CN113882016A

CN113882016A - Method for manufacturing nitrogen-doped P-type monocrystalline silicon

Info

Publication number: CN113882016A
Application number: CN202111153124.XA
Authority: CN
Inventors: 徐鹏; 兰洵
Original assignee: Xian Eswin Material Technology Co Ltd
Current assignee: Xian Eswin Material Technology Co Ltd
Priority date: 2021-09-29
Filing date: 2021-09-29
Publication date: 2022-01-04
Also published as: TWI825959B; WO2023051347A1; TW202300717A

Abstract

The embodiment of the invention discloses a method for manufacturing nitrogen-doped P-type monocrystalline silicon, which comprises the following steps: putting the high-doped nitrogen single crystal and polycrystalline silicon into a crucible and heating a crystal pulling furnace to obtain a first silicon melt; adding a highly doped boron single crystal into the first silicon melt to obtain a second silicon melt; and pulling nitrogen-doped P-type single crystal silicon in the second silicon melt by Czochralski method. According to the manufacturing method of the nitrogen-doped P-type monocrystalline silicon, nitrogen and boron are completely melted into silicon melt at a lower furnace chamber temperature, and the technical problems of high furnace chamber temperature and long heating time are solved. Meanwhile, the melting order of the highly doped boron single crystal and the highly doped nitrogen single crystal is changed to suppress the occurrence of a situation in which the formation of Boron Nitride (BN) causes a crystal pulling failure.

Description

Method for manufacturing nitrogen-doped P-type monocrystalline silicon

Technical Field

The invention relates to the field of semiconductor silicon wafer production, in particular to a method for manufacturing nitrogen-doped P-type monocrystalline silicon.

Background

With the global development of informatization, the size of devices in the field of silicon chip application is continuously reduced, and meanwhile, under the condition that the integration level of the devices is gradually improved, the application field of power devices is wider and wider, the main characteristics of the power devices are high voltage resistance, and the resistance of a substrate has great influence on the performance of the devices, so that the substrate is required to have high resistivity and small resistivity change. Because the crystal lattices of the doping elements are not matched with the crystal lattices of the silicon elements, a segregation phenomenon exists in the monocrystalline silicon growing process, namely the concentration of the doping elements crystallized in the monocrystalline silicon ingot is less than that in a melt (raw material), so that the concentration of the doping elements in the crucible is continuously increased, the concentration of the doping elements in the monocrystalline silicon ingot is also continuously increased, the concentration of the doping elements in the monocrystalline silicon ingot is rapidly increased at the last stage of the monocrystalline silicon ingot growing, the carrier density is obviously increased, the resistivity of the ingot is rapidly reduced, and the axial resistivity change gradient of the ingot is larger. In recent years, as electronic components, especially power devices, require a gradient of a resistivity change rate to be smaller, it is important to reduce the change of the resistivity by changing a doping method. In addition, since the presence of heavy metal impurities on the silicon wafer causes poor device characteristics, it is also remarkable to reduce the heavy metal impurities. As one of the techniques for reducing heavy metal impurities, the importance of gettering techniques is increasing, and therefore it is advantageous to use a P-type silicon wafer having a high gettering effect and a low resistivity. Meanwhile, under the condition that the integration level of a device is gradually improved, the stress borne by the silicon wafer can be obviously improved, the setting of processing and manufacturing process parameters can be influenced due to the low mechanical strength of the silicon material, and the silicon wafer is extremely easy to damage and break in the product assembling process, so that the production cost of the silicon wafer is increased, and the improvement of the mechanical strength of the silicon wafer has important significance.

Silicon wafers for producing semiconductor electronic components such as the integrated circuits are mainly produced by slicing a single crystal silicon rod drawn by a Czochralski (Czochralski) method. The czochralski method comprises melting polycrystalline silicon in a crucible made of quartz to obtain a silicon melt, immersing a single crystal seed crystal in the silicon melt, and continuously lifting the seed crystal to move away from the silicon melt surface, thereby growing a single crystal silicon rod at a phase interface during the movement.

