JP6496280B2 - Method for manufacturing diffusion wafer - Google Patents

Description

  The present invention relates to a diffusion wafer having a non-diffusion layer and a diffusion layer, and a manufacturing method for producing the diffusion wafer, which are used in manufacturing processes of semiconductor elements and semiconductor devices.

Conventionally, the manufacturing process of this type of semiconductor element is roughly divided into a wafer preparation process for forming a semiconductor wafer from a rod-shaped semiconductor single crystal and an element formation process for forming a semiconductor element on the main surface portion of the semiconductor wafer. . In the case of producing a semiconductor element from a rod-shaped silicon single crystal, first, in the wafer creating step, a silicon wafer sliced from the rod-shaped silicon single crystal is subjected to a diffusion treatment, and diffusion layers are formed on both main surfaces thereof. Next, the intermediate part of the silicon wafer after the diffusion treatment is cut and separated by slicing processing, and then the mirror polishing process is performed on each cut surface so that the thickness other than the diffusion layer becomes the target thickness, There is a method of manufacturing a semiconductor device that obtains two semiconductor wafer pieces having substantially the same thickness (see, for example, Patent Document 1).
Finally, a p-type layer and an n-type layer are formed on a part of the non-diffusion layer of the semiconductor wafer piece, which is cut and separated into pellets, thereby completing the element formation step.

Japanese Patent Laid-Open No. 54-041665

Incidentally, the wafer creation process and the element formation process are generally performed at different places. For example, a common semiconductor wafer (diffusion wafer) is created regardless of the type of semiconductor elements at a first factory where the wafer creation process is performed. This semiconductor wafer (diffusion wafer) is shipped to a second factory, and various kinds of semiconductor elements are efficiently manufactured in the second factory by performing various element forming processes.
In addition, by increasing the diameter of a silicon wafer, it becomes possible to make more semiconductor devices from one wafer, and the productivity is greatly increased.
Similarly, in the diffusion wafer, it has been studied to improve productivity by switching from the conventional small diameter size to the large diameter size.
However, since the diffusion wafer is a single crystal plate, it is very easy to chip. When the diameter of the diffusion wafer is increased while maintaining the same size as the conventional small diameter size, the mechanical strength further decreases. Therefore, there is a problem that it is broken by handling in the subsequent element formation process.
Therefore, in order to solve such problems, a non-diffusion having the same thickness as the conventional small-diameter size is performed by performing a diffusion process on the silicon wafer that is thicker and larger than the conventional small-diameter size in the wafer preparation process. It is conceivable to produce a diffusion wafer having a layer and a diffusion layer thicker than the conventional diffusion layer. If the diffusion wafer has a large diameter and is thick, it will not break even if it is handled in the subsequent element formation step, and finally, an excess portion of the diffusion layer may be scraped off to a target thickness. As a result, even for a diffusion wafer having a large diameter, the mechanical strength is improved and cracking in the element forming process can be prevented, and a semiconductor element can be formed in the same manner as a conventional small diameter size. It is possible to prevent a decrease in sex.
However, in this case, since the additional thickness of the diffusion layer thickened by increasing the mechanical strength is scraped off, the impurity concentration distribution in the diffusion layer of this large-diameter diffusion wafer is the diffusion of the conventional small-diameter diffusion wafer. This is different from the impurity concentration distribution in the layer.
Specifically, the impurity concentration distribution at the time of performing the diffusion process in the wafer preparation process is the same for the diffusion layer of the thick and large-diameter diffusion wafer and the diffusion layer of the conventional small-diameter diffusion wafer. Although there is a large-diameter diffusion wafer, a portion having a high impurity concentration distribution is subsequently scraped off from the diffusion surface of the diffusion layer.
Compared to the conventional small-diameter diffusion wafer, the large-diameter diffusion wafer whose impurity concentration distribution differs due to scraping is different from the existing small-diameter diffusion wafer, which is an existing product, in product specifications. However, there is a problem that it cannot be used as a substitute for the conventional small-diameter diffusion wafer.

