RU2054749C1 - Method of contactless measurement of temperature of silicon plate - Google Patents

Abstract

FIELD: radiation pyrometry. SUBSTANCE: it is proposed to take into account change of component of extraneous radiation reflected from plate tied up with change of illumination brightness of heating sources to increase accuracy of temperature measurement. Summary radiation (inherent P and reflected) PP₁ reflected from plate, extraneous radiation RP₂ not hitting plate are measured with the aid of silicon photodiode. P is found by formula P=(P₁-P₂)J_(1-0)/J_(2-0) and plate temperature is calculated by its value, where J_(1-0)/J_(2-0) is constant coefficient, J_(1-0) is value of current of silicon photodiode at point where value of first derivative of this current has grown positive after its passing through zero; J_(2-0) is value of current of additional photodiode measured simultaneously with J_(1-0). EFFECT: increased accuracy of temperature measurement. 15 dwg

Description

 The invention relates to radiation pyrometry, and more particularly to non-contact methods for measuring the temperature of silicon wafers irradiated to heat incandescent spirals, gas discharge lamps, lasers, etc. with external irradiators.

 In the manufacture of semiconductor structures, it is necessary to conduct their heat treatment, which is carried out by heating third-party emitters. Such processes are diffusion, annealing after ion implantation, etc. In these processes, the main parameter is temperature, and it can be measured practically only by the non-contact method, especially when conducting pulsed thermal processes.

 When measuring the temperature of an object heated by radiation from third-party sources, one of the main tasks is to reduce the influence of third-party radiation on the results of measurements of the object’s own radiation.

 Non-contact temperature measurement methods that take into account the influence of external radiation are known. For example, there is a known method [1] for non-contact measurement of the temperature of a mirror surface of a semiconductor wafer heated from two sides of radiation from third-party sources with reflectors, including measuring the amount of intrinsic radiation of a heated wafer using a radiation detector located on the optical axis inclined to the perpendicular to the center of the measured surface , and reducing the influence of external radiation by using an absorbing element in the form of the equivalent of a completely black body located sim etrichno radiation detector with respect to the perpendicular to the center of the surface to be measured.

 However, this method provides high accuracy of temperature measurements only for perfectly smooth surfaces, since in this method only external radiation is eliminated, which falls on the radiation detector as a result of specular reflection from the measured surface of the plate. The diffuse component of the external radiation reflected from the plate surface is not taken into account. At a high intensity of external radiation, which is characteristic, for example, for pulsed heating, the diffuse component of the reflected external radiation in the direction of the radiation detector can be comparable and even greater than the value of the intrinsic radiation of the measured plate, especially when the surface treatment quality is low.

The known method adopted for the prototype, non-contact temperature measurement of silicon is not ideal mirror structures in the process of pulsed heat treatment [2] This method measures the temperature of a silicon wafer with a silicon photodiode in the range of 900-1400 ^about With its heating by external radiation, for example by radiation of halogen lamps with a power density 40-200 W / cm ² . Using this method, the total radiation is measured, including the intrinsic radiation of the plate and the third-party radiation transmitted through the plate and reflected from it, then the third-party radiation is measured, and the intrinsic radiation of the plate is determined by the difference of the total and third-party radiation.

 The influence of extraneous radiation transmitted through the plate in the aforementioned measurement range is eliminated by selecting the temperature measurement start time, at which the heating plate does not transmit external radiation, the subtracted amount of the remaining external radiation reflected from the plate is measured after a certain time from the moment the lamps are turned on. However, this method has a significant drawback, namely that it is applicable only when heated by a single rectangular pulse of radiation. The need to maintain or change the temperature over time requires a change in the intensity of the external radiation, for example, a change in the incandescent lamp. In this case, the magnitude of the external radiation reflected from the plate changes, and this method, in which, when calculating as the deductible value, only the once-measured constant value of the external radiation is used leads to a large error in determining the temperature.

 The aim of the proposed technical solution is to increase the accuracy of measuring the temperature of a silicon wafer when it is heated by external radiation measured in intensity.