In the production of a low-resistivity P-type silicon wafer having a high gettering effect by the above-mentioned czochralski method, boron is generally selected as a dopant for controlling resistivity, and the mechanical strength is enhanced by removing the Void defect by nitrogen doping. However, in the process of boron and nitrogen doping, boron content of the prepared low-resistivity P-type silicon wafer is low because boron as a boron dopant is easy to volatilize and forms refractory Boron Nitride (BN) with nitrogen, or dislocation is caused to cause crystal pulling failure. Meanwhile, the melting temperature of the boron dopant and the nitrogen dopant is high, so that the quartz crucible is softened, oxygen precipitation is uncontrollable, and the quality of the prepared P-type silicon wafer with low gettering effect and low resistivity is influenced.

Disclosure of Invention

In order to solve the above-mentioned problems, embodiments of the present invention are directed to a method for manufacturing a nitrogen-doped P-type single crystal silicon, which can overcome the problem of boron volatility and avoid boron nitride generation, thereby manufacturing a high-quality nitrogen-doped P-type single crystal silicon with high gettering capability at a high yield and a low cost.

The technical scheme of the invention is realized in such a way, and comprises the following steps:

putting the high-doped nitrogen single crystal and polycrystalline silicon into a crucible and heating a crystal pulling furnace to obtain a first silicon melt;

adding a highly doped boron single crystal into the first silicon melt to obtain a second silicon melt;

and pulling nitrogen-doped P-type single crystal silicon in the second silicon melt by Czochralski method.

Preferably, the charging of the highly doped nitrogen single crystal and the polycrystalline silicon into the crucible and heating the crystal pulling furnace to obtain the first silicon melt comprises:

after quantitative calculation, the highly doped nitrogen single crystal and the polycrystalline silicon are put into a crucible together;

vacuumizing the crystal pulling furnace and introducing protective gas;

and opening the heater, raising the temperature of the furnace chamber and preserving the heat for a period of time until the highly doped nitrogen single crystal and the polycrystalline silicon are completely melted to obtain the first silicon melt.

Preferably, the first silicon melt is a low nitrogen-doped silicon melt.

Preferably, the protective gas is argon.

Preferably, the adding of the highly doped boron single crystal to the first silicon melt to obtain the second silicon melt comprises:

allowing the first silicon melt to stand for a period of time;

adding the highly doped boron single crystal to the first silicon melt after standing;

and increasing the temperature of the furnace chamber to enable the highly doped boron single crystal to be completely melted to the first silicon melt, so as to obtain the second silicon melt.

Preferably, the second silicon melt is a low boron-doped nitrogen silicon melt.

Preferably, pulling nitrogen-doped P-type single crystal silicon in the second silicon melt by czochralski method comprises: adjusting the temperature of the furnace chamber; slowly placing seed crystals into the second silicon melt, and completing the growth process of the crystals through seeding, necking, shouldering, isodiametric growth and ending; and preparing the nitrogen-doped P-type monocrystalline silicon.

Preferably, the highly doped nitrogen single crystal and the highly doped boron single crystal are particles or powders.

Drawings

FIG. 1 is a schematic diagram of a method for manufacturing nitrogen-doped P-type single crystal silicon according to an embodiment of the present invention;

FIG. 2 is a schematic view of a process for producing nitrogen-doped P-type single crystal silicon in which highly doped nitrogen single crystal and polycrystalline silicon are charged into a crucible and a crystal pulling furnace is heated to obtain a first silicon melt in accordance with an embodiment of the present invention;

fig. 3 is a schematic view of a second silicon melt obtained by adding a highly doped boron single crystal to the first silicon melt in a method for manufacturing nitrogen-doped P-type single crystal silicon according to an embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below.

The manufacturing method of the nitrogen-doped P-type monocrystalline silicon is based on a Czochralski method and is matched with a solid phase doping mode to prepare the nitrogen-doped P-type monocrystalline silicon. That is, polycrystalline silicon and a dopant are placed together in a crucible and heated to form a melt, and the melt is pulled to grow single crystal silicon. In the invention, a high-doped nitrogen monocrystal and polycrystalline silicon are put into a crucible and heated in a crystal pulling furnace to obtain a first silicon melt; adding a highly doped boron single crystal into the first silicon melt to obtain a second silicon melt; and pulling nitrogen-doped P-type single crystal silicon in the second silicon melt by Czochralski method.

