JP7790626B2 - Epitaxial silicon wafer and method for manufacturing the same, and method for manufacturing semiconductor device - Google Patents

Description

The present invention relates to an epitaxial silicon wafer and a method for manufacturing the same, as well as a method for manufacturing a semiconductor device.

Epitaxial silicon wafers, which have an epitaxial layer of single-crystal silicon formed on a silicon wafer, are used as device substrates for fabricating various semiconductor devices, such as MOSFETs (Metal-Oxide-Semiconductor Field-Effect Transistors), DRAMs (Dynamic Random Access Memory), power transistors, and BSI (Back Side Illumination) CISs (CMOS Image Sensors).

Heavy metal contamination of the epitaxial layer can increase the dark current of the CIS and cause defects known as white defects, degrading the characteristics of the semiconductor device. Therefore, to suppress such heavy metal contamination, there is a technology that forms gettering sites in the silicon wafer to capture heavy metals. One known method is to implant ions into the silicon wafer and then form the epitaxial layer. In this method, the ion-implanted region functions as the gettering site.

Patent Documents 1 and 2 describe methods for producing epitaxial silicon wafers, which include _the steps of irradiating the surface of a silicon wafer with cluster ions, such as _C3H5 , whose constituent elements are carbon and hydrogen, to form a modified layer in the surface layer of the silicon wafer, in which the constituent elements of the cluster ions are dissolved, and forming a silicon epitaxial layer on the modified layer of the silicon wafer.

Patent document 1 shows that a modified layer formed by irradiating cluster ions whose constituent elements are carbon and hydrogen exhibits higher gettering capabilities than an ion-implanted region obtained by implanting carbon monomer ions.

Patent Document 2 describes an improved technique for the technique described in Patent Document 1, in which the gettering capability of heavy metals can be improved by irradiating a high dose of cluster ions composed of carbon and hydrogen so that a portion of the modified layer in the thickness direction becomes amorphous. Patent Document 2 also describes that when irradiating with such a high dose of cluster ions, tiny black dot-like defects caused by implanted carbon, etc., are visible in cross-sectional TEM images of the modified layer after epitaxial growth, and that these black dot-like defects may contribute to improved gettering capability.

On the other hand, if the carbon dose is increased, the high concentration of carbon implanted in the surface layer (modified layer) of the silicon wafer diffuses into the silicon epitaxial layer during epitaxial growth and the device formation process, forming carbon-induced point defects in the silicon epitaxial layer (i.e., the device formation region), which may affect the device characteristics. Therefore, Patent Document 3 describes a method for manufacturing an epitaxial silicon wafer, in which a modified layer is formed by irradiating the surface of the silicon wafer with a beam of cluster ions containing SiH _x (x is one or more selected from integers of 1 to 3) ions and C ₂ H _y (y is one or more selected from integers of 2 to 5) ions, and a silicon epitaxial layer is formed on the modified layer. It describes that this method allows the dose of C ₂ H _y ions to be lowered, reducing the carbon concentration in the modified layer, while ensuring gettering capability by implanting SiH _x ions.

International Publication No. 2012/157162 International Publication No. 2015/104965 International Publication No. 2022/044562

Hydrogen implanted into the surface layer (modified layer) of a silicon wafer diffuses into the epitaxial layer through heat treatment during the device formation process of forming semiconductor devices on the epitaxial layer, passivating (inactivating) interface state defects within the epitaxial layer and contributing to improved device characteristics such as reduced leakage current. However, Patent Documents 1 to 3 all examine the irradiation conditions of cluster ions, focusing only on the gettering ability of epitaxial silicon wafers, and do not consider the passivation effect of hydrogen implanted into epitaxial silicon wafers. Therefore, the inventors conducted research and found that there is room for improvement in the passivation effect of hydrogen in epitaxial silicon wafers.

In view of the above problems, the present invention aims to provide an epitaxial silicon wafer with a high hydrogen passivation effect and a method for manufacturing the same.

To solve the above problems, the present inventors conducted extensive research and obtained the following findings. First, when irradiating only cluster ions composed of carbon and hydrogen as constituent elements, as in Patent Documents 1 and 2, reducing the carbon dose to suppress the formation of carbon-induced point defects in the silicon epitaxial layer not only fails to ensure gettering capability, but also reduces the amount of hydrogen implanted in the modified layer and remaining after the formation of the epitaxial layer, making it impossible to obtain the passivation effect of hydrogen. Therefore, as in Patent Document 3, it was decided to irradiate with cluster ions containing at least SiH _x ions and C ₂ H _y ions. However, even in this case, if the carbon dose is reduced too much, as in Patent Document 3, the amount of hydrogen remaining in the modified layer decreases, and the passivation effect of hydrogen is not sufficiently obtained. Therefore, in order to obtain the passivation effect of hydrogen, it was necessary to ensure a certain carbon dose. Furthermore, by adjusting the ratio of the number of Si atoms to the number of C atoms irradiated within a predetermined range, the passivation effect of hydrogen could be sufficiently obtained. By using this manufacturing method, an epitaxial silicon wafer was obtained that had micro-defects consisting of carbon aggregates and EOR (End of Range) defects caused by Si implantation in the surface layer (modified layer) of the silicon wafer, and that had a high passivation effect due to hydrogen.