In order to solve such a problem, a method for manufacturing a diffusion wafer according to the present invention includes a non-diffusion layer in which a semiconductor element is formed in the wafer, and a diffusion layer in which impurities are diffused from the entire surface of the wafer. A method for producing a diffusion wafer, wherein a deposition step for adhering and penetrating the impurities on the surface of the wafer, and a removal step for removing a compound layer generated on the surface of the wafer during the deposition step; A drive-in step of diffusing the impurities into the wafer to form a concentration attenuation layer portion in which the concentration of the impurities attenuates toward the non-diffusion layer as the diffusion layer. this performing step and after said removing step, the drive-in process while the compound layer that is generated by the last of the deposition process remained Then, the concentration of the impurity along the exposed diffusion surface of the diffusion layer is substantially the same as the maximum concentration of the impurity in the concentration attenuation layer portion, and a scraped portion that makes the diffusion layer a target thickness is included. A uniform density layer portion is formed.

It is explanatory drawing which shows the whole structure of the diffusion wafer which concerns on embodiment of this invention, (a) is a concentration distribution graph of an impurity, (b) is a vertical front view. It is explanatory drawing which shows the manufacturing method of the diffusion wafer which concerns on embodiment of this invention, (a)-(e) is the vertical front view which showed the manufacturing process in process order. It is explanatory drawing which shows the modification of the manufacturing method of the diffusion wafer which concerns on embodiment of this invention, (a)-(d) is the vertical front view which showed the manufacturing process in process order. It is explanatory drawing which shows the modification of the manufacturing method of the diffusion wafer which concerns on embodiment of this invention, (a) (b) is the vertical front view which showed the manufacturing process in process order.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
The diffusion wafer A according to the embodiment of the present invention is a diffusion process performed on a wafer sliced from a rod-shaped semiconductor single crystal (ingot) used in the manufacturing process of semiconductor elements, semiconductor devices, etc. Impurities (dopants) are diffused.
Manufacturing processes of semiconductor elements, semiconductor devices, and the like are roughly divided into a wafer creation process for creating a semiconductor wafer from a rod-shaped semiconductor single crystal and an element formation process for forming a semiconductor element on the main surface portion of the semiconductor wafer. When a semiconductor element is formed using a rod-shaped silicon single crystal, a diffusion wafer A is formed by first performing a diffusion process on a silicon wafer sliced from the rod-shaped silicon single crystal in a wafer forming process.
Specifically, as shown in FIGS. 1A and 1B, the diffusion wafer A according to the embodiment of the present invention includes a non-diffusion layer 1 and a diffusion layer 2 in which impurities are diffused as main components. As prepared.
Although not shown, it is possible to add another layer in addition to the non-diffusion layer 1 and the diffusion layer 2.

The diffusion layer 2 has a concentration attenuation layer portion 21 that attenuates as the impurity concentration goes toward the non-diffusion layer 1 side, and a concentration equalization layer portion 22 formed along the exposed diffusion surface 2a.
The concentration attenuation layer portion 21 has a depth (hereinafter referred to as “depth”) from the diffusion surface 2 a side opposite to the non-diffusion layer 1 to the adjacent surface 2 b adjacent to the non-diffusion layer 1 in the diffusion layer 2. As shown in FIG. 4, the impurity concentration is gradually reduced.
The concentration uniform layer portion 22 does not attenuate the impurity concentration from the diffusion surface 2a to the boundary surface 2c between the concentration attenuation layer portion 21 and the concentration attenuation layer portion 21 regardless of the depth. 21 is a portion formed so as to be substantially the same as the maximum concentration of impurities in 21.
Here, “substantially the same” is not limited to the fact that the impurity concentration from the diffusion surface 2a to the boundary surface 2c is exactly the same (identical) as the maximum impurity concentration in the concentration attenuation layer portion 21, but in addition to this, The impurity concentration from the boundary surface 2c toward the diffusion surface 2a is slightly higher than the maximum impurity concentration in the concentration attenuation layer portion 21.
Further, the concentration uniform layer portion 22 can be used to scrape the diffusion layer 2 to a target thickness in a subsequent process of the wafer creation process. That is, the concentration equalization layer portion 22 has a scraped portion 22a for making the diffusion layer 2 a target thickness on the diffusion surface 2a side.
Examples of the post-process of the wafer creation process include the element formation process for producing a semiconductor element for the non-diffusion layer 1 and the device formation process for producing a semiconductor device.

