JP2009170938A - Silicon fixture for wafer processing, and manufacturing method thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a silicon tower for wafer processing which reduces shadowing and contamination while correctly keeping the wafer in a stable state. <P>SOLUTION: This silicon tower for supporting a plurality of wafers is structured to removably support the plurality of wafers, and characterized by including: a plurality of leg parts each formed in accordance with a non-circular cross-sectional leg part shape, and formed of silicon; and two bases formed of silicon and each having a hole part having the non-circular shape for housing and fixing the leg part. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates generally to a fixture used to support a wafer during semiconductor integrated circuit manufacture, and more particularly to a silicon fixture that supports the wafer.

  As commercial production of silicon wafers develops, increasingly larger wafers are processed in increasingly larger batches, while the individual sizes are shrinking to less than 0.18 μm. Because of this process, requirements for processing equipment and wafer handling and transport mechanisms required to move, transport, and hold wafers during processing are becoming increasingly demanding. It has become to. These requirements include keeping the temperature constant and contamination with impurities or particles.

  During many chemical and thermal processing operations, the wafer is often accurately positioned during various processing steps, particularly annealing, dopant diffusion, or chemical vapor deposition, which are often performed simultaneously on many wafers. Must hold on. For this purpose, relatively large and complex structures such as “boats” or “towers” are usually used. Patent Document 1 discloses an example of such a structure. Patent Document 1 discloses a vertical boat for holding a plurality of semiconductor wafers. The boat includes two end members and a plurality of support members. In one embodiment, the support member is formed from a vertically cut pipe member to provide a long plate member having a quarter arc cross section. In another embodiment, the support member is formed from a vertically cut pipe member to provide a long plate member having a half arc cross section. Patent Document 1 lists quartz glass, silicon carbide, carbon, single crystal silicon, polycrystalline silicon, and silicon-impregnated silicon carbide as materials that can be used for the boat. When manufactured from quartz glass, the various elements must be fused together. Otherwise, “the device can be assembled in a predetermined manner”.

Patent Document 2 also relates to a vertical boat for holding semiconductor wafers. The boat includes a plurality of rods having slots formed along their entire length. The slots are for reducing shadowing on wafers installed in the boat during processing. The rod is cylindrical and is designated to be manufactured from fused quartz, but any well-known material suitable for holding the wafer can also be used.
Patent Document 3 discloses a reinforced carrier for silicon wafers and the like. The carrier includes a side element consisting of an annular rail with teeth protruding therefrom to hold the wafer at regular intervals. The rail is made of plastic and may include a hard insert for reinforcement. The teeth can be molded integrally with the rail or welded to the rail.

U.S. Pat. No. 6,057,059 relates to a wafer boat that includes a plurality of rods arranged to support a ring member. The plurality of wafer support pieces are connected to the ring member and include protrusions having corners for contacting the wafer. U.S. Pat. No. 6,057,089 also discloses a wafer boat that includes a plurality of cylindrical quartz rods having a wafer support recess therein.
The theoretical advantages of pure silicon structures are well known. Conventional towers and boats are usually made of quartz or silicon carbide, but these materials contain contaminants and become unstable at higher temperatures. Manufacturing the wafer holding structure from the same material as the wafer itself reduces the possibility of contamination and deformation to a minimum. Silicon structures react with processing temperatures, conditions and chemistry in exactly the same way as wafers, significantly improving the overall useful life of the structure.

  Unfortunately, the standard assembly of silicon structures in the “predetermined method” disclosed in US Pat. No. 6,057,049 is why pure silicon was not widely accepted as a material for structures such as boats and towers. One. Since single crystal silicon, polycrystalline silicon, and virgin polysilicon are difficult to process, the structure described in the patent of Patent Document 1 has been developed, but when single crystal silicon is selected as the material When considering the above, there is no description about the connection between the support member and the target member, and the above-described manufacturing method of the support structure only includes the cutting out of the extruded annular member. Such support structures are inherently less stable than support structures made from materials that have been used for a long time, such as quartz or silicon carbide, and can be more easily processed.

Similarly, the patents of U.S. Pat. Nos. 6,057,036, and 5,048, do not solve the specific problem of providing a robust and reliable wafer support structure that reduces shadowing and contamination. The protrusions and slots described in the above patent are effective to some extent, but are not suitable for manufacturing from materials such as silicon, and to provide a stable and accurate wafer support. A relatively large cross-sectional area is required.
Silicon has been found to be very brittle and difficult to melt. For this reason, the well-known silicon structures are considered to be delicate, at best, and very brittle at best. Therefore, the silicon structure has not been widely accepted commercially.

