JP2003152041A - Method for evaluating cleanness of silicon wafer - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for evaluating the cleanness of a silicon wafer in which the cleanness can be accurately evaluated with high reliability by obtaining a sufficient signal strength for the measurement of an SPV method for an N-type silicon wafer. SOLUTION: The method for evaluating the cleanness of the silicon wafer comprises the steps of treating the N-type silicon wafer with an SC-1 liquid used in an RCA cleaning method, then housing the wafer in a wafer case made of resin, preserving the wafer at 20 to 30 deg.C for 5 days or more, thereafter evaluating the cleanness of the wafer by measuring the surface photoelectric voltage.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for evaluating the cleanliness of a silicon wafer, and more particularly, to a surface photovoltage (N-type silicon wafer).
e; SPV) method for evaluating cleanliness of silicon wafers.

[0002]

2. Description of the Related Art Cleanliness of a wafer is one of the important quality evaluation items required for a silicon wafer. For the evaluation of the cleanliness of the wafer, for example, when the contamination state due to metal impurities is evaluated, the electrical characteristics such as lifetime are measured, or a chemical method such as IPC-MS (inductively coupled plasma / mass spectrometry) is used. , The impurity metal element concentration was measured.

However, in the above electrical characteristic measurement, it is necessary to form an oxide film, an electrode, etc. on the surface of the silicon wafer before the measurement, and a complicated process must be performed. Moreover, the measured value was influenced by the formation process of the oxide film, the electrode and the like. Further, in the above-mentioned chemical method, after the acid is used to dissolve the extreme surface layer of the silicon wafer, a concentration step and the like must be performed, which requires time and requires skill to perform the work. there were.

On the other hand, SPV is an easy and simple method for evaluating cleanliness of silicon wafers, which does not require pretreatment such as oxidation treatment and does not require skill in analysis work.
Method can be used. In this SPV method, minority carriers introduced by light irradiation or the like move to the wafer surface by the electric field of the depletion layer existing in the wafer surface layer portion, and the SPV generated as a result is measured. In the actual measurement, the wavelength is changed and light is irradiated, and the SPV value is measured for each wavelength. Then, the minority carrier diffusion length can be obtained from the relationship between the penetration depth of the irradiated light and the generated SPV value, and the lifetime can be calculated from the minority carrier diffusion length.

The minority carrier diffusion length in a silicon wafer is affected by impurity metals and crystal defects in the surface layer of the wafer. Therefore, from the measured value of the minority carrier diffusion length of the silicon wafer obtained as described above, it is possible to evaluate the contamination state of the impurity metal in the wafer surface layer portion, in particular, heavy metal such as Fe. Moreover, the SPV method has an advantage that the time required for measurement is short and the measurement can be performed without contact with the wafer. Since the wafer is not contaminated or destroyed during measurement, This is an excellent cleanliness evaluation method from the viewpoint of cost.

[0006]

However, the above S
In the case of a P-type silicon wafer doped with boron (B), the PV method can measure the concentration of heavy metals such as Fe in the bulk with sufficient accuracy.
In the case of N-type silicon wafer doped with
It was difficult to measure. That is, in the case of a P-type silicon wafer, Fe-B formed in the wafer surface layer part
Since a signal of sufficient intensity can be obtained by utilizing the relatively shallow donor level of the pair, it is possible to evaluate the concentration of heavy metal impurities such as Fe with sufficient accuracy. On the other hand, N
In the case of a type silicon wafer, measurement is possible as a function of the SPV measuring device, but in reality, since it does not have the shallow donor level as described above, the signal strength is weak,
Even if it could not be measured or could be measured, sufficient accuracy could not be obtained, and the measurement result was unreliable.

Therefore, conventionally, in the evaluation of the N-type silicon wafer by the SPV method, after the boiling treatment with hydrogen peroxide (H ₂ O ₂ ) water is performed as a pretreatment, an oxide film is formed on the surface of the silicon wafer. , The measurements were being made. However, even with this method, the signal strength is weak and cannot be measured, or even when the measured value is obtained, its reliability is low. For this reason, there has been a demand for a highly accurate and reliable wafer cleanliness evaluation method capable of obtaining a sufficient SPV signal intensity in an N-type silicon wafer.

The present invention has been made in order to solve the above technical problems. With respect to an N type silicon wafer, a signal intensity sufficient for measurement by the SPV method is obtained, and cleaning is performed with high accuracy and high reliability. An object of the present invention is to provide a silicon wafer cleanliness evaluation method capable of evaluating the cleanliness.