ここで、Ｎｄは、拡散ウエハ内の拡散による不純物の濃度（拡散によってウエハ内に追加された不純物の濃度）、ｔは、拡散ウエハの拡散時間、ｘは、拡散ウエハ内における深度、Ｄは、拡散係数である。
数式１を満たすｘとｔの関数をＦ（ｘ，ｔ）、拡散ウエハの露出した拡散面における拡散による不純物の濃度（拡散によってウエハ内に追加された不純物の表面濃度）をＮｓとすると、数式２に示すようになる。

さらに、拡散ウエハ内における不純物の最小濃度（拡散前からウエハ内に存在する不純物の濃度）をＮｂとすると、拡散ウエハの全体における全不純物の濃度Ｎの分布は、数式３に示すように表せる。

On the other hand, the diffusion wafer A according to the embodiment of the present invention is a large-diameter diffusion wafer that can be used as an alternative to the conventional small-diameter diffusion wafer B that is an existing product.
In the case of the example shown in FIGS. 1A and 1B, the conventional small-diameter diffusion wafer B that is an existing product is obtained by changing the concentration uniform layer portion 22 from the large-diameter diffusion wafer A according to the embodiment of the present invention. The removed boundary surface 2c corresponds to a structure exposed as a diffusion surface.
By the way, the distribution of the impurity concentration in the conventional small-diameter diffusion wafer B, which is an existing product, and the distribution of the impurity concentration in the large-diameter diffusion wafer A are in accordance with the basic law of diffusion phenomenon (Fick's law). It is expressed by the diffusion equation shown below.

Here, Nd is the concentration of impurities by diffusion in the diffusion wafer (concentration of impurities added in the wafer by diffusion), t is the diffusion time of the diffusion wafer, x is the depth in the diffusion wafer, and D is Diffusion coefficient.
Assuming that the function of x and t satisfying Equation 1 is F (x, t), and the concentration of impurities due to diffusion on the exposed diffusion surface of the diffusion wafer (surface concentration of impurities added in the wafer by diffusion) is Ns. As shown in 2.

Furthermore, if the minimum impurity concentration in the diffusion wafer (concentration of impurities existing in the wafer before diffusion) is Nb, the distribution of the concentration N of all impurities in the entire diffusion wafer can be expressed as shown in Equation 3.

ここで、ｘｍは、図１（ａ）（ｂ）に示されるように、拡散層２の拡散面２ａから任意の深度である。

Thus, when the distribution of the concentration of all impurities in the conventional small-diameter diffusion wafer B that is an existing product has the diffusion profile expressed by Equation 3, it can be used as an alternative to the conventional small-diameter diffusion wafer B. The distribution of the concentration N of all impurities in the large-diameter diffusion wafer A has a diffusion profile approximately expressed by Equations 4 and 5.

Here, xm is an arbitrary depth from the diffusion surface 2a of the diffusion layer 2, as shown in FIGS.

ここで、ｘｊは、図１（ａ）（ｂ）に示されるように、拡散層２の拡散面２ａから非拡散層１と隣り合う隣接面２ｂまでの深度である。
数式６から、従前の小口径な拡散ウエハＢの代替品として使用可能な大口径な拡散ウエハＡの全不純物の濃度Ｎの分布を、近似的に数式７に示される拡散プロファイルを有するように変更することも可能である。

数式４及び数式５又は数式７により、拡散層２において拡散面２ａから任意の深度ｘｍまでの箇所には、その全不純物の濃度Ｎの分布が、その深度に関係なく濃度減衰層部２１へ向かって減衰せず、濃度減衰層部２１における不純物の最大濃度と略同じ又は同一になる、濃度均等層部２２が形成される。
すなわち、この実施形態に係る拡散ウエハＡにおいても、前述した実施形態に係る拡散ウエハＡと同様になる。 In particular, as an example of Equation 3, “diffusion coefficient D does not depend on x and t”, “the total amount of impurities used for diffusion is constant”, “when diffusion starts (t = 0), it is used for all diffusion. In the case where all of the “concentrated impurities are concentrated on the diffusion surface (x = 0)” is satisfied, the distribution of the concentration N of all impurities in the entire diffusion wafer can be expressed as shown in Equation 6.

Here, xj is the depth from the diffusion surface 2a of the diffusion layer 2 to the adjacent surface 2b adjacent to the non-diffusion layer 1, as shown in FIGS.
From Equation 6, the distribution of the concentration N of all impurities in the large-diameter diffusion wafer A that can be used as an alternative to the conventional small-diameter diffusion wafer B is changed to have a diffusion profile approximately represented by Equation 7. It is also possible to do.