Further, due to the crystal structure of single crystal silicon, blanks extruded from crystalline silicon have a clear “grain” that runs substantially vertically in the blank. Silicon blanks are usually cut transversely across the grain by a thread saw. Unfortunately, when used for longitudinal cutting, conventional cutting techniques tend to divide the silicon blank along the grain, which makes the blank useless.
Monocrystalline silicon structure member and polycrystalline silicon for use in the manufacture of semiconductor wafers and the like that can eliminate the disadvantages of known silicon structures while retaining the advantages of silicon as a structural material It can be seen that a method for manufacturing a structural member is awaited.

  Czochralski (CZ) single crystal silicon is of the type used as a wafer in semiconductor integrated circuits, essentially having a horizontal length of 200 mm and 300 mm stretched from molten silicon as a large ingot. Made of single crystal of silicon. Czochralski polycrystalline silicon, often referred to as quasi-single crystalline silicon, has virtually the same local structure as single crystalline silicon, but is made up of individual crystals. Crystallite has a size of about 1 mm to more than 100 mm and is separated by grain boundaries. Such Czochralski polysilicon is considered to be conventional polysilicon in terms of structural members. Whether Czochralski silicon grows in a single crystal form or a polycrystalline form largely depends on the drawing speed. Czochralski silicon may grow with 1 ppm heavy metal impurities, but carbon and nitrogen may also be present at concentrations of 1-7 ppm, while oxygen is at concentrations of 10-25 ppm. Exists. Crystallites usually have orientations that are very similar to each other. Further, polysilicon is often grown as a thin layer in silicon integrated circuits by chemical vapor deposition, but such films cannot be applied directly to the present invention.

Virgin polysilicon (hereinafter referred to as virgin poly) is a special type of polysilicon that is widely manufactured for use within the semiconductor industry. Virgin poly is formed into relatively large ingots (with a maximum diameter of about 15 cm) by thermal chemical vapor deposition (CVD) using various silanes as precursors that agglomerate on heated seed rods. Such precursors include SiCH ₄ , SiClH ₃ , SiCl ₂ H ₂ , SiCl ₃ H, and SiCl ₄ . Of these compounds, SiHCl ₃ H is the most commonly used commercially, but has been used in float zone deposition so far, in some cases monosilane (SiH ₄ ) is used. In some cases. Virgin poly grows with very high purity, with an impurity concentration of 10 ⁻¹² cm ⁻³ or less. Normal standards, virgin poly is, 1 ppt (impurity levels of all of the expected contaminants containing oxygen, 1 × 10 ^-12) is a level containing about impurity. Although there is some variation, virgin poly has an impurity level of 10 ppt or less. This is in contrast to Czochralski polysilicon, which contains various impurities of at least 1 ppm heavy metal and up to 25 ppm or more of dissolved gas. Virgin poly is commercially produced with high internal stress so that it can be easily crushed. Semiconductor silicon wafers are usually grown by the Czochralski method, in which the polysilicon is crushed and then slightly above 1416 ° C., the melting point of silicon under atmospheric pressure. At a temperature in the vicinity of, perhaps together with the intentionally introduced dopant material. Single crystals agglomerate on small seed crystals placed on the surface of the melt, and the growing single crystal ingot is drawn very slowly from the melt.

Whether the silicon by the Czochralski method becomes single crystal or polycrystal depends largely on the pulling speed, and crystallite tends to grow greatly at a random size. Virgin poly, on the other hand, aggregates from hot seed rods and attempts to form crystalline arms that protrude radially from the rod.
To the best of our knowledge, virgin poly has never been used in silicon fixtures to support wafers during semiconductor processing.
Therefore, it can be seen that there is a need for the development of a strong and reliable support member for wafer processing fixtures that reduces shadowing and contamination while holding the wafer accurately in a stable state.

US Pat. No. 5,492,229 US Pat. No. 5,534,074 U.S. Pat. No. 4,872,554 US Pat. No. 5,752,609

  It is an object of the present invention to provide a silicon tower for wafer processing that reduces shadowing and contamination while holding the wafer accurately in a stable state.

The present invention includes a silicon wafer processing fixture. The fixture includes a uniformly elongated silicon support member with an attachment element extending outwardly from its terminal end. The fixture also includes a substantially flat silicon base. The base includes an attachment element receiving portion for securely securing the attachment element of the support member therein.
In the case of the first embodiment, at least a part of the fixture is made of virgin polysilicon. In one more specific embodiment, the elongated member is machined from virgin polysilicon and the monolithic base is single crystal silicon. In other embodiments, the base is composed of several parts, all of which are formed from virgin polysilicon.