[0009]

A cleanliness evaluation method for a silicon wafer according to the present invention is a cleanliness evaluation method for an N-type silicon wafer, wherein the wafer is treated with SC-1 solution used in an RCA cleaning method. Then, the wafer is made to be housed in a resin wafer case, stored at 20 to 30 ° C. for 5 days or more, and then the cleanliness of the wafer is evaluated by measuring the surface photovoltage. As a result, the signal strength in SPV measurement is improved, and it is possible to evaluate the cleanliness level with sufficient accuracy and reliability even with an N-type silicon wafer.

The resin forming the wafer case is polyethylene, polypropylene, polystyrene, A
It is preferably made of at least one selected from BS resin, PET and polycarbonate. These resin wafer cases have an appropriate emission concentration of organic gas from a trace amount of additives contained in the resin,
It is possible to form a charged film made of an organic substance that exerts a suitable effect for improving the signal strength.

The evaluation method according to the present invention is also effective when the N type silicon wafer is an N / N ⁻ type epitaxial wafer having a silicon epitaxial film formed on the surface thereof. Since the epitaxial film is almost defect-free, and since it does not contain oxygen, recombination of excess carriers in the epitaxial film is small, and therefore the SPV value measured in the epitaxial wafer is substantially the same as that obtained from the substrate wafer. Since it reflects the above, it can be evaluated similarly to the prime wafer.

[0012]

BEST MODE FOR CARRYING OUT THE INVENTION The present invention is described in more detail below.
To do. The evaluation method according to the present invention is an N-type silicon wafer.
When evaluating the cleanliness of the wafer, pretreatment of the wafer
₂ O₂ Instead of boiling treatment with water, SC-1 liquid (NH _Four 
OH + H₂ O₂ + H₂ O) treatment point, and
The pretreated wafer in the resin wafer case.
After receiving and storing at 20 ~ 30 ℃ for 5 days or more, SPV measurement
Is characterized in that By this, SP
The signal strength in V measurement is improved, and N type silicon
Impurity gold with sufficient accuracy and reliability even in
It is possible to evaluate the cleanliness of metal pollution.

As described above, in the present invention, SPV is used.
Before the measurement, the wafer is treated with the SC-1 solution. SC
The oxide film formed on the surface of the silicon wafer by the treatment with the -1 solution has an extremely thin film thickness of about 10 to 100 nm, and is homogeneous and dense, so that it has excellent potential characteristics. Further, by forming the oxide film, recombination of excess carriers (minority carriers) excited by light irradiation on the wafer surface is suppressed, and the excess carrier density is made uniform in the wafer thickness direction.

Further, in the present invention, after the oxide film is formed, it is stored in a resin wafer case for a predetermined period. As a result, an extremely small amount of organic gas released from the resin forming the wafer case is gradually adsorbed on the surface of the oxide film on the silicon wafer. Then, this organic substance gas adheres in a state close to a monomolecular layer, and an extremely thin organic substance film is formed on the surface of the oxide film. The organic film is charged with the lapse of the storage period to generate an electric charge. Therefore, for a certain period (5
By storing for more than a day), a large depletion layer is generated in the surface layer of the wafer, the minority carriers are easily moved to the wafer surface, and the SPV value is increased, so that the S / N (signal / noise) ratio is improved. Therefore, it becomes possible to stably measure the minority carrier diffusion length. As a result, it becomes possible to obtain a sufficient signal strength in SPV measurement, and it is possible to highly reliably evaluate the impurity metal contamination even in the N-type silicon wafer.

In the present invention, the silicon wafer to be measured is an N-type prime wafer such as a mirror-finished wafer doped with a Group 15 element such as phosphorus (P), arsenic (As), antimony (Sb), or the like. An N type silicon wafer such as an N / N ⁻ type epitaxial wafer in which a silicon single crystal film is epitaxially grown on the surface of these wafers as a substrate is used.

In the case of an N / N ^- type epitaxial wafer, the epitaxial layer is almost defect-free and contains no oxygen, and therefore recombination of excess carriers in the epitaxial layer is considered to be small. Therefore, SP
The contribution of the epitaxial layer to the V value is very small, and the measured SPV value substantially reflects the information from the wafer substrate itself. As a result, the influence of the non-uniform dopant concentration distribution on the epitaxial wafer is reduced, the above-mentioned preconditions for the measurement by the SPV method are satisfied, and the minority carrier diffusion length can be accurately measured. Therefore, the measured minority carrier diffusion length value mainly reflects metal contamination and crystal defects of the wafer substrate.

