JP3008901B2 - Impurity distribution simulation method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for simulating impurity distribution in consideration of point defects in ion implantation in a semiconductor manufacturing process.

[0002]

2. Description of the Related Art Conventionally, in a simulation of impurity distribution in a semiconductor, a simple one-dimensional diffusion equation of an impurity is solved. However, it is expanded to a two-dimensional one so that a complicated model can be handled. It has become to.

In particular, a model for explaining the accelerated diffusion of an impurity after the oxidation step or ion implantation (a phenomenon in which the diffusion is much larger than a normal diffusion) has been actively studied.
These phenomena involve point defects inside the semiconductor, and it has become mainstream to treat diffusion equations in consideration of the interaction between these point defects and impurities.

Usually, the distribution of point defects generated by ion implantation is calculated by the Monte Carlo method. However, since the Monte Carlo method takes a long calculation time, an analytical model that can reproduce the results of the Monte Carlo method is created,
The distribution of point defects is determined by using this.

It is also believed that when the ion-implanted impurity stops, the substrate atoms are pushed out of the lattice position to generate ε interstitial defects. Therefore, the following Equation 1
Is used to determine the concentration of interstitial defects.

[0006]

(Equation 1)

Here, I is the concentration of interstitial defects, F is the concentration of interstitial defects based on the result of the Monte Carlo method, C is the concentration of impurities, and ε is a parameter. Ε is set to 1.

When performing an impurity distribution simulation, an impurity concentration distribution after thermal diffusion is obtained by solving a predetermined diffusion equation with the concentration of the interstitial defect obtained by Equation 1 as an initial value.

[0009]

By the way, when boron is to be diffused as an impurity in a semiconductor manufacturing process,
In ion implantation, there are a case where boron ions (B ⁺ ) are implanted and a case where boron fluoride ions (BF ₂ ⁺ ) are implanted. However, in the thermal process such as the diffusion in the simulation, boron in the substrate is implanted in the form of boron ions (B ⁺ ) or boron fluoride ions (BF ₂
⁺ ) It cannot be distinguished whether it was typed in.

For this reason, the calculation result of the diffusion of boron after implanting boron ions (B ⁺ ) and the calculation result of the diffusion of boron after implantation of boron fluoride ions (BF ₂ ⁺ ) are the same. There is a problem that a fitting failure occurs with respect to an experimental value obtained by experimenting with parameters.

For example, when the same diffusion parameters are used, the calculation results of diffusion when boron ions (B ⁺ ) are implanted and the calculation results of diffusion when boron fluoride ions (BF ₂ ⁺ ) are implanted. And only one SI
It will not be fitted with experimental values such as MS.

[0012]

SUMMARY OF THE INVENTION The present invention is a method for simulating the distribution of impurities made to solve such a problem. That is, the present invention provides a method of simulating impurity distribution in a diffusion step using a model for analytically calculating the distribution of point defects generated by ion implantation into a substrate and a diffusion model in which the interaction between the point defects and impurities is considered. When calculating the distribution of interstitial defects generated by ion implantation, when calculating by the sum of the concentration of interstitial defects generated by ion implantation and the value obtained by multiplying the impurity concentration by a predetermined coefficient, the target ion This is a method of weighting the coefficient between the case where is implanted with atomic ions and the case where is implanted with molecular ions.

As described above, by weighting the coefficient multiplied by the impurity concentration in the case where the target ions are implanted with atomic ions and the case where the target ions are implanted with molecular ions, when the atomic ions are implanted and stopped in the substrate, The difference between the number of atoms pushed out and the number of substrate ions pushed out when the molecular ions are driven into the substrate and stopped can be adjusted. Good fitting with the experimental value can be performed both in the case of driving.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the impurity distribution simulation method of the present invention will be described below with reference to the drawings. 1 to 3 are diagrams for explaining the impurity distribution simulation method according to the present embodiment. This impurity distribution simulation method uses a model that analytically calculates the distribution of point defects generated by ion implantation into a substrate such as silicon, and a diffusion model that considers the interaction between the point defects and the impurities in the diffusion process. This method calculates the distribution of impurities. The sum of the concentration of the interstitial result generated by ion implantation and the value obtained by multiplying the impurity concentration by a predetermined coefficient is used as the distribution of interstitial defects generated by ion implantation at the beginning of the simulation. Value.

In particular, in the present embodiment, the distribution of interstitial defects set as the initial value is weighted by a coefficient to be multiplied by the impurity concentration between the case where target ions are implanted with atomic ions and the case where molecular ions are implanted. There is a characteristic in the point of going.

That is, it is considered that when the impurities ion-implanted into the substrate stop, the substrate atoms are pushed out of the lattice position to generate ε interstitial defects. The concentration of is determined.