In other words, the gist of the present invention is as follows:

[1] A first step of irradiating a surface of a silicon wafer with a beam of cluster ions containing SiH _x (x is one or more selected from integers of 1 to 3) ions and C ₂ H _y (y is one or more selected from integers of 2 to 5) ions to form a modified layer in a surface layer portion of the silicon wafer, in which constituent elements of the cluster ions are solid-dissolved;
a second step of forming a silicon epitaxial layer on the modified layer of the silicon wafer;
and
the total dose of the cluster ions irradiated in the first step is 6.00×10 ¹³ ions/cm ² or more and 1.00×10 ¹⁵ ions/cm ² or less;
the dose of the C ₂ H _ions irradiated in the first step is more than 1.00×10 ¹⁴ ions/cm ² and not more than 3.00×10 ¹⁴ ions/cm ² ;
a ratio of the number of Si atoms to the number of C atoms implanted in the first step, [Si/C], of 0.3 to 1.6;

[2] The method for producing an epitaxial silicon wafer according to the above [1], wherein the dose of the C ₂ H _ions irradiated in the first step is 1.25×10 ¹⁴ ions/cm ² or more.

[3] a silicon wafer;
a modified layer formed on a surface layer portion of the silicon wafer, the modified layer containing a solid solution of at least one of carbon and hydrogen;
a silicon epitaxial layer formed on the modified layer;
and
In a defect evaluation of the modified layer using a cross-sectional TEM image, first defect regions were observed in the modified layer, in which microdefects composed of carbon aggregates having a size of 3 nm or more and ¹⁰ nm or less existed at a density of 1.00× ¹⁰ /cm ^{or more} and 8.00× ¹⁰ /cm or less, and second defect regions were observed in which EOR defects having a maximum width of 50 nm or more and 250 nm or ^less existed at a density of 2.00× ¹⁰ /cm or more and less than 5.00× ¹⁰ /cm2,
the amount of carbon distributed in the silicon epitaxial layer and the modified layer is greater than 2.00×10 ¹⁴ atoms/cm ² and not more than 6.00×10 ¹⁴ atoms/cm ² ;
An epitaxial silicon wafer, wherein a peak hydrogen concentration in a depth direction of the modified layer measured by SIMS is 5.00×10 ¹⁶ atoms/cm ³ or more and 1.00×10 ¹⁸ atoms/cm ³ or less.

[4] The epitaxial silicon wafer according to the above [3], wherein, in defect evaluation using a cross-sectional TEM image of the modified layer, the EOR defects are present in ^{the second} defect region at a density of 2.50 × ¹⁰ /cm or more and 4.50 × ¹⁰ /cm or less.

[5] The epitaxial silicon wafer according to the above [3] or [4], wherein the amount of carbon distributed in the silicon epitaxial layer and the modified layer is 2.50×10 ¹⁴ atoms/cm ² or more.

[6] The method for producing an epitaxial silicon wafer according to [1] or [2] above;
forming a semiconductor device in the silicon epitaxial layer of the epitaxial silicon wafer;
A method for manufacturing a semiconductor device comprising:

[7] A method for manufacturing a semiconductor device, which comprises forming a semiconductor device in the silicon epitaxial layer of the epitaxial silicon wafer described in [3] or [4] above.

[8] A method for manufacturing a semiconductor device, which comprises forming a semiconductor device in the silicon epitaxial layer of the epitaxial silicon wafer described in [5] above.

The epitaxial silicon wafer manufacturing method of the present invention makes it possible to manufacture epitaxial silicon wafers with a high passivation effect due to hydrogen. Furthermore, the epitaxial silicon wafer of the present invention has a high passivation effect due to hydrogen.

1A to 1C are schematic cross-sectional views illustrating a method for manufacturing an epitaxial silicon wafer 100 according to an embodiment of the present invention. 1 is a graph (mass spectrum) showing mass fragments of various cluster ions obtained from diethylsilane (SiC ₄ H ₁₂ ) as a source gas. 1A and 1B are graphs showing (a) a carbon concentration profile and (b) a hydrogen concentration profile by SIMS after forming an epitaxial layer in an example of the present invention. 1 shows cross-sectional TEM images (magnification: 200,000 times) of modified layers of (a) Comparative Example No. 9, (b) Inventive Example No. 6, and (c) Inventive Example No. 4 in the examples of the present invention. 1A and 1B are equivalent circuit diagrams of a MOS structure during low-frequency CV measurement and high-frequency CV measurement, respectively, in a QSCV (Quasi-Static Capacitance-Voltage) method. 10 is a graph showing the calculation results of the interface state density _Dit in an example of the present invention.

Embodiments of the present invention will be described in detail below with reference to the drawings. Note that for ease of explanation, in Figure 1, the thicknesses of the modified layer 14 and silicon epitaxial layer 16 are exaggerated relative to the silicon wafer 10, rather than representing their actual thickness ratios.

(Method for manufacturing epitaxial silicon wafers)
1, a method for producing an epitaxial silicon wafer 100 according to one embodiment of the present invention includes a first step (steps A and B in FIG. 1) of irradiating a surface 10A of a silicon wafer 10 with a beam of cluster ions 12 containing SiH _x ions 12A (x is one or more selected from integers of 1 to 3) and C ₂ H _y ions 12B (y is one or more selected from integers of 2 to 5) to form a modified layer 14 in the surface layer portion of the silicon wafer 10, in which the constituent elements of the cluster ions 12 are solid-dissolved, and a second step (step C in FIG. 1) of forming a silicon epitaxial layer 16 on the modified layer 14 of the silicon wafer 10. The silicon epitaxial layer 16 becomes a device layer for manufacturing semiconductor devices such as BSI-type CIS.

