JPH0422019B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an apparatus for nondestructively measuring the lifetime of current carriers in a semiconductor wafer using the photovoltage effect.

In recent years, planar type solid-state circuit elements have become mainstream, and specifically, they are being formed on semiconductor wafers having a thickness of about 400 μm.
Naturally, solid-state circuit devices use the transport phenomenon of current carriers (hereinafter simply referred to as carriers) in semiconductors, so in the semiconductor device manufacturing industry, the lifetime of carriers in wafers is important. but,
This has a large effect on the characteristics of the element.

Furthermore, in order to improve the good product rate (yield) of a large number of devices cut out from a single wafer,
It is an essential condition that the carrier life within the wafer is uniform regardless of location.

In addition, in the case of solar cells, in order to increase the conversion efficiency of solar energy, it is necessary to
There is a requirement that the carrier have a long service life.

Therefore, not only in the case of circuit elements but also in the case of solar cells, it is extremely important to measure and evaluate the carrier life in a wafer serving as a substrate in advance.

For the above purpose, there is a commercially available device that excites a carrier in a semiconductor wafer with light and measures the attenuation state of this carrier using microwaves.
It is effectively used as a carrier life measuring device for wafers before bonding.

However, in order to operate a semiconductor device as an active device, whether it is a circuit element or a solar cell, a junction, typically a p-n junction, must be formed on the wafer. , including the ion implantation process and its annealing process, which often undergo a thermal process at about 800 to 1000°C. As is well known, when a wafer is subjected to a thermal process, the lifetime of the carrier is often reduced compared to that of the wafer alone before heating due to the precipitation and diffusion of oxygen and other trace impurities. make smaller

Therefore, the lifetime of the carrier measured for a single wafer does not ultimately represent the actual state of the completed device. Here, it becomes necessary to measure and evaluate the carrier life after the bond forming process is completed. Naturally, if a junction is present, the conventional optical excitation-microwave measurement method of carrier lifetime measurement is no longer applicable. This is because the microwave response is masked by the newly formed bonding layer and becomes undetectable.

Therefore, several methods have been proposed for measuring the carrier life time of a wafer substrate in a state in which a bond is formed. Among these methods, the photocurrent method that is closest to the present invention will be described below. [References: Solid-state electronics,
1975, Volume 18, Pages 453-458 (Solid-State
Electronics, 1975, Vol. 18, pp. 453-458)].

FIG. 1a shows the basic configuration of the photocurrent method. As an example, a pn junction is shown in which an n-type layer 2 is formed on a p-type Si substrate (wafer) 1. When this junction is irradiated with a photon beam 3, a photocurrent is generated in the junction, as is well known. Metal electrodes 4 and 4 are provided on the wafer substrate 1 and the n-type layer 2, respectively.
By forming 4', the photocurrent excited by the photon beam 3 can be drawn out of the junction, amplified by the amplifier 5, and accurately read by the meter 6.

Here, when the photon beam is pulsed and the photocurrent is read by changing the pulse frequency, the photocurrent I _ph changes as shown in Figure 1b. That is, the angular frequency ω
= 2π × pulse frequency, the photocurrent I _ph is almost constant in a small range of angular frequency ω, but for example, when ω exceeds ₀ , the photocurrent I _ph decreases in proportion to ω ^-1/2 . It's coming. The changing point (bending point) of this photocurrent I _ph
P ₀ can be easily determined graphically.

If it is assumed that most of the photocurrent is generated in the substrate 1, then ω ₀ and carrier lifetime τ
The relationship is simply expressed by the following equation.

ω ₀ τ=1 (1) Once ω ₀ is determined from Fig. 1b, the life time τ of the carrier can be easily calculated from equation (1).

However, when the junction already has metal electrodes 4, 4', as in the case of solar cells,
The photocurrent method described above is a very convenient tool.

However, even in the case of a solar cell, there is no metal electrode immediately after the bond is formed, and in such a state, the conventional photocurrent method cannot be applied. on the other hand,
In the case of circuit element processes, since the junction formation is done at an early stage of the process, it is not permissible to form electrodes on the junction for decontamination for some subsequent processes. Therefore, if you forcefully apply the conventional photocurrent method, you will end up abandoning the next step (in other words, disposing of it as a defective product), and in that sense, the photocurrent method becomes a kind of destructive inspection. .

In addition to p-n junctions, there are other types of junctions, such as shot junctions between metal and Si, and p ⁺ -p type junctions based on so-called resistivity change. In either case, once metal electrodes are formed for measurement, destructive inspection This is the same as in the case of a pn junction.

