JPH07249666A - Iron concentration measurement method of silicon wafer - Google Patents

Abstract

PURPOSE:To simplify measurement and improve precision by a method wherein a diffusion length of a minority carriers in a state of activating a p type silicon wafer is measured by a SPV method and a diffusion length in a re-coupling state after re-coupling is measured, and iron concentration of defect layers with high density is obtained from a difference in a value based on a measure value, etc. CONSTITUTION:A silicon wafer (S21) is a p type that boron is made dopant and a defect layer is formed inside a thickness direction and a non-defect layer is formed on both end surface. A part having density different in the defect layer is removed by chemical etching (S22). Next, iron-boron in a region being a measurement object of the defect layer is separated by an optical excitation of a light-emitting, a heating furnace or the like and Fe-B is activated in a state of Feint+B and promptly cooled (S23). From the surface of the defect layer in a separation state being Feint+B by a SPV method, a diffusion length La of a minority carries is measured (S24). Successively, until the separated Feint+B is re-coupled to Fe-B (S25), it is left freely to measure a diffusion length Lb1 (S26) to obtain a difference.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for measuring iron concentration in a silicon wafer, and more particularly to a method for measuring iron concentration in a silicon wafer in which defects are generated at high density.

[0002]

2. Description of the Related Art Conventionally, the following methods have been known as methods for measuring the iron concentration of silicon wafers. (A) Schottky junction or P on the silicon wafer surface
/ N junction is formed, and DLTS method (Deep level)
(transparent spectroscope method)
To measure the density of the deep level created by Fe. This method is, for example, a Diagnostic Technique.
s for Semiconductor Device
e and Materials (180th ECS
Meeting, p. 113-118 (1991),
Published in 1992).

(B) The surface of a silicon wafer is activated by optical excitation by light irradiation or heating at 200 ° C.,
Before and after this activation, a change in diffusion length due to iron-boron dissociation or the like is measured, and the concentration of iron is quantified based on the measurement. The surface photovoltage method (Surface photovolta)
The ge method, hereinafter referred to as SPV method. ) Is known, and is shown in, for example, heavy metal contamination measurement by SPV method (workshop data, November 16, 1993, Japan AD Corporation). This SPV is shown in FIG.
The procedure outline (step 41 to step 45) of the iron concentration measurement of the silicon wafer by the method is shown. First, a silicon wafer in an Fe-B bonded state after a predetermined process (as processed) is prepared (step 41), and the minority carrier diffusion length Lb of the silicon wafer is measured by the surface photovoltage method (step 42). After causing the dissociation of Fe-B → Fe _int + B by heating the silicon wafer at 200 ° C. or the like, the silicon wafer is rapidly cooled to room temperature (step 43), and the diffusion length La of the dissociated minority carriers is measured ( In step 44), the iron concentration is calculated (step 45). This iron concentration [Fe] is [Fe] ≈
It is calculated from 1 × 10 ¹⁶ (La ⁻² −Lb ⁻² ). Here, the unit of iron concentration [Fe] is cm ⁻³ , and the units of La and Lb are μm. By this method, the iron concentration can be measured with high sensitivity (10 ⁸ cm ⁻³ order).

[0004]

However, the above-mentioned prior art has the following problems. That is, (a)
In the above, there is a problem that it takes a long time to form a junction and to measure DLTS, and for example, it takes 1 hour to form a junction by vacuum vapor deposition, and the time required to measure one DLTS is 1 hour or more, which is a large number of steps. Next, in (b), there is a problem that the iron concentration [Fe] is calculated as a value larger than the actual value in a silicon wafer in which defects such as oxygen precipitation-induced defects are generated at a high density. That is, when the distribution of carrier recombination centers (oxygen precipitation induced defects, Fe-B, interstitial Fe, etc.) in the thickness direction of the silicon wafer is not uniform, accurate diffusion lengths La and Lb of minority carriers are obtained. I can't. For example, in a silicon wafer that has been subjected to a three-step IG (internal gettering) process, defect-free layers are formed on both end faces, and the distribution of defects such as oxygen precipitation-induced defects is non-uniform. The diffusion lengths La and Lb affected by the layers will be measured. Further, even if the defect-free layer is removed by etching or the like to ensure the uniformity of the distribution of recombination centers in the thickness direction, a silicon wafer in which defects are generated at a high density is activated by heating at 200 ° C. or the like. This is because the recombination center that exists exists in addition to Fe-B.

