JP6307347B2 - Semiconductor wafer temperature controller - Google Patents

Description

  The present invention relates to a temperature control device for a semiconductor wafer that controls the temperature of the plurality of temperature adjusting means in order to adjust the temperature of the semiconductor wafer by the plurality of temperature adjusting means.

The process of processing a semiconductor wafer such as a silicon wafer includes a process in which the temperature of the silicon wafer must be controlled to a target temperature and the temperature distribution in the surface of the silicon wafer must be controlled to a desired distribution.
For this reason, a method is known in which a control loop is provided independently for each temperature adjusting means, and the temperature of the semiconductor wafer is simultaneously controlled by a plurality of temperature adjusting means.
In such semiconductor wafer temperature control, it is necessary to maintain the target value while maintaining a certain deviation with respect to the reference control amount, and to maintain the target value even under disturbance. Conventionally, a master-slave control method is known (see, for example, Patent Document 1).
In the master-slave control method, one of a plurality of control loops is controlled as a master, and the remaining loops follow the control amount of the master loop. This is a method of calculating and controlling the deviation.
The master loop is usually a control loop with the slowest response, and is usually a slave loop that follows other control loops.

By the way, when such a master-slave control method is used in a plate-type semiconductor temperature control apparatus having a plurality of heating and cooling zones, depending on the application, not only the temperature of all zones is uniform, The target temperature of the zone may be changed to give the plate a temperature gradient. For example, when the plate is covered with a chamber, the semiconductor wafer may be affected by the heat of the chamber wall, and the end portion may be heated more easily than the center portion.
In this case, it is necessary to set the target temperature of the zone at the center of the plate high and the target temperature of the zone at the end of the plate to be low. It is only necessary to set an appropriate offset temperature in the temperature adjusting means and adjust the temperature.

Japanese Patent Laid-Open No. 7-200076

However, as will be described in detail later, in such a master-slave control method, when heating and cooling is performed with a temperature gradient in the plate, there are the following problems.
That is, when the master loop is set to the highest target temperature during heating, the slave loop is set to a target temperature having a certain offset with respect to the target temperature of the master loop. When the master loop starts to move in the heating direction, there is a problem that the slave loop temporarily acts on the master loop in the cooling direction so as to have a certain offset.
Further, for example, after the temperatures of the three zones are stabilized at the respective target temperatures (SV1, SV2, SV3) = (10 ° C., 20 ° C., 30 ° C.), the target temperatures are set to (SV1, SV2, SV3). = (30.degree. C., 20.degree. C., 10.degree. C.), there is a problem that the temperature in the middle zone is shaken at the moment of switching although the target temperature SV2 does not change. These problems will adversely affect throughput because it takes more time to settle to the target temperature.

  An object of the present invention is to provide a temperature control device for a semiconductor wafer capable of improving throughput in a temperature control device based on a master-slave control method.

According to a first aspect of the present invention, there is provided a temperature control device for a semiconductor wafer that performs temperature control of the plurality of temperature adjusting means in order to adjust the temperature of the semiconductor wafer by the plurality of temperature adjusting means.
A master loop for controlling the temperature of the reference temperature adjusting means;
A slave loop for controlling the temperature of other temperature adjusting means so as to follow this master loop;
Master temperature detection means for detecting the temperature of the semiconductor wafer whose temperature is adjusted by the temperature adjustment means of the master loop;
Slave temperature detecting means for detecting the temperature of the semiconductor wafer whose temperature has been adjusted by the slave loop temperature adjusting means;
Based on the temperature detected by the master temperature adjusting means and the temperature detected by the slave temperature detecting means, the operation amount given to the temperature adjusting means of the master loop and the operation amount given to the slave loop are calculated. Operating amount calculation means for
The operation amount calculation means includes
Master-slave switching means for switching between the master loop and the slave loop according to the status of the target temperature setting before and after the target temperature change,
And a master-slave canceling unit that cancels the settings of the master loop and the slave loop.

