JP2000028518A - Nondestructive test method for semiconductor epitaxial film - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a nondestructive test method for a semiconductor epitaxial film, capable of evaluating a sheet carrier concentration from a measurement of a photoluminescence spectrum near a room temperature without destroying the semiconductor epitaxial film. SOLUTION: This nondestructive test method of a semiconductor epitaxial film has a process for measuring a photoluminescence spectrum from the semiconductor epitaxial film, a process for obtaining a first energy value at the maximum intensity peak position of the spectrum, a process for obtaining a second energy value at the position, where the spectrum becomes a prescribed value in the range from 20% to 80% of the peak value on the high energy side of the peak, a process for calculating the difference between the first energy value and the second energy value and a process for estimating a sheet carrier concentration of the semiconductor epitaxial film from the difference by using a relation obtained beforehand by a measurement.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a nondestructive inspection method for a semiconductor epitaxial film. More specifically, the present invention relates to a nondestructive inspection method for a semiconductor epitaxial film, which can evaluate a sheet carrier concentration from a photoluminescence spectrum measurement near room temperature without breaking the semiconductor epitaxial film. In particular, the present invention is suitably used as a method for evaluating a semiconductor epitaxial film functioning as a channel layer used in a heterostructure modulation-doped field effect transistor which has been spotlighted as an ultra-high-speed device in recent years.

[0002]

2. Description of the Related Art In recent years, the speed of an optical transmission communication system or the speed of a wireless communication system such as a microwave has been increased.
Semiconductor devices are required to operate at higher speeds. In such fields, heterostructure modulation doped field effect transistors (hereinafter abbreviated as HFETs) are one of the fastest semiconductor devices. In particular, an InAlAs / InGaAs-based HFET on an InP substrate has attracted attention as an ultra-high-speed device. In a device having such a configuration, a semiconductor epitaxial film including a layer functioning as a channel is provided on a substrate in advance, and an InAlAs / InGaAs-based HFET is formed using the film. Therefore, if the semiconductor epitaxial film can be screened by non-destructive evaluation to determine the sheet carrier concentration (hereinafter abbreviated as Ns) of the semiconductor epitaxial film before it is subjected to the device manufacturing process, the HF to be subsequently manufactured
It is considered that desired device characteristics such as a threshold voltage (hereinafter abbreviated as Vth) in ET can be obtained with high yield. For example, it has been reported that Ns and Vth have a certain relationship [H. Hida, T. Tsukada,
Y. Ogawa, Toyoshima, M. Fujii, K. Shibahara, M. Ko
hno, and T. Nozaki, IEEE Trans. Electron Device, 3
6, 223 (1989)].

Conventionally, Ns evaluation of a channel layer which is an epitaxial film for HFET has been performed as follows.

(1) Measuring Method Using Hall Effect In an apparatus capable of growing a plurality of film wafers (a wafer having a semiconductor epitaxial film on a substrate) at the same time, one of the simultaneously grown film wafers is sampled and a Hall element is measured. A sample was prepared and Ns was measured. The Ns of the remaining film wafer was considered to be the same as the measured value. In an apparatus that can grow only a single wafer, a sample for measurement such as a Hall element is formed on a film wafer grown before or after a film used for manufacturing an HFET.
Was measured. Ns of film wafer used for HFET fabrication
Was considered the same as the measurement.

(2) Measuring Method Using Eddy Current Method In an HFET epitaxial film having no conductive layer (for example, a contact layer) other than the channel layer, the sheet resistance and the mobility are obtained by nondestructive evaluation by the eddy current method, and eventually Ns Was calculated.

(3) Measurement method using photoluminescence at low temperature In the research stage, a method for obtaining Ns from photoluminescence (hereinafter abbreviated as PL) measurement at a low temperature of 20K has been reported. In this PL measurement, the electron e ₁ at the ground level of the quantum level and the hole h at the ground level in the HFET channel layer
A PL peak energy position E ₁ corresponding to the bond with
Electron e _{F at the} Fermi energy position and hole h at the ground level
Seeking energy difference delta ^* E between energy position E _F of the shoulder of the corresponding high energy end for binding to. Its Δ
^* Utilizes that E has a linear relationship with Ns.

However, the above measuring method has the following problems.

