JPS6212855B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to charge trapping centers (hereinafter referred to as traps) in an insulator sandwiched between a conductive electrode and a semiconductor, such as a silicon oxide film in a metal-oxide-silicon (MOS) structure. The present invention relates to a method for measuring the distribution of

As the density of integrated circuits progresses from LSI to VLSI, the channel length of the MOS transistors used in them is also becoming shorter, but as the source-drain electric field becomes stronger, charge carriers in the channel are Obtaining high energy increases the probability of impact ionization. When the charge carriers multiplied by this impact ionization are injected into the silicon oxide film near the channel surface and captured in a trap, they act as fixed charges and cause fluctuations in the threshold voltage of MOS transistors, making MOS/LSI and VLSI unstable. Causes qualitative problems. Therefore, knowing the distribution and other characteristics of the traps in the silicon oxide film has become an industrially important measurement and evaluation method.

The conventionally proposed method for nondestructively measuring the trap density distribution in a silicon oxide film is to find the center of gravity x of the distribution, as described in DJ DiMaria's ``Journal of Applied Physics” (Journal
of Applied Physics), Vol. 47, No. 9 (September 1976)
It is known from the paper published on pages 4073 to 4077.

Here, the center of gravity x is as shown in Figure 1, when the origin is set at the surface of the silicon oxide film and the x coordinate is taken in the direction perpendicular to the surface toward the oxide film/silicon interface, x = ∫ ^L _O xNt (x) It is expressed as dx／∫ ^L _O Nt(x)dx (1). Nt(x) is the trap volume density distribution, L is the thickness of the oxide film, and the denominator of equation (1) ∫ ^L _O Nt(x)dx is the trap areal density Qt. This is the quantity determined together with x by the photocurrent-voltage method described. however,
In order to know the state of the trap distribution Nt(x), it is not enough to simply find x. In other words, with this method alone, it is not possible to know how far the trap has spread in the thickness direction of the oxide film.
The conventionally proposed method for measuring the trap distribution in the thickness direction is to remove the oxide film little by little by etching and measure the amount of charge in the film, but this is a destructive test and is time consuming. There is also a great risk of sample contamination.

On the other hand, the present invention uses the following equation (2) σ=[∫ ^L _O (x-x) ² Nt(x)dx/Qt] in order to obtain knowledge regarding the spread of the trap distribution by non-destructive means. 〓
This is to find the standard deviation σ of the trap distribution expressed by (2). Equation (2) is σ ² =∫ ^L _O x ² Nt(x)dx/Qt−(x) ²
......(3), so using the method in the above paper, x
If and Qt are known, σ can be found by knowing the integral ∫ ^L _O x ² Nt(x)dx. Here, the above
The basis for the establishment of equation (3) will be explained.

Volume density distribution of traps in silicon oxide film
The center of gravity x of Nt(x) can be expressed by equation (1) as described above, but if this is expressed as the area (area) density Qt per unit area of the film surface of the silicon oxide film, Qt= ∫ ^L _O Nt(x)dx ......(1a). Moreover, the center of gravity can also be expressed as x=∫ ^L _O xNt (x) dx/Qt (1b).

By the way, when Nt(x) is a distribution function based on variable x, by statistical definition, the variance σ ²
(σ is the standard deviation) is σ ² =∫ ^L _O (x-x) ² Nt(x)dx /∫ ^L _O Nt(x)dx =∫ ^L _O (x-x) ² Nt(x)dx/ Qt
......(1c) This equation (1c) is equivalent to equation (2). Equation (3) is derived from Equation (1c) as follows.