According to a second aspect of the present invention, in the first aspect,
The master slave switching means is
In the heating control, the loop having the lowest target temperature is switched to the master loop, and in the cooling control, the loop having the highest target temperature is switched to the master loop.
According to a third aspect of the present invention, in the first aspect or the second aspect,
There are two or more slave loops, and three or more temperature gradients are set between the master loop and the two slave loops,
The master-slave canceling unit cancels the master loop setting when the set three or more temperature gradients are reversed during temperature control.

  According to the first aspect of the present invention, the master-slave switching unit and the master-slave canceling unit are provided, so that the master-slave switching unit and the master-slave canceling unit have a uniform temperature or a temperature gradient between the master loop and the slave loop. Since the relationship can be switched and the master / slave setting can be canceled, the temperature of the master loop can be adjusted at the start of temperature control. It is possible to prevent the slave loop from responding to the temperature rising direction with the start of temperature lowering of the means, and the throughput of the temperature control device can be improved.

  According to the second aspect of the present invention, since the master-slave switching means is provided, the loop with the lowest target temperature is switched to the master loop in the heating control, and the loop with the highest target temperature in the cooling control. Is switched to the master loop, the target temperature response does not respond in the reverse direction in the slave loop at the start of control, and the throughput of the temperature control device can be reliably improved.

  According to the third aspect of the present invention, since the master-slave canceling unit cancels the master-slave even if the set three or more temperature gradients are reversed during the temperature control, the maximum value of the target temperature is There is no fluctuation in the target temperature response of the loop of the target temperature set between the minimum value.

The block diagram which shows the temperature control apparatus which concerns on embodiment of this invention. Sectional drawing and the top view showing arrangement | positioning of the temperature adjustment means and temperature sensor in the said embodiment. The block diagram showing the structure of the controller which controls the temperature control apparatus in the said embodiment. The schematic diagram for demonstrating the matrix transformation | conversion in the operation amount conversion part in the said embodiment. The flowchart for demonstrating the effect | action of the said embodiment. The schematic diagram showing the control system used for the simulation for confirming the effect of the said embodiment. The top view showing the control system used for the simulation for confirming the effect of the said embodiment. The block diagram showing the structure of the conventional master slave control. The graph showing the simulation result about the conventional problem. The graph showing the simulation result about the conventional problem. The graph showing the simulation result by this embodiment. The graph showing the simulation result by this embodiment.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
[1] Configuration of Temperature Control Device 1 FIG. 1 shows a temperature control device 1 according to a first embodiment of the present invention. The temperature adjusting device 1 is a device for controlling the temperature of the silicon wafer W placed on the plate-like stage 2 to a target temperature and controlling the temperature distribution in the surface of the silicon wafer W. This temperature adjusting device 1 is used for a dry process, for example.
The temperature adjustment device 1 is a plate-like stage 2 and includes temperature adjustment means 3. As the temperature adjusting means 3, a chiller device or a thermoelectric element is preferably used when heating / cooling control is performed, and a heater for heating can be used when only heating control is performed.

The stage 2 is disposed in the vacuum chamber 4, and a silicon wafer W is placed on the stage 2. The silicon wafer W is held on the stage 2 by static electricity. Note that helium gas may be allowed to flow between the stage 2 and the silicon wafer W to increase the efficiency of heat transfer between the stage 2 and the silicon wafer W.
During the dry process, the vacuum chamber 4 is evacuated and maintained at a predetermined low pressure state.
In the stage 2, as shown in FIGS. 2A and 2B, a plurality of temperature adjusting means 3 are provided so that the in-plane temperature distribution of the silicon wafer W placed on the stage 2 can be adjusted. Has been placed.