[First step]
The silicon wafer 10 may be, for example, a bulk single-crystal silicon wafer having no epitaxial layer on its surface. Carbon and/or nitrogen may be added to the silicon wafer to obtain higher gettering capability. Furthermore, a predetermined concentration of any dopant may be added to the silicon wafer to form a so-called n+ type or p+ type, or n- type or p- type substrate.

The silicon wafer 10 may also be an epitaxial silicon wafer, in which a silicon epitaxial layer is formed on the surface of a bulk single-crystal silicon wafer. The silicon epitaxial layer can be formed using typical CVD conditions. The epitaxial layer preferably has a thickness in the range of 0.1 to 10 μm, and more preferably in the range of 0.2 to 5 μm.

In the first step, a surface 10A of a silicon wafer 10 is irradiated with a beam of cluster ions 12 including SiH _x ions 12A (x is one or more integers selected from 1 to 3) and C ₂ H _y ions 12B (y is one or more integers selected from 2 to 5). The term "cluster ions" as used herein refers to ions obtained by electron bombardment, which involves colliding electrons with gaseous molecules to dissociate the bonds of the gaseous molecules, thereby forming atomic aggregates of various numbers of atoms, fragmenting the atomic aggregates, ionizing the ionized atomic aggregates of various numbers of atoms, and then mass-separating the ionized atomic aggregates of various numbers of atoms to extract ionized atomic aggregates of specific mass numbers. In other words, the term "cluster ions" as used herein refers to clusters formed by aggregating multiple atoms, which are ionized by imparting a positive or negative charge to the clusters, and is clearly distinguishable from monoatomic ions such as carbon ions and monomolecular ions such as carbon monoxide ions. Cluster ions typically contain approximately 5 to 100 atoms. As a cluster ion implantation device using this principle, for example, CLARIS (registered trademark) manufactured by Nissin Ion Equipment Co., Ltd. can be used.

When a silicon wafer 10 is irradiated with a beam of cluster ions 12 containing SiH _x ions 12A and C ₂ H _y ions 12B, the irradiation energy instantaneously heats the silicon in the surface layer of the silicon wafer 10 to a high temperature of approximately 1350 to 1400°C, causing it to melt. The silicon is then rapidly cooled, and carbon, hydrogen, and silicon derived from the cluster ions 12 are solid-dissolved in the surface layer of the silicon wafer. That is, the term "modified layer" used herein refers to a layer in which at least one of carbon, hydrogen, and silicon, which are constituent elements of the irradiated cluster ions, is solid-dissolved in interstitial or substitutional positions of the crystals in the surface layer of the silicon wafer. However, it is difficult to distinguish Si derived from SiH _x ions implanted in the surface layer of the silicon wafer from Si that constitutes the silicon wafer. Therefore, the term "modified layer" used herein is specified as a region in which the concentration of either element is detected to be higher than the background in the SIMS concentration profile of carbon and hydrogen in the depth direction of the silicon wafer. After the implantation of cluster ions and before the formation of the epitaxial layer, the surface layer of the silicon wafer that is approximately 500 nm or less from the surface becomes a modified layer. After the formation of the epitaxial layer, carbon diffuses from the surface to the interior of the silicon wafer, so that the surface layer of the silicon wafer that is approximately 2 to 4 μm from the surface (the interface between the epitaxial layer and the silicon wafer) becomes a modified layer.

Details will be explained based on experimental results in the examples, but in this embodiment, it is essential to use cluster ions 12 including SiH _x ions 12A and C ₂ H _y ions 12B. x is one or more selected from integers of 1 to 3, i.e., the SiH _x ions 12A include one or more selected from the group consisting of SiH ions, SiH ₂ ions, and SiH ₃ ions. y is one or more selected from integers of 2 to 5, i.e., the C ₂ H _y ions 12B include one or more selected from the group consisting of C ₂ H ₂ ions, C ₂ H ₃ ions, C ₂ H ₄ ions, and C ₂ H ₅ ions.

Without limiting the present invention, the present inventors consider the effects of using cluster ions 12 containing SiH _x ions 12A and C ₂ H _y ions 12B as follows. First, by irradiating the modified layer with C ₂ H _y ions 12B, micro-defects consisting of carbon aggregates are formed in the modified layer. These micro-defects serve as hydrogen trapping sites, so that a high concentration of hydrogen remains in the modified layer even after the formation of the epitaxial layer, thereby achieving a passivation effect by hydrogen. Furthermore, by simultaneously irradiating the modified layer with SiH _x ions 12A, since silicon atoms have a larger mass number than carbon atoms, significant damage (implantation defects) is introduced into the surface layer of the silicon wafer, and relatively large implantation defects (EOR defects) are formed in the modified layer 14. Then, implanted carbon derived from the C ₂ H _y ions 12B aggregates in these implantation defects, and hydrogen is further trapped in these implantation defects, forming defect regions that can exhibit a high passivation effect.