According to Equation 4, Equation 5, or Equation 7, the distribution of the concentration N of all impurities in the diffusion layer 2 from the diffusion surface 2a to an arbitrary depth xm is directed toward the concentration attenuation layer portion 21 regardless of the depth. Thus, the concentration equalization layer portion 22 is formed which is not attenuated and is substantially the same as or the same as the maximum impurity concentration in the concentration attenuation layer portion 21.
That is, the diffusion wafer A according to this embodiment is the same as the diffusion wafer A according to the above-described embodiment.

And the manufacturing method for producing the diffusion wafer A according to the embodiment of the present invention is a wafer creation step of creating a semiconductor wafer from a rod-shaped semiconductor single crystal, with respect to a wafer sliced from the rod-shaped semiconductor single crystal, This is a manufacturing method for performing diffusion treatment. When the semiconductor single crystal is a silicon single crystal, the wafer is a silicon wafer.
More specifically, as shown in FIGS. 2 to 4, the manufacturing method of the diffusion wafer A according to the embodiment of the present invention includes a deposition process in which impurities such as phosphorus adhere to and penetrate into the surface 11 of the wafer 10. The main process includes a drive-in process for diffusing impurities adhering and permeating in the deposition process into the wafer 10. In addition, it is possible to include a removing step of removing the compound layer 3 generated on the surface 11 of the wafer 10 during the deposition step after the deposition step.
The compound layer 3 generated on the surface 11 of the wafer 10 during the deposition process is made of, for example, an oxide such as phosphorus glass or another silicon compound.
In the drive-in process, by diffusing impurities into the wafer 10, a concentration attenuation layer portion 21 is formed as the diffusion layer 2 that attenuates as the concentration of the impurities moves toward the non-diffusion layer 1.
By performing a drive-in process after the deposition process, a concentration uniform layer portion 22 is formed in which the impurity concentration is substantially the same as the maximum impurity concentration in the concentration attenuation layer portion 21 along the diffusion surface 2a of the diffusion layer 2. doing. The concentration uniform layer portion 22 includes a scraped portion 22a for setting the diffusion layer 2 to a target thickness in a subsequent process of the wafer creation process.

In the process from the deposition process to the drive-in process, in the process from the deposition process to the drive-in process, the drive-in process is performed after removing the compound layer 3 formed on the surface 11 of the wafer 10 in the deposition process. There are a first manufacturing method to be performed and a second manufacturing method to perform the drive-in process while leaving the compound layer 3 generated in the last deposition process.
In the case of the example shown in FIGS. 2A to 2E as the first manufacturing method, the deposition process of FIGS. 2A and 2C and the removal process of FIGS. 2B and 2D. Is repeated once or a plurality of times (twice or more), and then the drive-in process of FIG. 2E is performed in a state where the compound layer 3 is finally removed.
In the case of the example shown in FIGS. 3A to 3D as the second manufacturing method, the deposition process of FIG. 3A and the removal process of FIG. After repeating the process twice or more), the drive-in process of FIG. 3D is performed in a state where the compound layer 3 generated in the last deposition process of FIG. Yes.
In the case of the example shown in FIGS. 4 (a) and 4 (b) as the second production method other than that, after the deposition step of FIG. 4 (a) is repeated once or a plurality of times (twice or more). In the state where the compound layer 3 is finally left, the drive-in process of FIG. 4B is performed.
In the case of driving in with the compound layer 3 finally left as in the examples shown in FIGS. 3A to 3D and the examples shown in FIGS. Even after the implantation, the compound layer 3 remains on the surface 11 of the wafer 10. However, the compound layer 3 remaining after the drive-in can be easily removed by washing or the like. In addition, it is possible to scrape off the compound layer 3 when a part (scraped portion 22a) of the concentration uniform layer portion 22 is scraped off in a subsequent step such as the wafer creation step.