FIG. 2A is a cross-sectional view of the stage 2, in which the temperature adjusting means 3 is disposed on the base plate 7, and the plate 5 is further placed thereon. A temperature sensor 6 is provided in the plate 5 as temperature detecting means.
FIG. 2B is a plan view of the stage 2. The stage 2 is concentrically divided into three zones 2A, 2B, and 2C, and the temperature adjusting means 3 is disposed in each region. Further, the temperature sensor 6 in the plate 5 is arranged at a position corresponding to the temperature adjusting means 3.

  When the temperature adjusting means 3 is energized, each zone 2A, 2B, 2C of the stage 2 can be individually heated. Therefore, by adjusting the energization to each temperature adjusting means 3 and controlling the temperature adjusting means 3, the in-plane temperature distribution of the silicon wafer W on the stage 2 can be adjusted, and the zones 2A, 2B, 2C Each temperature adjusting means 3 is controlled by a controller 24.

[2] Configuration of Controller 24 As described above, the controller 24 controls the temperature adjusting means 3 (master side 3M, slave side 3S) based on the temperature detected by the temperature sensor 6, as shown in FIG. Functionally configured as shown in the block diagram.

The controller 24 includes a master loop ML for controlling the temperature adjusting means 3M for heating the zone 2A in FIG. 2, a slave loop SL for controlling the temperature adjusting means 3S for heating the zones 2B and 2C, and the temperature of the temperature adjusting means 3M. The master side temperature sensor 6M for detecting the temperature, the slave side temperature sensor 6S for detecting the temperature of the temperature adjustment means 3S, and the operation amount calculation means 30 for calculating the operation amounts of the master loop ML and the slave loop SL. Although there are two slave loops SL corresponding to the zones 2B and 2C, the configuration following the master loop ML is the same, and therefore the illustration is omitted in FIG.
Further, in the master-slave control system, the object is to perform temperature distribution control by making the control amount on the slave side follow the control amount on the master side. Therefore, since the maximum response speed of the control system is usually limited by the loop with the slowest response, when the temperature of each zone 2A, 2B, 2C is uniformly controlled with respect to the same target temperature, the loop with the slowest response is selected. Set to master loop ML.

The operation amount calculation means 30 gives the operation amounts MVm and MVs to the temperature adjustment means 3M and 3S based on the master side control target value SVm and the slave side control target value SVs.
The operation amount calculation means 30 includes a master deviation calculation unit 31M, a slave deviation calculation unit 31S, a master control calculation unit 32M, a slave control calculation unit 32S, an operation amount conversion unit 33, a master operation amount restriction unit 34M, and a slave operation amount restriction unit. 34S, zone 2B target value setting unit 35, zone 2A target value setting unit 36, master slave release unit, master slave switching unit, and supervisor 37 as an internal target temperature calculation unit. Here, the internal target temperature is a target temperature for master-slave control calculated by the supervisor 37 using the target temperature of each zone set by the target temperature setting unit. When the master slave is released, the target temperature and the internal target temperature are the same.

The master deviation calculation unit 31M and the slave deviation calculation unit 31S use the internal target temperature calculated by the supervisor 37 and the temperature sensor detection values PVm and PVs of each zone, and deviations em and es for the master loop and the slave loop. Is calculated.

The master control calculation unit 32M is, for example, a PID controller, and outputs the calculation result Um to the operation amount conversion unit 33.
Similarly, the slave control calculation unit 32 </ b> S outputs the calculation result Us to the operation amount conversion unit 33.

  The operation amount conversion unit 33 uses the input calculation result Um in the master control calculation unit 32M and the calculation result Us in the slave control calculation unit 32S so that mutual interference between the master loop ML and the slave loop SL is reduced. This is the part that converts the operation amount. The conversion from the 2-input Um, Us to the 2-output Vm, Vs is performed when the control target is represented by a transfer function matrix P (s), for example, a steady gain matrix Gp = P (0) and a master-slave manipulated variable conversion matrix Gm. The master operation amount Vm and the slave operation amount Vs are output with the transformation matrix H obtained by the above. The manipulated variable conversion matrix H is obtained by the following equation (1) when the transfer function matrix P (s) has 2 rows and 2 columns.

