A voltage control based temperature control method for radiant heating lamps

Abstract

The application provides a kind of temperature control method of radiation heating lamp based on voltage control, specifically relates to the field of semiconductor technology.The method first calculates the voltage value and measured temperature of each group of heating lamp by thermal simulation software, and obtains the regression matrix by the above two groups of data;Then lock the emissivity of optical pyrometer, obtain the correction coefficient of each temperature measuring point, and then obtain the corrected temperature value of each temperature measuring point;The thermal effective coefficient of each group of heating lamp is obtained by simulation, and the real voltage of each group of heating lamp is obtained;Finally, according to the input value of the first iteration obtained by the regression matrix, the ideal voltage value of each group of heating lamp is obtained by iterative method.This method does not need manual calculation, only needs to calculate the initial voltage value of each group of heating lamp in the initial first cycle stage by simulation, and can gradually optimize the algorithm by automatic iteration, meet the requirements of wafer heating on time efficiency and temperature uniformity.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of semiconductor technology, and in particular to a radiation heating lamp temperature control method based on voltage control. BACKGROUND

[0002] In the semiconductor field, chips are the most important, and as raw materials for producing chips, the production process quality of silicon wafers will directly determine the pros and cons of chip quality. In all process steps of wafer production, the most important is heat treatment.

[0003] The most common heat treatment method at present is to use hydrogen passivation treatment technology: anneal the wafer for preparing semiconductor devices in a hydrogen environment, so that the dangling bonds are combined with hydrogen, thereby reducing the number of dangling bonds and minimizing the adverse effects on semiconductor devices. According to various researches and practices, heat treatment at a temperature of 1100 degrees Celsius will achieve the most ideal state. After a period of exploration and practice, the industry has summarized the following two steps for heating wafers: the first step is the heating algorithm, that is, fifteen groups of heating lamps in the device are respectively given different percentages of power, so that the temperature of the center point of the wafer is raised from room temperature (about 20 degrees Celsius) to 600 degrees, while the temperature difference between the center and the edge of the wafer is ensured to be not more than 20 degrees Celsius; compared with the method of directly heating the wafer to 1100 degrees Celsius in a room temperature environment, the advantage of this method is that the wafer is heated while the cavity is heated, and the structure of the wafer is prevented from being broken due to the large temperature difference with the cavity in the case of rapid temperature increase. Only when all the necessary conditions of the first step are met, the second step of closed loop heating is carried out: a closed loop heating algorithm based on a PID model is used to heat the wafer at a high heating rate, so that the temperature of the wafer is raised from 600 degrees Celsius to 1100 degrees Celsius (the temperature difference is plus or minus 0.75 degrees Celsius) within ten seconds, and then the temperature is stabilized, and then heat treatment is carried out.

[0004] Although the heating process based on the above steps in the domestic semiconductor industry is quite mature, the existing heating process, especially the heating treatment link, still needs to solve the problem of how to ensure real-time temperature uniformity and time efficiency when heating the wafer in the first step. SUMMARY

[0005] The technical problem to be solved by the present application is to solve the deficiencies of the prior art, and to provide a radiation heating lamp temperature control method based on voltage control.

[0006] To solve the above technical problems, the technical solution adopted by the present application is:

[0007] The basic configuration of a radiant heating lamp includes a heating lamp directly above the wafer and an optical thermometer below the wafer. Each group of heating lamps consists of several different small heating lamps, each with the same size, model, specifications, and power. The optical thermometer has several measuring points placed on the bottom of the wafer to measure the temperature and emissivity at corresponding locations on the wafer.

[0008] Step 1: Determine the required memory allocation, select a suitable microcontroller, allocate memory addresses for different function blocks, and initialize related data;

[0009] Step 2: Conduct preliminary heating simulation experiments;

[0010] Calculate the voltage that each heating lamp j should be set to at a given initial temperature to ensure that the wafer reaches the final specified temperature within the specified heating time, and that the temperature difference between each temperature measuring point i and the center point is not too large. Record the voltage value of each heating lamp after each simulation experiment.