As a specific example of the manufacturing method of the large-diameter diffusion wafer A according to the embodiment of the present invention, a conventional single-crystal diffusion wafer B is formed from a semiconductor single crystal having a larger diameter than the conventional small-diameter diffusion wafer B. It is preferable that the wafer 10 is sliced more thickly and subjected to a diffusion treatment by a vapor phase diffusion method or the like on the large diameter and thick wafer 10.
More specifically, a wafer 10 that has been subjected to pre-processing such as lapping or etching after slicing is placed in a diffusion furnace (not shown) heated to a predetermined temperature, and phosphorus, which is an n-type impurity, is contained. Phosphoryl chloride (phosphorus oxychloride) and the like, nitrogen, argon, helium, or any mixed gas thereof, and oxygen are poured simultaneously, and deposition is performed at about 1200 ° C. or less.
As a result, the wafer 10 in the diffusion furnace takes in n-type impurities such as phosphorus from the surface 11 and the impurities 20 penetrate from the surface 11 to a predetermined depth. At the same time, since the surface 11 of the wafer 10 is oxidized by the influence of oxygen and heat, the compound layer 3 made of an oxide such as phosphorus glass is generated on the surface 11 of the wafer 10.
Thereafter, the deposited wafer 10 is taken out from the diffusion furnace and the compound layer 3 is removed and then put into the diffusion furnace and driven in, or the deposited wafer 10 is taken out from the diffusion furnace without taking out the compound layer 3. Drive in with the remaining. Drive-in is generally performed by pouring nitrogen, argon, helium, or any mixed gas into the diffusion furnace and heating to a temperature higher than the deposition temperature. In addition, the diffusion furnace used for drive-in may be different from the diffusion furnace used for deposition, or may be the same.
Thereby, the diffusion layer 2 having a predetermined concentration and a predetermined depth is generated from the surface 11 of the wafer 10 toward the inside.
In the example shown in FIGS. 2 to 4, the intermediate portion of the wafer 10 subjected to the diffusion process in the diffusion furnace is cut and separated (divided into two parts) by slicing so that the thickness other than the diffusion layer 2 becomes the target thickness. In addition, each cut surface is polished. As a result, two diffusion wafers A having the diffusion layer 2 only on one surface 11 and having the same overall thickness are produced. In the illustrated example, only one of the two diffusion wafers A is shown, and the other is omitted.
Although not shown in the drawings as other examples, a diffusion method different from the vapor phase diffusion method is used, an impurity other than phosphorus is used, the deposition conditions are changed each time the deposition process is repeated, or diffusion is performed. Instead of dividing the processed wafer 10 into two parts, it is possible to create one diffusion wafer A in which the thickness other than the diffusion layer 2 becomes the target thickness by grinding and polishing the non-diffusion layer 1 side.

On the other hand, in order to perform batch processing of a large number of wafers 10 at once, it is preferable to place a plurality of wafers 10 in close contact with each other in a diffusion furnace so as to stand in parallel.
In this case, it is preferable that the wafers 10 are stacked with a powder (not shown) having excellent heat resistance such as silicon or aluminum oxide sandwiched between the wafers 10 and driven in this state. As a result, the created diffusion wafers A can be easily separated from each other, and the workability is excellent.
Further, as another example, the drive-in is performed with the compound layer 3 finally left as in the example shown in FIGS. 3A to 3D and the example shown in FIGS. 4A and 4B. In this case, it is preferable to arrange the plurality of wafers 10 in the diffusion furnace so that each of them has a predetermined interval, specifically about 500 μm or more.

According to the diffusion wafer A according to the embodiment of the present invention, the impurity concentration is the same as that of the diffusion layer of the conventional small-diameter diffusion wafer B while the diameter is larger than that of the conventional small-diameter diffusion wafer B. By forming the concentration uniform layer portion 22 having the impurity concentration substantially the same as the maximum impurity concentration in the concentration attenuation layer portion 21 along the concentration attenuation layer portion 21 thus formed, the diffusion wafer B having a small diameter as before is formed. It becomes thicker than.
For this reason, even if it is handled in a subsequent process for manufacturing a semiconductor element, a semiconductor device, etc., it does not break, and even if part of the concentration uniform layer portion 22 (the scraped portion 22a) is finally scraped off. Thus, it is possible to ensure the same impurity concentration distribution as the diffusion layer of the conventional diffusion wafer B having a small diameter.
That is, the same impurity concentration distribution as that of the diffusion layer of the conventional small-diameter diffusion wafer B is obtained even after the concentration uniform layer portion 22 is scraped off while reducing the mechanical strength (preventing cracking) as the diameter increases. As a result, the product standard of the conventional small-diameter diffusion wafer B that becomes an existing product can be met.
Thereby, the freedom degree at the time of shaving off a part of diffusion layer 2 becomes high.
Therefore, a desired impurity concentration distribution can be obtained regardless of the increase in the diameter or the partial removal of the diffusion layer 2.
As a result, the large-diameter diffusion wafer A can be used as an alternative to the conventional small-diameter diffusion wafer B that is an existing product.
For this reason, it is possible to prevent cracking due to a decrease in mechanical strength accompanying the increase in diameter, and it is possible to significantly increase productivity and improve economic rationality by increasing the diameter.
In addition, since the impurity concentration in the deep layer portion near the non-diffusion layer 1 in the diffusion layer 2 can be precisely adjusted, it is possible to improve the conductivity immediately below the region of the semiconductor device.
Therefore, a semiconductor device with little power loss can be manufactured.

