DE102018125531A1 - Device and method for controlling the temperature in a CVD reactor - Google Patents

Abstract

The invention relates to a device and a method for temperature control in a CVD reactor (1), in which a first temperature (.) And a temperature (.) Are measured with a first temperature measuring device (2) at a first measuring point (17) on a substrate (8) with a second temperature measuring device (3) at a second measuring point (18) outside the substrate (8) a second measured value (T) of a temperature is measured. To regulate a heater, a recalibration factor is applied to the temperature (T), the recalibration factor being a quotient of a large number of first measured values (T) and second measured values (T) measured in the past. In particular, it is the quotient of two mean values.

Description

Technical field

The invention relates to a method for temperature control in a CVD reactor and a CVD reactor, in which a substrate temperature is measured with a first temperature measuring device at a first measuring point and the temperature of a susceptor or a substrate holder is measured at a second measuring point. The CVD reactor has a control device for controlling the substrate temperature.

State of the art

Methods for measuring the temperature of substrate surfaces or of susceptor surfaces, in which temperatures are measured by means of two pyrometers at different points on the susceptor, are known, for example, from US Pat US 8,888,360 B2 ; 9,200,965 B2 or US 6,398,406 B1 previously known.

The formation of target values from different temperatures is also in the DE 10 2015 100 640 A1 described.

Summary of the invention

When layers are deposited on substrates in a CVD reactor, first measured values of a substrate temperature are measured at first measuring points and second measured values of a susceptor temperature are measured at second measuring points outside the substrate. The measurement is usually carried out using two different temperature measuring devices, which temperature measuring devices can be pyrometers. The temperature measuring devices can deliver qualitatively different measured values, the measured values differing qualitatively in that, for example, only the second measured value is technically suitable for regulation and the first measured value is not technically suitable for regulation. For example, the first measured value cannot be used technically for a control because it is a lagging measured value that is only available with a time delay and because the first measured value is subject to fluctuations or is technically difficult to determine due to surface properties, emission properties or reflection properties of the substrate . However, the technologically relevant temperature is not the temperature measured with the second measured value, but the surface temperature of the substrate, since chemical or physical reactions take place on this surface. For example, a semiconductor layer consisting of several components is deposited in a CVD reactor according to the invention. The CVD reactor can be used, for example, to deposit GaN layers or AlN layers. These layers can be deposited on silicon substrates but also on sapphire substrates. The material of the layers or of the substrate can be transparent to infrared light, so that the first measured values cannot be determined with an IR pyrometer.

Proposals have already been made from the above-mentioned prior art on how to use mathematical functions, the arguments of which are several measured values measured at different times, to generate an actual temperature value which largely corresponds to the substrate temperature and which is used to control a heating device can be used to heat the susceptor and the substrate carried by the susceptor.

The object of the invention is to further improve the above-mentioned method for generating an actual temperature and to provide a CVD reactor which can be used for this purpose.

The object is achieved by the invention specified in the claims, the subclaims not only being advantageous developments of the solution specified in the independent claims, but also representing independent solutions to the object.