The gaseous molecules that serve as the raw material for the cluster ions are not particularly limited as long as they can simultaneously generate the above-mentioned _SiHx ions and _{C2Hy ions} , and examples thereof include diethylsilane ( _SiC4H12 ), _butylsilane ( _SiC4H12 ), _{methylpropylsilane} ( _{SiC4H12 )} , pentylsilane ( _SiC5H14 ), _{methylbutylsilane} ( _SiC5H14 ), ethylpropylsilane ( _SiC5H14 ) _{, etc.} However, cluster ions of various sizes can be generated from each of these raw material gases. For example, as shown in FIG. 2 and Table 1, diethylsilane (SiC ₄ H ₁₂ ) generates C ₂ H 2 ions, C ₂ H ₃ ions, C ₂ H ₄ ions, and C ₂ H ₅ ions as C ₂ H _y ions in the _mass number range of 26 to 31, and SiH ions, SiH ₂ ions, and SiH ₃ ions as SiH _x ions. Therefore, diethylsilane is optimal as the source gas used in this embodiment. By extracting cluster ions (fragments) within a desired mass number range from this, cluster ions of a desired ion species can be generated. For example, by extracting fragments with mass numbers of 29 to 31, cluster ions including SiH ₃ ions, SiH ₂ ions, and SiH ions as SiH _x ions and C ₂ H ₅ ions as C ₂ H _y ions can be generated.

The total dose of cluster ions can be adjusted by controlling the ion irradiation time as an apparatus setting value. In this embodiment, if the total dose of cluster ions irradiated in the first step is less than 6.00×10 ¹³ ions/cm ² , implantation defects that serve as hydrogen trapping sites and heavy metal gettering sinks are not formed, and therefore the hydrogen passivation effect and gettering ability are not obtained. Therefore, the total dose of cluster ions irradiated in the first step is set to 6.00×10 ¹³ ions/cm ² or more, and preferably 8.00×10 ¹³ ions/cm ² or more. On the other hand, if the total dose of cluster ions irradiated in the first step exceeds 1.00×10 ¹⁵ ions/cm ² , defects may occur in the epitaxial layer. Therefore, the total dose of cluster ions irradiated in the first step is set to 1.00×10 ¹⁵ ions/cm ² or less, and preferably 8.00×10 ¹⁴ ions/cm ² or less.

If the dose of _{C2Hy ions} irradiated in the first step is ^1.00x1014 ions/ ^cm2 or less, the passivation effect by hydrogen is not sufficiently obtained. Therefore, the dose of _{C2Hy ions} irradiated in the first step is more than ^1.00x1014 ions/ ^cm2 , and preferably ^1.25x1014 ions/ ^cm2 or more. On the other hand, if the dose of _{C2Hy ions} irradiated in the first step is ^3.00x1014 ions/ ^cm2 or less, carbon diffusion into the epitaxial layer during epitaxial growth and the device formation process can be suppressed. Therefore, the dose of _C2Hy ions irradiated in the first step is ^3.00x1014 ions/ _cm2 or ^less , and preferably ^2.50x1014 ions/ ^cm2 or less.

Regarding the dose of SiH _x ions irradiated in the first step, the ratio to the dose of C ₂ H _y ions is set to a predetermined range. That is, if the ratio [Si/C] of the number of Si atoms to the number of C atoms to be implanted is less than 0.3, the effect of the present invention by implanting SiH _x ions cannot be fully obtained. Therefore, [Si/C] is set to 0.3 or more, and preferably 0.5 or more. To satisfy the above, the dose of SiH _x ions irradiated in the first step is preferably 2.50 × 10 ¹⁴ ions/cm ² or more. On the other hand, if [Si/C] exceeds 1.6, the passivation effect by hydrogen becomes insufficient. Therefore, [Si/C] is set to 1.6 or less, and preferably 1.5 or less. To satisfy the above, the dose of SiH _x ions irradiated in the first step is preferably 4.00 × 10 ¹⁴ ions/cm ² or less.

The total dose can be grasped as an apparatus setting value. However, since the dose of C ₂ H _y ions and the dose of SiH _x ions cannot be grasped individually, they are calculated as follows. That is, for a silicon wafer after irradiating with cluster ions, a carbon concentration profile in the depth direction from the surface of the silicon wafer is measured by SIMS measurement, and the amount of carbon implanted into the modified layer is calculated from the carbon concentration profile. Since the carbon number of _{C 2} H _y ions is 2, the value obtained by dividing the implanted carbon amount calculated above by 2 can be considered as the "dose of C ₂ H _y ions." Furthermore, the "dose of SiH _x ions" can be calculated by subtracting the dose of C ₂ H _y ions calculated as above from the total dose. The ratio of the dose of _{C2Hy ions} to the dose of _SiHx ions in the total dose can be controlled by the resolution of a mass separator that selects ions, the set value of the mass number of the ions to be implanted, the amount of raw material gas introduced, the energy of electrons irradiated during ionization, etc. in the ion implantation apparatus.

The acceleration voltage of the cluster ions, along with the ion species, affects the peak position of the depth-direction concentration profile of the constituent elements in the modified layer. In this embodiment, the acceleration voltage of the cluster ions can be set to more than 0 keV/ion and less than 200 keV/ion, preferably 100 keV/ion or less, and more preferably 80 keV/ion or less. Two methods are commonly used to adjust the acceleration voltage: (1) electrostatic acceleration and (2) radio-frequency acceleration. The former method involves arranging multiple electrodes at equal intervals and applying equal voltages between them to create a uniform axial acceleration field. The latter method involves the linear linac method, in which ions are accelerated using radio frequency while traveling in a straight line.

The beam current value of the cluster ions is not particularly limited, but can be appropriately determined, for example, from the range of 50 to 5000 μA. The beam current value of the cluster ions can be adjusted, for example, by changing the decomposition conditions of the raw material gas in the ion source.

