JP7250216B2 - Method for measuring carbon concentration in silicon substrate - Google Patents

Description

The present disclosure relates to a method for measuring carbon concentration in silicon substrates.

A small amount of impurities are mixed into the silicon substrate when the crystal is grown. For example, it is known that when a silicon substrate is used for manufacturing a power device, the device characteristics change depending on the amount of impurities contained in the silicon substrate in the lifetime control process. In particular, among impurities contained in a silicon substrate, carbon attracts attention, and reduction of carbon contamination during growth of silicon crystals is being studied.

As a method for measuring the carbon concentration in a silicon substrate, FT-IR (Fourier Transform Infrared Spectroscopy) is commonly used because of the simplicity of measurement. However, in FT-IR, the lower limit of measurement of carbon concentration in a silicon substrate is about 2×10 ¹⁵ atoms/cm ³ , and it cannot be applied to measurement of carbon concentration below that.

Therefore, in order to detect low-concentration carbon, attempts have been made to activate carbon (convert to complex defects) using electron beam irradiation or chemical treatment and detect the trap levels of the complex defects. For example, in Patent Document 1 (Japanese Patent Application Laid-Open No. 2015-222801), electron beam irradiation is used to activate carbon in a silicon substrate, and a PL (Photoluminescence) method is used to detect carbon (complex defects). .

JP 2015-222801 A

The measurement method described in Patent Document 1 requires a calibration curve and a reference sample to measure the trace carbon concentration. However, even if the concentration of carbon or other impurities (for example, oxygen concentration) changes when making a calibration curve or reference sample, the efficiency of converting carbon into complex defects by the activation treatment is constant. assumptions and calibration curves may not be exact.

Accordingly, an object of the present disclosure is to provide a carbon concentration measuring method capable of accurately measuring a minute amount of carbon concentration in a silicon substrate.

A method for measuring the carbon concentration of a silicon substrate according to the present disclosure includes:
A step of subjecting each of a p-type region having a Schottky junction and an n-type region having a Schottky junction formed on one surface of a silicon substrate to at least one of electron beam irradiation, charged particle beam irradiation and ion implantation. and,
measuring the concentration of the first trap level in the p-type region and the concentration of the second trap level in the n-type region using a transient capacitance method;
quantifying the carbon concentration using the measured concentration of the first trap level and the concentration of the second trap level;
including.

In the conventional measurement method using a calibration curve to measure the carbon concentration, the calibration curve may be inaccurate due to the influence of oxygen impurities, but the measurement method of the present disclosure does not use a calibration curve. , it is possible to accurately measure trace amounts of carbon concentration in silicon substrates.

4 is a flow chart showing each step of carbon concentration measurement according to Embodiment 1. FIG. It is a cross-sectional schematic diagram of the silicon substrate used as a measuring object. 3 is a graph showing the relationship between the dose of electron beams and the concentration of trap level Ev: +0.37 eV. 2 is a graph showing the relationship between the dose of electron beams and the concentration of trap level Ec: -0.17 eV. Fig. 10 is a graph showing the relationship between the DLTS signal intensity of Ec: -0.17 eV and the filling pulse width;

Embodiments of the present disclosure will be described below. In addition, in the drawings of the present disclosure, the same reference numerals represent the same or corresponding parts. Also, dimensional relationships such as length, width, thickness, and depth are appropriately changed for clarity and simplification of the drawings, and do not represent actual dimensional relationships.

Embodiment 1.
The method for measuring the carbon concentration of the silicon substrate according to the present embodiment includes
At least one of electron beam irradiation, charged particle beam irradiation, and ion implantation is applied to each of the "p-type region having a Schottky junction" and the "n-type region having a Schottky junction" formed on one surface of a silicon substrate. (first step: activation of carbon);
measuring the concentration of the first trap level in the p-type region and the concentration of the second trap level in the n-type region using the transient capacitance method (second step: concentration measurement of the trap level);
a step of quantifying the carbon concentration using the measured first trap level concentration and second trap level concentration (third step: quantifying the carbon concentration);
including.

Specific procedures of the method for measuring the carbon concentration of the silicon substrate according to the present embodiment will be described below with reference to the drawings.

The object to be measured is a silicon substrate. A silicon substrate is a semiconductor substrate containing silicon as a main component. Here, containing silicon as a main component means that the content of silicon in the semiconductor substrate is more than 50% by mass. The content of silicon in the semiconductor substrate is preferably 90% by mass or more, more preferably 95% by mass or more, and still more preferably 98% by mass or more. In this embodiment, as an example, a silicon substrate grown by the FZ (Floating-Zone) method and having n-type conductivity will be described.