First and foremost, it is proposed that a recalibration factor be used to determine an actual value regulated against a target value, in particular an actual temperature. The recalibration factor is obtained from at least several first measured values, which can be obtained in a time interval of at least 10 seconds in the past. The current second measured value is multiplied by the recalibration factor. In a first variant, the recalibration factor is formed from an average of first measured values measured in a time interval. The time interval can contain measured values that are measured at a time that is one time behind the current second measured value. The time course of the substrate temperature, that is to say the first measured value, can be delayed in relation to the time course of the susceptor temperature, that is to say the second measured value. The time delay is of the order of 10 to 30 seconds. The time delay is due to various factors, for example the inertia of the system, the different heat flow paths, the signal processing times and a rotation of the susceptor about an axis of rotation. Typically, the susceptor rotates at 5 rpm. The used to form the mean Measured values in particular contain measured values that are a time behind the time of the measurement of the current second measured value, which corresponds, for example, to one third, half or a whole of the time by which the time course of the first measured value is delayed compared to the time course of the second measured value is. In particular, it is provided that the first measurement values are used to obtain the recalibration factor, which lie back by a time that corresponds at least to the time of one revolution of the susceptor. These times are typically above 4 seconds or above 12 seconds. In a preferred variant, the recalibration factor contains a quotient from the first value, in particular the mean value discussed above, and a second value, the second value being formed from second measurement values in the past. Like the first value, the second value can be an average of a plurality of first measured values measured in a time interval. Here, too, the time interval is preferably at least the rotation time of the susceptor or a delay time by which the two measured values change with a time delay and in particular assume a steady state after a temperature change, or at least 10 seconds. The first temperature values or the second temperature values can be used directly to form the mean value. However, it is also provided to filter the temporal course of the temperature measured values with a low-pass filter beforehand. As an alternative to averaging, the first value for generating the recalibration factor and / or the second value for generating the recalibration factor can also be obtained in each case via a low-pass filtered temperature profile over time. The cut-off frequency of the digital low-pass filter used here, in particular, can correspond to the time mentioned above, that is to say, for example, the round trip time of the susceptor or 10 seconds or more. The cut-off frequency of the low-pass filter can also be the reciprocal time by which the times are apart at which the two measured values again reach a steady state after a temperature change. The limit frequency is in particular a maximum of 0.1 Hz. The second temperature measured outside the substrate at the susceptor is therefore not used for the control, but rather a mixed temperature which is calculated from a product of a first mean value and the second measured value. To form the first average, a large number of first measured values are integrated within an integration interval of at least 10 seconds. The first average thus forms an average over time of the first temperature for a certain past time. The recalibration factor is preferably calculated as follows: $$R_c=\frac{M_1}{M_{\mathrm{2nd}}}$$

The first value depends M ₁ preferably exclusively from first values measured in a previous time interval, that is to say in particular from the substrate temperatures. The second value preferably depends M ₂ exclusively from second values measured in a previous time interval, that is to say preferably from the susceptor temperatures or substrate holder temperatures. In a further development of the method according to the invention, it can be provided that the time interval or as the reciprocal time the cut-off frequency of a low-pass filter lasts longer than 10 seconds, for example at least 15 seconds, at least 20 seconds, at least 40 seconds, at least 60 seconds, at least 80 seconds , be at least 100 seconds or at least 120 seconds long. The longer the time interval, the slower the control reacts to changes in the measured values supplied with the first temperature measuring device. It is envisaged that a measured value is obtained approximately every second. It is further provided that each mean value is formed from at least ten measured values, preferably at least 20 or more than 30 measured values.

wobei τ ein Zeitintervall angibt, das zwischen 10 Sekunden und 120 Sekunden liegen kann oder auch länger andauern kann. Mit T₁ ist der Messwert der ersten Temperatur bezeichnet, der beispielsweise in Sekundenabständen gemessen wird. In einer Variante der Erfindung ist der zweite Wert ein Mittelwert M₂ , der wie folgt gebildet wird: $$M_2=\frac{1}{\tau }{\displaystyle {\int }_{t-\tau }^t{T}_{2}dt}$$