In this embodiment, the parameter ε multiplied by the impurity concentration C in Equation 1 is weighted depending on whether ions are implanted with atomic ions or molecular ions.

For example, when boron (B) is used as an impurity, there are a case where boron ions (B ⁺ ) are implanted into the substrate and a case where boron fluoride ions (BF ₂ ⁺ ) are implanted into the substrate. Parameters for determining the concentration of interstitial defects when ions (BF ₂ ⁺ ) are implanted:
ε is a parameter that determines the concentration of interstitial defects when boron ions (B ⁺ ) are implanted;
Weight twice.

That is, boron fluoride ion (BF ₂ ⁺ )
When boron is implanted, it immediately decomposes into boron (B) and fluorine (F) as soon as it enters the substrate. Therefore, it is considered that when boron (B) and fluorine (F) stop, each ε substrate atoms are pushed out. . Therefore, if the parameter for determining the concentration of the interstitial defect: ε is set to be about three times that in the case where boron ions (B ⁺ ) are implanted, even if a simulation is performed using the same diffusion parameter, an experimental value such as SIMS is obtained. And good fitting can be realized.

As described above, the case where boron ions (B ⁺ ) are implanted into the substrate and the case where boron ions (BF ⁺ )
_{In the} case where ²⁺ ) is implanted, after weighting a parameter ε for determining the concentration of interstitial defects, the impurity distribution is calculated by a predetermined diffusion equation.

Here, the diffusion equation applied in the present embodiment will be described. Inside a semiconductor (for example, silicon), there are two types of defects: interstitial, which has extra semiconductor atoms at positions other than the lattice position of the semiconductor, and point defects, which are vacancy, in which semiconductor atoms do not exist at lattice positions where they should exist. There are types.

Although these point defects exist even in a thermal equilibrium state (interstitial concentration at that time is I ^* and vacancy concentration is V ^* ), they also occur inside the semiconductor due to oxidation, ion implantation and the like. Oxidation occurs at the reaction site when the semiconductor surface changes to an oxide. In ion implantation, when implanted ions collide with the lattice inside the semiconductor and give a certain energy or more, the semiconductor atoms are repelled from the lattice to form an interstitial-vacancy pair.

It is considered that impurities inside the semiconductor are diffused by forming pairs with these point defects (impurities alone do not diffuse). Therefore, oxidation processes with more point defects than normal thermal equilibrium conditions, and much more impurity-point defect pairs formed after ion implantation than normal, resulting in diffusion much faster than normal. become.

The activated impurities are represented by A, interstitial
Where I is V and vacancy is V, it is considered that a reaction as shown in Equation 2 occurs inside the semiconductor.

[0025]

(Equation 2)

In Equation 2, M is an impurity-inters
Indicates a titial pair, and N indicates an impurity-vacancy pair. The first and second reactions are reactions in which impurities and point defects form a pair (and vice versa). The third and fourth reactions are reactions of a point defect and an impurity-point defect pair, in which the impurity returns to the lattice position. The fifth reaction is a point defect annihilation reaction.

In a general model, diffusion equations are considered for five of I, V, M, N, and A in consideration of all these reactions. Alternatively, assuming that the formation reaction of the impurity-point defect pair is sufficiently faster than the diffusion, M and N can be eliminated from the diffusion equation, so that only three diffusion equations need to be considered. Equation 3 shows these three diffusion equations.

[0028]

(Equation 3)

In Equation 3, C is the (chemical) concentration of the impurity, I is the concentration of the isolated interstitial, and V is the isolated va
The concentration of cancy, M is the concentration of the impurity-interstitial pair, and N is the concentration of the impurity-vacancy pair.

The first equation in Equation 3 is expressed as
terstitial, or isolated interstitial and impurities-
This is the diffusion equation of the sum of interstitial pairs. The second equation is the total vacancy, that is, the isolated vacancy.
And the diffusion equation of the sum of the impurity and the vacancy pair. Furthermore, the third equation shows an impurity diffusion equation.

Then, J _I , J _V ,
J _M and J _N are isolated interstitial, isolated vacancy, impurity-interstitial pair, and impurity-va, respectively.
It shows the flow of cancy pairs, and is generally defined by the equation shown in Equation 4.

[0032]

(Equation 4)

Here, X indicates any of I, V, M, and N, S _X indicates the electrical state of X, and φ indicates the potential inside the substrate. G indicates generation / disappearance by the reaction represented by Formula 2. Further, R _loop is an interstiti
It shows absorption of al.