The present invention focuses on the above-mentioned problems of the prior art, and it is possible to accurately and easily measure the iron concentration of a silicon wafer having a high density of defects such as oxygen precipitation-induced defects. It is intended to provide a method for measuring concentration.

[0006]

In order to achieve the above-mentioned object, in the method for measuring the iron concentration of a silicon wafer according to the present invention, the first invention is the F concentration in a p-type silicon wafer.
The diffusion length Lb of the minority carrier in the e-B bond state and the diffusion length La of the minority carrier in the activated and dissociated into Fe _int + B and then rapidly cooled were measured by the surface photovoltage method, and the difference based on the measured value was used. In the method for measuring the iron concentration of a silicon wafer for determining the iron concentration, first, the diffusion length La in the rapidly cooled state after activation is measured and recombined, and the diffusion length L in the recombined state is measured.
b1 is measured, and the iron concentration of the high-density defect layer of the p-type silicon wafer is obtained from the difference between the values based on these measured values. The second invention is to measure the diffusion length Lb of a minority carrier in a Fe—B bonded state and the diffusion length La of a minority carrier in an activated and dissociated state into Fe _int + B and then rapidly cooled in a p-type silicon wafer by a surface photovoltage method. Then, in the iron concentration measuring method of the silicon wafer to obtain the iron concentration from the difference of the value based on the measured value, first after activation, rapidly cooled,
The diffusion length Lb1 in the recombined state and the recombined state is measured, the diffusion length La1 in the rapidly cooled state after activation is measured again, and the difference in the values based on these measurement values is used to measure the high-density defect layer of the p-type silicon wafer. It is characterized in that the iron concentration is obtained. A third invention is the first or second invention, wherein the defect layer is
It is a defect layer exposed by removing the defect-free layer by etching.

[0007]

The operation of the present invention having the above construction will be described. After measuring the diffusion length La of a silicon wafer having a high-density defect layer, the diffusion length L is measured by the SPV method in the state of recombining with Fe-B.
Since b1 is measured, the influence of recombination centers existing in addition to Fe—B can be eliminated, and changes in the diffusion lengths La and Lb1 of minority carriers for iron and boron can be accurately measured. That is, in a silicon wafer in which defects are generated at a high density, oxygen precipitation-induced defects and inter-lattice recombination centers such as Fe exist in addition to Fe-B, and they are activated by heating at 200 ° C. -B (inactive state)
It remains active after being left to recombine with. Therefore, the present invention eliminates the effect of recombination centers other than Fe-B,
Accurate measurement is possible. This gives (La ^-2
An accurate value is obtained because the iron concentration is calculated by multiplying -Lb1 ^-2 ) by a constant.

Further, after measuring the diffusion length Lb1, the diffusion length L
By measuring a1, a large number of silicon wafers to be measured are collectively activated, and then recombined with Fe-B, which takes a relatively long time, are also collectively taken, and then diffusion takes a relatively short time individually. Since the lengths Lb1 and La1 are measured, it is useful for reducing the total number of measurement steps. Moreover,
In the case of a silicon wafer or the like, which is subjected to a three-step IG process or the like to form a defect-free layer on both end surfaces, the defect-free layer is removed by etching to expose the defect layer, so that the defect distribution of the measurement portion by the SPV method is uniform. Therefore, it becomes possible to perform accurate measurement in the same manner as above.

[0009]

Embodiments of the method for measuring the iron concentration of a silicon wafer according to the present invention will be described below in detail with reference to the drawings. FIG. 1 shows an outline of the procedure for measuring the iron concentration of a silicon wafer according to this embodiment (steps 21 to 27), and FIG.
The following is an outline of the state of the silicon wafer in the main steps of. In FIG. 2, (a) is a silicon wafer whose iron concentration is to be measured, (b) is a silicon wafer after being removed by etching, (c) is a measurement of diffusion length La, and (d) is a diffusion length Lb.
1 shows the measurement of 1.