wobei τ auch hier eine Zeit von größer 10 Sekunden ist, die insbesondere im Bereich zwischen 15 Sekunden und 120 Sekunden liegen kann. Mit T₂ ist ein Messwert der zweiten Temperatur bezeichnet der in Sekundenabständen gewonnen werden kann.The CVD reactor used can be a CVD reactor as is known from the prior art. The CVD reactor has a gas-tight, evacuable housing in which a process chamber is located. The bottom of the process chamber is formed by a susceptor. The susceptor can be an optionally coated graphite disc, which can be heated from below with a heating device, for example an RF or IR heating device. One or more substrates can be arranged on the broad side surface of the susceptor facing the process chamber. Cover plates can be provided between the substrates, which cover the susceptor surface. The one or more substrates can be arranged on substrate holders which lie in pockets of the susceptor. The susceptor can be rotated around its figure axis. The substrate holders lie rotatably in the pockets and can rest on a gas cushion, which can set the substrate holder in rotation about its axis. A gas inlet element can be assigned to the process chamber ceiling. It can be a central gas inlet element arranged in the center of the process chamber. However, it is also provided that the gas inlet element is formed by a showerhead, which extends essentially over the entire broad side surface of the susceptor and has a large number of gas outlet openings. Through the gas inlet element, gaseous starting materials are together with a carrier gas is introduced into the process chamber, where the starting materials, which are hydrides of elements of the 5th main group and organometallic compounds of the III . Main group act, disassemble. The temperature profile within the process chamber influences the layer growth of decomposition products of the gaseous starting materials, for example GaN or AlN. The substrates can be sapphire or silicon substrates. Above the process chamber ceiling there are two pyrometers, one of which has the first measured value at a first measuring point on a substrate and a second pyrometer at a second measuring point outside the substrate, for example on the susceptor surface or on the bottom of a pocket in which a substrate holder is stored, deliver a second reading. According to a variant of the invention, the first value is an average M ₁ which is calculated as follows $${\mathrm{<figure-callout\; id="1"\; label="M"\; filenames="DE102018125531A1\_0008.png,DE102018125531A1\_0009.png"\; state="\{\{state\}\}">M</figure-callout>}}_1=\frac{1}{\tau }{\displaystyle {\int }_{t-\tau }^t{T}_{1}dt}$$

where τ indicates a time interval that can be between 10 seconds and 120 seconds or can last longer. With T ₁ denotes the measured value of the first temperature, which is measured, for example, at intervals of seconds. In a variant of the invention, the second value is an average M ₂ , which is formed as follows: $$M_{\mathrm{2nd}}=\frac{1}{\tau }{\displaystyle {\int }_{t-\tau }^t{T}_{\mathrm{2nd}}dt}$$

where τ is also a time of more than 10 seconds, which can be in particular in the range between 15 seconds and 120 seconds. With T ₂ is a measured value of the second temperature that can be obtained in intervals of seconds.

For averaging the averages M ₁ , M ₂ can't just measure the temperature T ₁ , T ₂ be used. It is also provided, initially from the measured temperatures T ₁ , T ₂ low pass filtered temperatures T ₁ ' and or T ₂ ' form to get out of these dynamically filtered temperatures T ₁ ' , T ₂ ' To form averages.

ein Ist-Wert einer Temperatur berechnet, welcher Ist-Wert die Temperatur ist, die eine Regeleinrichtung gegen einen Soll-Wert regelt.According to the first variant, the first mean value becomes M ₁ a recalibration factor Rc and from it and a second reading T ₂ as follows $$T_R\left(t\right)=Rc\ast T_{\mathrm{2nd}}\left(t\right)$$

an actual value of a temperature is calculated, which actual value is the temperature that a control device controls against a target value.

Bei den Werten M₁ und M₂ handelt es sich gewissermaßen um geglättete, in der Vergangenheit zurückliegende erste und zweite Temperaturen. Je kleiner das Zeitintervall ist, innerhalb dessen die Glättung stattfindet, desto mehr verhält sich die Temperatur T_R im kurzfristigem Regime wie die Substrattemperatur T₁ , je größer die Zeit ist desto mehr verhält sich T_R wie die Suszeptortemperatur T₂ ; Mittelfristig konvergiert sie immer gegen die Substrattemperatur T₁ .According to a second, preferred variant of the invention, the actual temperature is calculated as follows $$T_R\left(t\right)=\frac{1}{M_{\mathrm{<figure-callout\; id="1"\; label="2nd\; \ast \; M"\; filenames="DE102018125531A1\_0008.png,DE102018125531A1\_0009.png"\; state="\{\{state\}\}">2nd\; </figure-callout>}}}\ast M_{}{}_1\ast T_{\mathrm{2nd}}\left(t\right)$$

With the values M ₁ and M ₂ they are, to a certain extent, smoothed past and first temperatures in the past. The smaller the time interval within which the smoothing takes place, the more the temperature behaves T _R in the short-term regime like the substrate temperature T ₁ , the longer the time, the more it behaves T _R like the susceptor temperature T ₂ ; In the medium term, it always converges to the substrate temperature T ₁ .