  The present invention provides a method for manufacturing an elongated structural member from a single blank of crystalline material. The blank has a predetermined length, width and depth. A first substantially flat cut is formed in the blank that extends along substantially the entire length of the blank and a distance that is significantly shorter than the entire width of the blank. Within the blank, at least one additional cut is formed, which extends in the same plane as the first cut and cuts the blank into two parts.

Forming the at least one additional cut may include forming a plurality of additional cuts. In some embodiments, this number of additional cuts is at least three. The cut can be formed by a rotating saw comprising a blade with a diamond-coated cutting surface. When the saw operates, the blade speed is in the range of 50 to 50,000 rpm, but preferably about 4,000 rpm.
The blank of crystalline material can be supplied as a single blank of silicon material. The method of the present invention can be carried out with a single crystal silicon material or a polycrystalline silicon material.

  In some embodiments, the blank can be supplied as a generally cylindrical blank, i.e., an ingot. The step of forming the first substantially flat cut and the step of forming at least one additional cut are performed iteratively for each of the two parts to form four parts from the original cylinder. Can be performed iteratively as well to form eight parts from the original cylinder. Each of the eight parts has a substantially wedge-shaped cross section.

  The present invention discloses a method for securing a first silicon member to a second silicon member to form at least a portion of the silicon wafer processing fixture as well as the silicon wafer processing fixture itself. ing. The method includes providing an attachment member extending outwardly on the first silicon member. At least partially, a second silicon member is provided that includes an attachment element receiving portion surrounding the attachment member. Thereafter, the attachment element is securely fixed in the attachment element receiving portion.

  The present invention provides a method of manufacturing a support member for a wafer processing fixture. In the first step of the method, an elongated support member base mold is formed. The basic mold has a substantially wedge-shaped cross section and an angular edge. The base edge of the support member is then machined to form a generally arcuate edge instead of an angular edge. A plurality of wafer holding slots are formed along one face of the base mold of the support member.

  The basic form of the support member can include a front face and a rear face, with at least one angular edge on each face. The step of machining the edges of the base of the support member is to machine the edges of each face to a radius in the range of 0.25 inches (6.3 mm) to 5.25 inches (133.3 mm). Can be executed. In certain embodiments, the at least one angular edge of the rear surface can be machined to a radius of about 1.5 inches (38 mm), and the at least one angular edge of the front surface is about. It can be machined to a radius of 35 inches (8.9 mm). Conveniently, all other corners of the support member have a radius that is smaller than the radius of the front surface.

At least one mounting structure can be formed on at least one terminal end of the basic form of the support member. The mounting member is configured such that the support member can be easily mounted on the substantially flat base member. In some embodiments, the mounting structure can be provided as a set of cylindrical pegs, each extending from each end of the base type of the support member.
The basic type of elongated support member can be made from an inert crystalline material such as polycrystalline silicon or single crystal silicon.

  The step of forming a plurality of wafer holding slots along one side of the support member basic mold can be performed to form a plurality of cuts substantially perpendicular to the longitudinal axis of the support member basic mold. The cut can be extended by a suitable distance the depth of the basic form of the support member and can be formed by a rotating saw with a blade having a diamond-coated surface. The wafer holding slot can be formed perpendicular to the front face of the basic mold of the support member.

  The present invention also discloses a support member for a wafer processing fixture. The support member can include an elongate body portion comprising a pair of opposing terminations, an arcuate front surface having a first radius of curvature, and an arcuate rear surface having a second radius of curvature. The first radius of curvature can be much smaller than the second radius of curvature. A plurality of mutually parallel wafer holding slots are formed in the front surface of the body portion.

  In some embodiments, a set of mounting structures can extend from each end of the support member. The mounting element is generally cylindrical and the mounting element receiving portion is formed as a cylindrical bore having a diameter and length corresponding to the diameter and length of the mounting element. However, other shapes such as blades are possible. The first and second silicon members, including the attachment members, can be formed from single crystal silicon, polycrystalline silicon, or virgin polysilicon.

  According to one method, embodiment, the step for securely securing the mounting element forms a first lateral bore in the mounting element and a second lateral bore in the mounting element receiving portion. Can be executed. The first and second lateral bores are coaxial with each other and can accommodate locating pins. When the first and second lateral bores are coaxially aligned with each other, the locating pins are secured within the first and second lateral bores. The pin length can be slightly longer than the combined length of the first and second bores, and the pin can be secured in the following manner. Initially, the locating pin is inserted into the alignment bore such that a portion of the locating pin extends outwardly from the outer limits of the first and second bores. The extension portion of the locating pin is then machined until it is flush with the outer limits of the first and second bores. Alternatively, the outer diameter of the locating pin can be approximately equal to the inner diameter of the first and second bores. In this case, the pin can be fixed by the following method. First, the positioning pin is frozen at a cryogenic temperature, and the positioning pin is contracted. The locating pin is then inserted into the aligned bore while maintaining the bore at ambient or higher temperature. Thereafter, the temperature of the positioning pin is returned to ambient temperature and expanded.