где I_1-0 значение тока кремниевого фотодиода в точке, где величина первой производной этого тока по времени стала положительной после прохождения ее через ноль; а I_2-0 значение тока дополнительного фотодиода, измеренное одновременно с I_1-0.The goal is achieved by the fact that by a non-contact method for measuring the temperature of a silicon wafer in the range of 900-1400 ° C when it is heated by external radiation, including measuring the optical radiation P _{1 of the} silicon radiation photodiode that is optically coupled ^{to the} wafer and consists of the intrinsic radiation of the wafer P reflected from it radiation P _neg. and passing through it of external radiation P _pr. , the choice of time to start the temperature measurement to eliminate the influence of P _ave. , measurement P _neg. and calculating the temperature of the plate by the value of P, defined as the difference P ₁ P _neg. , measured by an additional photodiode that does not have an optical connection with the plate, external radiation P ₂ not falling onto the plate, and the magnitude of P _neg. defined as P _neg. P ₂

where I _{1-0 the} value of the current of the silicon photodiode at the point where the magnitude of the first time derivative of this current has become positive after passing through zero; and I _{2-0 the} value of the current of the additional photodiode, measured simultaneously with I _1-0 .

K const (2)
На фиг. 2-9 изображены зависимости токов I₁ и I₂ во времени, а также первые производные I₁ по времени на начальном участке нагрева пластин кремния разных толщин и концентраций носителей заряда.In FIG. 1 shows a setup diagram in which the proposed method is implemented. In this setup, the photodiode 1 measures the total radiation P ₁ incident on it, consisting of the intrinsic radiation P of the plate 3, as well as the reflection from it of the _Pg. and passing through it, R _pr. third-party radiation. It follows that
P P ₁ (P _neg. + P _off ). (one)
By the value of P, the temperature of the plate is calculated according to the known laws of radiation. The photocurrent I _{2 of the} additional photodiode 2 is due only to external radiation P ₂ , since the diode is mounted in such a way that radiation from the plate does not fall on it. Since the radiation of P ₂ and P _neg. due to the same source, then at any moment of time there is equality

K const (2)
In FIG. Figures 2–9 show the time dependences of currents I ₁ and I ₂ , as well as the first derivatives of I ₁ with respect to time in the initial section of heating of silicon wafers of different thicknesses and concentrations of charge carriers.

In FIG. Figure 2 shows the time dependence of the photodiode currents when measuring the temperature of an n-type silicon wafer with a thickness of h 300 μm and a carrier concentration of N 1 · 10 ¹⁷ cm ^-3 . In FIG. 3 shows the time dependence of the derivative of the photodiode current for the silicon wafer of FIG. 2. In FIG. Figure 4 shows the time dependence of the current of the main photodiode when measuring the temperature of an n-type silicon wafer with a thickness of h 100 μm and a charge carrier concentration of N 1 · 10 ¹⁷ cm ^-3 . In FIG. 5 shows the time dependence of the derivative of the photodiode current for the silicon wafer of FIG. 4. In FIG. Figure 6 shows the time dependence of the current of the main photodiode when measuring the temperature of an n-type silicon wafer with a thickness of h 300 μm and a charge carrier concentration of N 5 · 10 ¹⁹ cm ^-3 . In FIG. 7 shows the time dependence of the derivative of the photodiode current for the silicon wafer of FIG. 6. In FIG. Figure 8 shows the time dependence of the current of the main photodiode when measuring the temperature of an n-type silicon wafer with a thickness of h 100 μm and a carrier concentration of N 5 · 10 ¹⁹ cm ^-3 . In FIG. 9 shows the time dependence of the derivative of the photodiode current for the silicon wafer of FIG. eight.

 In FIG. 10-15 show how the components of the photodiode currents and the plate temperature change over time. In FIG. 10 shows the change in the intensity of the external radiation in time. In FIG. 11 shows the change in current of the additional photodiode in time. In FIG. 12 shows the change in current of the main photodiode in time. In FIG. 13 shows the change in the subtracted part from the current of the main photodiode in time. In FIG. 14 shows the change in the current portion of the main photodiode due to the intrinsic radiation of the plate. In FIG. 15 shows the temperature change of the plate over time.

Section SD on the curve I ₂ f (t) (Fig. 2) corresponds to the time of heating of the lamp spirals and the emitters go to a stationary mode in terms of radiation intensity. This time for halogen lamps of the type KG-220-1000 with a tungsten spiral is 0.25 s. In the area SA (Fig. 2), the intrinsic radiation of the plate is very small and the current I ₁ is due only to the diffuse reflection of the external radiation from the plate P _neg. and third-party radiation passing through the plate to the photodetector R _pr . Section AB (Fig. 2) shows a decrease in the transmission of the plate due to a shift in the edge of the absorption band to the long-wavelength region of the spectrum. The decrease in transmittance associated with the heating of the plate causes a decrease in the fraction of external radiation passing to the photodetector through the plate by a value of P _a.s. Intrinsic radiation of the plate at this portion is also too low due to its low temperature (less than 900 ° ^C) and does not contribute to the total signal. In the VO region (Fig. 2), the photocurrent does not change, since the transmission of radiation through the plate is practically absent, the intensity of the external radiation has already become constant, and the intrinsic radiation of the plate is still small for indication by its silicon photodiode. With further heating of the plate, the intensity of its own radiation increases and, starting from point O (Fig. 2), a current component I ₁ appears due to its own radiation. Thus, a spurious signal can be measured in the horizontal portion of the VO, which is a photoresponse to third-party radiation reflected from the plate, and this signal is not associated with the plate's own radiation. In FIG. 4, relating to a thinner silicon wafer, points B and O practically coincide (the wafer heats up much faster).