σ ² =∫ ^L _O (x-x) ² Nt(x)dx/Qt =∫ ^L _O x ² Nt(x)dx/Qt −∫ ^L _O 2x・xNt(x)dx/Qt +∫ ^L _O ( x) ² Nt(x)dx/Qt =∫ ^L _O (x) ² Nt(x)dx/Qt −2x∫ ^L _O xNt(x)dx/Qt +(x) ² ∫ ^L _O Nt(x)dx /Qt =∫ ^L _O x ² Nt (x) dx / Qt - 2xx + (x) ² ∫ ^L _O Nt (x) dx / Qt = ∫ ^L _O x ² Nt (x) dx / Qt - 2xx + (x) ² =∫ ^L _O x ² Nt(x) dx/Qt-(x) ² Hereinafter, embodiments of the present invention will be described with reference to FIG.

The measurement sample has an oxide film 2 grown on the surface of a silicon substrate 1, and a gate electrode 3 is further provided on the surface of the oxide film 2. For reasons to be described later, the gate electrode 3 is composed of a conductive antireflection film 3a and a conductor 3b, and the portion of the silicon substrate 1 facing the gate electrode 3 is made thin so that light can easily pass through. . 4 is a constant current pulse power supply,
5 is a capacitance-voltage characteristic measuring device, 6 is a DC power supply, 7 is
8 is the irradiation light, and 8 is the changeover switch. The measurement procedure will be described below, taking as an example the case where the traps in the oxide film 2 are electron traps.

First, in an initial state, the flat band voltage of the MOS diode consisting of the gate electrode 3, oxide film 2, and silicon substrate 1 is measured using the capacitance-voltage characteristic measuring device 5, and this is set as (V _FB ) ₀ .
Next, electrons are uniformly captured by traps in the oxide film 2. The method may use avalanche injection using a constant current pulse power source 4, or may use injection using an internal photoelectron emission effect by irradiation with light or ionizing radiation. Assuming that electrons are captured in all traps by this operation, when the flat band voltage is measured again after the operation, its value (V _FB ) ₁ will be (V FB ) 1 when measured with (V _FB ) ₀ as the reference point, (V _FB ) ₁ = q/Co∫ ^L _O xNt(x)dx (4) Takes the value expressed by the equation. Co here
is the capacitance per unit area of the oxide film 2, and q is the electronic charge. Next, irradiation light 7 is irradiated onto the oxide film 2 through the silicon substrate 1 . This irradiation light 7 has sufficient photon energy to emit electrons from the trap, and has a wavelength that satisfies the condition αL<<1, where α is the absorption coefficient in the oxide film 2. choose something. Specifically, light with a photon energy slightly smaller than the band gap of the oxide film 2 is suitable, for example, a photon energy of 8.3 eV.
(wavelength 1500 Å), α = 10 ³ cm ^-1 , so L
= 0.1 μm, αL = 10 ^-2 , which satisfies the above condition. Note that the absorption coefficient of silicon for light at this wavelength is 6.9×10 ⁵ cm ^-1 , so the thickness of the portion of silicon substrate 1 directly below the gate is 500 Å.
Then, 3% of the incident light can pass through the silicon substrate 1. During this irradiation, the light intensity distribution I(x) in the oxide film 2 is expressed as I(x)=I ₀ e ^- 〓 ^(Lx) ......(5) where αx≪1 Therefore, this can be approximately expressed as I(x)≒I ₀ e ^- 〓 ^L (1+αx) (6), as shown by the dashed line in Figure 1. Here, I ₀ is the intensity of light when it enters the oxide film 2. This light causes electrons to be emitted from the trap at a rate proportional to the intensity of the light. Also,
It has been confirmed that one trap captures one electron in a silicon oxide film, so
The distribution of traps that emitted electrons is βI(x)Nt
It is expressed as (x). Here, β is the number of electrons emitted by light of unit intensity. Therefore, the flat band voltage (V _FB ) ₂ measured by the capacitance-voltage characteristic measuring device 5 after light irradiation is (V _FB ) ₂ = q/Co∫ ^L _O x{Nt(x)−βI(x )Nt(x
)} dx ...(7) Therefore, if the change in flat band voltage before and after irradiation is ΔV _FB , then ΔV _FB is calculated by equations (4), (6), and (7) as follows: ΔV _FB = βI ₀ It is expressed as e ^- 〓 ^L [(V _FB ) ₁ + αq/Co∫ ^L _O x ² Nt(x) dx]...(8). I ₀ and α are determined by measuring the light intensity, and β can be determined by performing two or more measurements with slightly different light intensities or wavelengths. In this way, ∫ ^L _O x ² Nt (x) dx can be found from the measured value of ΔV _FB using equation (8), so using x and Qt, which were found separately,
The standard deviation σ of the trap distribution can be obtained from equation (3). The above explanation applies equally to the case where the charge carriers to be captured are holes.