In the present invention, the composition of the SC-1 solution used for forming the oxide film on the wafer surface is NH ₄ OH: H ₂ O.
₂ = 1: 1 to 2: 5, especially NH ₄ OH: H ₂ O ₂ : H
₂ O = 1: 2: 10 to 24 are preferable. By using the SC-1 solution having the above composition, a thickness of 1
It is possible to surely form a silicon oxide film having an extremely thin thickness of about 0 to 100 nm, a homogeneous and dense structure, and an excellent potential characteristic within an appropriate processing time.

The process for forming the oxide film on the wafer surface is, for example, pure water cleaning → SC-1 liquid processing → pure water cleaning, or hydrofluoric acid (for example, 25% HF aqueous solution) processing → pure water cleaning. → SC-1 liquid treatment → Pure water cleaning process sequence, etc.
It is preferable to perform the process by combining the SC-1 liquid treatment and pure water cleaning from the viewpoint of further improving the reliability of SPV measurement values.

In the pretreatment for measurement, it is important not to perform hydrofluoric acid treatment after the oxide film formation treatment with the SC-1 solution. When hydrofluoric acid water treatment is performed after the SC-1 solution treatment, even if the treated wafer is stored in a resin-made wafer case and stored for a long period of time, signal strength does not improve in SPV measurement and reliability is high. It is difficult to obtain accurate evaluation results. It is presumed that this is because the surface state of the formed oxide film is changed by the hydrofluoric acid treatment such that the surface of the oxide film is made hydrophilic and the charging of the adhering organic material layer is hindered.

The wafer that has undergone the above processing steps is housed in a resin wafer case. The components of this wafer case are
It is necessary to be made of an organic resin that gradually releases an extremely small amount of organic gas over a long period of time, and it is preferable to be made of a synthetic resin material such as a hydrocarbon type or an oxygen-containing hydrocarbon type. As described above, this organic substance gas forms an extremely thin organic substance film on the surface of the oxide film of the silicon wafer and charges it. Therefore, it is possible to obtain a signal intensity sufficient for SPV measurement. Become.

As the resin constituting the wafer case, polyethylene, polypropylene, polystyrene, A
Examples include BS resins, polyesters such as PET, and polycarbonates. Among them, polyethylene, polypropylene, and polycarbonate are particularly preferable. In these synthetic resins that make up the wafer case,
Additives such as a plasticizer, an antioxidant, an ultraviolet absorber, a polymerization initiator or a crosslinking agent are contained, and these are released as an organic gas. The main components of the released organic gas are plasticizer components such as dibutyl phthalate, tri- (2-ethylhexyl) phosphate, di-2-ethylhexyl phthalate, bis (t-butylperoxy-isopropyl) benzene, 1- ( 2-t-butylperoxy-
Examples include a polymerization initiator such as isopropyl) -3-isopropenylbenzene or a crosslinking agent component.

The storage temperature in the resin wafer case is preferably 20 to 30 ° C. 20-30 ° C
Then, the vapor pressure of the above organic gas component is low, and it is gradually released in trace amounts, and although it has an extremely low concentration in the atmosphere inside the wafer case, many of them have polar groups and gradually increase on the oxide film surface of the wafer. It is adsorbed and forms a thin film close to a monomolecular film.

It should be noted that the above-mentioned organic gas generating substance is introduced into a wafer case made of metal, glass, ceramics or the like to fill the storage atmosphere of the wafer with the organic gas or apply the organic material. Although it is conceivable, it is difficult to uniformly control the release amount of the organic substance gas and the application amount of the organic substance to an extremely small amount, and it is not always possible to obtain the same effect as when the wafer is made of resin.

In the present invention, the SPV is applied to the wafer stored in the resin wafer case and stored at 20 to 30 ° C. for 5 days or more, more preferably 7 days or more.
Take a measurement. That is, for the N-type silicon wafer in which the charged adsorbed organic compound layer is formed on the surface of the oxide film,
A PV measuring device is used to scan the optical probe and measure the generated photovoltage. Specifically, a plurality of lights having different wavelengths are individually irradiated and scanned, and the SPV value is measured for each wavelength. Then, the penetration depth of the irradiated light and the generated SP
The minority carrier diffusion length can be obtained from the relationship with the V value.