[Second step]
The silicon epitaxial layer 16 can be formed under typical conditions. For example, hydrogen is used as a carrier gas, and a source gas such as dichlorosilane or trichlorosilane is introduced into a chamber. The growth temperature varies depending on the source gas used, but epitaxial growth can be performed on the modified layer 14 of the silicon wafer 10 by CVD at a temperature generally in the range of 1000 to 1200°C. The silicon epitaxial layer 16 preferably has a thickness in the range of 1 to 15 μm. A thickness of less than 1 μm may change the resistivity of the silicon epitaxial layer 16 due to out-diffusion of dopants from the silicon wafer 10, while a thickness of more than 15 μm may affect the spectral sensitivity characteristics of the CIS.

The manufacturing method of this embodiment described above ensures gettering capability while suppressing the diffusion of carbon into the epitaxial layer during epitaxial growth and the device formation process, making it possible to produce epitaxial silicon wafers with a high hydrogen passivation effect.

After the first step and before the second step, the silicon wafer 10 may be subjected to a recovery heat treatment to restore crystallinity. In this case, the recovery heat treatment may involve holding the silicon wafer 10 in an atmosphere of, for example, nitrogen gas or argon gas at a temperature of 900°C to 1100°C for 10 minutes to 60 minutes. The recovery heat treatment may also be performed using a rapid temperature increase/decrease heat treatment device, such as RTA (Rapid Thermal Annealing) or RTO (Rapid Thermal Oxidation), separate from the epitaxial device.

(epitaxial silicon wafer)
Referring to FIG. 1 , an epitaxial silicon wafer 100 according to one embodiment of the present invention is obtained by the above-described manufacturing method, and comprises a silicon wafer 10, a modified layer 14 formed in a surface layer portion of the silicon wafer 10 and containing a solid solution of at least one of carbon and hydrogen, and a silicon epitaxial layer 16 formed on the modified layer 14.

[Defect evaluation using cross-sectional TEM images]
In the epitaxial silicon wafer 100, defect evaluation using a cross-sectional TEM image of the modified layer reveals that first defect regions in which microdefects composed of carbon aggregates having a size (diameter) of 3 nm to ¹⁰ nm exist at a density of 1.00× ¹⁰ /cm to 8.00× ¹⁰ /cm, and ^second defect regions in which EOR defects having a maximum width of 50 nm to 250 nm exist at a density of 2.00× ¹⁰ /cm ^to less than ^5.00 × ¹⁰ /cm. In this specification, the term "cross-sectional TEM image" refers to an image obtained by cleaving the epitaxial silicon wafer 100 in the thickness direction and observing the cleavage cross section of the modified layer using a TEM.

In the modified layer of the epitaxial silicon wafer 100, a first defect region is observed, in which microdefects consisting of carbon aggregates with a size (diameter) of 3 nm to 10 nm are present. This is because black dot-like microdefects are formed by the implanted carbon. If the microdefect density is less than 1.00 × 10 ¹⁶ /cm ² , the passivation effect by hydrogen is not sufficiently obtained. Therefore, the density of microdefects consisting of carbon aggregates is set to 1.00 × 10 ¹⁶ /cm ² or more, and preferably 1.50 × 10 ¹⁶ /cm ² or more. On the other hand, if the microdefect density exceeds 8.00 × 10 ¹⁶ , since the microdefects are composed of carbon aggregates, the concentration of carbon diffusing from the defects during device heat treatment increases, affecting the electrical characteristics of the device. Therefore, the density of microdefects consisting of carbon aggregates is set to 8.00 × 10 ¹⁶ /cm ² or less, and preferably 7.60 × 10 ¹⁶ /cm ² or less.

Furthermore, second defect regions, in which EOR defects with a maximum width of 50 nm to 250 nm are present, are observed in the modified layer of the epitaxial silicon wafer 100. It is presumed that the EOR defects are defects resulting from the implantation of _SiHx ions. That is, the presence of the second defect regions allows for sufficient gettering capability to be exhibited despite the small amount of implanted carbon. In the present invention, the term "EOR defects" refers to a general term for defects in the form of {111} stacking faults, dislocation loops, {311} defects, and the like, which are formed when atoms (in the present invention, silicon atoms in the silicon wafer) pushed out of the crystal lattice by the ion-implanted elements aggregate at a position deeper than the implantation range (the peak position of the carbon concentration profile measured by SIMS) due to heat treatment. The "maximum width" of an EOR defect refers to the maximum width of each EOR defect in a TEM image.

If the density of EOR defects is less than 2.00×10 ⁷ /cm ² , the passivation effect by hydrogen cannot be sufficiently obtained. Therefore, the density of EOR defects is set to 2.00×10 ⁷ /cm ² or more, and preferably 2.50×10 ⁷ /cm ² or more. On the other hand, if the density of EOR defects is 5.00×10 ⁷ /cm ² or more, this causes epitaxial defects during epitaxial growth, making it impossible to maintain the crystalline perfection of the epitaxial layer. Therefore, the density of EOR defects is set to less than 5.00×10 ⁷ /cm ² , and preferably 4.50×10 ⁷ /cm ² or less.