First, in step 21 and FIG. 2A, the silicon wafer 10 to which the present embodiment is applied is applied.
Is a p-type silicon wafer 10 using boron as a dopant, which has a defect layer 1 inside in the thickness direction and defect-free layers 2 formed on both end faces. The defect layer 1 has defects such as oxygen precipitation induced defects at a high density, for example, about 10 ⁷ / cm ³ or more which is usually called high density. The silicon wafer 10 was single-crystallized by the Czochralski (CZ) method or a new CZ method such as a magnetic field pulling (MCZ) method, a recharge pulling (RCCZ) method, and a continuous charge pulling (CCCZ) method. The ingot is processed into a mirror-polished wafer by wafer processing such as slicing, and then subjected to a 3-step IG process. Since the silicon wafer 10 may have the high-density defect layer 1 inside, in whole or in part, it may have a structure such as an epitaxial layer or a buried layer for semiconductor devices, The IG treatment may also be performed under heat treatment conditions, such as a two-step IG treatment, as necessary, and may not be performed. In addition, the silicon wafer 1 is measured by the SPV method described later.
From the relationship of measuring the iron concentration of the defect layer 1 of 0, the thickness of the defect layer 1 is preferably a thickness which is the measurement depth, usually about 200 μm or more. For the same reason, the resistivity of the defect layer 1 is 0.0
It is preferably about 5 Ω · cm or more.

Next, as shown in step 22 and FIG. 2B, in order to expose the defect layer 1, a portion of the silicon wafer 10 having a defect density different from that of the defect layer 1 such as the defect-free layer 2 is chemically etched. To remove. In order to surely expose the defect layer 1, a part of the surface of the defect layer 1 may be etched. As the etching solution used here, a general etching solution for silicon such as a mixed solution of hydrofluoric acid and nitric acid or a mixed solution of hydrofluoric acid and nitric acid with water or acetic acid is used. Silicon wafer 10 after this etching
Is preferably in a smooth mirror-like state, and it is preferable to perform temperature control and stirring with a mixed liquid of hydrofluoric acid and nitric acid to which water or acetic acid is added. Since the measurement by the SPV method described later measures the exposed defect layer 1, the defect-free layer 2 to be removed by etching may be only one side. From before this step,
Silicon wafer 1 having exposed high-density defect layer 1
In the case of 0, it goes without saying that this step may be omitted.

Next, in step 23, activation processing is first performed. In this activation treatment, iron-boron in the region of the defect layer 1 to be measured is dissociated by optical excitation by light irradiation, a heating furnace, or the like. After this dissociation is completed or can be regarded as completed, it is rapidly cooled. That is, the activation is a state in which dissociation of Fe-B → Fe _int + B occurs, and a dissociation temperature of about 200 ° C. or higher is preferable, and further, the surface of the defect layer 1 due to the supersaturated solid solution of Fe _int , To prevent precipitation into
More preferred is about 200 ° C to about 220 ° C. The above dissociation is
It is said that the process is completed in about 3 minutes at 200 ° C., and it is preferable to keep the activated state within about 30 minutes, although it depends on the thickness of the silicon wafer 10. After this activation and holding, the silicon wafer 10 is rapidly cooled to a temperature range where the rate of recombination with Fe-B is small, generally about room temperature. Step 24
And, as shown in FIG. 2C, after the rapid cooling, the defect layer 1
The exposed silicon wafer 10 can be used for SPV within a short time.
The diffusion length La of minority carriers is measured from the surface 1c of the defect layer 1 in the dissociated state 1a which becomes Fe _int + B by the measuring device 3 by the method.

Subsequently, in step 25, the silicon wafer 10 is left until the dissociated Fe _int + B is recombined with Fe-B. This standing may be carried out at a temperature lower than the dissociation temperature, but at room temperature it takes about 1 week,
It is said that the recombination is almost completed in about 20 minutes at ℃, room temperature to about 100 ℃, about 20 minutes to 1 week is preferable, and in consideration of work efficiency, about 80 to 100 ℃ is more preferable. Sometimes preferred. Step 26 and FIG.
As shown in (d), after standing, the silicon wafer 10 recombined with Fe-B has a temperature that can be measured by the measuring device 3,
Generally, the diffusion length Lb1 of minority carriers is measured by the measuring device 3 from the surface 1c of the defect layer 1 in the recombination state 1b with Fe-B at about room temperature.