  In another securing step, energy may be supplied to at least one of the attachment element and the attachment element receiving portion to fuse the attachment element to the attachment element receiving portion. In some embodiments, an access hole is provided in the mounting element receiving portion of the second silicon member. Laser energy is supplied from the access hole in order to tack weld the mounting element and the mounting element receiving portion. Alternatively, the mounting element and the mounting element receiving portion can be approximately the same size. Laser energy can be supplied to an area adjacent to both the mounting element and the mounting element receiving portion.

Furthermore, in other embodiments, a peripheral convex portion can be provided at the terminal portion of the first silicon member. A peripheral recess can be provided at the end edge of the second silicon member. Laser energy is supplied so that the peripheral convex portion of the first silicon member is melted into the peripheral concave portion of the second silicon member. Laser energy is supplied to heat the protrusion to the silicon melting point of 1,416 degrees, preferably about 1,450 degrees, for about 3 minutes or until the silicon melt fills the peripheral recess. Is done.
Furthermore, in other embodiments, the step for securely securing the mounting element within the mounting element receiving portion includes a strong, contamination-free bond between the mounting element and the mounting element receiving portion. Applying an agent. The mounting element is substantially cylindrical and the mounting element receiving portion is formed as a cylindrical bore having a diameter and length corresponding to the diameter and length of the mounting element. The first and second silicon members can be formed from single crystal silicon, polycrystalline silicon, or virgin polysilicon.

By using the present invention, a structural member of single crystal silicon, polycrystalline silicon, or virgin polysilicon for use in manufacturing a semiconductor wafer or the like can be manufactured. It can be applied to any large and / or complex fixtures, or parts thereof, used in the processing. If components using the structural members of the present invention are used, deformation during high temperature processes can be avoided. Since the raw material is of the same quality as the wafer material, particle contamination and “crystal slip” inherent in well-known materials such as silicon carbide can be virtually avoided. In addition, there is no shadowing. This is because the raw material copies the physical properties and important constants of the processed wafer on a one-to-one basis. Single crystal fixtures and parts thereof provide tolerances and expected lifetimes that cannot be achieved with those made from commonly used materials such as quartz or silicon carbide. Using the present invention, advantageous silicon parts and fixtures can be manufactured when the manufacturing process moves to 300 mm and higher wafer diameters.
Although the invention has been described with reference to particular embodiments, those skilled in the art will recognize that the invention can be variously modified without departing from the spirit and scope of the invention. Will be able to.

1 is an orthographic view of a silicon wafer processing fixture incorporating the principles of the present invention. FIG. A blank for use in the method of the present invention. Fig. 6 is a flow chart illustrating steps of an embodiment of the present invention. FIG. 6 is a side cross-sectional view of a blank during the cutting process. It is an axial sectional view of the blank corresponding to FIG. FIG. 6 is a side cross-sectional view of a blank during the cutting process. It is an axial sectional view of the blank corresponding to FIG. It is a figure which shows each step included in the method of this invention. It is a figure which shows each step included in the method of this invention. It is a figure which shows each step included in the method of this invention. It is a figure which shows the basic type of the supporting member used with the method of this invention. It is an end elevation view of the basic mold of the support member in the first stage of the manufacturing process. FIG. 6 is a side elevational view of a basic form of a support member at a subsequent stage in the manufacturing process. It is an orthographic view of 1st Embodiment of the combination of a leg part and a base. FIG. 6 is a side elevational view of a basic form of a support member at another stage of the manufacturing process. It is an orthographic view of 2nd Embodiment of the combination of a leg part and a base. It is sectional drawing which shows the technique for fixing the structural material of a silicon wafer processing fixture. It is sectional drawing which shows the technique for fixing the structural material of a silicon wafer processing fixture. It is sectional drawing which shows the technique for fixing the structural material of a silicon wafer processing fixture. It is sectional drawing which shows the technique for fixing the structural material of a silicon wafer processing fixture. It is sectional drawing which shows the technique for fixing the structural material of a silicon wafer processing fixture. It is sectional drawing which shows the technique for fixing the structural material of a silicon wafer processing fixture. It is sectional drawing which shows the technique for fixing the structural material of a silicon wafer processing fixture.