Referring to FIG. 1, the method for measuring the carbon concentration of a silicon substrate according to the present embodiment includes at least the first to third steps described above. First, the steps before the first step will be described.

In this embodiment, a case where silicon substrate 1 is an n-type semiconductor substrate will be described.

(Formation of p-type region and n-type region)
Referring to FIG. 2, when silicon substrate 1 is an n-type semiconductor substrate, p-type region 3 is formed on one surface (first surface 2) of silicon substrate 1 . Formation of the p-type region can be performed by ion implantation and activation annealing. Examples of p-type dopants include boron. These elements (dopants) are ionized, accelerated and implanted into the silicon substrate. Furthermore, in order to activate the implanted elements (move them to substitution positions), annealing is performed at 600 to 1000° C. using, for example, a batch furnace or RTA (rapid thermal anneal).

When the silicon substrate is p-type, an n-type region is formed on one surface (first surface 2) of silicon substrate 1. As shown in FIG. Examples of n-type dopants include phosphorus.

Alternatively, a p-type region (n-type region in the case of a p-type silicon substrate) may be formed by a diffusion method without using ion implantation. For example, a p-type region may be formed by placing a silicon substrate in a temperature-controlled quartz tube and annealing at 800 to 1200° C. while flowing a mixed gas containing the necessary dopants.

The method of forming the p-type region (n-type region) mentioned here is just an example, and methods used in general semiconductor manufacturing processes can be used without particular limitation. However, since the carrier concentration affects the lower detection limit of the trap level in the transient capacitance method described later, the carrier concentration of the formed p-type region is low within the range where the conductivity type of the silicon substrate is inverted. is preferred.

It should be noted that the step of forming the p-type region (n-type region) need not be performed for a silicon substrate that already has an n-type region and a p-type region.

(Formation of Schottky junction)
Next, a Schottky junction is formed in each of p-type region 3 and n-type region (region other than p-type region 3) of first surface 2 .

Specifically, first, in order to form an electrode on the first surface 2 of the n-type silicon substrate 1, patterning and formation of a metal film are performed. As patterning, a mask having openings above each of the p-type region and the n-type region (regions other than the p-type region) is formed by resist coating, photolithography, or the like. In this state, metal films (the first electrode 4 and the second electrode 5) are formed in the openings of the mask, thereby forming a Schottky junction between the p-type region 3 and the first electrode 4 to form an n-type electrode. An ohmic contact is formed between the region and the second electrode 5 . Note that the resist is removed after the first electrode 4 and the second electrode 5 are formed.

As a method for forming the metal film, a vapor deposition method or a CVD (Chemical vapor deposition) method used in general semiconductor manufacturing processes can be applied. As the metal to be used, a metal with a small work function such as In is suitable. Metals are also applicable.

Next, another electrode is formed on the first surface 2 of the n-type silicon substrate 1 . Patterning and formation of metal films (third electrode 6 and fourth electrode 7) are performed again in the same manner as the first electrode 4 and the second electrode 5 . Thereby, an ohmic junction is formed between the p-type region 3 and the third electrode 6 and a Schottky junction is formed between the n-type region and the fourth electrode 7 .

As the metal used here, a metal with a large work function such as Au or Pt is suitable. Other metals are also applicable.

In the first step, a Schottky junction is formed in the p-type region and an ohmic junction is formed in the n-type region (or an ohmic junction is formed in the p-type region and a Schottky junction is formed in the n-type region). ) sometimes used the same metal. Here, the same metal is used to reduce the number of steps, but there is no problem even if different metals are used.

Also, although the ohmic junction of the n-type region is formed on the first surface 2 of the silicon substrate 1, forming an ohmic electrode on the second surface opposite to the first surface 2 poses no problem.

Further, when the carrier concentrations of the n-type region and the p-type region are high, it is possible to obtain conduction between the silicon substrate and the metal member in contact with the surface of the silicon substrate, and the transient capacitance method, which will be described later, can be measured. It is also possible to omit the formation of the ohmic junction, since there is no effect on However, the formation of a Schottky junction is necessary for the measurement of the transient capacitance method.