In order to reduce the technical problem of resistivity variation by changing the doping manner, the conductivity of the single crystal can be significantly changed by doping two or more trace elements as impurities in the prior art. The doped impurities are mainly trivalent or pentavalent elements, and can be generally divided into N-type monocrystalline silicon wafers and P-type monocrystalline silicon wafers according to the difference in conductivity caused by different dopants. When a small amount of impurities are doped into a single crystal silicon wafer, a periodic potential field near impurity atoms is disturbed and forms an additional bound state, and an additional impurity level is generated in a forbidden band. Impurities that can provide electron carriers are called Donor (Donor) impurities and the corresponding energy levels, called Donor levels, are located above the forbidden band near the conduction band bottom. For example, when impurity atoms such as phosphorus, arsenic, antimony, etc. which are pentavalent elements are doped into silicon crystals of tetravalent elements, the impurity atoms serve as one molecule of a crystal lattice, four of the pentavalent electrons form covalent bonds with surrounding silicon atoms, and the extra one electron is bound near the impurity atom to generate a shallow hydrogen-like level-donor level. The energy required for the electrons on the donor level to jump to the conduction band is much less than the energy required for the electrons to be excited from the valence band to the conduction band, and the electrons are easily excited to the conduction band to become electron carriers, so that for the monocrystalline silicon wafer doped with donor impurities, the conduction carriers are mainly the electrons excited into the conduction band, and belong to the electron conduction type, and the monocrystalline silicon wafer is called an N-type monocrystalline silicon wafer. Since an intrinsically excited electron-hole pair always exists in a single crystal silicon wafer, electrons are majority carriers and holes are minority carriers in an N-type single crystal silicon wafer.

Accordingly, an impurity that can provide a hole carrier is called an Acceptor (Acceptor) impurity, and the corresponding energy level is called an Acceptor level, and is located below the forbidden band near the top of the valence band. For example, when a trace of impurity atoms such as boron, aluminum, gallium and the like, which are trivalent elements, are doped into a silicon crystal, the impurity atoms and four surrounding silicon atoms form covalent bonds and lack one electron, so that a vacancy exists, and the energy state corresponding to the vacancy is an acceptor level. Because the acceptor level is close to the top of the valence band, electrons in the valence band are easily excited to the acceptor level to fill the vacancy, so that the acceptor impurity atoms become negative charge centers. At the same time, a vacancy is left in the valence band due to ionization of an electron, and a free hole carrier is formed. Therefore, holes are the majority carriers, and the impurity monocrystalline silicon wafer is mainly based on hole conduction, namely, the hole conduction type, and is called p-type monocrystalline silicon wafer. In a P-type single crystal silicon wafer, holes are majority carriers and electrons are minority carriers. Among the various effects of semiconductor devices, minority carriers often play an important role. The method is characterized in that the more impurities are doped, the higher the concentration of the majority (holes) is, and the stronger the conductivity is.

In practical use, boron is often used as a resistivity-controlling dopant. Therefore, it is very advantageous to provide a P-type silicon wafer having a region without crystal Defects (DZ) extending from the front side, which refers to the surface of the P-type silicon wafer where electronic components are to be formed, into the body and a region containing Bulk Micro Defects (BMD) adjacent to the DZ and further extending into the body. The DZ is important because in order to form an electronic component on a P-type silicon wafer, it is required that no crystal defect exists in the formation region of the electronic component, otherwise, a circuit break or other faults occur, and the formation of the electronic component in the DZ can avoid the influence of the crystal defect; the BMD has an Intrinsic Gettering (IG) effect on metal impurities, so that the metal impurities in the P-type silicon wafer are kept away from the DZ, thereby preventing adverse effects such as an increase in leakage current and a decrease in film quality of a gate oxide film due to the metal impurities.