  FIG. 1 is a silicon wafer processing fixture 10. The silicon wafer processing fixture 10 includes a plurality of uniformly elongated support members 12 (hereinafter referred to as legs) that are secured between a set of generally flat base members 14, hereinafter referred to as a base. A plurality of slots 16 are typically formed in each leg 12 at regular intervals and are used to support a plurality of wafers within the assembled tower 10. The tower is typically installed semi-permanently in a semiconductor processing reactor configured for one of many different processes. Multiple wafers are placed in the tower 10 and then processed simultaneously. This process involves moderate temperatures in the range of 400-700 ° C, or high temperatures in the range of 1,000-1,380 ° C, depending in part on whether chemical vapor deposition, annealing, or thermal diffusion is performed. Can do.

  Although the tower 10 in the figure has four leg portions 12, the number of leg portions 12 may be three, two, or one. The most typical method is that the legs 12 firmly support the wafer, but the wafer is linearly placed in the slot 16 by an automated robot that supports the wafer on a paddle that moves transversely to the axis of the tower 10. The plurality of legs are attached to the base 14 within a range slightly wider than 180 degrees on the outer periphery of the base 14 so that the base 14 can be inserted into the base 14.

Leg 12 and base 14 can be made of any suitable method from an inert crystalline material such as single crystal silicon or polycrystalline silicon. Preferably, the legs 12 are preferably formed from virgin poly because the level of impurities is low. This is because the leg 12 directly contacts the wafer. Also, for many applications, the legs 12 need to be relatively long, and long virgin poly is readily available. However, the base 14 is easily formed from single crystal silicon. Conveniently the diameter of the base 14 formed in a substantially circular shape needs to be larger than the diameter of the wafer to be processed. Currently, most wafers have a diameter of 200 mm, but 300 mm wafers have also been manufactured. Even virgin poly ingots greater than 300 mm or even 200 mm are not readily available, but super large CZ crystalline ingots are available for special applications.
The manufacturing of the legs is divided into two main steps. That is, (1) forming a wedge-shaped basic mold, and (2) machining the above basic mold into legs having a novel cross section.

A basic mold is machined from the cylindrical blank 20 of FIG. The blank 20 is substantially cylindrical and has a length L and a diameter D. However, the present invention can be applied to any suitable blank of almost any shape.
The blank 20 can be made from a crystalline material such as single crystal silicon or polycrystalline silicon, but virgin poly blanks are currently preferred. Silicon blanks are widely available commercially. One supplier of suitable silicon blanks is SILICON CRYSTALS INC. Blanks can be made to any size, but are typically 4 inches (10 cm) to 80 inches (200 cm) in length and 0.75 inches (2 cm) to 36 inches (91 cm) in diameter.

As schematically shown in FIG. 3, a method according to one aspect of the present invention uses a series of progressive cuts along the longitudinal axis of the blank 20 to cut the blank into small pieces. In step 22, a first substantially flat cut C ₁ is formed in the blank 20. As shown in the side cross-sectional view of FIG. 4 and the axial cross-sectional view of FIG. 5, the first cut C ₁ extends substantially over the entire length L of the blank 20 and much more than the full width of the blank 20 (here, the diameter D). Extend a short distance.
In step 24 of FIG. 3, an additional cut C ₂ is formed in the blank 20. As shown in the side sectional view of FIG. 6 and the axial sectional view of FIG. 7, the cut C ₂ extends in the same plane as the first cut C ₁ .

If the diameter of the blank 20 is relatively small, two cuts are sufficient to divide the blank into two parts H ₁ and H ₂ as shown in FIG. Otherwise, to completely separate into the _two parts H ₁ and H ₂ in FIG. 8, (N−2) additional cuts, C ₃ to C _N , as shown in step 26 of FIG. It may be necessary. In the case of a regular blank with a diameter of 3 inches (76 mm), it has been found that three cuts give good results. However, it has been found that high pressure fluid cutting is effective when cutting all blanks 20 in one step. Furthermore, if the surface finish can be compromised, a one-step cut using a diamond blade is effective.

  The blank 20 can be cut by any suitable technique. At present, it is considered that the cutting can be effectively performed using a rotating saw such as the M4K34F21G model manufactured by MK. The saw can be equipped with a blade with a diamond at the tip, such as part number 10125D22 or 10125D100 from National Diamond Lab. During the cutting of the blank, the saw can be operated at a speed of 50 to 50,000 rpm. A speed of about 4,000 rpm has been found to be particularly effective. Although the use of a rotating saw is effective, it should be understood that the rotating saw is merely exemplary. Other cutting devices can also be used to achieve acceptable results. Examples of such devices include, but are not limited to, saws using a blade without diamond, lasers, wire saws, polishing saws, reciprocating saws, and abrasive fluid cutting devices.