(First step: activation of carbon)
In the first step, each of the p-type and n-type Schottky structures formed on one surface of the silicon substrate is subjected to at least one of electron beam irradiation, charged particle beam irradiation and ion implantation. This activates carbon in both the n-type and p-type regions of the silicon substrate.

The electron beam irradiation damages the silicon crystals forming the silicon substrate, generating interstitial silicon and silicon atomic vacancies (hereinafter simply referred to as "vacancies"). In a silicon crystal in a room temperature environment, interstitial silicon diffuses easily and reaches sites where impurity atoms, other interstitial silicon, vacancies, etc. exist in the silicon crystal. At this time, the interstitial silicon reacts with impurity atoms, other interstitial silicon, and the like to form some complex defects (in the case of vacancies, they are pair-annihilated).

For example, interstitial carbon is produced by diffusion of interstitial silicon to carbon sites that occupy lattice sites. Further, it is known that interstitial carbon diffuses into interstitial oxygen sites to form complex defects called interstitial carbon-interstitial oxygen pairs (C _i O _i ). It is also known that interstitial carbon diffuses to lattice-position carbon sites to form complex defects called interstitial carbon-lattice-position carbon pairs (C _i C _s ). Here, "C" means carbon and "O" means oxygen. "i" means interstitial site and "s" means substitutional site.

Since the electron beam easily penetrates the silicon substrate, the acceleration voltage of the electron beam is not limited.
It is preferable to select the irradiation amount of the electron beam in consideration of the carbon concentration of the silicon substrate in order to activate all the carbon.

For example, in experiments by the inventors, an FZ-grown n-type silicon substrate having a carbon concentration of about 1×10 ¹⁵ atoms/cm ³ was irradiated with an electron beam having an energy of 750 keV from 2.3×10 ¹⁴ e/cm ² to 9 When the trap levels of C _i O _i and C _i C _s were evaluated by irradiation in the range of 3×10 ¹⁴ e/cm ² , it was confirmed that the trap level concentrations did not depend on the electron beam dose. bottom.

The results are shown in FIG. 3 (C _i C _s trap level) and FIG. 4 (C _i O _i trap level). The silicon substrate A is a silicon substrate having a high oxygen concentration, specifically, the oxygen concentration is on the order of 1×10 ¹⁷ (atoms/cm ³ ). The silicon substrate B is a silicon substrate having a low oxygen concentration, specifically, the oxygen concentration is on the order of 1×10 ¹⁵ (atoms/cm ³ ). In addition, in FIGS. 3 and 4, "arbitrary unit" means the trap level concentration estimated by the PL emission intensity.

From the results shown in FIGS. 3 and 4, in the case of a silicon substrate with a carbon concentration of about 1×10 ¹⁵ atoms/cm ³ , when the electron beam dose is 2.3×10 ¹⁴ e/cm ² or more, all It can be seen that the carbon of is activated and converted to complex defects.

3 and 4 show data for two types of silicon substrates with different oxygen concentrations. Silicon substrate A, which has a high oxygen concentration, has a large amount of C _i O _i (small amount of C _{i Cs} ) and oxygen Silicon substrate B with a low concentration has less C _i O _i (more C _i C _s ).

In this way, electron beam irradiation activates carbon in the silicon _substrate to either _{C i} O _i or C _{i C s .} Carbon concentration in the substrate can be measured. However, basically, activation of all carbon is a precondition for carbon concentration measurement. The wire irradiation dose ratio is preferably 4 atoms/e·cm or more.

In addition, the formation of C _i O _i and C _i C _s proceeds spontaneously even when the silicon substrate is left at room temperature after electron beam irradiation. Annealing the substrate is preferred. The annealing temperature is preferably 100° C. or less in consideration of the decomposition _{of CiCs} , which is weak against heat.

Instead of electron beam irradiation, charged particle beam irradiation, ion implantation, or the like can also be used to activate carbon in the silicon substrate and convert it to C _i C _s or C _i O _i .

In the case of ion implantation, if the implanted element and carbon form complex defects, the above-mentioned measurement accuracy of the carbon concentration is lowered. Therefore, it is preferable to implant ions of an inert element that hardly reacts with carbon.

In addition, when heavy ions are implanted, in addition to activating carbon, clusters of interstitial silicon and vacancies are formed, reducing the measurement accuracy of the trap level. For this reason, for example, the charged particle beam is preferably α rays, and the ion implantation is preferably Si ion implantation with a small dose.