Next, in step 27, the iron concentration [Fe] of the defect layer 1 is obtained based on the diffusion lengths La and Lb1 measured as described above. Generally, the iron concentration [F] is obtained.
e] = constant × (La ⁻² −Lb 1 ⁻² ). further,
The iron concentration [Fe] with reference to the iron concentration in the DLTS method
(Cm ⁻³ ) ≈1 × 10 ¹⁶ (La ⁻² −Lb 1 ⁻² ) is obtained, where the units of La and Lb 1 are μm.

The difference between the diffusion lengths La and Lb1 measured in this embodiment and the diffusion lengths La and Lb measured by the conventional SPV method will be described with reference to FIG. That is, in the prior art, assuming that the silicon wafer 10 (see FIG. 2B) that exposes the high-density defect layer 1 is measured, the diffusion length Lb of the minority carriers in the Fe—B bonded state is measured at time T _0. Then, the diffusion length La of minority carriers is measured in a state of being dissociated into Fe _int + B by activation (time T ₁ ), and the iron concentration [Fe] (c
m ⁻³ ) ≈1 × 10 ¹⁶ (La ⁻² −Lb ⁻² ) is obtained.
On the other hand, in the present invention, the diffusion length La of minority carriers in the state of being dissociated into Fe _int + B by activation (time T ₁ )
Was measured, by measuring the diffusion length Lb1 of minority carriers in the recombination state (time T ₂₎ to the Fe-B on standing, the iron concentration ^{[Fe] (cm -3) ≒} 1 × 10 16 (La -2 -Lb1 ^-2 )
Ask for. As is clear from the figure, the diffusion length La to be measured
Has the same value, but the diffusion length Lb and the diffusion length Lb1 are different values. In the prior art, in the case of a silicon wafer in which a layer having a density different from that of the defect layer, for example, a defect-free layer or the like is formed on the surface, the measured diffusion lengths La and Lb include layers other than the target layer. More accurate.

This difference is based on the assumption that in the prior art, there is no change other than Fe-B → Fe _int + B at 200 ° C. heating. That is, recombination centers other than Fe-B (oxygen precipitation-induced defects, interstitial Fe, etc.) exist in a silicon wafer in which oxygen precipitation-induced defects and the like are generated at high density, and Fe- at the time of measuring the diffusion length Lb. Although the recombination centers other than B are inactive, the recombination centers other than Fe-B are also activated by activation treatment such as heating at 200 ° C. Therefore, the diffusion length Lb in the prior art also includes changes in recombination centers other than Fe-B. On the other hand, the diffusion length Lb1 of the present invention measures the change of only Fe-B, and an accurate iron concentration can be obtained.

Next, regarding another embodiment according to the present invention,
The outline of the procedure for measuring the iron concentration of the silicon wafer of this embodiment shown in FIG. 4 (steps 31 to 38) will be described with reference to FIGS. 2 and 3. This embodiment is basically the same as the above-mentioned embodiment, but the diffusion length La of the minority carrier is set to the time T _3.
The difference is that it is measured with. That is, the silicon wafer 10 having the defect layer 1 exposed by chemical etching (steps 31 and 32) is activated to a state in which Fe-B is dissociated, as in the above-described embodiment, and the dissociation is completed or can be regarded as completed. Rapidly cools down (step 33)
However, the diffusion length is not measured. After that, dissociated Fe _int +
The silicon wafer 10 is allowed to stand until B is recombined with Fe-B (step 34), and then the diffusion length Lb1 of the minority carriers is measured (time T in the same manner as in the above-described embodiment) in step 35.
₂ ) Do. Next, Fe-B is activated again to a state where it dissociates, and after this dissociation is complete or can be regarded as complete, it is rapidly cooled (step 36), and the diffusion length La1 is measured in the same manner as in the above-mentioned embodiment (step 36). 37), iron concentration [F
e] (cm ⁻³ ) ≈1 × 10 ¹⁶ (La1 ⁻² −Lb1 ⁻² ) is calculated (step 38). It should be noted that the diffusion length La1 and the diffusion length La of the above-described embodiment have the same values, although the measuring times are different.