Furthermore, as a type of junction, there are so-called field-induced junctions (Field-induced junctions) represented by SiO ₂ -p-type Si.
Induced Junction). For this connection,
Even if a metal electrode could be formed on the SiO ₂ film,
Unless the SiO ₂ film is as thin as 50 Å, the photocurrent generated at the junction cannot be detected. Because,
This is because SiO ₂ is an insulator and generally does not conduct current. Here, there are junctions to which the conventional photocurrent method cannot be applied at all. However, as is well known in FET transistors, field-induced junctions, like p-n junctions, are now very widely used, and the inability to apply them is a fatal flaw as a method for measuring semiconductor characteristics. It means that you have it.

The present invention has been proposed to compensate for the drawbacks of the photocurrent method described above. That is, an object of the present invention is to be able to measure the life time of a carrier using the bond itself without intentionally forming a metal electrode for measurement on the bond, and thus to improve semiconductor device manufacturing after forming the bond. An object of the present invention is to provide a carrier life measuring device suitable for non-damaging inspection of semiconductor wafers during the process.

Another object of the present invention is that it can be effectively applied to the measurement of the carrier lifetime of semiconductor wafers having electric field-induced junctions, which could not be measured using the conventional photocurrent method. The objective is to provide a measurement method that far exceeds the applicable scope of the law.

In order to achieve the above object, in the present invention, instead of photocurrent, a photovoltage generated at a junction is detected using a capacitive coupling method, and the carrier life of a semiconductor wafer having a junction is determined from this photovoltage. It is characterized by the fact that it is measured by destruction.

In the following, the principle of the present invention will be described in detail in the case of a pn junction.

As is well known, a pn junction has a junction capacitance C _j and a junction resistance R _j defined, and these two factors are considered to be connected in parallel. We have already mentioned that when a junction is excited with a photon beam, a photocurrent I _ph is generated, but if no external circuit is provided to this junction, the photocurrent will be as shown in Figure 2a.
This will flow to the junction capacitance C _j and the junction resistance R _j . As a result, a voltage is generated across the junction, which becomes the photovoltage V _ph .

Therefore, the photovoltage V _ph is given by the following equation.

V _ph =R _j /1+jωC _j R _j・I _ph (2) In the above equation, j=√−1 indicates the imaginary part.

In equation (2), the junction capacitance C _j and the junction resistance R _j are constant independent of the angular frequency ω of the pulsed photon beam,
And if the photocurrent I _ph also does not depend on ω, then the photovoltage V _ph
As shown in Fig. 2b, the amplitude change of is constant in a small range of ω, and when ω ₁ is exceeded, it decreases in proportion to ω ^-1 . As can be easily determined from equation (2), it is clear that the following relationship holds true: ω ₁ C _j R _j =1 (3).

However, as already mentioned in the explanation of the photocurrent method, I _ph changes depending on ω ^-1/2 when ω ₀ is crossed, as shown in Figure 1b, so the amplitude change of V _ph is , as shown in Figure 2b, when ω ₀ is removed, it begins to decrease in proportion to ω ^-3/2 . Then, from this value of ω ₀ , based on the relationship τ=1/ω ₀ ,
As already mentioned, it is possible to determine the carrier life time τ of the substrate.

The reason why I _ph decreases in proportion to ω ^-1/2 as shown in Figure 1b is because the lifetime time τ(ω) of the carrier excited by pulsed light is apparently as follows. This is due to frequency dependence, which is a well-known fact.

τ(ω)=τ/1+jωτ (4) Now, in the present invention, it is assumed that C _j and R _j are constant regardless of ω, but this assumption is not always satisfied. isn't it. As can be seen from FIG. 2b, V _ph depends on ω, but both C _j and R _j change depending on V _ph . That is, when V _ph exceeds the equivalent voltage of thermal energy (approximately 25 mV at room temperature), R _j becomes V _ph
It changes significantly depending on , and C _j also changes slightly. vice versa,
If V _ph is sufficiently smaller than the equivalent voltage of thermal energy, the thermal equilibrium state is maintained, and R _j and C _j at this time are
It is constant regardless of V _ph and therefore regardless of ω.
At this time, the ideal curve shown in FIG. 2b is obtained.

As described above, in the present invention, the photovoltage V _ph is detected instead of the photocurrent I _ph . Generally, voltage can be detected by capacitive coupling, unlike current detection, and this capacitive coupling is an important focus of the present invention.

FIG. 3a shows the basic configuration of the device of the present invention. Normally, the back side of the wafer 1 is allowed to come into contact with metal, so here, the wafer 1 with a bond is placed on the metal electrode 4'' which also serves as a sample stage.Wafer 1
A spacer 7 made of a transparent insulator with a thickness of approximately 10 μm to 100 μm is placed on the n-type layer 2 . Ideally, it would be an air layer (or even a vacuum layer), but any material that does not damage or contaminate the surface of the bond and is an insulator that is transparent to the irradiating light beam is especially suitable. Regardless of the material. A transparent electrode 8 is placed on top of this spacer 7. In this way, when the junction is irradiated with the photon beam 3, a photovoltage is generated between the front and back sides of the junction, and this photovoltage is applied to the electrode 4''.
(ground) and electrode 8. Since the voltage thus generated is usually 1 mV or less, it is amplified by the lock-in amplifier 9 and measured.