In many cases, the manufacture of the structural support member is facilitated by the step of supplying small pieces smaller than half of the blank of FIG. 8, H ₁ , H ₂ . In such a case, in step 26 of FIG. 3, to make _four pieces Q ₁ -Q 4 from the original blank 20 as shown in FIG. 9, or as shown in FIG. to make a ₁ to E _8, for each piece of two halves, can be iteratively performing progressive cutting steps of the above. The resulting basic leg 30 shown as an orthographic view in FIG. 11 has an apex angle of 45 degrees. The depth of the cut in step 26 is shallow, and as a result, the value of N becomes small.
The above procedure can be modified as necessary to produce solid segments with selected values of apex angles of less than 90 degrees, but to minimize shadowing, 45 degrees or less The following angles are preferred.

The basic leg mold 30 in FIG. 11 has a substantially wedge-shaped cross section. The basic leg mold 30 has a front surface 32 that includes an angular edge 34 at the front of the wedge. The base leg 30 has an arcuate rear surface 36 that includes a pair of angular edges 38 at the rear of the wedge, and two flat sides that are inclined relative to each other.
12 to 14 show the steps of building the basic mold 30 of the support member into a completed support member. As shown in FIG. 12, the first step is the removal of the angular edges 34, 38, thereby forming a rounded trapezoidal member 40 (shown in broken lines). Machining the front and back surfaces. The surface can be machined to a radius between 0.25 (6 mm) and 5.25 inches (133 mm). In the illustrated example, the front surface of trapezoidal member 40 is machined to a radius of about 0.35 inches (9 mm) and the rear surface is machined to a larger radius of about 1.5 inches (38 mm). A preferred ratio of the radius of the rear surface to the radius of the front surface is at least 3, more preferably 4. The advantage of a significantly different ratio is that the formation of wafer support slots on the sides with smaller radii results in less wafer shadowing that occurs during processing, while larger radii far from the wafer are mechanically rigid. Is to give.

The rounded wedge shape of FIG. 12 has two further advantages. The first advantage is that material that is wasted during machining can be minimized. A second advantage is that there are no sharper (smaller radius) shapes than the front surface on which the slot is formed. As a result, there is no sharp corner that tends to scrape off the accumulated deposit and increase the number of particles. That is, the rounded shape increases the adhesion to any film deposited on the fixture during wafer processing.
Other embodiments have a rounded wedge shape at the front and a rounded rectangular shape at the rear of the leg, all surfaces equal to or equal to the radius of the front corner where the slots are formed. Joins a bend with a larger radius.

  The radius of FIG. 12 can be machined using any machine tool. Examples of such machine tools include a plated diamond router manufactured by National Diamond Lab, or a resin-bonded diamond wheel equipped with a water swivel. Such a machine tool has been found to achieve excellent results when used with a bit such as a custom resin bonded diamond wheel from National Diamond Lab.

  The basic leg mold 40 is machined to form a plurality of wafer holding slots 16 as shown in FIG. The slot 16 is formed along one side of the basic leg mold 40, here the front face, to form the slotted leg 42. The slots 16 are parallel to each other and are substantially perpendicular to the longitudinal axis A of the basic leg mold 36 of FIG. The slot 16 extends along a substantial portion of the length L except for the end length E. The slot 16 may extend longer or shorter than half the depth D of the base type of the support member, depending on the particular application. The illustrated slot 16 has two flat parallel faces, but more complex shapes may be advantageous, and such complex shapes can be easily formed by standard machining techniques. it can.

  The slot 16 can be formed by any suitable diamond tool, such as a commercially available diamond saw blade. An example of a suitable cutting device is a 3-4 inch resin bonded wheel manufactured by National Diamond Lab. Such a slotter is particularly effective when used in conjunction with a blade having a diamond-coated cutting surface, such as a 16-400 grain size blade, manufactured by National Diamond Lab. During the formation of slot 16, the slotter can be operated at a speed in the range of 5-125,000 rpm. A speed of about 4500 rpm has been found to be particularly effective. While it would be advantageous to use the specified cutting device, it should be understood that the cutting device is merely exemplary. Other cutting devices can also be used to achieve acceptable results. Examples of such devices include, but are not limited to, slotters that use diamond-free blades, lasers, and abrasive fluid cutting devices.