In the measurement method using the Hall effect, in order to produce a Hall element for measuring the Hall effect, one wafer having an epitaxial film must be broken or broken by etching. Therefore, Ns
It was difficult to manufacture and commercialize an HFET in the same wafer region where was measured. There is also a problem that it takes several hours to make a Hall element.

In the measurement method using the eddy current method, an HFE having a conductive layer (for example, a contact layer) in addition to a channel layer is used.
There is a drawback that it cannot be applied to a T epitaxial film. This is because the Ns of the channel layer and the Ns of the other conductive layers are measured at the same time, and it is impossible to separate the two Ns. In IAlAs / InGaAs system HFET on InP substrate, can not source drain electrode alloy heat treatment to avoid fluorine pollution material degradation, using a highly conductive contact epitaxial layer, such as n ⁺ InGaAs is to ohmic contact non-alloy It is necessary to attach electrodes. Therefore, the method using the eddy current cannot be used.

In a measuring method using PL at low temperature,
There is no need to crack or etch the wafer,
During the cooling process, the surface of the wafer may be stained or damaged, and the epitaxial film may be distorted. Therefore, the measurement is not completely non-destructive, and is not accepted in a practical production line of devices such as HFETs. In addition, there is a problem that it takes a long time to return to room temperature after measurement due to cooling and the like. Note that this PL evaluation was applied only to an HFET epitaxial film having no conductive layer other than the channel layer, and the effect of the conductive layer was not clarified. On the other hand, the PL spectrum around room temperature without fear of wafer contamination and distortion, energy position E _F can not be identified, also Ns is 2 × 10 ¹² c
Since m ^-2 vicinity or more regions difficult to identify the PL peak energy position E ₁ has a problem that it can use the method. This means that the probability of the state occupation of electrons near the Fermi energy near room temperature is
In addition, since the PL peak becomes broad, the e ₁ hPL peak corresponding to the coupling between the electron e ₁ at the ground level of the quantum level and the hole h at the level e is the electron e at the excited level of the quantum level. E ₂ hPL peak corresponding to the bond between ₂ and the hole h at the ground level (this peak increases with an increase in Ns, and is approximately 2 ×
Above 10 ¹² cm ^-2 , the peak becomes much more dominant than the e ₁ hPL peak) and a considerable partial overlap.

[0011]

SUMMARY OF THE INVENTION It is an object of the present invention to provide a non-destructive semiconductor epitaxial film capable of evaluating a sheet carrier concentration by measuring a photoluminescence spectrum near room temperature without destroying the semiconductor epitaxial film. The purpose is to provide an inspection method.

[0012]

A nondestructive inspection method for a semiconductor epitaxial film according to the present invention comprises a step of measuring a photoluminescence spectrum from the semiconductor epitaxial film, and a first energy value at a maximum intensity peak position of the spectrum. And determining that the spectrum has 20% to 80% of the peak value on the high energy side of the peak.
% Determining a second energy value at a position where the energy value becomes a predetermined value in the range of%.
And a step of estimating a sheet carrier concentration of the semiconductor epitaxial film from the difference by using a relationship obtained by measurement in advance from the difference.

The inventor of the present invention conducted an PL measurement at room temperature on a semiconductor epitaxial film having an HFET structure, and obtained the energy position Ep of the largest PL peak from the HFET channel layer and the half value of the peak on the high energy side. The inventors have found that the energy difference ΔE between the energy position Eh of the position and Ns obtained from the Hall effect has a linear positive correlation, and devised the present invention having the above configuration.

It has been found that this positive correlation is maintained regardless of the presence or absence of the contact layer. Therefore, if the correspondence data between ΔE and Ns is stored in advance and the correlation between them is stored in a computer, the temperature of the wafer provided with the semiconductor epitaxial film having the HFET structure before the device is manufactured is kept at room temperature. By measuring the PL, it was found that Ns of the channel layer can be determined nondestructively, and that the wafer can be screened.

The inventor has considered that the above-mentioned correlation is due to the following reasons.

The energy difference ΔE is (Eh−Ep)
It is. Here, Ep is the binding energy E ₁ between the electron e ₁ at the ground level of the quantum level and the hole h at the ground level (when the carrier concentration of the channel layer is low) or the electron at the excited level of the quantum level. binding energy E ₂ between e ₂ and ground level hole h
(When the carrier concentration of the channel layer is high).