The master operation amount restricting unit 34M is a part that restricts the operation amount so that the temperature adjusting means 3M does not exceed the minimum output and the maximum output. When it is determined that the operation amount is saturated, a determination signal awm indicating that is provided. The data is output to the master control calculation unit 32M. The output of the master operation amount limiting unit 34M is output as the operation amount MVm to the temperature adjusting means 3M.
Similarly, the slave operation amount restriction unit 34S is a portion that restricts the operation amount so that the temperature adjusting unit 3S does not exceed the minimum output and the maximum output. If it is determined that the operation amount is saturated, a determination to that effect is made. The signal aws is output to the slave control calculation unit 32S. The output of the slave operation amount limiting unit 34S is output as the operation amount MVs to the temperature adjusting unit 3S. The determination signals awm and aws are used as antiwindup activation signals in the respective control calculation units.

Based on the current target temperature, the changed target temperature, and the current temperature information of each zone, the supervisor 37 sets the internal target temperature to the deviation calculation units 31M and 31S of each zone after the change according to the change pattern. This is a part for determining / outputting output, master / slave switching, and release command, and determining / outputting a conversion matrix switching command in the operation amount converter 33.
The supervisor 37 ends the master-slave switching / cancellation process and the conversion matrix switching process before outputting the internal target temperature to the deviation calculation units 31M and 31S, as shown in the flowchart described later. The control calculation units 32M and 32S perform normal output calculation based on the deviations calculated by the deviation calculation units 31M and 31S, but add a process of switching to the bumpless in accordance with the master-slave switching and the release command.

In addition, the supervisor 37 also performs conversion matrix switching in the operation amount conversion unit 33. The conversion matrix is switched as follows. The transformation matrix is the product of the state quantity transformation matrix S and the non-interacting matrix D that mitigates plant interference. H = S × D
That is, the supervisor 37 switches the conversion matrix S. For example, in the case of 3 zones, it is as follows.

If zone 1 is the master, it is equivalent to adding the output of zone 1 to the other slave outputs,

If zone 3 is the master, it is equivalent to adding the output of zone 3 to the other slave outputs,

となる。操作量変換部３３は、この３つをスーパーバイザー３７からの指令によって切り替える。 In the case of master-slave cancellation, each output is handled independently.

It becomes. The operation amount conversion unit 33 switches between the three according to a command from the supervisor 37.

Next, bumpless switching will be described.
As shown in FIG. 4 (A), the output of the 3-input 3-output manipulated variable conversion unit 33 is set to zone 1 for the master loop, zone 2 to zone 3 for the slave loop, and each controller in the steady state before switching. Assuming that the outputs are U1, U2, and U3, the output of the manipulated variable conversion unit 33 is made non-interfering by the non-interacting matrix D after state conversion inside the conversion matrix H. When the state of the outputs U1, U2, U3 of the operation amount conversion unit 33 before switching is changed, the output of each zone is as follows.
Zone 1: U1
Zone 2: U2 + U1 (5)
Zone 3: U3 + U1

When the master loop is switched from zone 1 to zone 3, as shown in FIG. 4B, if the manipulated variable conversion unit 33 can maintain the steady state even after switching, the outputs are V1, V2, and V3, respectively. The output of each zone after state conversion is as follows.
Zone 1: V1 + V3
Zone 2: V2 + V3 (6)
Zone 3: V3

If (5) = (6) before and after switching, the output will not change, so zone 3
V3 = U3 + U1
It becomes.
Zone 1 is from V1 + V3 = U1,
V1 = U1-V3 = U1-(U3 + U1) =-U3 (7)
It becomes.
Zone 2 has V2 + V3 = U2 + U1,
V2 = U2 + U1-V3 = U2 + U1- (U3 + U1) = U2-U3 (8)
It can be expressed as. The bumpless switching can be realized with the same concept for other switching.