  There are two different designs for attaching the legs 12 to the base 14. As shown in the side elevation view of FIG. 13 and the orthographic view of FIG. 14, the slotted leg 42 has a cross section that extends to its rounded wedge-shaped end 44 except for the slot 16. To accommodate the rounded wedge-shaped ends of the legs 42, complementary, blind, round, wedge-shaped holes 46 are formed in the two opposing bases 14 by machining. By doing in this way, there exists an advantage that rotational rigidity increases.

  Alternatively, as shown in the side elevation view of FIG. 15 and the orthographic view of FIG. 16, at least one mounting structure 50 is machined on each end 52 of the leg 54 with the pile. Formed. The mounting structure 50 is shown as a set of opposed cylindrical pegs. The peg is formed by machining the end of the basic mold 30 of the support member. This machining can be performed by any suitable cutting mechanism. An example of a suitable cutting device is a vertical or horizontal milling machine or CNC machine made by a manufacturer such as NOVA, JET or PRESTO. Mounting structure 50 is adapted and assembled to facilitate attachment of legs 54 to a generally flat base 14 having a corresponding blind cylindrical mortise 56. For example, the mounting structure 50 may have other shapes, such as a blade shape corresponding to the groove shape of the base. It should be understood that the legs can be attached to the base in any suitable manner.

  The finished product of the leg 22 can be made to any desired size. For example, the illustrated leg having a wedge angle of about 22 degrees 60 seconds can be formed from a basic leg mold having a width of about 0.475 inches (12 mm) and a length of about 45 inches (114 mm). . The slot 16 in the illustrated embodiment has a depth of about 0.25 inch (6 mm) and extends a distance of about 43 inches (109 cm) along the entire length of the leg 22. This distance can be longer. The cylindrical peg 50 comprising the mounting structure of FIG. 16 extends from the end of the leg 54 about 0.6 inches (15 mm) and has a diameter of about 0.4 inches (10 mm).

  17-23 show various techniques for attaching the legs to the base. In each of these examples, as shown in FIG. 17, a first silicon member 60, which is a support member in this view, includes an attachment member or tenon 62 that extends outward. However, it should be understood that an extension of the first silicon member 60 could be used instead, as shown in FIG. The second silicon member 64 which is a base member in this figure includes a mortise 66. The mortise 66 may surround at least a part of the mortise 62.

In some embodiments, a first lateral bore 70 is drilled into the mortise 62 and a second lateral bore 72 is mortiseed to securely secure the mortise 62 within the mortise 66. A portion of the second silicon member 64 adjacent to 66 is drilled. The first and second sideways bores 70, 72 are sized to accommodate the locating pin 74 when the bores are coaxially aligned with each other.
The locating pin 74 can be secured within the first and second lateral bores 70, 72 in several ways. In the embodiment of FIG. 17, the locating pin 74 is slightly longer than the sum of the lengths of the first and second bores 70,72. The locating pin 74 is inserted into the aligned bores 70, 72 such that a portion 76 of the locating pin 74 (shown in dashed lines) extends outwardly from the outer limit 78 of the second bore 74. Thereafter, the portion 76 extending outwardly of the locating pin 74 is machined to be flush with the outer limit 78 of the second bore 72.

FIG. 18 shows another fixation technique. The locating pin 80 has an outer diameter D ₁ that is substantially equal to the inner diameter D ₂ of the first and second bores 82, 84 in a portion of the second silicon member 64 adjacent to the tenon 62 and tenon hole 66. In this example, the positioning pin 80 is cooled to −100 ° C. at a very low temperature, and therefore the positioning pin 80 contracts. The cooled positioning pin 80 is inserted into the alignment bores 82, 84 with the bore maintained at ambient temperature. Thereafter, the temperature of the positioning pin 80 returns to the ambient temperature and the positioning pin 80 expands. The diameter of the locating pin 80 shrinks by about 0.001% when cooled at cryogenic temperatures.
As an alternative to the technique of using locating pins, energy can be supplied to the mortise, mortise or both to fuse the mortise and mortise.

FIG. 19 shows an example of the above technique. In this example, the access hole 88 is
The second silicon member 64 is drilled to reach the bottom of the mortise 66. Laser energy is supplied through the access hole 88 to perform a tack weld W between the mortise 62 and the mortise 66. It has been found that to obtain a favorable result by the laser energy from CO ₂ lasers 250W pulse width 30ns and pulse period 0.001 seconds irradiation 1-5 minutes. It can be used Coherent Inc. of CO ₂ any suitable laser energy source such as a laser.