Thus, over the entire area of VO (Fig. 2), equality (1) has the form P P ₁ P _{neg.article (VO)} or taking into account equality (2)
P R ₁ KR _{2 (BO)} , (3) where P _{bg (BO)} and P _{2 (BO) the} value of the radiation P _{b bg} and P ₂ in this section.

From consideration of FIG. 4, 6 and 8, it can be concluded that the current value I ₂ corresponding to P _{2 (BO)} (FIG. 2) should be measured at points O of FIG. 4 and 8 and in the portion BO of FIG. 6.

на фиг. 3, 5 и 7 следует, что наиболее удобно во всех случаях измерять величину I₂ в точках О, характерным для которых является появление сигнала от Р (т. е. начала фиксируемого фотодиодом собственного излучения), а математически то, что величина

становится положительной после прохождения ее через ноль. При этом выражение (3) для всех случаев (на фиг. 2, 4, 6) приобретает вид Р Р₁ K R_2(О) или для соответствующих токов
I I₁ KI_2(О), (4) где I_2(О) величина тока I₂ в точках О.From dependency analysis

in FIG. 3, 5 and 7 it follows that in all cases it is most convenient to measure the value of I ₂ at points O, characteristic of which is the appearance of a signal from P (i.e., the beginning of the intrinsic radiation fixed by the photodiode), and mathematically, that the quantity

becomes positive after passing through zero. Moreover, the expression (3) for all cases (in Fig. 2, 4, 6) takes the form P P ₁ KR _{2 (O)} or for the corresponding currents
II ₁ KI _{2 (О)} , (4) where I _{2 (О)} is the current value I ₂ at points O.

и далее определяя истинную величину I из равенства, полученного из выражения (4):
I I₁ KI₂, (5) где I₂ значение тока дополнительного фотодиода в любые последующие моменты времени.If it is necessary to maintain the temperature at a constant predetermined level or to adjust its value, it is necessary to change the intensity of the external radiation, maintaining the value at the corresponding preset level. When this is measured, the value of P _neg . This can be _{taken into account by} measuring P _neg. and P ₂ once at the point O, thus calculating K

and further determining the true value of I from the equation obtained from expression (4):
II ₁ KI ₂ , (5) where I ₂ is the current value of the additional photodiode at any subsequent time points.

In FIG. 10 shows the change in radiation intensity P _meas. external sources from time t: in the EF section, the change in the third-party radiation is carried out in order to maintain the temperature at one predetermined level, and in the section after the F point at another, lower level. In FIG. 11 and 12 show the corresponding change in the currents of the additional and main photodiodes. Unsharp changes in current I _{1 are} due to the thermal inertia of the plate. In FIG. 13 shows the change in the subtracted part from the current value of the main photodiode I ₁ (solid curve B). As can be seen from FIG. 13, this part changes when the magnitude of the external radiation changes. The same figure shows the deductible part if its change were not taken into account (dashed line A).

а на фиг. 15 измене- ние температуры t^о пластины, рассчитанное по этой величине I (сплошная кривая).In FIG. 14, the solid line shows the time variation of the calculated current component of the main photodiode II ₁ I ₂

and in FIG. 15 is the change in temperature t ^{о of the} plate, calculated from this value of I (solid curve).

In FIG. 14 and 15, the dotted line shows the change in I and t ^o calculated without taking into account changes in the deductible part. The difference between them (Δ t in Fig. 15) shows the error in determining the temperature, if you do not use the proposed method for adjusting the deductible part. The experiments showed that at a temperature level of 1000-1300 ^о С this error was 50 ^о or more when the unit was operating in the mode of maintaining the temperature at a given level (region EF).

PRI me R 1 specific implementation of the method. An n-type silicon wafer with a charge carrier concentration of N 1 · 10 ¹⁷ cm ^-3 and a thickness of h 300 μm was placed in the Impulse-5 installation, in the hole of the wall of which an additional photodiode of the brand ФД 265Б 3368.92 TU was installed, fenced off from the camera by a diaphragm to prevent ingress radiation reflected from the silicon wafer onto it. The Impulse-5 installation is electrically connected through an operational amplifier of the K 140 UD8A type with a Micro-86 computer, which contains the program for processing signals from photodiodes of the Impulse-5 installation in accordance with the claimed claims.