  FIG. 1 shows a case where zone 2A is zone 1, zone 2B is zone 2, and zone 2C is zone 3, and the switching process in the three-zone temperature control system is executed by the PID algorithm. However, for simplicity, the differential term is omitted (PI control).

<Explanation of symbols>
Each symbol in the following PID algorithm means the following.
MV1k_1, Kp_1, errk_1: Zone 1 proportional term value, proportional gain, deviation
MV1k_2, Kp_2, errk_2: Zone 2 proportional term value, proportional gain, deviation
MV1k_3, Kp_3, errk_3: Zone 3 proportional term value, proportional gain, deviation
MV2k_1, MV3k_1, Ki_1: Current integration term value, previous integration term value, integration time in zone 1
MV2k_2, MV3k_2, Ki_2: The current integral term value, the previous integral term value, and integration time in zone 2
MV2k_3, MV3k_3, Ki_3: Current integration term value in zone 3 Previous integration term value Integration time
istop1, istop2, istop3: Anti-windup command for each zone
buf: Buffer (temporary save)
MVfinal_1: Zone 1 manipulated variable
MVfinal_2: Zone 2 manipulated variable
MVfinal_3: Zone 3 manipulated variable

<Algorithm>
% P term calculation
MV1k_1 = Kp_1 * errk_1;
MV1k_2 = Kp_2 * errk_2;
MV1k_3 = Kp_3 * errk_3;
% I term calculation
MV2k_1 = MV3k_1 + Kp_1 * Ki_1 * errk_1;
MV2k_2 = MV3k_2 + Kp_2 * Ki_2 * errk_2;
MV2k_3 = MV3k_3 + Kp_3 * Ki_3 * errk_3;
% Anti-windup
if (istop_1 == 1)
MV2k_1 = MV3k_1;
end
if (istop_2 == 1)
MV2k_2 = MV3k_2;
end

% Master-slave switching / cancellation processing Manipulation output
if (when releasing the master / slave of zone 3)
MV2k_1 = MV3k_1 + MV3k_3;
MV2k_2 = MV3k_2 + MV3k_3;
MV2k_3 = MV3k_3;
MVfinal_1 = MV2k_1;
MVfinal_2 = MV2k_2;
MVfinal_3 = MV2k_3;
if (when canceling master slave of zone 1)
MV2k_1 = MV3k_1;
MV2k_2 = MV3k_2 + MV3k_1;
MV2k_3 = MV3k_3 + MV3k_1;
MVfinal_1 = MV2k_1;
MVfinal_2 = MV2k_2;
MVfinal_3 = MV2k_3;

else if (when zone 3 is the master from the release state)
MV2k_1 = MV3k_1-MV3k_3;
MV2k_2 = MV3k_2-MV3k_3;
MV2k_3 = MV3k_3;
MVfinal_1 = MV2k_1;
MVfinal_2 = MV2k_2;
MVfinal_3 = MV2k_3;
else if (when zone 1 is the master from the release state)
MV2k_3 = MV3k_3-MV3k_1;
MV2k_2 = MV3k_2-MV3k_1;
MV2k_1 = MV3k_1;
MVfinal_1 = MV2k_1;
MVfinal_2 = MV2k_2;
MVfinal_3 = MV2k_3;

else if (from zone 3 to zone 1 as master)
buf = MV3k_1;
MV2k_1 = MV3k_1 + MV3k_3;
MV2k_2 = MV3k_2-buf;
MV2k_3 = -buf;
MVfinal_1 = MV2k_1;
MVfinal_2 = MV2k_2;
MVfinal_3 = MV2k_3;
else if (when zone 1 to zone 3 is the master)
buf = MV3k_3;
MV2k_3 = MV3k_3 + MV3k_1;
MV2k_2 = MV3k_2-buf;
MV2k_1 = -buf;
MVfinal_1 = MV2k_1;
MVfinal_2 = MV2k_2;
MVfinal_3 = MV2k_3;
else (when there is no switching)
MVfinal_1 = MV1k_1 + MV2k_1;
MVfinal_2 = MV1k_2 + MV2k_2;
MVfinal_3 = MV1k_3 + MV2k_3;
end