FIG. 3b shows an equivalent circuit for this measurement method. In the figure, C _n is the capacitance due to the spacer 7, and R _l and C _l are the input resistance and input capacitance of the lock-in amplifier 9. Letting the observed voltage be ΔV, ΔV=V _ph jωC _n R _l (1+jωC _j R _j )/1−ω ² (C _n C _l
+C _j C _l +C _n C _j )R _j R _l +jω{(C _n +C _j )R _j +(C _l +C _n )R
_l }......(5).

As can be seen from equation (5), if R _l is large and C _l is small (that is, the input impedance of lock-in amplifier 9 is large), ΔVV _ph will be obtained over a wide frequency range, so capacitive coupling Since V _ph can be measured and, in turn, the curve shown in FIG. 2b can be obtained, we can know ω ₀ from this and calculate the lifetime τ of the carrier at the photon beam irradiation point from τ = 1/ω ₀ .

FIG. 4 shows the overall configuration of an embodiment of the device of the present invention. The light source 12 for emitting the photon beam 3 may be a gas laser, a semiconductor laser, a lamp, etc., but here an example using a light emitting diode will be shown. Light emitting diodes are inexpensive, small, and convenient to use. This light emitting diode 12
irradiates pulsed light 3 by being supplied with an alternating current (pulsed) current from a drive current source 10. The driving current signal is detected by the detection resistor 11 and used as a reference signal for the lock-in amplifier 9.

The photovoltage is detected by a transparent electrode 8 and amplified by a lock-in amplifier 9. The output of the lock-in amplifier 9 is supplied to a signal processing circuit 15. Signal processing circuit 15
From here, an instruction to change the frequency of the pulsed light 3 is given to the variable frequency oscillator 14, and the signal processing circuit 15 monitors the oscillation frequency of the oscillator 14. The output of the oscillator 14 is supplied to a light emitting diode drive circuit 10, where it is converted into a current pulse of a predetermined frequency and supplied to the light emitting diode 12.

The emitted light pulse 3 from the light emitting diode 12 is
After the radiation angle is limited by the aperture 13, the radiation is focused by the lens 17 and irradiated onto the pn junction.

In the signal processing circuit 15, the change mode is evaluated according to a predetermined sequence set in advance from the frequency and photovoltage information taken in in this way, and from this, ω ₀ is known, and the life time τ of the carrier is determined. =1/ω ₀ is calculated, and the result is digitally displayed on the display unit 16.

Although redundant explanation will be omitted, the above process is
The same holds true not only for -n junctions but also for electric field induced junctions, and it is clear that the method can be applied non-destructively to a wide range of samples beyond the scope of the conventional photocurrent method.

As is clear from the above explanation, according to the present invention, the carrier life can be measured non-destructively without the need to take the trouble of forming a metal electrode for measurement on the junction. It is possible to check whether the carrier life of semiconductor wafers is suitable or not at any time, which greatly contributes to improving the yield of the semiconductor manufacturing process, and what is more, it does not require complicated and expensive measurement methods such as those used in conventional microwave measurement equipment. Since the method does not use any of the following, it can be easily manufactured and used, and its effects are extremely large in practical use.

[Brief explanation of drawings]

Figures 1a and b are diagrams explaining the principle of the carrier life measurement method using the conventional well-known photocurrent method, Figures 2a and b are diagrams explaining the principle of the carrier life measurement method based on the photovoltage method according to the present invention, and Figure 3. 4A and 4B are explanatory diagrams of the principle structure of a carrier life measuring device according to the present invention, and FIG. 4 is a schematic block diagram showing the overall structure of a carrier life measuring device according to an embodiment of the present invention. DESCRIPTION OF SYMBOLS 1... Wafer substrate, 2... N-type layer, 3... Photon beam, 4''... Metal electrode, 7... Spacer, 8... Transparent electrode, 9... Lock-in amplifier, 10... Drive current source,
11...Detection resistor, 12...Light emitting diode, 13...
aperture, 14... oscillator, 15... signal processing circuit, 16... display unit.

Claims (1)

[Claims] 1 means for irradiating a semiconductor wafer having a bond with a pulsed photon beam, and an electric voltage V _ph generated in the bond by irradiation with the photon beam;
means for detecting the angular frequency ω of the pulsed photon beam via capacitive coupling; and means for changing the angular frequency ω of the pulsed photon beam;
Find the angular frequency ω ₀ when the amplitude of V _ph changes from a region where it changes in proportion to ω ^-1 to a region where it changes in proportion to ω ^-3/2 , and from this angular frequency ω ₀ τ=
A carrier life measuring device comprising means for determining the carrier life time τ at the photon beam irradiation position of the semiconductor wafer based on the relationship 1/ω ₀ .