(Second step: trap level concentration measurement)
In the second step, the concentration of the first trap level in the p-type region and the concentration of the second trap level in the n-type region are measured for the above sample (silicon substrate for measurement) using the transient capacitance method. Note that the Schottky junctions of the n-type region and the p-type region are used for measuring the concentration of the trap level.

The concentration of the first trap level of defects in the p-type region 3 is measured.
Specifically, first, the capacitances of the first electrode 4 and the third electrode 6 are measured. For example, probes are brought into contact with the first electrode 4 and the third electrode 6 . In this state, the transient capacitance method is used to modulate the width of the first depletion layer 8 extending in the p-type layer to measure the trap level. At this time, a C _i O _i trap level peak is obtained near Ev: +0.37 eV. Here, E _v means the energy at the top of the valence band of silicon. The trap level obtained in this manner is defined as a first trap level.

As the transient capacitance method, for example, a DLTS (Deep Level Transient Spectroscopy) method, an ICTS (Isothermal Capacitance Transient Spectroscopy) method (isothermal transient capacitance method), or the like can be suitably used. Hereinafter, DLTS will be used for explanation.

For the measurement of the DLTS method, for example, the bias voltage and filling pulse voltage are 1 to 10 V, the filling pulse width is 100 ms to 10 seconds, the rate window is 100 ms to 1 second, and the temperature sweep range is 50 K to 300 K. Therefore, known techniques can be applied without any particular limitation. Also, as the analysis method for converting the DLTS signal into the trap concentration, a known technique such as Lang's analysis method can be applied without particular limitation.

Next, the concentration of the second trap level of defects in the n-type region is measured.
Specifically, first, the capacitances of the second electrode 5 and the fourth electrode 7 are measured. For example, probes are brought into contact with the second electrode 5 and the fourth electrode 7 . In this state, the DLTS method is used to modulate the width of the second depletion layer 9 extending in the n-type layer to measure the trap level.

At this time, a signal obtained by superimposing the signal of the vacancy-oxygen pair (VO) and the signal of C _i C _s is obtained near E _c : -0.17 eV. Note that Ec means the energy at the bottom of the conduction band of silicon. Moreover, "V" means Vacancy.

E _c : In order to remove the VO signal from the peak near -0.17 eV, measurement is performed twice using the DLTS method with different application times of the pulse voltage.

FIG. 5 shows the results of measuring the DLTS signal intensity at Ec: -0.17 eV by changing the pulse width in the rate window and at a constant temperature. When ^the pulse width is lengthened, the trap level of the response increases and ^the signal intensity becomes stronger. In addition to VO, C _i C _s with a small capture cross-section also responded in the interval longer than 10 ⁻⁴ seconds. That is, when the pulse width is 1×10 ⁻⁵ seconds or more and less than 1×10 ⁻⁴ seconds, the signal of C _i C _s is excluded and only the signal of VO is detected. Next, when the pulse width is increased to 1×10 ⁻⁴ sec or more, the signal of E _c : −0.17 eV includes both VO and C _i C _s signals.

In this ^embodiment , ^C Measure the concentration of _{the i} C _s trap level. The trap level obtained in this manner is used as a second trap level.

Compared to the measurement of the first trap level, the work added to the measurement of the second trap level is only the adjustment of the pulse width and the difference of the data, as described above, and the other measurement conditions are as described above. Measurement conditions can be applied to the first trap level.

In the second step, the p-type region of the silicon substrate was used to measure the concentration of the C _i O _i trap level (Ev: +0.37 eV). It is also possible to apply the photoexcitation DLTS method in which the capacitance of 7 is measured while being excited by a light pulse. In this case, there is no need to form a Schottky junction in the p-type region, which has the advantage of simplifying the process. However, since there is no guarantee that the occupancy of the trap level during optical pulse excitation is 1, there is a risk that the measurement accuracy of the second trap level will be lower than in the above-described procedure of measurement using the p-type region. be.

(Third step: quantification of carbon concentration)
In the third step, the carbon concentration is quantified using the measured first trap level concentration and second trap level concentration.

Specifically, the concentration of the first trap level (Ev: +0.37 eV) and the concentration of the second trap level (Ec: -0.17 eV) obtained in the second step are substituted into Equation 1 below. Then, the carbon concentration ([carbon]) contained in the silicon substrate is measured.

[Carbon]=[First trap level]+2×[Second trap level] (Formula 1)
In Equation 1, the coefficient of the concentration of the first trap level is 1 because one defect contains one carbon in the concentration of C _i O _i . The coefficient of the concentration of the second trap level is 2 because two carbon atoms are contained in one defect at the concentration of C _i C _s .