In the present invention, the "microdefect density" and "EOR defect density" are determined as follows. First, microdefects consisting of carbon aggregates are confirmed at the same position as the carbon concentration peak position detected by SIMS measurement. Furthermore, as is clear from FIGS. 4(b) and (c), EOR defects are confirmed at a position slightly deeper than the position where microdefects occur densely. Therefore, a TEM evaluation sample is cut out from around the depth position of the carbon concentration peak observed in SIMS measurement so as to include the region where microdefects and EOR defects occur, and this evaluation sample is observed with a TEM. Then, as shown in FIGS. 4(b) and (c), a density calculation area (i.e., a defect area) is set to a vertical length (depth) of 300 nm so as to include the microdefects and EOR defects. The number of microdefects observed in the area is counted, and the number of defects is divided by the area of the area to obtain the microdefect density (/ ^cm2 ). Furthermore, the number of EOR defects with a maximum width of 50 to 250 nm observed in the area is counted, and the number of defects is divided by the area of the area to determine the EOR defect density (/cm ² ). In the examples of Figures 4(a) to 4(c), the density calculation area is a region of 300 nm length x 200 nm width, but the width is not particularly limited.

[SIMS profile]
In the epitaxial silicon wafer 100, if the amount of carbon distributed in the silicon epitaxial layer 16 and the modified layer 14 is 2.00×10 ¹⁴ atoms/cm ² or less, the passivation effect by hydrogen is not sufficiently obtained. Therefore, the amount of carbon distributed in the silicon epitaxial layer 16 and the modified layer 14 is more than 2.00×10 ¹⁴ atoms/cm ² , and preferably 2.50×10 ¹⁴ atoms/cm ² or more. On the other hand, if the amount of carbon distributed in the silicon epitaxial layer 16 and the modified layer 14 is more than 6.00×10 ¹⁴ atoms/cm ² , carbon diffuses into the epitaxial layer during epitaxial growth and the device formation process. Therefore, the amount of carbon distributed in the silicon epitaxial layer 16 and the modified layer 14 is set to not more than 6.00× ¹⁰ atoms/cm ² , and preferably not more than 5.50× ¹⁰ atoms/cm ^2. In the present invention, this "amount of carbon" can be determined by measuring a carbon concentration profile in the depth direction from the surface of the silicon epitaxial layer by SIMS measurement for the epitaxial silicon wafer, and integrating the profile from the surface of the epitaxial layer to the end of the modified layer (the position in the silicon wafer where the carbon concentration profile becomes flat).

In this embodiment, as in the example of the invention shown in Figure 3(a), the SIMS carbon concentration profile in the depth direction of the silicon epitaxial layer and the modified layer has a gentle first peak that exists across the silicon epitaxial layer and the modified layer, and a steep second peak that branches off from this first peak and exists at a position near the interface of the modified layer with the epitaxial layer. In this embodiment, even though the amount of injected carbon is small, the carbon concentration profile has such a steep second peak. This allows sufficient gettering ability to be exhibited. The peak concentration of the steep second peak in the carbon concentration profile is preferably 5.00 x ¹⁰ atoms/cm ^or more, and preferably 2.00 x ¹⁰ atoms/cm ^or less.

In this embodiment, as shown in the example of FIG. 3B , a SIMS hydrogen concentration profile in the depth direction of the modified layer shows a peak in the modified layer (at a position near the interface with the epitaxial layer) with a peak concentration of 5.00×10 ¹⁶ atoms/cm ³ or more. If the hydrogen peak concentration is 5.00×10 ¹⁶ atoms/cm ³ or more, hydrogen remaining in the modified layer can be sufficiently diffused into the epitaxial layer during heat treatment in the device formation process for forming a semiconductor device on the epitaxial layer, thereby passivating defects in the epitaxial layer. Therefore, the hydrogen peak concentration is set to 5.00×10 ¹⁶ atoms/cm ³ or more, and preferably 8.00×10 ¹⁶ atoms/cm ³ or more. In contrast, in this embodiment, the hydrogen peak concentration is generally 1.00×10 ¹⁸ atoms/cm ³ or less.

(Method for manufacturing semiconductor devices)
A method for manufacturing a semiconductor device according to one embodiment of the present invention includes the steps of the method for manufacturing the epitaxial silicon wafer 100 described above, and a step of forming a semiconductor device in the silicon epitaxial layer 16. Furthermore, a method for manufacturing a semiconductor device according to another embodiment of the present invention includes a step of forming a semiconductor device in the silicon epitaxial layer 16 of the epitaxial silicon wafer 100. These manufacturing methods can suppress the formation of point defects caused by carbon in the device formation region of the epitaxial layer while ensuring gettering capability, and can obtain a sufficient passivation effect by hydrogen.

The semiconductor devices formed on the silicon epitaxial layer 16 are not particularly limited, and examples include MOSFETs, DRAMs, power transistors, and back-illuminated solid-state imaging devices.

[Silicon wafer preparation]
An n-type silicon wafer (diameter: 300 mm, thickness: 775 μm, dopant type: phosphorus, resistivity: 10 Ω·cm) obtained from a CZ single crystal silicon ingot was prepared.

[Cluster ion irradiation]
Nine experiments (Nos. 1 to 9) were carried out under different conditions for cluster ion irradiation, as shown in Table 2. Furthermore, as a comparative example, an epitaxial silicon wafer (No. 10) was produced in which epitaxial growth was carried out without implanting cluster ions.