 The plate 3 was located in the installation chamber on quartz needles 4 between two panels of tubular halogen lamps 5 of the type KG-220-1000 3 so that the axis of the halogen lamps 5 are at a distance of 10-12 mm from the plate 3. The installation is equipped with water-cooled reflectors 6 , which serve to increase the heating rate and its uniformity. The signals from the photodiodes are fed to the control unit 7. The nature of changes in currents and temperature over time is displayed on display 8.

The signal of the main photodiode was preliminarily calibrated as a function of the plate temperature by the melting points of various metals. After the plate was installed in the chamber, halogen lamps were turned on, and the change in the signal of the main photodiode was continuously analyzed on a computer and reflected on the display. This change corresponded to FIG. 2, and at the point indicated in FIG. 2 as O, the computer recorded the values of the currents of the photodiodes, calculated the value of K and the current temperature I and the calibration dependence I f (t ^o ).

Then the plate continued heating at a constant radiation power level with continuous measurement of currents I ₁ and I ₂ and calculation on a computer of the magnitude of the component I of the current of the main photodiode according to formula (5). Upon reaching this current component equal to the value obtained at calibration, t ^o 1100 ^о С, the computer gave a signal to a change in the intensity of the side radiation in order to maintain the achieved value I. Moreover, in accordance with formula (5), the change in the level of intensity of the side radiation was taken into account. The experiments showed that if this accounting is not carried out, the error in the maintained temperature is ≈50 ^о С.

PRI me R 2. The heating of the plate of silicon n-type with a thickness of 100 μm with a concentration of charge carriers N 1 · 10 ¹⁷ cm ^-3 . The plate temperature was measured analogously to example 1, the curves of the response of the photoresponse signal and its first time derivative are similar to those shown in FIG. 4 and 5. In FIG. 4 and 5, in contrast to FIG. 1 and 2, there is no BO shelf; point O corresponds to it. This is due to the fact that at such a plate thickness its very rapid heating occurs. Curve analysis and calculations were carried out on a Micro-86 computer.

PRI me R 3. The n-type silicon wafer with a thickness of 300 μm was heated with a concentration of charge carriers N 5 · 11 ¹⁹ cm ^-3 . The plate temperature was measured analogously to Example 1; the nature of the curves of the change in the photoresponse signal and its first time derivative is shown in FIG. 6 and 7, where, in contrast to FIG. 2 and 3, peak A is absent, since the silicon wafer is highly resistive and does not pass radiation through itself. Curve analysis and calculations were carried out by the Micro-86 computer.

PRI me R 4. The heating of the plate of silicon n-type with a thickness of 100 μm with a concentration of charge carriers N 5 · 10 ¹⁹ cm ^-3 . The plate temperature was measured analogously to example 1, the nature of the curves of the signal and its first time derivative is shown in FIG. 8 and 9, where, in contrast to FIG. 1 and 2, there is no VO shelf and peak A, since a plate with such a high doping level initially does not pass through the radiation of halogen lamps and with a small thickness, the plate heats up very quickly.

The method of contactless measuring the wafer temperature in the temperature range 900-1400 ^{° C} by heating them by changing the intensity of the radiation pulses party allowed to eliminate the error of temperature measurement, which was exclude varying radiation reflected from an outside of the plate and was approximately 50-80 ^o. As a result, the accuracy of the measurement is determined only by the accuracy of the preliminary calibration, which can be realized with an acceptable error for practice by many modern methods (thermocouple, pyrometry, melting points of metals, etc.).

Claims (1)

METHOD contactless TEMPERATURE MEASURING silicon wafer in the range of 900 - 1400 ^o C in heating outside radiation comprising measurement having an optical coupling with plate a silicon photodiode total radiation P ₁ consisting of self-radiation P plate, reflected from it sided emission P _{of m p. with t} and passing through it of external radiation P _{p p . with t} , the choice of time to start the temperature measurement to eliminate the effect of P _{p p . with t} , the definition of P _{about t r . with t} and calculating the temperature of the plate by the value of P, defined as the difference P ₁ - P _{about t p . with t,} characterized in that an additional photodiode is measured, having no connection with the optical plate without impinging on the side of radiation plate P _2, and P value _{of r p. with t} determine how

where I _{1 - 0} is the current value of the silicon photodiode at the point where the value of the first time derivative of this current became positive after passing through zero, and I _{2 - 0} is the current value of the additional photodiode measured simultaneously with I _{1 - 0} .

Images