[0011] Step 3: Using the voltage values ​​of each heating lamp and the temperature values ​​of each temperature measuring point in the simulation experiment, calculate the regression matrix B using the linear regression method.

[0012] Step 4: The optical thermometer inside the heating lamp device will produce some errors due to temperature changes, therefore a new coefficient C is introduced. i , is used to represent the correction coefficient for each temperature measurement point i; by forcibly locking the emissivity for each temperature measurement point, the ideal and true values ​​of several temperature measurement points at different temperatures are obtained, and the correction coefficient C for each temperature measurement point is fitted based on the above data. i Finally, the corrected temperature value T was calculated. i =t i *C i , where t i The original temperature value measured at temperature measurement point i;

[0013] Step 5: Given that the locations of the temperature measuring points are not equidistant and the positions of the heating lamps in different groups are asymmetrical, which will lead to a certain degree of heat loss, a thermal efficiency coefficient K is introduced. j , representing the actual heating efficiency of the j-th group of heating lamps. Based on multiple simulations, the thermal efficiency coefficient of each group of lamps is calculated. The actual voltage V of each group of heating lamps is... j The calculation formula is as follows:

[0014] V j =v j *K j (1)

[0015] In equation (1) v j The voltage setting value for the j-th group of heating lamps obtained in step 2 of the simulation is given.

[0016] Step 6 initializes the voltage values of each group of heating lamps for the first iteration;

[0017] The voltage values of each group of heating lamps for the first iteration are calculated:

[0018] First, the expected temperature vector T mat of each temperature measuring point i is defined exp is:

[0019]

[0020] where T exp i is the final expected temperature of the i-th temperature measuring point, and n is the total number of temperature measuring points. Multiply T mat exp by the regression matrix B obtained in step 3 to obtain the voltage V mat1 of each group of heating lamps in the first iteration cycle:

[0021]

[0022] where V 1,j is the voltage value of the j-th group of heating lamps in the first iteration cycle, m is the total number of groups of heating lamps, and since each group of heating lamps needs to be multiplied by the corresponding thermal effective coefficient K j in step 5, V mat1 also needs to be multiplied by the optimization matrix K in front:

[0023]

[0024] where K j is the thermal effective coefficient of the j-th group of heating lamps in step 5, so the corrected voltage vector should be:

[0025] V max 1，校正后 = K V mat1 (5)

[0026] and this is used as the input for the first iteration.

[0027] Step 7 iteratively optimizes the voltage values of each group of heating lamps according to the temperatures of the corresponding temperature measuring points;

[0028] The temperature of each temperature measuring point is affected by one or more groups of heating lamps, so we introduce a mapping

[0029] F: {1, 2…m}→{1, 2…n}

[0030] This mapping represents the mapping from the set of 1 to m to the set of 1 to n, and F(j) = i represents that the j-th group of heating lamps affects the i-th temperature measuring point;

[0031] According to the data in the previous steps, perform optimization iteration and obtain the voltage data of the m groups of lamps for the next set of actual measurements, and the iteration formula is:

[0032]

[0033] wherein, T F(j) is the temperature of the temperature measuring point i controlled by the jth group of heating lamps after heating, T 中心 is the temperature of the temperature measuring point at the center of the wafer after a unit of time, V j is the rated voltage of the jth group of heating lamps at present, V j is the rated voltage of the jth group of heating lamps in the next cycle;

[0034] It is specified that the temperature difference between the temperature measuring point at the center of the wafer and the temperature measuring point at the edge of the wafer after the fth iteration cycle is Δt f , that is,

[0035] Δt f = (T 中心 ) f - (T 边缘 ) f (7)

[0036] wherein, (T 中心 ) f is the final temperature of the temperature measuring point at the center of the wafer in the fth cycle, (T 边缘 ) f is the final temperature of the temperature measuring point at the edge of the wafer in the fth cycle;

[0037] When the number of cycles is enough, the Δt f of the fth iteration cycle at present tends to be equal to the Δt f-1 of the iteration cycle last time, that is:

[0038]

[0039] In formula (8), when the iteration number f is continuously increased, the Δt f tends to be a fixed value, at this time, the iteration of step 7 can be stopped, and the voltage value of each group of lamps in this cycle is saved as the final result for subsequent use.