In Nos. 1 to 8, diethylsilane (SiC ₄ H ₁₂ ) was used as the source gas. FIG. 2 shows the mass spectrum of diethylsilane. Table 1 lists the ion species corresponding to mass numbers 26 to 31 in the mass spectrum shown in FIG. 2. The peak at mass number 31 corresponds to SiH ₃ ions. The lower peak at mass number 30 corresponds to SiH ₂ ions. The highest peak at mass number 29 corresponds to SiH ions and C ₂ H ₅ ions. The peaks at mass numbers 28, 27, and 26 correspond to C ₂ H ₄ ions, C ₂ H ₃ ions, and C ₂ H ₂ ions, respectively. In this example, a cluster ion generator (CLARIS (registered trademark), manufactured by Nissin Ion Equipment Co., Ltd.) was used to extract ion species with mass numbers in the range of 29 to 31 from the various ion species corresponding to the mass spectrum shown in FIG. 2 to obtain cluster ions, and a beam of these cluster ions was irradiated onto the surface of a silicon wafer at an acceleration voltage of 80 keV/ion and a beam current value of 800 μA. The cluster ions mainly contained SiH ₃ ions as SiH _x ions, with trace amounts of SiH ₂ ions and SiH ions, and further contained C ₂ H ₅ ions as C ₂ H _y ions. Since the cluster ion generator can set the total dose of all ion species, different levels of total dose were set as the device setting value in each example, as shown in Table 2.

In No. 9, only cluster ions of C ₂ H ₃ extracted from cyclohexane (C ₆ H ₁₂ ) were injected as the source gas, as shown in Table 2.

For each example after irradiation with cluster ions, the amount of implanted carbon, the dose of _{C2Hy ions} , and the dose of _SiHx ions were determined by the methods described above. Table 2 shows the cluster ion irradiation conditions for each example.

[Epitaxial growth]
Next, the silicon wafer after the cluster ion irradiation was transferred into a single-wafer epitaxial growth apparatus (manufactured by Applied Materials, Inc.), where it was subjected to a hydrogen bake treatment at a temperature of 1120°C for 30 seconds, and then a silicon epitaxial layer (thickness: 5 μm, dopant type: phosphorus, resistivity: 10 Ω cm) was epitaxially grown on the surface of the silicon wafer on which the modified layer had been formed by a CVD method at 1120°C using hydrogen as a carrier gas and trichlorosilane as a source gas, thereby obtaining an epitaxial silicon wafer.

[SIMS analysis]
For Nos. 1 to 9, the carbon and hydrogen concentration profiles in the depth direction from the surface of the silicon epitaxial layer were measured by SIMS. As a representative example, the carbon concentration profile for No. 5 is shown in Figure 3(a) and the hydrogen concentration profile in Figure 3(b). In Figure 3(a), the carbon concentration profile was higher than the background in the range of approximately 3.5 to 7.0 μm from the surface of the silicon epitaxial layer. On the other hand, in Figure 3(b), the hydrogen concentration profile was higher than the background in the range of approximately 4.7 to 4.9 μm from the surface of the silicon epitaxial layer. Therefore, in No. 5, the thickness of the epitaxial layer formed on the silicon wafer was approximately 4.7 μm, and the range of approximately 4.7 to 7.0 μm in depth from the surface of the silicon epitaxial layer was identified as a modified layer formed in the surface layer portion of the silicon wafer.

The amount of carbon after epitaxial growth for each example was determined by integrating the carbon concentration profile from the surface of the epitaxial layer to the end of the modified layer (the point where the carbon concentration profile becomes flat), and the results are shown in Table 2. As is clear from Figure 3(a), the carbon concentration profile for No. 5 shows a gentle first peak across the silicon epitaxial layer and modified layer. Furthermore, a steep second peak branched off from this gentle first peak and appeared near the interface of the modified layer with the epitaxial layer. This type of second peak was also observed in Nos. 3 to 9.

3(b), a peak with a peak concentration of 1.0× ¹⁰ atoms/ ^cm or more appeared near the interface between the modified layer and the epitaxial layer in the hydrogen concentration profile of No. 5. Such a peak was also observed in Nos. 3 to 9. The peak concentration of this peak is shown in Table 2 as the "hydrogen peak concentration."

[Cross-sectional TEM observation]
The cross section of the modified layer (near the interface with the epitaxial layer) of each epitaxial silicon wafer was observed using a TEM. Figure 4 shows TEM images (magnification: 200,000 times) obtained for (a) No. 9, (b) No. 6, and (c) No. 4. In the TEM image of No. 9, black dot-like microdefects with a diameter of approximately 5 nm were observed. These are carbon aggregates and are known to be caused by the implantation of _{C2Hy ions} . Meanwhile, in Nos. 4 and 6, in addition to the microdefects, stacking faults (EOR defects) with a maximum width of 50 to 250 nm were observed. These are presumed to be defects caused by the implantation of _SiHx ions. For each example, the microdefect density and EOR defect density were determined using the methods described above. Table 2 shows the microdefect density and EOR defect density for each example.

[Evaluation of passivation effect]
To evaluate the passivation effect of hydrogen, an oxide film (25 nm) was formed on the epitaxial silicon wafer of each example at 800°C for 4 hours, followed by an additional heat treatment at 700°C for 30 minutes. The samples after the additional heat treatment were measured by the QSCV method, and the interface state density D _it was calculated from the obtained capacitance value.

The method for calculating the interface state density using the QSCV method will be explained below. Figure 5 shows the equivalent circuits of the MOS structure during (a) low-frequency CV measurement and (b) high-frequency CV measurement. C _ox is the gate oxide film capacitance, C _it is the interface state capacitance, and C _s is the capacitance on the silicon substrate side, which represents the sum of the depletion layer capacitance and the inversion layer capacitance. From this equivalent circuit, the minimum capacitance C _LF of the MOS capacitor during low-frequency CV measurement is expressed by the following equation (1). Furthermore, the interface state density _Dit is expressed by the following equation (2).