Therefore, the impurity metal concentration in the N-type silicon wafer can be evaluated from the minority carrier diffusion length obtained as described above. For example, the minority carrier diffusion length of the wafer to be evaluated is measured by the method according to the present invention by previously obtaining the correlation between the concentration of the impurity metal such as Fe in the wafer and the minority carrier diffusion length. From the measured value, the impurity metal (F
e) Contamination can be easily evaluated.

As described above, the method for evaluating the cleanliness of a silicon wafer according to the present invention is easy to pretreat and uses the SPV method for the measurement, so that the measurement can be performed without contact with the wafer. Since no skill is required, highly reliable evaluation can be easily performed, and it can be suitably used for wafer quality control in a semiconductor manufacturing process.

[0027]

EXAMPLES The present invention will be described in more detail based on the following examples, but the invention is not intended to be limited by the following examples. [Example 1] A 6-inch diameter P-doped N-type CZ silicon wafer (resistivity 10.5 to 11.3 Ωcm) was used as a substrate, and a silicon epitaxial film having a thickness of 46 μm (film resistivity 27.5 Ωcm) was formed on the substrate surface. To form N /
An N ^- type epitaxial silicon wafer was prepared.
The wafer was stored in a polypropylene wafer case and stored for 215 days, and then SPV measurement was performed. This wafer was washed with pure water → SC-1 liquid (NH ₄ OH: H ₂ O
₂ : H ₂ O = 1: 2: 10) Treatment → purified water washing, then SPV measurement is performed, and this wafer is stored in a polypropylene wafer case and stored at 25 ° C. for 15 days. Then, SPV measurement was performed. The results of these measurements are shown in Table 1. In the SPV measurement, seven kinds of optical filters were used to measure 177 points on the wafer surface, and the correlation (linearity) between the penetration depth of light at each wavelength and the reciprocal of the SPV signal was obtained. Are also shown in Table 1.

[Embodiment 2] P-doped N-type CZ silicon wafer having a diameter of 6 inches (resistivity: 14.6 to 26.7Ω).
cm) as a substrate, a silicon epitaxial film (film resistivity 27.5 Ωcm) having a thickness of 46 μm was formed on the surface of the substrate to prepare an N / N ⁻ type epitaxial silicon wafer. This wafer was washed with pure water / SC-1 solution (N
H ₄ OH: H ₂ O ₂ : H ₂ O = 1: 2: 10) After treatment / purified water washing, the same procedure as in Example 1 was performed.
PV measurement was performed. The results of this measurement are shown in Table 1.

[Comparative Example 1] A wafer which has been subjected to a series of SC-1 solution treatments in Test 1 was subjected to isopropyl alcohol coating treatment, and then sealed in a polypropylene wafer case containing a small amount of isopropyl alcohol at the bottom. After being stored and stored for 15 days, SPV measurement was performed in the same manner as in Example 1. The results of these measurements are shown in Table 1.

[0030]

[Table 1]

As shown in Table 1, long-term (215 days)
The stored N / N ^- type epitaxial silicon wafers (Sample Nos. 1 and 7) are considered to have a natural oxide film formed on the surface of the wafer and organic substances spontaneously adsorbed thereon, and SPV measurement is not possible. However, the signal strength is 2m
V was weak and insufficient. Here, the signal strength means Sensitiv in the SPV measuring device used in the above test.
This means the ity value (mV).
If the Sensitivity value is 5 mV or more, it can be said that the signal strength is sufficient.

In addition, in Examples 1 and 2, after forming an oxide film on the wafer surface by the SC-1 solution treatment, the longer it was stored in the resin-made wafer case, the more the signal intensity increased, and the signal strength was stored for 5 days or more. Then (Sample No. 5, 6, 1
1, 12), the signal intensity was 5 mV or more, and it was confirmed that sufficient signal intensity was obtained. Also, the linearity was 0.995 or more, and it was confirmed that a reliable measured value was obtained.

On the other hand, when isopropyl alcohol treatment was performed (Comparative Example 1), SPV measurement was difficult even when stored for 5 days or longer (Sample No. 16), and further after 15 days storage (Sample No. 16). Also in 17),
The signal strength was weak, the measured values varied widely, and the linearity was less than 0.99, making it difficult to obtain reliable measurement results.