Figure list

• 1 einen Querschnitt durch einen CVD-Reaktor,
• 2 eine Darstellung gemäß 1 eines zweiten Ausführungsbeispiels der Erfindung,
• 3 eine Darstellung gemäß 1 eines dritten Ausführungsbeispiels der Erfindung,
• 4a schematisch den Temperaturverlauf T₁ der Substrattemperatur bei einer Verminderung der Heizleistung,
• 4b schematisch den Verlauf der Suszeptortemperatur bzw. Substrathaltertemperatur nach dem Absenken der Heizleistung,
• 4c den zeitlichen Verlauf der unter Verwendung eines Rekalibrierungsfaktors gebildeten Ist-Wert, der zur Temperaturregelung verwendet wird und
• 4d schematisch den zeitlichen Verlauf des Rekalibrierungsfaktors, der beim Ausführungsbeispiel ein Quotient aus einem ersten Mittelwert von in einem Zeitintervall gemessenen Substrattemperaturen und einem zweiten Mittelwert aus in einem Zeitintervall gemessenen Suszeptor- oder Substrathaltertemperaturen ist.

An embodiment of the invention is explained below with reference to the accompanying drawings. It shows:

• 1 a cross section through a CVD reactor,
• 2nd a representation according to 1 a second embodiment of the invention,
• 3rd a representation according to 1 a third embodiment of the invention,
• 4a schematically the temperature curve T ₁ the substrate temperature when the heating power is reduced,
• 4b schematically the course of the susceptor temperature or substrate holder temperature after the heating power has been reduced,
• 4c the time profile of the actual value formed using a recalibration factor, which is used for temperature control, and
• 4d schematically the time course of the recalibration factor, which in the exemplary embodiment is a quotient of a first mean value of substrate temperatures measured in a time interval and a second mean value of susceptor or substrate holder temperatures measured in a time interval.

Description of the embodiments

The one in the 1 to 3rd CVD reactor shown 1 consists of a gas-tight housing, in particular made of stainless steel, in which there is an axis of rotation 16 rotary driven susceptor 6 made of graphite or coated graphite. Below the susceptor 6 there is a heater 5 with which the susceptor 6 can be heated.

On the top of the susceptor 6 there are pockets 13 , in each of which substrate holder 7 are arranged. The substrate holder 7 can rest on a gas cushion and around axes of rotation 16 are driven. Any substrate holder 7 carries at least one substrate to be coated 8th . The susceptor 6 has a circular disc shape. The substrate holder 7 are ring-shaped around the axis of rotation 16 arranged.

In the exemplary embodiment, the CVD reactor has a central gas inlet element 12th , through which the process gases mentioned in the beginning can flow into the process chamber, down through the susceptor 6 and up through a process chamber ceiling 9 is limited.

The process chamber ceiling 9 has openings 10th , 11 . Above the openings 10th , 11 there are two temperature measuring devices 2nd , 3rd , which can be pyrometers that supply measurement signals to a control device 4th be forwarded. The control device 4th uses that from the temperature measuring devices 2nd , 3rd obtained first and second temperature readings to the heater 5 to regulate.

The first temperature measuring device 2nd measures along a first optical path 14 through the opening 11 through at a first measuring point 17th a surface temperature of the substrate 8th . The second temperature measuring device 3rd measures along a second optical path 15 through the opening 10th through at a second measuring point 18th a temperature of the susceptor 6 . The second temperature measuring device 3rd measures along a second optical path 15 through the opening 10th through at a second measuring point 18th a temperature at which in the 1 illustrated embodiment and in which in the 3rd illustrated embodiment, the temperature of the susceptor 6 and the one in the 2nd illustrated embodiment, the temperature of the substrate holder 7 is.