However, in the process of manufacturing the above-described P-type silicon wafer having the BMD region, it is very advantageous to dope the P-type silicon wafer with nitrogen. When the P-type silicon wafer is doped with nitrogen, the formation of BMDs having nitrogen as a core can be promoted, thereby making the BMDs reach a certain density, making the BMDs effectively function as a metal gettering source, and also making the density distribution of the BMDs favorably influenced, for example, by making the distribution of the BMD density in the radial direction of the P-type silicon wafer more uniform, for example, by making the BMD density higher in a region near DZ and gradually lower toward the inside of the silicon wafer. The nitrogen impurity doped in the P-type silicon wafer prepared by the CZ method can also reduce the size of Voids, so that the Voids can be eliminated by annealing at high temperature, and the nitrogen can also improve dislocation and improve the mechanical property of the P-type silicon wafer. In addition, nitrogen impurities are doped in the P-type silicon wafer drawn by the CZ method, nitrogen and oxygen can interact with each other to form a nitrogen-oxygen complex with shallow thermal donor performance, and the electrical performance of the P-type silicon wafer is further improved.

In the preparation of nitrogen-doped P-type single crystal silicon, in the prior art, a nitrogen dopant, a boron dopant and a polycrystalline silicon raw material are generally added into a quartz crucible at the same time for heating, the raw materials are all melted to obtain a silicon melt doped with nitrogen and boron, and then, the nitrogen-doped P-type single crystal silicon is pulled by a CZ method. In the preparation process, the dopant is added by using a solid material in order to accurately control the addition concentration of the dopant, such as Si3N4 raw material and boron as the dopant. In the process of preparing nitrogen-doped P-type single-crystal silicon, polycrystalline silicon, Si3N4 raw material, and elemental boron in a quartz crucible need to be heated to a molten state to obtain a silicon melt doped with nitrogen and boron. Polycrystalline silicon can be completely melted at 1450 ℃, however, Si3N4 raw material and boron single crystal respectively need to be kept at 1900 ℃ and 2100 ℃ for a period of time to be completely melted, and incompletely melted dopants can cause dislocation defects in the crystal bar, so that the crystal pulling fails.

However, keeping the furnace chamber temperature at a high temperature for a long time causes softening of the quartz crucible, and the reaction product of the molten Si and the crucible raw material SiO2 is gaseous SiO, and reacts with the graphite product in which the crucible is placed after escaping to form CO gas, which easily enters the silicon melt, and introduces carbon and oxygen into the silicon, so that the oxygen content in the silicon melt increases. Excessive oxygen atoms are the main cause of defect formation, and oxygen precipitation in the active region of the device can cause breakdown or leakage, and during annealing, the generation of oxygen precipitates lowers the yield of the device. Meanwhile, the volatilization amount of the boron simple substance as the dopant is extremely large, and particularly the boron dopant can be further volatilized in a high-temperature crystal pulling furnace. In addition, boron is easy to oxidize, boron and nitrogen are easy to form refractory Boron Nitride (BN) at the ambient temperature of 1600 ℃, the boron nitride has good high-temperature resistance and stable chemical property, is not reflected with silicon at high temperature, is not infiltrated with silicon and is not adhered with silicon, and when the boron nitride occurs in silicon melt, dislocation defects are easily caused in the crystal pulling process to cause crystal pulling failure.

The method mainly comprises the steps of replacing silicon nitride serving as a nitrogen source with a high-doped nitrogen single crystal and replacing elemental boron serving as a boron source with a high-doped boron single crystal in the process of preparing the nitrogen-doped P-type monocrystalline silicon, so that the technical problems of high furnace chamber temperature and long heating time are solved. In addition, changing the melting order of the boron dopant and the nitrogen dopant suppresses the formation of Boron Nitride (BN).