Next, points to be noted when implementing this invention will be described. In the measurement method of the present invention, it is necessary that the intensity distribution of the irradiated light 7 in the oxide film 2 is like the one-dot chain line in FIG. 7 reflections make measurement difficult. In order to avoid this, an antireflection film 3a is provided between the conductor 3b of the gate electrode 3 and the oxide film 2. Specifically, when polycrystalline silicon (refractive index 3.5) is used as the conductor 3b, indium oxide In ₂ O ₃ with an appropriate composition and a refractive index of about 2.3 is used as the anti-reflection film 3a to prevent light If the wavelength is 1500 Å, the film thickness should be 160 Å.
To a certain degree, the reflection at the interface between the oxide film 2 and the gate electrode 3 can be almost ignored. next,
During the irradiation with the irradiation light 7, electrons excited from the silicon substrate 1 or the gate electrode 3 are injected into the oxide film 2, or electrons excited from the valence band to the conduction band in the oxide film 2 are If these are captured by the trap, it may cause an error, but the former can be pushed back by the electric field created in the oxide film 2 by the bias from the DC power supply 6 or the charged trap. Furthermore, the latter can be suppressed by using the irradiation light 7 having a photon energy slightly smaller than the bandgap of the oxide film 2, and by mainly exciting excitons or the like that do not involve free electrons. Note that the electric field in the oxide film 2 described above quickly drives electrons excited from the traps toward the silicon substrate 1 and the gate electrode 3, so these electrons may be captured by the traps again and cause errors. There are also few.

Note that the method of the present invention is applicable not only to silicon and its oxide film, but also to structures that generally have a combination of a semiconductor and an insulating thin film.

As explained above, the present invention irradiates an insulator with light or ionizing radiation that passes through the insulator while attenuating approximately in proportion to the distance passed through the insulator, and changes the charge at the charge trapping center caused by this. Since the standard deviation from the center of gravity of the distribution of the charge trapping center in the thickness direction in the insulator is obtained from the measurement, the trap distribution in the thickness direction in the insulator can be measured non-destructively and in a relatively short time. It has the advantage of being able to do a lot of things, and has important industrial value.

[Brief explanation of the drawing]

FIG. 1 is a conceptual diagram for explaining various quantities related to trap distribution Nt, light intensity distribution I, etc. in a silicon oxide film, and FIG. It is a schematic diagram. In the figure, 1 is a silicon substrate, 2 is an oxide film, 3 is a gate electrode, 3a is an antireflection film, 3b is a conductor, and 4
5 is a constant current pulse power supply, 5 is a capacitance-voltage characteristic measuring device, 6 is a DC power supply, 7 is an irradiation light, and 8 is a changeover switch.

Claims (1)

[Claims] 1. When measuring the distribution of charge trapping centers in an insulator sandwiched between a conductor electrode and a semiconductor, light or ionizing radiation that passes through the insulator is attenuated approximately in proportion to the distance it passes through the insulator. The insulator is characterized in that the change in charge of the charge trapping center caused by this is measured, and the standard deviation from the center of gravity of the distribution of the charge trapping center in the thickness direction in the insulator is obtained from this. A method for measuring the distribution of charge trapping centers inside.