The one in the 1 illustrated embodiment is the second measuring point 18th on the bottom of a bag 13 so the optical path 15 through an annular gap between the substrate holder 7 and the pocket wall runs through.

The one in the 2nd illustrated embodiment runs the optical path 15 through that on the substrate holder 7 overlying substrate 8th through it. The temperature measuring device 3rd is an IR pyrometer. The substrate 8th is transparent to infrared light, so that with the IR pyrometer the surface temperature of the substrate holder 7 can be determined. Alternatively, the measuring point 18th but also next to the substrate 8th on the top of the substrate holder 7 lie.

The one in the 3rd Embodiment shown is with the second temperature measuring device 3rd the surface temperature of the susceptor 6 measured. The measuring point 18th is right next to the substrate holder. The measuring point 18th can be both radially inside and radially outside of the substrate holder 7 lie. However, it can also be located at a location on the susceptor surface that is between two adjacent substrate holders 7 is arranged. It is also provided that the measuring point 18th for measuring the susceptor temperature also on the underside of the susceptor 6 can be arranged. Pyrometers or thermocouples or the like can be used to measure the susceptor temperature.

In the second temperature measuring device 2nd it can be a UV pyrometer, with which the surface temperature of the substrate 8th is measured.

On the susceptor 6 can also cover plates. The second measuring point can also be arranged on one of the cover plates.

The control device 4th is designed in such a way that, within predetermined time intervals, which are at least 10 seconds, the first temperature measured values T ₁ mathematically linked to an algebraic mean M ₁ the first temperature T ₁ to form over the interval. It can also be provided that the control device is set up in such a way that it comprises a plurality of measured values of the second temperature T ₂ a second algebraic mean over an interval of at least 10 seconds M ₂ the second temperature T ₂ forms.

The control device 4th is also set up to be from the first mean M ₁ and the currently measured second measured value T ₂ the second temperature at the second measuring point 18th an actual value T _R forms the regulation of the heating device 5 is used, the modified actual temperature T _R at least from a product of the first mean M ₁ and the current second measured value T ₂ consists.

In a variant of the invention it can be provided that the modified actual value T _R not just the product of the first mean M ₁ and the current second measured value T ₂ exists, but also by a second mean M ₂ the second temperature has been divided.

To determine an average, at least ten measured values of either the first temperature are preferred T ₁ or the second temperature T ₂ used.

The 4a shows the reaction of temperature T ₁ of the substrate 8th when the heat output to reduce the temperature at the time t ₁ is reduced. The substrate temperature reaches during a cooling time of approx. 20 to 30 seconds T ₁ at a time t ₃ their minimum to then take a lower value after an overshoot caused by the regulation.

The 4b shows the temperature over time T ₂ of the susceptor 6 or the substrate holder 7 after a decrease in the setpoint temperature. The temperature T ₂ reached earlier, namely at the time t ₂ their minimum in order to then assume a substantially constant value after an overshoot caused by the control algorithm. From the 4a and 4b it can be seen that the minimum of the substrate temperature T ₁ at a later time t ₃ is reached as the minimum of the susceptor temperature T ₂ which is already at a time t ₂ is achieved. The time difference between the two times t ₂ and t ₃ is in the range of 10 to 20 seconds.

The 4a and 4b show that the temperature T ₁ slightly delayed after the heating output has decreased at the time t ₁ versus temperature T ₂ sinks. If you look at the time after time t ₃ observed overshoot, it can be seen that the system has a generic time in the form of the time difference t ₃ minus t ₂ possesses, i.e. the time within which the two temperatures after a temperature change T ₁ , T ₂ return to their steady state.