[Embodiment 3] A P-doped N-type electrode having a diameter of 6 inches
Ip CZ silicon wafer (resistivity 35.40-40.0
3 Ωcm, oxygen concentration 1.57 × 10¹⁸atoms / cm
³ ) (Sample A) and P-doped N-type 6 inch diameter
Ip CZ silicon wafer (resistivity 39.19-43.0
6 Ωcm, oxygen concentration 1.56 to 1.59 × 10¹⁸ato
ms / cm ³ ) (Sample B) as a substrate
The surface of these substrates has a thickness of 10 μm
Forming a axial film (film resistivity 20 Ωcm), N / N^- 
Type epitaxial silicon wafer (Sample A ',
B ') was produced. These four types of wafer samples
25 in each polypropylene wafer case
Stored at 7 ° C for 7 days. And two types of N / N^- Type epi
For the axial silicon wafers (Samples A ', B')
Then, SPV measurement was performed. As a result, sample A ',
For both B ', the signal strength (Sensitivity value) is 1 mV or less
Was weak. In addition, SPV measurement is performed with 7 types of optics.
Measure 9 points on the wafer surface using a filter
The reverse of the penetration depth of light and SPV signal at each wavelength
The correlation (linearity) with the number was also obtained.

Next, the above-mentioned four types of wafers (Sample A,
B, A ', B'), in the following four steps, S
C-1 liquid treatment was performed. 0.2% HF cleaning → pure water cleaning (twice) → SC-1 solution (NH ₄ OH: H ₂ O ₂ : H ₂ O = 1: 2: 24) treatment (twice) → pure water cleaning (2 Times) → 0.017% HF cleaning →
Pure water washing (twice) SC-1 solution _{(NH 4 OH: H 2 O} 2: H 2 O = 1:
2:24) Treatment (twice) → Pure water washing (twice) → 0.01
7% HF cleaning → pure water washing (twice) pure water washing → SC-1 solution _{(NH 4 OH: H 2 O} 2: H 2
O = 1: 2: 10) treatment → pure water cleaning (twice) 25% HF cleaning tank → pure water cleaning (twice) → SC-1 liquid (NH ₄ OH: H ₂ O ₂ : H ₂ O = 1) : 2:10) Treatment tank → Washing with pure water (twice) The wafers that have undergone the above-mentioned SC-1 solution treatment steps are respectively housed in polypropylene wafer cases and stored at 25 ° C. for 1 hour.
The wafer was stored for 8 days, and SPV measurement was performed on each wafer.

As a result, the samples A'and B'processed in the above steps had weak signal intensities of 2 mV and could not be said to have sufficient intensity even after being stored for 18 days. From this, it was confirmed that when the HF treatment was performed after the SC-1 liquid treatment, the improvement of the signal intensity in the SPV measurement was hindered. Also, similarly,
Samples A and B also had a weak signal intensity of 1 to 2 mV.

On the other hand, the samples treated in the process and step had a signal intensity of 5 mV or more when stored for 7 days or more. Moreover, it was confirmed that the linearity of each of these wafers was 0.995 or more, and reliable SPV measurement values were obtained. In addition, when the steps were compared, no significant difference was observed regarding the presence or absence of HF treatment before the SC-1 solution treatment.

[0038]

As described above, according to the present invention, with respect to the N type silicon wafer, a sufficient signal strength for the measurement by the SPV method can be obtained, and with high accuracy and high reliability,
Cleanliness can be evaluated. In addition, the method for evaluating cleanliness of a silicon wafer according to the present invention is easy to perform pretreatment for measurement and uses the SPV method for measurement, so that measurement can be performed without contact with the wafer, and no skill is required. For,
Highly reliable evaluation can be easily performed, and it can be suitably used for wafer quality control in a semiconductor manufacturing process.

Claims (3)

[Claims] 1. A method for evaluating the cleanliness of an N-type silicon wafer, wherein the wafer is treated with an SC-1 solution used in an RCA cleaning method, and then stored in a resin wafer case at 20 to 30 ° C. A method for evaluating the cleanliness of a silicon wafer, which comprises storing the wafer for more than a day and then measuring the surface photovoltage to evaluate the cleanness of the wafer. 2. The resin constituting the wafer case is made of at least one selected from polyethylene, polypropylene, polystyrene, ABS resin, PET and polycarbonate.
A method for evaluating cleanliness of a silicon wafer as described above. 3. The N-type silicon wafer is an N / N ⁻ type epitaxial wafer having a silicon epitaxial film formed on the surface thereof.
Alternatively, the cleanliness evaluation method for a silicon wafer according to claim 2.