According to the manufacturing method of the diffusion wafer A according to the embodiment of the present invention, by performing a drive-in process after the deposition process, the concentration of impurities as the diffusion layer 2 on the surface 11 side of the wafer 10 is shifted to the non-diffusion layer 1 side. A concentration attenuation layer portion 21 that attenuates toward the surface is formed. At the same time, the impurity concentration from the diffusion surface 2a to the boundary surface 2c of the diffusion layer 2 is the same as the maximum impurity concentration in the concentration attenuation layer portion 21 regardless of the depth, and the diffusion layer 2 is targeted in a later step. A concentration uniform layer portion 22 including a shaved portion 22a for increasing the thickness is formed.
For this reason, even if it is handled in a subsequent process for manufacturing a semiconductor element, a semiconductor device, etc., it will not break, and even if a part of the concentration uniform layer portion 22 (the scraped portion 22a) is finally scraped off, The same impurity concentration distribution as that of the diffusion layer of the small-diameter diffusion wafer B can be ensured. Thereby, the freedom degree at the time of shaving off a part of diffusion layer 2 becomes high.
Therefore, a desired impurity concentration distribution can be obtained regardless of the increase in the diameter or the partial removal of the diffusion layer 2.
As a result, the large-diameter diffusion wafer A can be used as an alternative to the conventional small-diameter diffusion wafer B that is an existing product.
For this reason, it is possible to prevent cracking due to a decrease in mechanical strength accompanying the increase in diameter, and it is possible to significantly increase productivity and improve economic rationality by increasing the diameter.
In addition, since the impurity concentration in the deep layer portion near the non-diffusion layer 1 in the diffusion layer 2 can be precisely adjusted, it is possible to improve the conductivity immediately below the region of the semiconductor device.
Therefore, a semiconductor device with little power loss can be manufactured.

In particular, as shown in FIGS. 2A to 2E, it is preferable to perform the drive-in process after performing the deposition process and the removal process.
In this case, since the next deposition is performed in a state where the compound layer 3 generated by the previous deposition is removed, the impurity 20 penetrates deeper in a short time.
Therefore, a desired impurity concentration distribution can be obtained by short-time heat treatment regardless of the increase in diameter or partial scraping of the diffusion layer 2.
As a result, the impurity concentration from the diffusion surface 2a to the boundary surface 2c is independent of the depth, compared to the case where the next deposition was performed without removing the compound layer 3 generated in the previous deposition. The uniform concentration layer portion 22 having the same concentration as the maximum impurity concentration in the concentration attenuation layer portion 21 can be easily created in a short time.
The number of repetitions of the deposition process and removal process, the deposition conditions such as the deposition temperature and deposition time in each deposition process, the supply amount of the diffusion source, the drive-in temperature and the drive-in time in the drive-in process, etc. By adjusting either or both of the drive-in conditions, it is possible to arbitrarily set an arbitrary depth xm from the diffusion surface 2a of the diffusion layer 2.

Further, as shown in FIGS. 3A to 3D, after the deposition process and the removal process are performed, the drive-in process is performed with the compound layer 3 generated in the last deposition process remaining. Is preferred.
In this case, since the next deposition is performed in a state where the compound layer 3 generated by the previous deposition is removed, the impurity 20 penetrates deeper in a short time. It is possible to drive in continuously without removing the compound layer 3 after the last deposition.
Therefore, a desired impurity concentration distribution can be obtained more reliably and with a short heat treatment regardless of the increase in the diameter or the partial removal of the diffusion layer 2.
As a result, compared to the production method shown in FIGS. 2A to 2E in which the deposition and the removal of the compound layer 3 are repeatedly performed, deep diffusion at a high concentration can be achieved even if the number of depositions is reduced or the number of removals is reduced. realizable. This simplifies and reduces costs at the same time.
Either the number of repetitions of the deposition process and removal process and the deposition conditions such as the deposition temperature, deposition time, and supply amount of the diffusion source, or the drive-in conditions such as the drive-in temperature and the drive-in time, or By adjusting both, it is possible to arbitrarily set an arbitrary depth xm from the diffusion surface 2a of the diffusion layer 2.