Here, the lower measurement limit of the above measuring method is determined by the lower measurement limit of the trap concentration of the transient capacitance method such as the DLTS method. Generally, the lower measurement limit of the trap concentration of the transient capacitance method is 10 ⁻⁴ to 10 ⁻³ times the carrier concentration of the object to be measured. For example, for an n-type silicon substrate with a resistivity of 50 Ωcm (corresponding to a carrier concentration of about 2.6×10 ¹⁴ /cm ³ when the dopant is phosphorus), 2.6×10 ¹⁰ atoms/cm ³ to 2.6×10 10 atoms/cm 3 . It is also possible to measure carbon concentrations of .6×10 ¹¹ atoms/cm ³ .

For the silicon substrate to be measured, the carbon concentration measured by the above measuring method is less than 2×10 ¹⁵ atoms/cm ³ . According to the above measuring method, it is possible to measure the carbon concentration of a silicon substrate containing a low concentration of carbon, which is difficult to measure by the FT-IR method.

In the present embodiment, the n-type silicon substrate was described as an example, but the carbon concentration measuring method of the present disclosure can also be applied to a p-type silicon substrate.

When the object to be measured is a p-type silicon substrate, an n-type region is formed on the surface of the p-type silicon substrate by ion-implanting group V ions such as phosphorus or arsenic, and an appropriate metal material is used to form the n-type region. By forming a Schottky junction in the p-type region and the n-type region on the first surface, forming an ohmic electrode on the second surface, and measuring the trap level by the above-described transient capacitance method, the p-type silicon substrate The carbon concentration of can be measured.

In addition, the silicon substrate to be evaluated by the carbon concentration measuring method of the present embodiment is not limited to those grown by the FZ method, for example, liquid phase grown silicon substrates such as the CZ method (Czochralski), It can also be applied to silicon substrates laminated by vapor phase growth such as epitaxial growth.

The size, thickness, shape, etc. of the silicon substrate to be measured are not particularly limited. For example, it is possible to measure the carbon concentration of a silicon substrate that is not a wafer-shaped silicon substrate.

Further, in the present embodiment, all carbon is activated, and the concentration of trap levels of complex defects generated thereby is quantitatively measured. For this reason, calibration of measured values on a silicon substrate with a known carbon concentration and measurement procedures using a calibration curve are not required, and the work is simple. Furthermore, it is possible to accurately measure the trace amount of carbon concentration in the silicon substrate to be measured without being affected by impurities (oxygen and dopants) other than carbon in the silicon substrate.

It should be considered that the embodiments disclosed this time are illustrative in all respects and not restrictive. The scope of the present disclosure is indicated by the scope of claims rather than the above description, and is intended to include all changes within the meaning and scope of equivalence to the scope of claims.

1 n-type silicon substrate, 2 first surface, 3 p-type region, 4 first electrode, 5 second electrode, 6 third electrode, 7 fourth electrode, 8 first depletion layer, 9 second depletion layer.

Claims (8)

A step of subjecting each of a p-type region having a Schottky junction and an n-type region having a Schottky junction formed on one surface of a silicon substrate to at least one of electron beam irradiation, charged particle beam irradiation and ion implantation. and,
measuring the concentration of the first trap level in the p-type region and the concentration of the second trap level in the n-type region using a transient capacitance method;
quantifying the carbon concentration using the measured concentration of the first trap level and the concentration of the second trap level;
A method for measuring carbon concentration in a silicon substrate, comprising: 2. The measuring method according to claim 1, wherein in the electron beam irradiation, the ratio of the electron beam irradiation amount to the carbon concentration is 4 atoms/e·cm or more. 3. The measuring method according to claim 1, wherein said charged particle beam is an alpha ray. The measuring method according to any one of claims 1 to 3, wherein said ion implantation is silicon ion implantation. The measuring method according to any one of claims 1 to 4, wherein the first trap level is Ev: +0.37 eV. The measuring method according to any one of claims 1 to 5, wherein the second trap level is Ec: -0.17 eV. The measuring method according to any one of claims 1 to 6, wherein the carbon concentration is calculated by the following formula.
Carbon concentration = (concentration of first trap level) + 2 x (concentration of second trap level) The measuring method according to any one of claims 1 to 7, wherein the transient capacity method is the DLTS method or the ICTS method.

Images