[3] Operation of Embodiment Next, the operation of the present embodiment will be described based on the flowchart shown in FIG. In the following description, the zone 3 will be described as the master loop ML.
First, the supervisor 37 checks the target temperature in the target temperature setting unit 35 of the zone 2B and the target value setting unit 36 of the zone 2A (procedure S1), and determines whether or not the target temperature has been changed (procedure S2). .
If there is no change in the target temperature, the supervisor 37 maintains the previous state (step S11).
When the target temperature is changed, the supervisor 37 confirms the changed target temperature (step S3).

Next, the supervisor 37 confirms the change pattern of the target temperature (step S4).
When the change pattern of the target temperature in each zone is uniform to uniform, the supervisor 37 maintains the master loop setting of the zone 3 without changing the setting (step S5).
When the change pattern of the target temperature of each zone is changed from the uniform state to the gradient state or from the gradient to the uniform state, the supervisor 37 determines whether the control is heating control or cooling control ( Procedure S6).
In the case of heating control, the supervisor 37 sets the zone set at the lowest target temperature in the master loop ML (step S7). On the other hand, in the case of the cooling control, the supervisor 37 sets the zone set to the highest target temperature as the master loop ML (step S8).
When the change pattern of the target temperature in each zone is changed from the gradient state to the gradient state, the supervisor 37 determines whether or not the gradient is inverted (step S9).
When the gradient state is reversed, the supervisor 37 cancels the master / slave setting (step S10).
On the other hand, if the gradient state is non-inverted, the previous state is held (step S11).

When the above determination and master / slave setting process are completed, the supervisor 37 outputs a command corresponding to the setting to the control calculation unit 32 (step S12).
Subsequently, the supervisor 37 outputs a conversion matrix switching command to the operation amount conversion unit 33 (step S13).
Finally, the supervisor 37 calculates the internal target temperature in each zone, and outputs it to the deviation calculator 31 and the control calculators 32M and 32S (step S14).
The deviation calculation unit 31 of each zone calculates the deviation based on the internal target temperature output from the supervisor 37 (step S15), and the control calculation unit 32 calculates the P term, I term, and D term in PID control. (Procedure S16).
The control calculation unit 32 of each zone executes master-slave switching / cancellation processing based on the command output from the supervisor 37 (step S17).
The control calculation unit 32 of each zone calculates the operation amount and outputs it to the operation amount conversion unit 33 (step S18).
The manipulated variable conversion unit 33 switches the conversion matrix based on the conversion matrix switching command output from the supervisor 37 (step S19).

[4] Confirmation of effect by simulation
[4-1] Configuration of Control System in Simulation The master-slave control of the three-input three-output system will be described using a simulation result obtained by modeling the control system shown in FIG. 6 as an example. As shown in FIG. 7, this control system is a system for controlling the temperature of a 400 × 150 × t4 aluminum plate, and uses three thermo-modules that can be heated and cooled as actuators. The temperature of the aluminum plate is measured by three K thermocouples placed near the module. The thermo module and the thermocouple are intentionally arranged so as to be asymmetric with respect to the longitudinal direction of the plate, and the dimensional details are shown in FIG. Zone 1, 2, and 3 from the left. Table 1 shows the dynamic characteristics of each zone. However, in this description, in order to simplify the description, the gain on the cooling side is 1/3 of the gain on the heating side, the time constant, and the dead time are the same as those on the heating side.