Further, as shown in FIGS. 4A and 4B, it is preferable to perform the drive-in process with the compound layer 3 generated in the deposition process remaining after performing the deposition process.
In this case, it is possible to drive in continuously without removing the compound layer 3 after deposition.
Therefore, a desired impurity concentration distribution can be more easily obtained with the minimum number of steps regardless of the increase in the diameter or the partial removal of the diffusion layer 2.
As a result, compared with the manufacturing method including the removal process of the compound layer 3 shown in FIGS. 2A to 2E and FIGS. 3A to 3D, the number of depositions is reduced or the removal process is omitted. Can achieve high concentration and deep diffusion. As a result, simplification and cost reduction can be achieved at the same time.
The diffusion surface of the diffusion layer 2 is adjusted by adjusting one or both of the deposition conditions such as the deposition temperature, the deposition time, the supply amount of the diffusion source, and the drive-in conditions such as the drive-in temperature and the drive-in time. It is possible to arbitrarily set an arbitrary depth xm from 2a.

Furthermore, in the drive-in process, it is preferable to arrange a plurality of wafers 10 so as to leave a predetermined interval.
In this case, even if drive-in is performed without removing the compound layer 3 after deposition, a part of the impurities in the compound layer 3 is uniformly scattered from the surface of the wafer 10 to the outside. To diffuse into the wafer 10.
Therefore, impurities can be diffused in the wafer 10 in a uniform state regardless of the presence or absence of the compound layer 3.
As a result, compared to the manufacturing method in which powders having excellent heat resistance such as silicon are sandwiched between the wafers 10 and stacked and driven in, variation in the diffusion depth in the wafer 10 surface (in-plane distribution is not uniform) Can be prevented.
Further, when the drive-in is performed with a powder such as silicon sandwiched between the wafers 10, the surface 11 of the wafer 10 does not come into contact with the powder as compared with a manufacturing method in which the surface 11 of the wafer 10 may be contaminated. , Can prevent the occurrence of dirt.

Next, an embodiment of the present invention will be described with reference to the drawings.
As this example, a diffusion wafer A having a larger diameter than the conventional small-diameter diffusion wafer B and having a thickness equivalent to the concentration equality layer portion 22 than the thickness of the conventional small-diameter diffusion wafer B was produced. . After the diffusion wafer A according to this embodiment is created in the wafer creation process, a part of the concentration equality layer portion 22 (the scraped portion 22a) is scraped off in an element forming process that is a subsequent process, thereby reducing the conventional small diameter. It was verified whether or not the concentration distribution of all impurities in the diffusion wafer B was the same.
That is, it was verified whether or not the large-diameter diffusion wafer A can be used as an alternative to the existing small-diameter diffusion wafer B that is an existing product.

For this reason, as shown in FIGS. 1A and 1B, a conventional small-diameter diffusion wafer B, which is an existing product, has a diameter of 125 mm and a temperature of 1200 ° C. or less by a vapor phase diffusion method. No. 11 (hereinafter referred to as “existing product B”) and a large-diameter diffusion wafer A of this embodiment having a diameter of 150 mm and a concentration-uniform layer portion as compared with the existing product B Thickness equivalent to 22 and a diffusion wafer having a diameter of 150 mm as a comparison target and thicker than the existing product B by the concentration uniform layer portion 22 (hereinafter referred to as “comparative products C and D”) , Respectively.
Further, in order to easily compare these, in the impurity concentration distribution graph shown in FIG. 1A, the concentration distribution A1 of all impurities in the large-diameter diffusion wafer A of the embodiment is shown by a solid line, and the existing product B The concentration distribution B1 of all impurities in FIG. 2 is indicated by a thick broken line, the concentration distribution C1 of all impurities in the comparative product C is indicated by a two-dot chain line, and the concentration distribution D1 of all impurities in the comparison product D is indicated by a two-dot chain line.