In the present embodiment, three diffusion equations shown in Equation 3 are used to determine whether boron ions (B ⁺ ) are implanted into the substrate or boron fluoride ions (BF ₂ ⁺ ) are implanted into the substrate. The parameter ε for determining the concentration of interstitial defects due to ion implantation, which is an initial value, is weighted.

That is, FIG. 1 shows an energy of 15 ke.
V, impurity distribution simulation results and SIMS measurement results when boron ions (B ⁺ ) are ion-implanted into a silicon substrate and then annealed at a temperature of 850 ° C. for 81 minutes under ion implantation conditions of a dose of 5 × 1e13. I have.
Note that, in the example shown in FIG.
0.3.

FIG. 2 shows an impurity distribution simulation result when boron ions (B ⁺ ) are ion-implanted into a silicon substrate under the same ion implantation conditions as in the example shown in FIG. The measurement results are shown. In the example shown in FIG. 2, the parameter shown in Expression 1 is set to ε = 1.0.

FIG. 3 shows an energy of 30 keV,
FIG. 11 shows the results of an impurity distribution simulation and SIMS measurement in the case where boron fluoride ions (BF ₂ ⁺ ) are ion-implanted into a silicon substrate under the ion implantation conditions of a dose of 1 × 1e13 and then annealed at 850 ° C. for 40 minutes. ing. In the example shown in FIG. 3, the parameter shown in Expression 1: ε =
1.0 is set.

As shown in FIG. 3, when boron fluoride ion (BF ₂ ⁺ ) is ion-implanted and thermally diffused, the parameter shown by the equation 1 is set to ε = 1.0, which is good for the SIMS measurement result. (Depth 0-0.
About 15 μm).

Further, as shown in FIG. 2, when boron ions (B ⁺⁾ is ion-implanted is thermally diffused into the silicon substrate, the parameters shown in the case as well as the number 1 of the boron fluoride ions (BF ₂ ⁺⁾ : If ε = 1.0, a defect occurs in fitting with the SIMS measurement result.

[0040] Therefore, as shown in FIG. 1, boron ions when (B ⁺⁾ to be ion-implanted is thermally diffused into the silicon substrate, the parameters shown in Equation 1: a epsilon, boron fluoride ions (BF ₂ ⁺⁾ In the case of (3), by setting 0.3, which is about 1/3 of ε = 1.0, a good fit with the SIMS measurement result can be obtained.

As described above, by weighting parameters for determining the distribution of interstitial defects in the case where target ions are implanted with atomic ions and the case where molecular ions are implanted, diffusion using the same diffusion parameter is performed. Even if the impurity distribution is calculated by the equation, both will well fit the experimental values.

In this embodiment, an example in which boron ions (B ⁺ ) are used as atomic ions to be ion-implanted and boron fluoride ions (BF ₂ ⁺ ) are used as molecular ions has been described. It is not limited, and the same applies to the case where other atomic ions and molecular ions are implanted.
In this case, the parameters may be weighted according to the number of atoms of the atomic ions and the molecular ions.

[0043]

As described above, according to the impurity distribution simulation method of the present invention, the following effects can be obtained. That is, even when the same diffusion parameter is used, it is possible to satisfactorily perform the fitting with the actually measured values both in the case of implanting atomic ions and in the case of implanting molecular ions. This makes it possible to perform simulation with high accuracy even in a case where a process of implanting atomic ions using the same ions as impurities and a process of implanting molecular ions coexist.

[Brief description of the drawings]

FIG. 1 is a diagram (part 1) for explaining the present embodiment;

FIG. 2 is a diagram (part 2) for explaining the embodiment;

FIG. 3 is a diagram (part 3) for explaining the embodiment;

Claims (4)

(57) [Claims] 1. A method of simulating an impurity distribution in a diffusion step using a model for analytically calculating a distribution of point defects generated by ion implantation into a substrate and a diffusion model in which an interaction between the point defects and impurities is considered. When calculating the distribution of interstitial defects generated by the ion implantation, when calculating by the sum of the concentration of the interstitial defects generated by the ion implantation and the value obtained by multiplying the concentration of the impurity by a predetermined coefficient, A method of simulating impurity distribution, wherein the coefficient is weighted when ions are implanted with atomic ions and when ions are implanted with molecular ions. 2. The impurity distribution simulation method according to claim 1, wherein the weighting of the coefficient is set according to a difference in the number of atoms between the atomic ions and the molecular ions. 3. The impurity distribution simulation method according to claim 1, wherein said atomic ions are boron ions, and said molecular ions are boron fluoride ions. 4. The impurity distribution simulation method according to claim 3, wherein the weighting of the coefficient is set in accordance with a difference in the number of atoms between the boron ions and the boron fluoride ions.