Ns = 2 × often used in practical HFETs
Room temperature PL near 10 ¹² cm ^-2 or above
Generally corresponds to E ₂ . As the carrier concentration of the channel layer increases, electrons are present up to the high energy side in the conduction band. Since PL occurs due to the bond between the electron and the hole relaxed to the ground state after photoexcitation, the hem on the high energy side of the largest PL peak Ep expands and extends to the high energy side.

Accordingly, the half-value energy position Eh is also shifted to the higher energy side relative to Ep, so that (Eh-
The inventor considered that Ep) had a positive correlation with Ns.

In the above description, the half-value energy position E
Although the description has been given based on the relationship between h and Ep, it has been found that the same operation and effect can be obtained by using an energy position of 20% to 80% of the maximum peak value instead of the half-value energy position Eh.

Further, what is hardly affected by the contact layer is that the contact layer is usually a very low-resistance or very heavily doped layer, and the PL
The inventor has also found that the peak is very broad and very low in intensity.

[0021]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, a nondestructive inspection method for a semiconductor epitaxial film according to the present invention will be described in detail based on a specific procedure.

First, an HFET wafer having a layer structure shown in Table 1 having a channel layer made of InGaAs as a semiconductor epitaxial film was prepared as an evaluation object. That is, the wafer is InA on an InP substrate.
It has an lAs / InGaAs HFET structure. However, as InAlAs in Table 1, In _0.52 Al _0.48 As
And In _0.53 Ga _0.47 As as InGaAs.

[0023]

[Table 1]

Using 20 wafers having the constitutions shown in Table 1, the PL spectrum from the channel layer was measured at two places on each wafer at room temperature. FIG. 1 is a graph showing an example of a PL spectrum obtained by this measurement. And
ΔE was obtained as shown in FIG. In FIG. 1, the horizontal axis is the wavelength of the light used for the PL measurement, and the vertical axis is the PL intensity observed at each wavelength. However, the vertical axis in FIG. 1 is a numerical value normalized by setting the maximum peak intensity to 10.

After the above-mentioned PL measurement, the Hall effect was measured using a Hall element fabricated at the same location, thereby obtaining the sheet carrier concentration Ns.

FIG. 2 is a graph showing the relationship between ΔE obtained from the PL measurement and Ns obtained from the Hall effect measurement. In FIG. 2, a mark ● indicates a case where there is no diode region, and a mark ○ indicates a case where there is a diode region.

FIG. 2 shows that ΔE and Ns show a good linear positive correlation. Further, FIG. 2 shows that the data of the wafer having the diode region and the data of the wafer having no diode region are mixed, but it is also clear that the correlation is hardly affected by the presence or absence of the diode region. Therefore, the sheet carrier concentration Ns of the channel layer can be determined nondestructively from ΔE obtained by performing room temperature nondestructive PL measurement on the same wafer with the same semiconductor epitaxial film structure by using the correlation shown in FIG. it can.

In FIG. 2, there is a slight variation in the correlation between ΔE and Ns, which can be sufficiently reduced by improving the S / N ratio of the PL measurement system. For example, ΔE = 49me
In the case of V, Ns = about 2 × 10 ¹² cm ⁻² is obtained. Therefore, even if there is a contact layer, ΔE determined by room temperature PL measurement
From this, it has been clarified that the sheet carrier concentration Ns can be inspected nondestructively.

Further, at the room temperature PL, which is the nondestructive inspection described above,
Using the wafer after the measurement, the threshold voltage Vth of the HFET to be manufactured can also be estimated.

FIG. 3 is a schematic plan view of a 3-inch wafer, showing an example in which the room temperature PL measurement is performed at intervals of 5 mm on the surface and the numerical value of ΔE obtained nondestructively is mapped. In FIG. 3, the average value of ΔE is 51.54 meV
And the standard deviation of ΔE is 1.06 meV. Then, by using the relationship of FIG.
Is about 2 × 10 ¹¹ cm ⁻² . From these results, it was expected that the standard deviation of the threshold voltage Vth would be about 50 mV when the intended HFET was manufactured on this wafer.
Based on this expectation, the uniformity of this wafer could be determined to be almost acceptable for HFET fabrication.