[0040] The beneficial effects produced by the above technical scheme are that the present application provides a radiation heating lamp temperature control method based on voltage control. The method does not need manual calculation, only needs to calculate the initial voltage value of each group of heating lamps through simulation in the initial first cycle stage, and can gradually optimize the algorithm through automatic iteration to meet the requirements of time and temperature uniformity. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 is the algorithm flowchart provided by the embodiment of the present application;

[0042] Figure 2Temperature situation of each temperature measuring point after the first iteration cycle of the present application;

[0043] Figure 3 Temperature situation of each temperature measuring point after the third iteration cycle of the present application;

[0044] Figure 4 Temperature situation of each temperature measuring point after the last iteration cycle of the present application. DETAILED DESCRIPTION

[0045] The specific embodiments of the present application are described in further detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present application, but are not used to limit the scope of the present application.

[0046] As shown in the following table, the method of the present embodiment is as follows: Figure 1

[0047] Suppose that a wafer heating lamp device has 10 groups of heating lamps and 7 temperature measuring points. Each of the 10 groups of heating lamps can be set to a power percentage (any integer from 0 to 100), and the rated maximum voltage of each group of heating lamps is 6553.6V. Therefore, the voltage of each group of heating lamps is equal to the product of the rated maximum voltage and the power percentage of each group of heating lamps. We stipulate that the time of each iteration cycle is 70 seconds, and the final expected temperature of the wafer center is 600 degrees Celsius.

[0048] Step 1 determines the required memory allocation, selects a suitable single-chip microcomputer and distributes the memory addresses of different functional blocks, and initializes the related data. In the present embodiment, we use an arm architecture single-chip microcomputer (STM32 series, 64pin, 128KB flash memory), and the Keil uVision series software as the platform for software and hardware development.

[0049] Step 2 performs a preliminary heating simulation experiment.

[0050] The voltage of each group of heating lamps is calculated to be set to a value that can ensure that the wafer reaches the final specified temperature within the specified heating time, and the temperature difference between each temperature measuring point i and the center point is not too large, and the voltage value of each group of heating lamps is recorded after each simulation experiment.

[0051] In the present embodiment, we use ANSYS Fluent as the thermal simulation software, and simulate 8 heating experiments. The initial temperature of each experiment is set to 100 degrees Celsius, and the final temperature is set to a random temperature in the range of 600 to 615 degrees Celsius.

[0052] Step 3 calculates the regression matrix B by using the voltage value of each group of heating lamps and the temperature value of each temperature measuring point in the simulation experiment by means of linear regression method.

[0053] ​In this embodiment, B is finally obtained by using the final temperature of each of the 8 groups of temperature measuring points in step 2 and the voltage value of each group of heating lamps:

[0054]

[0055] Step 4: The optical thermometer inside the heating lamp device will produce some errors due to temperature changes, therefore a new coefficient C is introduced. i , is used to represent the correction coefficient for each temperature measurement point i; by forcibly locking the emissivity for each temperature measurement point, the ideal and true values ​​of several temperature measurement points at different temperatures are obtained, and the correction coefficient C for each temperature measurement point is fitted based on the above data. i Finally, the corrected temperature value T was calculated. i =t i *C i , where t i The original temperature value measured at temperature measurement point i;

[0056] In this embodiment, the correction coefficients for the seven temperature measurement points are: C1 = 0.994, C4 = 1.004, C5 = 1.001, C6 = 1.006, and the other three correction coefficients are all 1.

[0057] Step 5: Given that the locations of the temperature measuring points are not equidistant and the positions of the heating lamps in different groups are asymmetrical, which will lead to a certain degree of heat loss, a thermal efficiency coefficient K is introduced. j , representing the actual heating efficiency of the j-th group of heating lamps. Based on multiple simulations, the thermal efficiency coefficient of each group of lamps is calculated. The actual voltage V of each group of heating lamps is... j The calculation formula is as follows:

[0058] V j =v j *K j (1)

[0059] In equation (1) v j The voltage setting value for the j-th group of heating lamps obtained in step 2 of the simulation is given.