[4-2] Problems in conventional master-slave control As shown in Fig. 8, the controller that can realize conventional master-slave control sets the control loop with the slowest response as the master loop ML, and performs other control. The loop is set as a slave loop SL, and the slave loop SL is controlled to follow the master loop.
This is because when all target temperatures are set uniformly, a control loop with a slow response speed cannot follow other control loops.

In plate-type heating and cooling using such control, not only the temperature is made uniform depending on the application, but also the target temperature of each zone is changed to give the plate a temperature gradient.
In this case, an appropriate offset temperature may be added to or subtracted from the target value on the slave side. However, when the master loop ML is fixed and used as in the conventional type, the following problem occurs.

<First problem>
In FIG. 9A, the target temperature of each zone is changed from 0 ° C.
(SV1, SV2, SV3) = (10 ° C., 20 ° C., 30 ° C.)
The result of the target value response by simulation is shown.
Further, in FIG. 9B, the target temperature of each zone is changed from 30 ° C.
(SV1, SV2, SV3) = (30 ° C., 20 ° C., 10 ° C.)
The result of the target value response by simulation is shown.
In either case, zone 1 shows a reverse response at the start of heating (cooling), resulting in wasted rise time. This is because the zone with the highest (low) target temperature is set as the master loop ML during heating (during cooling). For example, when the master loop ML starts to move in the heating direction due to the start of heating, the slave loop SL tries to have an offset with respect to the master loop ML, so that the master loop ML temporarily acts in the cooling direction. To do. Normally, the master loop ML has a slower response than the slave loop SL, and the phenomenon becomes more remarkable.

<Second problem>
In FIG. 10, the target temperature of each zone is
(SV1, SV2, SV3) = (30 ° C., 20 ° C., 10 ° C.)
The target temperature is set to 1000 seconds at time
(SV1, SV2, SV3) = (10 ° C., 20 ° C., 30 ° C.)
Switch to the target temperature at 2000 seconds,
(SV1, SV2, SV3) = (30 ° C., 20 ° C., 10 ° C.)
The response when switching to is shown.
Master loop ML is zone 3. In zone 2, although the target temperature is always 20 ° C., a large fluctuation occurs every time the target temperature is switched.

This has a problem in the temperature setting method of the master slave. As described above, the master / slave target temperature setting sets its own target temperature for the master loop, but the slave side sets an offset for the master loop.
For example, the master loop ML is zone 3, and the target temperature of each zone is
(SV1, SV2, SV3) = (30 ° C., 20 ° C., 10 ° C.)
In this case, the internal target temperature of zone 3 is 10 ° C. for the master, but the internal target temperature of zone 2 that is the slave is + 10 ° C. because the offset amount is + 10 ° C. relative to the internal target temperature of zone 3 The internal target temperature of zone 1 is + 20 ° C. because the offset amount is + 20 ° C. relative to the internal target temperature of zone 3.

Where the target temperature is
(SV1, SV2, SV3) = (10 ° C., 20 ° C., 30 ° C.)
Since the internal target temperature of the zone 3 is 30 ° C. when switching to, the internal target temperature of the zone 2 is −10 ° C., and the internal target temperature of the zone 1 is −20 ° C.
For this reason, the internal target temperature is switched from + 10 ° C. to −10 ° C. even though the target temperature of the zone 2 remains 20 ° C. before and after the target temperature switching. That is, at the moment when the target temperature is switched, a deviation of +10 − (− 10) = 20 ° C. occurs, and the control target is affected as an operation amount.
As described above, in the conventional master-slave control, since the master loop and the slave loop are fixed, the above-mentioned problems occur, and the actual process leads to deterioration of the throughput.