  FIG. 20 shows a sectional view of another assembling method. In this embodiment, tenon 62 and tenon 66 extend through the second silicon member with substantially the same extent. Laser energy may be supplied to the region A1 adjacent to the interface between the mortise 66 and the tenon 66 to form a tack weld that can extend around all or part of the interface region A1. it can. Any suitable laser energy source can be used.

  In the other example shown in FIGS. 21 and 22, the tenon 62 extends through the second silicon member 64 and includes a peripheral protrusion 90 at its end 92. The second silicon member 64 includes a peripheral inclined portion 94 that surrounds the outer end portion of the mortise 66. Thermal energy is supplied to melt the peripheral convex portion 90 of the first silicon member 60 into the peripheral inclined portion 94 of the second silicon member 64, thereby fusing the tenon 62 with the tenon 66. Any suitable heat source can be used to supply the heat. One heat source that can be used to advantage is known to be the heat generated by the laser energy. Laser energy heats the protrusion 90 for about 3 minutes to a temperature above the melting point of silicon, 1416 ° C., preferably about 1450 ° C., or the molten silicon has a peripheral slope as shown in FIG. Used until part 94 is satisfied.

FIG. 23 shows another fixing technique. In this example, securing the tenon 62 within the tenon 66 includes applying an adhesive to the area A ₂ extending axially between the tenon 62 and the tenon 66. Typically, high temperature, non-staining adhesives must be used to avoid the contamination associated with the use of adhesives.
The mounting element of FIGS. 17 to 23 is substantially cylindrical and includes a mounting element receiving portion formed as a cylindrical bore. However, it should be understood that these shapes are exemplary only and that any suitable cooperating shape can be selected for the attachment elements and associated retaining portions.

The various fastening methods described above can be used not only for manufacturing wafer towers, but also for other structures where two or more pieces of silicon must be closely assembled.
In order to form the tower 10 of FIG. 1, one of the above fixation methods is applied between each leg 12 and the two bases 14. However, other fixing methods can also be used. Preferably, the number of the leg portions 12 is preferably four, but three are sufficient, and two are all right. A mortise is drilled in the two bases. It is preferable to open at an equal angular position at an angle slightly wider than 180 degrees. Since the slot 16 can widen the gap, the robot blade preferably slides the slot 16 without touching the sides of the slot 16 during insertion to insert the wafer horizontally into the tower 10; A stable state can be maintained on the bottom surface of the slot 16.

In the case of FIG. 1, the base 14 is a continuous, symmetrical circle, but the base may be more complex, although it is still substantially circular. For example, the base can include an outer flat portion and an inner opening. In addition, the monolithic base is preferably made from a single crystal of silicon, but multiple parts can be secured together to form a multiple-part silicon base. In this case, it is more preferable to use virgin poly for the base.
In the case of the above embodiment, the slot is formed in a wedge-shaped leg. However, the present invention is not limited to this. The legs and the wafer support may be otherwise shaped, such as a silicon arm that protrudes from a cylindrical or rectangular leg and supports the wafer at its far end.

10: Silicon wafer processing fixture 12: Support member (leg)
14: Base portion 16: Slot 20: Blank 22: Leg 30: Basic leg 32: Front surface 34/38: Edge 36: Rear surface 40: Member 42: Leg 44 with slot 44: End 46: Hole 50: Mounting structure 52: End portion 54: Leg 56: Cylindrical mortise 60: First silicon member 62: Tenon 64: Second silicon member 66: Mortise 70: First lateral bore 72: Second Lateral bores 74, 80: positioning pin 78: outer limit 82: first bore 84: second bore 88: access hole 90: peripheral convex portion 92: terminal end portion 94: peripheral inclined portion

Claims (5)

A silicon tower for supporting multiple wafers,
A plurality of legs made of silicon configured to removably support a plurality of wafers and formed according to a non-circular cross-sectional leg shape;
A silicon tower comprising: two bases made of silicon and provided with a hole portion having the non-circular shape for housing and fixing the leg portion. The silicon tower according to claim 1,
The silicon tower, wherein the non-circular shape includes a curved shape having a first radius and two linearly extending sides inclined with respect to each other. The silicon tower according to claim 1,
A silicon tower, wherein the legs are made of virgin polysilicon grown by a CVD method. A silicon tower for supporting multiple wafers,
A plurality of legs made of silicon configured to removably support a plurality of wafers and having a wedge-shaped cross section;
Two bases made of silicon and provided with a hole having a wedge-shaped cross section formed to receive an end of the leg,
2. The silicon tower according to claim 1, wherein the leg is fixed to the base by the hole. The silicon tower according to claim 4,
2. The silicon tower according to claim 1, wherein the legs are made of virgin polysilicon grown by a CVD method.

Images