In equation (2), C _ox C _LF /(C _ox - _{C LF} ) is the substrate-side capacitance obtained by measurement. The substrate-side capacitance _Cs of an ideal MOS structure can be calculated from a theoretical formula or from high-frequency CV measurement. In this experiment, _Cs was calculated using the results of high-frequency CV measurement.

Figure 6 shows the results of _Dit calculations for Nos. 3, 4, and 6. From the _Dit calculation results, the _Dit value at the midgap for each example was determined. Table 2 shows the _Dit value at the midgap for each example. Comparing the Dit _values at the midgap for each example, it can be seen that the Dit _values of the invention examples are lower than those of the comparative examples. This shows that implanting _cluster ions containing _SiHx ions and _C2Hy ions provides a higher passivation effect than conventional hydrogen molecule ion-implanted epiwafers.

[Considerations based on evaluation results]
Considering the evaluation results, it is presumed that the following phenomenon occurred in the inventive example. Microdefects were formed due to the implantation of _{C2Hy ions} , and EOR defects were further formed due to the implantation of _SiHx ions. Since the density of microdefects decreases with a decrease in the carbon implantation dose, it is believed that limiting the carbon implantation dose also reduces the hydrogen trapping concentration due to microdefects. However, since microdefects consisting of carbon aggregates are also thought to have a high hydrogen trapping ability (adsorption reaction), it is believed that the low microdefect density reduces the amount of hydrogen retrapped in the hydrogen adsorption/desorption reaction. Furthermore, simultaneous implantation of _SiHx ions increases the point defect density of vacancies and interstitial silicon in the mixed molecular ion implantation region, increasing the hydrogen diffusion rate. Therefore, it is believed that the decrease in hydrogen retrapping rate and the increase in hydrogen diffusion rate increase the amount of hydrogen reaching the _SiO2 /Si interface, leading to an improvement in the passivation effect by hydrogen.

The present invention provides an epitaxial silicon wafer with a high hydrogen passivation effect and a method for manufacturing the same.

100 epitaxial silicon wafer 10 silicon wafer 10A silicon wafer surface 12 cluster ions 12A SiH _x ions 12B C ₂ H _y ions 14 modified layer 16 silicon epitaxial layer

Claims (8)

a first step of irradiating a surface of a silicon wafer with a beam of cluster ions containing SiH _x (x is one or more selected from integers of 1 to 3) ions and C ₂ H _y (y is one or more selected from integers of 2 to 5) ions to form a modified layer in a surface layer portion of the silicon wafer, in which constituent elements of the cluster ions are solid-dissolved;
a second step of forming a silicon epitaxial layer on the modified layer of the silicon wafer;
and
the total dose of the cluster ions irradiated in the first step is 6.00×10 ¹³ ions/cm ² or more and 1.00×10 ¹⁵ ions/cm ² or less;
the dose of the C ₂ H _ions irradiated in the first step is more than 1.00×10 ¹⁴ ions/cm ² and not more than 3.00×10 ¹⁴ ions/cm ² ;
a ratio of the number of Si atoms to the number of C atoms implanted in the first step, [Si/C], of 0.3 to 1.6; 2. The method for producing an epitaxial silicon wafer according to claim 1, wherein a dose of the C ₂ H _ions irradiated in the first step is 1.25×10 ¹⁴ ions/cm ² or more. A silicon wafer;
a modified layer formed on a surface layer portion of the silicon wafer, the modified layer containing a solid solution of at least one of carbon and hydrogen;
a silicon epitaxial layer formed on the modified layer;
and
In a defect evaluation of the modified layer using a cross-sectional TEM image, first defect regions were observed in the modified layer, in which microdefects composed of carbon aggregates having a size of 3 nm or more and ¹⁰ nm or less existed at a density of 1.00× ¹⁰ /cm ^{or more} and 8.00× ¹⁰ /cm or less, and second defect regions were observed in which EOR defects having a maximum width of 50 nm or more and 250 nm or ^less existed at a density of 2.00× ¹⁰ /cm or more and less than 5.00× ¹⁰ /cm2,
the amount of carbon distributed in the silicon epitaxial layer and the modified layer is greater than 2.00×10 ¹⁴ atoms/cm ² and not more than 6.00×10 ¹⁴ atoms/cm ² ;
An epitaxial silicon wafer, wherein a peak hydrogen concentration in a depth direction of the modified layer measured by SIMS is 5.00×10 ¹⁶ atoms/cm ³ or more and 1.00×10 ¹⁸ atoms/cm ³ or less. 4. The epitaxial silicon wafer according to claim 3, wherein, in defect evaluation using a cross-sectional TEM image of the modified layer, the EOR defects are present in the ^second defect region at a density of 2.50× ¹⁰ /cm or ^more and 4.50× ¹⁰ /cm or less. 5. The epitaxial silicon wafer according to claim 3, wherein the amount of carbon distributed in the silicon epitaxial layer and the modified layer is 2.50×10 ¹⁴ atoms/cm ² or more. The method for producing an epitaxial silicon wafer according to claim 1 or 2;
forming a semiconductor device in the silicon epitaxial layer of the epitaxial silicon wafer;
A method for manufacturing a semiconductor device comprising: A method for manufacturing a semiconductor device, comprising forming a semiconductor device in the silicon epitaxial layer of the epitaxial silicon wafer according to claim 3 or 4. A method for manufacturing a semiconductor device, comprising forming a semiconductor device in the silicon epitaxial layer of the epitaxial silicon wafer described in claim 5.