The 4d shows one with the current temperature T ₂ the recalibration factor to be multiplied by the susceptor Rc , which is the inertia of the history of the substrate temperature T ₁ considered. In the exemplary embodiment, the recalibration factor Rc is formed by the quotient of two mean values, the mean of the first temperatures in the counter T ₁ and the mean of the second temperatures in the denominator T ₂ stands.

Die Integrationszeiten zur Bildung der Mittelwerte M₁ , M₂ beträgt hier zumindest die Zeit, die der Suszeptor für einen Umlauf um seine Drehachse benötigt. Anstelle der Mittelwerte M₁ , M₂ können aber auch tiefpassgefilterte Temperaturverläufe verwendet werden. Die Grenzfrequenz des dabei verwendeten, insbesondere digitalen Tiefpassfilters ist maximal der Kehrwert der Umlaufzeit des Suszeptors.The 4c shows the calculated actual temperature used for control T _R , which is calculated as follows: $$T_R\left(t\right)=\frac{1}{M_{\mathrm{<figure-callout\; id="1"\; label="2nd\; \ast \; M"\; filenames="DE102018125531A1\_0008.png,DE102018125531A1\_0009.png"\; state="\{\{state\}\}">2nd\; </figure-callout>}}}\ast M_{}{}_1\ast T_{\mathrm{2nd}}\left(t\right)$$

The integration times for averaging M ₁ , M ₂ is at least the time that the susceptor needs for one revolution around its axis of rotation. Instead of averages M ₁ , M ₂ low-pass filtered temperature profiles can also be used. The limit frequency of the, in particular digital, low-pass filter used here is at most the reciprocal of the round trip time of the susceptor.

The integration time for forming the mean values M ₁ , M ₂ but can also at least the time difference t ₃ minus t ₂ be. When using a low-pass filter, the maximum cut-off frequency is equal to the reciprocal of this time difference, the time difference being the time by which the first temperature T ₁ the second temperature T ₂ runs after.

Depending on the design of the CVD reactor, the temperature at the measuring points reacts T ₁ respectively. T ₂ different in time for a change in the heating power supplied. This leads to an overestimation or underestimation of the recalibration factor in dynamic situations. By means of suitable filtering of the signals T ₁ and T ₂ the temporal response of the filtered quantities can be T ₁ ' and T ₂ ' compensate for a change in heating output. A suitable filtering can be a low pass filter. In some embodiment variants of the recalibration method, the combination of the temperature signals is sufficient T ₁ and T ₂ ' respectively. T ₁ ' and T ₂ in order to obtain the recalibration factor in sufficient quality.

It is therefore also provided that averaging, as described above, is not carried out with the directly measured temperatures, but with previously filtered temperatures.

The above statements serve to explain the inventions covered by the application as a whole, which also independently further develop the prior art at least through the following combinations of features, it being possible for two, more or all of these combinations of features to also be combined, namely:

A process for temperature control in a CVD reactor 1 , in which with a first temperature measuring device 2nd at a first measuring point 17th on a substrate 8th first measured values T ₁ a temperature and with a second temperature measuring device 3rd at a second measuring point 18th outside or below the substrate 8th second readings T ₂ a temperature can be measured, with the determination of an actual value regulated against a target value T _R at least from the first measured values in the past T ₁ a recalibration factor Rc is obtained with the current second measured value T ₂ is multiplied.

A CVD reactor 1 with a first temperature measuring device 2nd , which is set up so that it is at a first measuring point 17th that on a substrate 8th is arranged, first measured values T ₁ supplies a temperature and with a second temperature measuring device 3rd , which is set up so that it is at a second measuring point 18th outside or below the substrate 8th measured second measured values T ₂ supplies a temperature with a control device 4th for temperature control, the control device 4th is set up in such a way that for determining an actual value regulated against a target value T _R at least from the first measured values in the past T ₁ a recalibration factor Rc is obtained with the current second measured value T ₂ is multiplied.