A method of manufacturing nitrogen-doped P-type single crystal silicon according to an embodiment of the present invention, as shown in fig. 1, includes the steps of:

Because the highly doped nitrogen single crystal dopant and the highly doped boron single crystal dopant are used, nitrogen and boron elements can be doped into the crystal bar at a lower temperature (the melting temperature of polycrystalline silicon), and a quartz crucible used for bearing crystal pulling materials is not required to be increased to a high temperature so as to uncontrollably separate out a large amount of oxygen. In addition, the adding sequence of the high-doped nitrogen single crystal dopant and the high-doped boron single crystal dopant is adjusted, a first silicon melt which is a low-doped nitrogen silicon melt is obtained through the high-doped nitrogen single crystal dopant and a polycrystalline silicon raw material, then the high-doped boron single crystal is added into the first silicon melt to obtain a second silicon melt which is a low-doped boron nitrogen silicon melt, and finally the nitrogen-doped P-type single crystal silicon is prepared in the second silicon melt through Czochralski method, so that the occurrence of Boron Nitride (BN) is avoided, and the situation of crystal pulling failure is prevented.

According to a method for manufacturing nitrogen-doped P-type single crystal silicon of one embodiment of the present invention, as shown in fig. 2, the steps of charging a highly doped nitrogen single crystal and polycrystalline silicon into a crucible and heating a crystal pulling furnace to obtain a first silicon melt include:

and quantitatively calculating the high-doped nitrogen single crystal, and putting the high-doped nitrogen single crystal and the polycrystalline silicon into a crucible. First, highly nitrogen-doped single crystal grains or powder and a polycrystalline silicon raw material are put into a quartz crucible in advance, and since the nitrogen content of the dopant added here directly determines the nitrogen content in the produced nitrogen-doped P-type single crystal silicon, the quality of the highly nitrogen-doped single crystal need only be determined in accordance with the nitrogen concentration of the target nitrogen-doped P-type single crystal silicon. Proper nitrogen doping amount can pin dislocation so as to enhance mechanical strength, inhibit gap generation, increase and decrease integrity (GOI) of a grid oxide layer of an integrated circuit device, and promote oxygen precipitation so as to enhance the internal gettering capability of single crystals.

And vacuumizing the crystal pulling furnace and introducing protective gas. In order to avoid oxidation of silicon during the crystal pulling process, it is necessary to carry out the process under the action of a vacuum environment and a protective gas. Protective gas (high-purity argon) is introduced into the crystal pulling furnace, the protective gas is injected from the top of the crystal pulling furnace, the vacuum pump is started at the bottom of the crystal pulling furnace to pump out the gas, so that the vacuum value in the furnace is kept in dynamic balance, gas flow in the furnace penetrates through a growth area of a crystal bar from top to bottom, and silicon oxide and impurity volatile matters produced at high temperature can be taken away in time. Preferably, the vacuum value in the furnace is kept stable by controlling the air inflow and maintaining the vacuum pumping efficiency of the vacuum pump.

And opening the heater, raising the temperature of the furnace chamber and preserving the heat for a period of time until the highly doped nitrogen single crystal and the polycrystalline silicon are completely melted to obtain the first silicon melt. In this process, the temperature of the furnace chamber is raised to the melting temperature of the polycrystalline silicon (about 1450-. At the temperature, the quartz crucible has no problems of softening, cracking and the like, and the precipitation of elements is within an acceptable range, so that the technical problem of excessive oxygen content is avoided. Because the highly doped nitrogen single crystal as the nitrogen source has high nitrogen content and low oxygen content, the dopant can be effectively prevented from carrying excessive oxygen for the first silicon melt.

According to the method for manufacturing nitrogen-doped P-type single-crystal silicon of one embodiment of the present invention, as shown in fig. 3, the step of adding highly doped boron single crystal to the first silicon melt to obtain the second silicon melt comprises the steps of:

the first silicon melt is allowed to stand for a period of time. After the fact that the high-doped nitrogen single crystal dopant and the polycrystalline silicon raw material in the first silicon melt are completely melted is confirmed, the heater is placed for a period of time after the power of the heater is reduced, so that the surface of the first silicon melt is solidified, the first silicon melt is prevented from splashing when other dopants are added subsequently, and safety accidents are avoided.