[0060] In this embodiment, the thermal efficiency coefficients of 10 lamps were obtained through multiple simulation tests using AnsysFluent as follows: K1 = 0.84, K2 = 0.98, K3 = 0.99, K7 = 0.98, and the rest were all 1.

[0061] Step 6: Initialize the voltage values ​​of each group of heating lamps in the first iteration;

[0062] Calculate the voltage value of each group of heating lamps during the first iteration:

[0063] First, define the expected temperature vector T_mat for each temperature measurement point i.exp T_exp = T_exp1 + T_exp2 + … + T_expn (1)

[0064]

[0065] T_exp = T_exp1 + T_exp2 + … + T_expn (1) i T_exp = T_exp1 + T_exp2 + … + T_expn (1) exp V_mat1 = B * T_exp (3)

[0066]

[0067] V_mat1 = B * T_exp (3) 1,j V_mat1 = B * T_exp (3) j V_mat1 = B * T_exp (3)

[0068]

[0069] K = K1, K2, …, Km (5) j K = K1, K2, …, Km (5)

[0070] V_max = KV_mat1 (5) 1，校正后

[0071]

[0072] In this embodiment, according to the initial settings, T_exp1 = T_exp2 = … = T_exp7 = 600℃, the initial voltage percentage of each heating lamp in the first cycle is finally obtained as follows: the first group is 30%, the second to third groups are 27%, the fourth to sixth groups are 25%, the seventh to eighth groups are 22%, the ninth group is 20%, and the tenth group is 18%.

[0073] Step 7: iteratively optimize the voltage value of each group of heating lamps according to the temperature of the corresponding temperature measuring point;

[0074] The temperature of each temperature measuring point is affected by one or more groups of heating lamps, so we introduce a mapping F: {1, 2…m}→{1, 2…n}

[0075] F: {1, 2…m}→{1, 2…n}

[0076] This mapping represents the mapping from the set of 1 to m to the set of 1 to n, and F(j) = i represents that the jth group of heating lamps affects the ith temperature measuring point.

[0077] ​​According to the data in the previous step, the optimization iteration is performed, and the next group of measured voltage data of the m groups of lamps is obtained, and the iteration formula is:

[0078]

[0079] wherein, T F(j) is the temperature of the temperature measuring point i controlled by the jth group of heating lamps after heating, T 中心 is the temperature of the wafer center temperature measuring point after a unit of time, V j is the rated voltage of the current jth group of heating lamps, V' j is the rated voltage of the jth group of heating lamps in the next cycle;

[0080] It is stipulated that after the fth iteration cycle, the temperature difference between the wafer center temperature measuring point and the wafer edge temperature measuring point is Δt f , that is,

[0081] Δt f = (T 中心 ) f - (T 边缘 ) f (7)

[0082] wherein, (T 中心 ) f is the final temperature of the wafer center temperature measuring point in the fth cycle, and (T 边缘 ) f is the final temperature of the wafer edge temperature measuring point in the fth cycle;

[0083] When the number of cycles is sufficient, the Δt f of the current fth iteration cycle and the Δt f-1 of the last iteration cycle tend to be 1, that is:

[0084]

[0085] In formula (8), when the iteration number f increases, Δt f tends to a fixed value, at which time the iteration of step 7 can be stopped, and the voltage value of each group of lamps in this cycle is saved as the final result for subsequent use.

[0086] In this embodiment, after the first iteration cycle, the final temperature image of the seven temperature measuring points is as shown in Figure 2 After experiencing three rounds of iteration signals, the final temperature image of the seven temperature measuring points is as shown in Figure 3 After experiencing the eighth iteration, that is, the last iteration of this embodiment, the final temperature image of the seven temperature measuring points is as shown in Figure 4 .

[0087] After going through several sets of iterative cycles, we finally got a set of ideal heating lamp voltage percentage data: the first set is 30%, the second to third set is 29%, the fourth to fifth set is 28%, the sixth set is 27%, the seventh set is 28%, the eighth set is 25%, the ninth set is 26%, and the tenth set is 24%.