In the large-diameter diffusion wafer A of this embodiment, as shown in FIGS. 2 to 4, impurities adhere to the surface 11 of the wafer 10 at 1200 ° C. or less by the vapor phase diffusion method as in the existing product B, and penetrate. Was made, followed by a drive-in.
More specifically, as shown in FIGS. 2A to 2E, the deposition step and the removal step of the compound layer 3 such as phosphorus glass (including impurities) generated by the deposition were repeated twice. I prepared a drive-in later. Furthermore, as Ru shown in FIG. 3 (a) ~ (d) , carried out deposition step with a compound layer 3 removal step once each, subsequent deposition steps compound layer 3 that is generated by the remaining We also prepared a drive-in in the state. Also, as shown in FIGS. 4A and 4B, a drive-in was prepared with the compound layer 3 left after deposition.
On the other hand, Comparative products C and D were each subjected to the deposition process and the compound layer 3 removal process only once.
Further, in the comparative product C, the impurity concentration distribution C1 in the drive-in process is copied to a part of the total impurity concentration distribution B1 in the existing product B as shown by a two-dot chain line in FIG. The diffusion is performed so as to be higher than the concentration distribution D1 of all impurities in the comparative product D.
In the comparative product D, the impurity concentration distribution D1 in the drive-in process is shown in FIG. 1A by the two-dot chain line, and the impurity concentration (surface concentration of all impurities, Ns + Nb) on the diffusion surface 2a is The diffusion is performed so as to be thinner than the concentration distribution C1 of all impurities in the comparative product C so as to be the same as the concentration of impurities in the exposed diffusion surface of the product B.

Comparing the large-diameter diffusion wafer A of the example and the comparative products C and D, the large-diameter diffusion wafer A of the example has a diffusion layer whose concentration distribution A1 is indicated by a solid line in FIG. In FIG. 2, the diffusion can be easily performed in a short time so as to be substantially in a straight line from the diffusion surface 2 a to an arbitrary depth xm (the boundary surface 2 c between the concentration attenuation layer portion 21 and the concentration equalization layer portion 22).
Further, the concentration distribution A1 of the large-diameter diffusion wafer A in the embodiment is the same as the impurity concentration distribution B1 in the existing product B from an arbitrary depth xm (boundary surface 2c) to the adjacent surface 2b with the non-diffusion layer 1. It was possible to diffuse easily in a short time.
As a result, by removing a part of the uniform concentration layer portion 22 (the scraped portion 22a), the impurity concentration distribution B1 in the existing product B is obtained, and it is verified that it easily matches the product standard of the existing product B. did it.
As a result, it was found that the large-diameter diffusion wafer A of the example can be used as an alternative to the existing product B.

On the other hand, the comparative products C and D have the same concentration distributions C1 and D1 as the impurity concentration distribution B1 in the existing product B as shown by the two-dot chain line in FIG. Could not spread.
Thereby, even if the diffusion surface 2a side that becomes the concentration uniform layer portion 22 of the large-diameter diffusion wafer A of the embodiment is scraped off, the impurity concentration distribution B1 in the existing product B cannot be obtained, and the product standard of the existing product B It was proved impossible to match.

  In the embodiment described above, the case where a part of the concentration uniform layer portion 22 is scraped off in the subsequent step after the diffusion wafer A is formed has been described. However, the present invention is not limited to this. The portion may be changed so as not to be scraped off, or may be changed so as to be further scraped off beyond the density uniform layer portion 22.

A Diffusion wafer 1 Non-diffusion layer 2 Diffusion layer 2a Diffusion surface 2b Adjacent surface 2c Boundary surface 20 Impurity 21 Concentration attenuation layer portion 22 Concentration equality layer portion 22a Cut-off site 3 Compound layer

Claims (2)

A method for producing a diffusion wafer comprising a non-diffusion layer on which a semiconductor element is formed on a wafer and a diffusion layer in which impurities are diffused from the entire surface of the wafer,
A deposition process for adhering and infiltrating the impurities on the surface of the wafer;
A removal step of removing a compound layer generated on the surface of the wafer during the deposition step;
A drive-in step of diffusing the impurities into the wafer to form a concentration attenuation layer portion in which the concentration of the impurities attenuates toward the non-diffusion layer as the diffusion layer, and
After performing the deposition step and the removal step, by performing the drive-in step while the compound layer generated in the last deposition step remains , along the exposed diffusion surface of the diffusion layer The diffusion is characterized in that the concentration of the impurity is substantially the same as the maximum concentration of the impurity in the concentration attenuation layer portion, and a concentration uniform layer portion including a scraped portion that makes the diffusion layer a target thickness is formed. Wafer manufacturing method. The drive is in process, the manufacturing method of the diffuse wafer according to claim 1, wherein placing the wafer so that each plurality frees predetermined intervals.

Images