following can be obtained:

u_j(k)＝u_j(k-1)+Δu_j(k)j＝1，2

wherein, Δ u_j(k) In response to the amount of change in the actuator control amount, d_jIs prepared from known A_jR and Q form a matrix, u_j(k) Corresponding to the control quantity of the actuator, finally solving the control quantity u of each actuator at each moment_j(k)；

S4, correcting continuous sampling of the prediction error is equivalent to superposition of pulse influence; the perturbation of the model parameters is derived from temperature fluctuation of input UPW and PCW, fuzzy rules are adopted, PCW temperature deviation Ep obtained through filtering and UPW temperature deviation Eu are used as input of the fuzzy rules to determine the model parameters, and a transfer function of the influence of the flow change of the servo valve of the heat exchanger on output temperature is simplified into a first-order time lag model;

performing sliding filtering on the sampled PCW and UPW, wherein the filtering adopts the average of sampling values for nearly ten times;

initializing Ep0 and Eu0 to be 0, judging that Ep and Eu after filtering at the current sampling moment are deviated from PCW temperature difference Ep0 and UPW temperature difference Eu0 when the parameters of the previous trigger model are adjusted, and judging whether the limited deviation Em is exceeded; and if the model parameters exceed Em, updating the model parameters according to fuzzy rules, recording the values of Ep and Eu at the time, and assigning the values to Ep0 and Eu0, otherwise, not performing fuzzy rule adjustment.

9. Temperature control of an immersion lithography machine according to claim 6The device is characterized in that when the difference between the UPW temperature after rough adjustment and the set value is smaller than a preset value Ec, the fine adjustment module starts to perform fine adjustment on the UPW; the influence of the environment temperature on the temperature of the remote transmission pipeline is realized, the UPW temperature fluctuation caused by the rise or the fall of the environment temperature is controlled by a semiconductor refrigerating sheet in the fine adjustment module, and the set value is determined by the set value of two paths of UPW liquid injection at the tail end, the heater power and the heating power of a servo valve for adjusting the flow; two paths of UPW liquid injection flow rates are respectively Q₁、Q₂(ii) a Determining the relationship between temperature rise and flow caused by heating of the servo valve as f (Q) by a priori experiment, knowing that the relationship between the temperature rise and the flow corresponding to the rated power of the heater is H (Q), wherein epsilon is a safety coefficient for ensuring the redundancy of the system; UPW temperature set value r adjusted by fine adjustment semiconductor refrigeration piece_yUpper limit of interval r_ymaxComprises the following steps:

r_ymax＝min(r_z1-εf(Q₁)，r_z2-εf(Q₂) ); lower limit r_yminEnsuring that the temperature output under the rated power of the heater is not less than the set value r of the temperatures of two UPW injection_z1，r_z2；r_ymin＝max(r_z1-εH(Q₁)，r_z2-εH(Q₂))；r_y∈[r_ymin，r_ymax](ii) a Taking thermal interference of the servo valve and the power range of the heater into consideration, and taking the average value of the adjustable interval as a set value of the UPW temperature after the semiconductor chilling plate is adjusted; two-way UPW liquid injection temperature set value needs to be combined with actual liquid injection flow Q₁、Q₂And decomposing the UPW temperature control target SV to obtain:

wherein r is_z1＝T₁(SV，Q₁)、r_z2＝T₂(SV，Q₂) Respectively obtaining set values r of two UPW injection temperatures obtained through priori knowledge_z1、r_z2And SV and each path of flow Q₁、Q₂The relation between; and the controllers for UPW fine adjustment all adopt PID controllers to complete the tracking adjustment of the set value.

10. The temperature control apparatus of an immersion lithography machine according to claim 8, wherein the first order time lag model for transfer function simplification of the effect of servo valve flow changes on output temperature in the coarse tuning module is:

k is the gain of the initialization state, T is the inertia duration, tau is the pure time delay, s is the Laplace operator, and G(s) is a transfer function model of the influence of the flow change of the servo valve on the output temperature.

Images