Claims

1. A voltage control based temperature control method for radiant heating lamps, characterized in that, The method comprises the following steps: Step 1: determining the required memory allocation, selecting a suitable single-chip microcomputer, distributing the memory addresses of different functional blocks, and initializing relevant data; Step 2: performing a preliminary heating simulation experiment; Step 3: calculating the voltage value of each group of heating lamps j that should be set to ensure that the wafer reaches the final specified temperature within the specified heating time at a given initial temperature, and recording the voltage value of each group of heating lamps after each simulation experiment ends; Step 4: calculating the regression matrix B by using the voltage value of each group of heating lamps and the temperature value data of each temperature measurement point in the simulation experiment through linear regression; Step 4 introduces the correction coefficient , which is used to represent the correction coefficient of each temperature measuring point i; by forcibly locking the radiation rate of each temperature measuring point, ideal values and real values of several temperature measuring points at different temperatures are obtained, and the correction coefficients of the temperature measuring points are fitted according to the above data , and finally the corrected temperature value is calculated = * , wherein is the original temperature value measured by the temperature measuring point i; Step 5 introduces the thermal effective coefficient , represents the actual heating efficiency of the jth group of heating lamps, and the thermal effective coefficient of each group of lamps is obtained according to the simulation of multiple groups; the true voltage of each group of heating lamps The calculation formula is as follows: (1) wherein, Vsetj is the voltage setpoint for the jth set of heating lamps found in the step 2 simulation; Step 6: initializing the voltage value of each group of heating lamps for the first iteration; Step 7: iteratively optimizing the voltage value of each group of heating lamps according to the temperature of the corresponding temperature measurement point.

2. A voltage control based temperature control method for radiant heating lamps as claimed in claim 1, wherein, The step 6 specifically comprises the following steps: calculating the voltage value of each group of heating lamps for the first iteration: First, the expected temperature vector of each temperature measuring point i is defined is: （2） wherein, T is the final expected temperature of the ith temperature measurement point, n is the total number of temperature measurement points, and is multiplied by the regression matrix B obtained in step 3 to obtain the voltage of each group of heating lamps in the first iteration cycle is: = (3) wherein, Vj is the voltage value for the jth group of heating lamps in the first iteration loop, m is the total number of groups of heating lamps, and since each group of heating lamps needs to be multiplied by the respective thermal effectiveness factor in step 5 , It is also necessary to multiply the preceding by the optimization matrix K: (4) wherein is the thermal effective coefficient of the jth group of heating lamps in step 5, so the corrected voltage vector should be: (5) and taking the voltage value as the input of the first iteration.

3. A voltage control based temperature control method for radiant heating lamps as defined in claim 1, wherein, The step 7 specifically comprises the following steps: The temperature of each temperature measurement point is affected by one or more groups of heating lamps, so a mapping ； This mapping represents a mapping from the set 1 to m to the set 1 to n and provides represents the influence of the jth group of heating lamps on the ith temperature measurement point; According to the data in the previous steps, the optimization iteration is performed, and the voltage data of the m groups of lamps in the next actual measurement are obtained, and the iteration formula is: （6） wherein, Tj(i) is the temperature of the temperature measuring point i controlled by the jth group of heating lamps after heating, Tj(i) is the temperature of the temperature measuring point i controlled by the jth group of heating lamps after heating, Vj is the rated voltage of the jth group of heating lamps, Vj is the rated voltage of the jth group of heating lamps; It is specified that the temperature difference between the wafer center temperature measuring point and the wafer edge temperature measuring point after the fth iteration cycle is i.e. （7） wherein, Tf is the final temperature of the wafer center temperature measurement point at the fth cycle, Tf is the final temperature of the wafer edge temperature measurement point at the fth cycle; When the number of iterations is sufficiently large, the ratio of the current fth iteration cycle to the previous iteration cycle tends to 1, i.e.: the ratio of the current fth iteration cycle to the previous iteration cycle （8） In formula (8), when the iteration number f is constantly increased, Tends to a fixed value, at which time the iteration of step 7 is stopped and the voltage value of each group of lamps in this cycle is saved as the final result.

Images