A process or a CVD reactor 1 which or which is characterized in that the recalibration factor Rc a quotient of a first measured value from the past T ₁ formed first value M ₁ and one from previous second measured values T ₂ formed second value M ₂ is.

A process or a CVD reactor 1 , which or which, is characterized in that a characteristic time around which the at least one first measured value T ₁ or at least a second reading T ₂ compared to the time of the determination of the actual value, the time of a rotation of the susceptor about its axis of rotation or a time difference t ₃ minus t ₂ around which the first measured value is concerned T ₁ delayed compared to the second measured value T ₂ changes and in particular returns to a steady state after a change in heating output or is at least 10 seconds, at least 15 seconds, at least 20 seconds, at least 40 seconds, at least 60 seconds, at least 80 seconds, at least 100 seconds or at least 120 seconds.

A process or a CVD reactor 1 , which or which, is characterized in that the mean values M ₁ , M ₂ be formed over the characteristic time.

A process or a CVD reactor 1 , which or which is characterized in that for determining the actual value T _R with a low-pass filtered first measurement T ₁ and / or second measured values T ₂ are used, the cut-off frequency of the low-pass filter being the reciprocal characteristic time.

A process or a CVD reactor 1 , which or which, is characterized in that the mean values M ₁ , M ₂ from low-pass filtered first and second measured values T ₁ , T ₂ are formed.

A process or a CVD reactor 1 , which or which, characterized in that to form the recalibration factor Rc a time-shifted mean M ₂ the second temperature T ₂ is used, and in particular an additional time-shifted mean M ₁ the first temperature T ₁ is used.

A process or a CVD reactor 1 , which or which, characterized in that the CVD reactor 1 one from its bottom with a heater 5 Heated susceptor, the second measuring point 18th a top of the susceptor 6 , an underside of the susceptor 6 , the bottom of a bag 13 in the susceptor 6 in which a substrate holder 7 is rotatably arranged, the at least one substrate 8th carries a dot on the top of the substrate holder 7 next to the substrate 8th or one below the substrate 8th lying place on the substrate holder 7 assigned.

A process or a CVD reactor 1 , which or which, characterized in that the first and second temperature measuring devices 2nd , 3rd Pyrometers are their optical paths 14 , 15 through openings 10th , 11 a process chamber ceiling 9 pass through and / or that the first measuring device 2nd for measuring the first measured value T ₁ , which corresponds to a substrate temperature, is a UV pyrometer and that the second temperature measuring device 3rd which is a measurement of the temperature of the substrate holder 7 or the susceptor 6 delivers, is an IR pyrometer and / or that for measuring the second temperature T ₂ a thermocouple in particular on the underside of the susceptor 6 is used.

Reference list

1

CVD reactor

2nd

Temperature measuring device

3rd

Temperature measuring device

4th

Control device

5

Heating device

6

Susceptor

7

Substrate holder

8th

Substrate

9

Process chamber ceiling

10th

opening

11

opening

12th

Gas inlet member

13

bag

14

optical path

15

optical path

16

Axis of rotation

17th

first measuring point

18th

second measuring point

M ₁

first mean

M ₂

second mean

T ₁

first reading

T ₂

second reading

T ₁ '

low pass filtered temperature

T ₂ '

low pass filtered temperature

t ₁

time

t ₂

time

t ₃

time

T _R

Actual temperature value

Rc

Recalibration factor

QUOTES INCLUDE IN THE DESCRIPTION

This list of documents listed by the applicant has been generated automatically and is only included for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent literature cited

• US 8888360 B2 [0002]
• US 9200965 B2 [0002]
• US 6398406 B1 [0002]
• DE 102015100640 A1 [0003]

Claims (11)