Adding the highly doped boron single crystal to the first silicon melt after standing. The highly doped boron single crystal particles or powder are added to the first silicon melt after being left to stand for a while, and since the boron content of the dopant added here directly determines the nitrogen content in the produced nitrogen-doped P-type single crystal silicon, it is only necessary to determine the quality of the highly doped nitrogen single crystal according to the nitrogen concentration of the target nitrogen-doped P-type single crystal silicon.

And increasing the temperature of the furnace chamber to enable the highly doped boron single crystal to be completely melted to the first silicon melt, so as to obtain the second silicon melt. In this step, the surface of the first silicon melt after standing may be solidified, and the temperature of the furnace stone is raised or maintained at the melting temperature of silicon (about 1450-. At the temperature, the quartz crucible has no problems of softening, cracking and the like, and the precipitation of elements is within an acceptable range, so that the technical problem of excessive oxygen content is avoided. Because the high-doped boron monocrystal used as the boron source has high boron content and low oxygen content, the dopant can be effectively prevented from removing redundant oxygen for the first silicon melt.

According to a nitrogen-doped P-type single-crystal silicon manufacturing method of an embodiment of the present invention, pulling nitrogen-doped P-type single-crystal silicon by czochralski method includes: adjusting the temperature of the furnace chamber; slowly putting down the seed crystal into the second silicon melt, and finishing the growth process of the crystal through seeding, necking, shouldering, equal-diameter growth and ending stages of the seed crystal in the furnace body; and preparing the nitrogen-doped P-type monocrystalline silicon.

According to the manufacturing method of the nitrogen-doped P-type monocrystalline silicon, the high-doped nitrogen monocrystalline and the high-doped boron monocrystalline are used as the doping agents for preparing the nitrogen-doped P-type monocrystalline silicon, so that nitrogen atoms and boron atoms can be completely melted into silicon melt at a lower furnace chamber temperature, redundant oxygen atoms cannot be separated out from a quartz crucible, the control of oxygen concentration is improved, and BMD control is improved. In addition, the generation of Boron Nitride (BN) in the process of preparing the nitrogen-doped P-type monocrystalline silicon is avoided through different melting sequences of the high-doped nitrogen monocrystalline and the high-doped boron monocrystalline, and the condition that crystal pulling fails due to impurities which are difficult to dissolve is prevented.

The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and all the changes or substitutions should be covered within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the appended claims.

Claims

1. A method for manufacturing nitrogen-doped P-type monocrystalline silicon is characterized by comprising the following steps:

2. The method of manufacturing nitrogen-doped P-type single crystal silicon according to claim 1, wherein the charging of the highly doped nitrogen single crystal and the polycrystalline silicon into the crucible and the heating of the crystal pulling furnace to obtain the first silicon melt comprises:

vacuumizing the crystal pulling furnace and introducing protective gas;

3. The method of manufacturing nitrogen-doped P-type single-crystal silicon according to claim 2, wherein the first silicon melt is a low-nitrogen-doped silicon melt.

4. The method of manufacturing nitrogen-doped P-type single crystal silicon of claim 2, wherein the protective gas is argon gas.

5. The method of manufacturing nitrogen-doped P-type single-crystal silicon according to claim 1, wherein the adding highly doped boron single crystal to the first silicon melt to obtain a second silicon melt comprises:

allowing the first silicon melt to stand for a period of time;

6. The method for manufacturing nitrogen-doped P-type single-crystal silicon according to claim 5, wherein the second silicon melt is a low boron-doped nitrogen silicon melt.

7. The method of manufacturing nitrogen-doped P-type single crystal silicon of claim 1, wherein pulling nitrogen-doped P-type single crystal silicon in the second silicon melt by czochralski method comprises: adjusting the temperature of the furnace chamber; slowly placing seed crystals into the second silicon melt, and completing the growth process of the crystals through seeding, necking, shouldering, isodiametric growth and ending; and preparing the nitrogen-doped P-type monocrystalline silicon.

8. The method for manufacturing nitrogen-doped P-type single-crystal silicon according to claim 1, wherein the highly doped nitrogen single crystal and the highly doped boron single crystal are particles or powders.