Further, in FIG. 3, it was found that ΔE is small at the center of the wafer and close to a large rotational symmetry on the outer peripheral side. This result is considered to reflect the in-plane non-uniformity of the epitaxial growth conditions, and it was also clarified that the results can be used as important information for improving the growth conditions.

In the above description, the energy at which the PL intensity is の of Ep is adopted as Eh. However, if the correlation between ΔE and Ns is determined at a fixed ratio, for example, 1/3 before and after that, for example, 1/3 It is clear from the above description that / 3 is also possible. As a specific range, the maximum peak value is 2
It has been confirmed that the same effect can be obtained by setting the range of 0% to 80%.

Further, in the above description, the semiconductor epitaxial film has a channel layer made of InGaAs,
InAlAs / InGaAs HFET on InP substrate
Although an example using a wafer having a structure has been described, the method according to the present invention is not limited to this configuration. For example, AlGaAs / InGaAs-based H on a GaAs substrate
The method according to the present invention is applicable to various electronic devices and optical devices using an epitaxial film having an FET structure, an epitaxial film having another HFET structure, or an epitaxial film having a quantum well structure other than the HFET structure. it can.

[0034]

As described above, by using the method for nondestructively inspecting a semiconductor epitaxial film according to the present invention, a semiconductor epitaxial film including a layer used as a channel can be nondestructively processed before being subjected to an HFET fabrication process. By the evaluation, the film can be screened, and as a result, a device having a desired threshold voltage Vth can be obtained with a high yield.

For example, a single HFET fabrication process requires considerably more time (several months) and labor than a single epitaxial film growth. On the other hand, even if the above screening is rejected and the film is grown again, it takes at most several days. Therefore, by employing the method according to the present invention, the time, labor, and cost required for device fabrication can be significantly reduced.

In addition, the method according to the present invention requires less time and labor to obtain the distribution of Ns in the wafer surface than the Hall effect measurement in the conventional destructive inspection. In addition, as shown in FIG. 3, Δ obtained by the method according to the present invention.
Since the distribution of E reflects in-plane non-uniformity of epitaxial growth conditions (for example, growth temperature, etc.), it can be used as an effective means for determining growth conditions for quick feedback and improving growth conditions.

[Brief description of the drawings]

FIG. 1 is a graph showing the result of measuring a PL spectrum from a channel layer at room temperature, wherein the horizontal axis represents the wavelength of light used for PL measurement, and the vertical axis represents the PL intensity observed at each wavelength.

FIG. 2 is a graph showing a relationship between ΔE obtained from PL measurement and Ns obtained from measurement of the Hall effect, where ● indicates a case where there is no diode region, and ○ indicates a case where there is a diode region. Show.

FIG. 3 is a schematic plan view of a 3-inch wafer, showing an example in which a room temperature PL measurement is performed at intervals of 5 mm and a numerical value of ΔE obtained in a non-destructive manner is mapped in the plane.

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 2G020 BA02 BA18 CA01 CD03 CD11 CD14 2G059 AA01 AA03 AA05 BB16 CC20 DD15 EE07 EE12 EE15 MM01 MM02 MM03 MM04 MM12 PP01 4M106 AB04 CA18 CA20 CB01 DJ12

Claims (4)

[Claims] 1. A step of measuring a photoluminescence spectrum from a semiconductor epitaxial film, a step of obtaining a first energy value at a maximum intensity peak position of the spectrum, and a step of calculating the peak value on a high energy side of the peak. Obtaining a second energy value at a position having a predetermined value in the range of 20% to 80% of
Calculating a difference between the first energy value and the second energy value, and estimating a sheet carrier concentration of the semiconductor epitaxial film from the difference using a relationship obtained by measurement in advance. And a non-destructive inspection method for a semiconductor epitaxial film. 2. The method according to claim 1, wherein the measurement of the photoluminescence spectrum from the semiconductor epitaxial film is performed at around room temperature. 3. The relation determined in advance by the measurement includes: a difference between the first energy value and the second energy value obtained from measurement of a photoluminescence spectrum from the semiconductor epitaxial film; 2. The non-destructive inspection method for a semiconductor epitaxial film according to claim 1, wherein the relation is a relationship with a sheet carrier concentration obtained from measurement of a Hall effect on the epitaxial film. 4. The method according to claim 1, wherein the semiconductor epitaxial film is a channel layer used for a heterostructure modulation doped field effect transistor.