In order to solve the problem 1, in the case of heating control, the zone with the lowest target temperature is used as the master loop ML, and in the case of cooling control, the zone with the highest target temperature is used as the master loop ML.
For example, in the example described in FIGS. 9A and 9B, the target temperature is the lowest in the heating control in FIG. 9A, and the target temperature is lower in the cooling control in FIG. 9B. The highest zone 1 is set as the master loop ML.
Thereby, in the case of heating control, the zone 1 is heated so as to push up the zones 2 and 3 from the bottom, and in the case of cooling control, cooling is performed so as to push down from above.
That is, since all zones try to respond in the same direction, a more stable response can be obtained.
FIGS. 11A and 11B show the results of making zone 1 the master loop ML under the same conditions as in FIGS. 9A and 9B. It can be seen that the response is more stable than when the zone 3 is the master loop ML.

In order to solve the problem 2, the master-slave relationship is canceled when the set temperature is such that the temperature gradient is reversed from the set temperature having a gradient.
By canceling, the target temperature and the internal target temperature coincide with each other, a sudden change in deviation that occurs at the time of gradient reversal is mitigated, and the influence on the response is reduced.
FIG. 9 shows the result when the master slave is released under the same conditions as in FIG. It can be seen that the overshoot and undershoot in Zone 2 have been greatly relaxed.
As described above, it has been confirmed by simulation that the reverse response during response to the target value and the disturbance of the response are alleviated by the above-described method, so that the throughput is improved in the actual process.

DESCRIPTION OF SYMBOLS 1 ... Temperature control apparatus, 2 ... Stage, 24 ... Controller, 2A, 2B, 2C ... Zone, 3, 3M, 3S ... Temperature adjustment means, 4 ... Vacuum chamber, 5 ... Plate, 6, 6M, 6S ... Temperature sensor, 7 ... Base plate, 30 ... Operation amount calculation means, 31M ... Master deviation calculation unit, 31S ... Slave deviation calculation unit, 32M ... Master control calculation unit, 32S ... Slave control calculation unit, 33 ... Operation amount conversion unit, 34M ... Master operation Amount limiter, 34S ... Slave operation amount limiter, 35 ... Target value setting unit for zone 2B, 35A ... Offset setting unit, 36 ... Target value setting unit for zone 2A, 37 ... Supervisor, ML ... Master loop, SL ... Slave loop, W ... silicon wafer

Claims (3)

A temperature control device for a semiconductor wafer, which is provided on a stage on which a semiconductor wafer is mounted, and performs temperature control of a plurality of temperature adjusting means for adjusting the temperature of the semiconductor wafer,
A master loop that is provided in the center of the stage and controls the temperature of the temperature adjusting means serving as a reference;
A slave loop that is arranged at least one concentric circle outside the master loop and controls the temperature of other temperature adjusting means so as to follow the master loop;
Master temperature detection means for detecting the temperature of the central portion of the semiconductor wafer temperature adjusted by the temperature adjustment means of the master loop,
Slave temperature detecting means for detecting the temperature of the outer periphery of the semiconductor wafer whose temperature is adjusted by the temperature adjusting means of the slave loop;
Based on the temperature detected by the master temperature detecting means and the temperature detected by the slave temperature detecting means, an operation amount for controlling the temperature adjusting means of the master loop, and a temperature adjusting means of the slave loop and an operation amount calculating means for calculating a manipulated variable control to,
The operation amount calculation means includes
Master-slave switching means for switching between the master loop and the slave loop based on the current target temperature and the changed target temperature ;
A master-slave releasing unit for releasing the setting of the master loop and the slave loop when the target temperature of the slave loop is set with a temperature gradient with respect to the target temperature of the master loop; A semiconductor wafer temperature control device. In the temperature control apparatus of the semiconductor wafer of Claim 1,
The master slave switching means is
A semiconductor wafer temperature control apparatus characterized in that a loop having the lowest target temperature is switched to a master loop in heating control, and a loop having the highest target temperature is switched to a master loop in cooling control. In the temperature control apparatus of the semiconductor wafer of Claim 1 or Claim 2,
There are two or more slave loops, and three or more temperature gradients are set between the master loop and the two slave loops,
The master-slave canceling unit cancels the master loop setting when the set three or more temperature gradients are reversed during the temperature control.

Images