The iron concentration [Fe] obtained in this example is the same as that in the above example, but a large number of silicon wafers 1
This is useful when measuring an iron concentration of 0. That is,
A large number of silicon wafers 10 after the chemical etching treatment are simultaneously activated, rapidly cooled and then left to be recombined, and the diffusion length Lb1 and the diffusion length La1 are individually measured. Therefore, the activation process is increased once, but it is sufficient to leave it to be recombined once, so that the working time is greatly reduced when measuring the iron concentration of a large number of silicon wafers.

[0019]

According to the present invention, a silicon wafer having a high-density defect layer is heated, Fe-B is dissociated and then rapidly cooled, the diffusion length La is measured by the SPV method, and then dissociated iron and boron. Diffuses the diffusion length Lb1 after recombining with Fe-B.
It is measured by the V method, or after the diffusion length Lb1 is measured first, the diffusion length La1 is measured, and from these, (La ⁻² −Lb
1 ⁻² ) or (La1 ⁻² −Lb1 ⁻² ) is obtained, and this is multiplied by a constant, for example, about 1 × 10 ¹⁶ to obtain the iron concentration. Therefore, the influence of recombination centers other than Fe-B is excluded and the diffusion length is measured, so that an accurate iron concentration is measured. Further, when the diffusion length Lb1 is measured first, a large number of silicon wafers are processed at once by a standing process that requires a relatively long time, so that man-hours can be reduced, which contributes to cost reduction. Further, in the iron concentration measuring method, the defect-free layer or the like of the silicon wafer is removed by etching as necessary to expose the defect layer, and the diffusion length is measured by the SPV method.
It is a simple measurement method that does not require highly skilled technology.
As described above, the degree of iron contamination of the silicon wafer can be measured accurately and easily, so that the quality evaluation such as the cleanliness of the silicon wafer during the manufacturing process or after the manufacturing can be accurately and easily performed.

[Brief description of drawings]

FIG. 1 is a diagram illustrating an outline of a procedure for measuring an iron concentration of a silicon wafer according to an example of the present invention.

2 is a diagram showing an outline of the state of the silicon wafer in the main steps of FIG.

FIG. 3 is a diagram illustrating a difference between a diffusion length measured by an example according to the present invention and a diffusion length measured by an SPV method according to a conventional technique.

FIG. 4 is a diagram for explaining the outline of the procedure for measuring the iron concentration of a silicon wafer according to another embodiment of the present invention.

FIG. 5 is a diagram illustrating an outline of a procedure for measuring the iron concentration of a silicon wafer by the SPV method according to the related art.

[Explanation of symbols]

1 defect layer, 2 defect free layer, 3 measuring device by SPV method, 10 silicon wafer, La, La1, Lb, L
b1 diffusion length.

Claims (3)

[Claims] 1. Fe-B in a p-type silicon wafer
Fe is activated by activating the diffusion length Lb of the minority carriers in the bound state.
_{In the} method for measuring the iron concentration of a silicon wafer, the diffusion length La of the minority carriers in the state of being rapidly cooled after dissociation into _int + B is measured by the surface photovoltage method, and the iron concentration is determined from the difference between the values based on the measured values. After that, the diffusion length La in the rapidly quenched state is measured, recombined, the diffusion length Lb1 in the recombined state is measured, and the iron concentration in the high-density defect layer of the p-type silicon wafer is determined from the difference between the values based on these measurement values. A method for measuring the iron concentration of a silicon wafer, which is characterized by obtaining. 2. Fe-B in a p-type silicon wafer
Fe is activated by activating the diffusion length Lb of the minority carriers in the bound state.
_{In the} method for measuring the iron concentration of a silicon wafer, the diffusion length La of the minority carriers in the state of being rapidly cooled after dissociation into _int + B is measured by the surface photovoltage method, and the iron concentration is determined from the difference between the values based on the measured values. Then, it is rapidly cooled, recombined, the diffusion length Lb1 in the recombined state is measured, the diffusion length La1 in the rapidly cooled state after activation is measured again, and the high density of the p-type silicon wafer is obtained from the difference between the values based on these measurement values. Method for measuring iron concentration in a silicon wafer, characterized in that the iron concentration in a defect layer is obtained. 3. The method for measuring iron concentration of a silicon wafer according to claim 1, wherein the defect layer is a defect layer exposed by removing the defect-free layer by etching.