Method for temperature control in a CVD reactor (1), in which first measurement values (T ₁ ) of a temperature are measured with a first temperature measurement device (2) at a first measurement point (17) on a substrate (8) and with a second temperature measurement device (3) At a second measuring point (18) outside or below the substrate (8), second measured values (T ₂ ) of a temperature are measured, with an actual value (T _R ) regulated against a target value being determined, at least from first measured values dating back in time ( T ₁ ) a recalibration factor (Rc) is obtained, which is multiplied by the current second measured value (T ₂ ). CVD reactor (1) with a first temperature measuring device (2), which is set up in such a way that it supplies first measured values (T ₁ ) of a temperature at a first measuring point (17), which is arranged on a substrate (8), and with a second temperature measuring device (3), which is set up in such a way that it delivers second measured values (T ₂ ) of a temperature measured at a second measuring point (18) outside or below the substrate (8), with a control device (4) for temperature control, wherein the control device (4) is set up in such a way that for determining an actual value (T _R ) regulated against a target value, a recalibration factor (Rc) is obtained at least from previous first measured values (T ₁ ), which is based on the current second measured value (T ₂ ) is multiplied. Procedure according to Claim 1 or CVD reactor (1) Claim 2 Characterized in that the Rekalibrierungsfaktor (Rc) is a quotient of a formed from temporally previous first measurement values (T ₁₎ the first value (M ₁₎ and one from the past second measured values (T ₂₎ the second value is formed (M _2). Method or CVD reactor (1) according to one of the preceding claims, characterized in that a characteristic time by which the at least one first measured value (T ₁ ) or at least one second measured value (T ₂ ) compared to the time of the determination of the actual Value, the time of one revolution of the susceptor around its axis of rotation or a time difference (t ₃ ) minus (t ₂ ) by which the first measured value (T ₁ ) changes with a time delay compared to the second measured value (T ₂ ) and in particular after one Change in heating power returns to a steady state or is at least 10 seconds, at least 15 seconds, at least 20 seconds, at least 40 seconds, at least 60 seconds, at least 80 seconds, at least 100 seconds or at least 120 seconds. Method or CVD reactor (1) according to Claim 4 , characterized in that the mean values (M ₁ , M ₂ ) are formed over the characteristic time. Method or CVD reactor (1) according to one of the preceding claims, characterized in that first measured values (T ₁ ) and / or second measured values (T ₂ ) are used to determine the actual value (T _R ) with a low-pass filter, wherein the cut-off frequency of the low-pass filter is the reciprocal characteristic time. Method or CVD reactor (1) according to one of the preceding claims, characterized in that the mean values (M ₁ , M ₂ ) are formed from low-pass filtered first and second measured values (T ₁ , T ₂ ). Method or CVD reactor (1) according to one of the Claims 4 to 7 , characterized in that a time-shifted mean value (M ₂ ) of the second temperature (T ₂ ) is used to form the recalibration factor (Rc) and in particular an additional time-shifted mean value (M ₁ ) of the first temperature (T ₁ ) is used. Method or CVD reactor (1) according to one of the preceding claims, characterized in that the CVD reactor (1) has a susceptor which can be heated from its underside with a heating device (5), the second measuring point (18) being an upper side of the susceptor (6), an underside of the susceptor (6), the bottom of a pocket (13) in the susceptor (6), in which a substrate holder (7) is rotatably arranged, which carries at least one substrate (8), a point on the top the substrate holder (7) next to the substrate (8) or a location below the substrate (8) on the substrate holder (7). Method or CVD reactor (1) according to one of the preceding claims, characterized in that the first and second temperature measuring devices (2, 3) are pyrometers whose optical paths (14, 15) through openings (10, 11) of a process chamber ceiling (9 ) and / or that the first measuring device (2) for measuring the first measured value (T ₁ ), which corresponds to a substrate temperature, is a UV pyrometer and that the second temperature measuring device (3), which measures the temperature of the substrate holder (7 ) or the susceptor (6), is an IR pyrometer and / or that a thermocouple is used in particular on the underside of the susceptor (6) to measure the second temperature (T ₂ ). CVD reactor (1) or method for temperature control in a CVD reactor (1), characterized by one or more of the characterizing features of one of the preceding claims.

Images