TWI251039B - Semi-pure monocrystalline silicon wafers, pure monocrystalline silicon wafers and a Czochralski puller - Google Patents

Abstract

A silicon ingot is manufactured in a hot zone furnace by pulling the ingot from a silicon melt in the hot zone furnace in an axial direction, at a pull rate profile of the ingot from the silicon melt in the hot zone furnace that is sufficiently high so as to prevent interstitial agglomerates but is sufficiently low so as to confine vacancy agglomerates to a vacancy rich region at the axis of the ingot. The ingot so pulled is sliced into a plurality of semi-pure wafers each having a vacancy rich region at the center thereof that includes vacancy agglomerates and a pure region between the vacancy rich region and the wafer edge that is free of vacancy agglomerates and interstitial agglomerates. According to another aspect of the present invention, the ingot is pulled from the silicon melt in the hot zone furnace at a pull rate profile of the ingot from the silicon melt in the hot zone furnace that is sufficiently high so as to prevent interstitial agglomerates, but is also sufficiently low as to prevent vacancy agglomerates. Accordingly, when this ingot is sliced into wafers, the wafers are pure silicon wafers that may include point defects but that are free of vacancy agglomerates and interstitial agglomerates.

Description

1251039 A7 B7

Ministry of Economic Affairs, Intellectual Property Bureau, employee consumption cooperative, printing

Regarding the case, the right to apply for US application No. 60/063,086 forms the method of semiconductor casting and the ingots and wafers formed by it," application Ε (4), January 24, and Korean application, 9'5 Na No. 2, October 24, the second case is incorporated herein by reference. In the field of the invention, the present invention relates to a microelectronic manufacturing method and apparatus, in particular to the method of cutting and casting Ingots and wafers. Inventive circuits are widely used in consumer and commercial materials. Integral circuits are usually made of early quartz. As the integrated density of integrated circuits continues to increase, for integrated products. The circuit provides high-quality single-crystal semiconductor materials, which is of great importance. The integrated circuit often cuts into large wafers by manufacturing single crystals, and then cuts into large numbers of microelectronics manufacturing processes. Then, the wafer is die-cut into a packaged individual integrated circuit. The purity and crystallinity of the wafer have a great impact on the performance of the final integrated circuit device, so that less than 1⁄4 of each ingot and crystal are produced. There have been many efforts on the circle. The method of the method of Shi Xizhuan. The retrospective of these methods = textbook "VLSI World Cut, _ volume, process technology, the first chapter, author wolf and Tauber, 1986 Page (', the chrome valley is included here for reference). In the manufacture of single crystal, the electron polycrystalline hair is converted into a single crystal ingot. Polycrystalline crystals such as quartzite are refined to form electronic grade polysilicon. Subsequently, the refined electronic grade polycrystalline silicon is grown into a chalcura (cz) and floating surface (FZ) technique.

The paper size is applicable to the Chinese National Standard (CNS) A4 specification (210 X 297 mm)

• I .^1 · ->1 (Please read the back of the article; tti means ^^ fill out this page} II _ Book--- 1251039 A7 B7 V. Invention Description (2 Zhihui Property Officer Fee

S crystal. Since the present invention is based on the pure ingot casting using the cz technique, the technique will now be described. The CZ growth is related to the crystallization solidification originating from the liquid phase at the interface; In particular, the electronic grade polycrystalline lithofate feed is loaded into the net and melted. Hai feed. Let the seed with the correct orientation tolerance (4) fall into the glare. The seed crystal is then discharged at a controlled rate in the axial direction. The seed crystal is usually rotated in the opposite direction during the pulling process. The initial pull rate is usually faster, resulting in a fine stone neck. Then, the T substrate stabilizes the melt temperature' to form the desired diameter. This path is usually maintained by controlling the pull rate. The pulling is continued until the feed is nearly exhausted, at which point the tail is formed. The first picture shows the CZ pulling the crying + standing solid renting. As shown in Fig. 1, the cz 2 device 100 includes a furnace, a crystal pulling mechanism, an environmental controller, and a computer-based control system. Cz^ and the heating element...4' quartz plug-in (four) and the yoke 112 made of the first direction, the second fault 106' graphite, which is suspected to rotate (as shown). Well 114 can provide additional heat distribution. The crystal pulling mechanism includes a crystal pulling axis m which, as shown, rotates in a direction 122 of the direction 112. b 八 124, Festival clock day 0 rape pull shaft 12G holding seed crystals Faraway species 124 series are pulled out in the hanging (3), forming a casting spin 128. W 进 Feed The nuclear control system consists of chamber enclosures i3Q, _ port and other flow controllers and vacuum ventilation systems (not shown). The main control system can be used to control heating elements, pullers and other electric devices and to pay attention to the 敎 • έ can be species 132 brain gas and ^ ft - middle _ fresh (CNS) A4 1251039

Ministry of Economic Affairs, Intellectual Property Bureau, employee consumption cooperative, printing

The real bismuth ingot is different from the ideal single crystal enamel because the real bismuth ingot contains defects or flaws. These defects are not expected in the manufacture of integrated circuit devices. These defects are usually classified as point or agglomerates (three-dimensional space 瑕疵). There are two main categories of points: gap points and gaps. In the void point, a germanium atom is lost by the normal position of the crystal lattice. This gap creates a void point. On the other hand, if an atom appears in the amorphous lattice position (interstitial position) of the germanium crystal, the interstitial point is caused. The spot is typically formed on the interface 130 between the molten crucible 126 and the solid crucible 128. However, as the ingot 128 continues to be pulled, the interface portion begins to cool. The void point 瑕疵 and the interstitial point 瑕疵 diffusion during cooling may cause enthalpy coalescence to form void aggregates or interstitial aggregates. The agglomerate is a three-dimensional (large) structure caused by the coalescence of the point. The gap filling point is also called translocation D or D-瑕疵. Sometimes, agglomerates are also named after the technique used to detect such defects. Therefore, void agglomerates are occasionally referred to as crystal-derived particles (COP), laser scattering tomography (LST), or flow pattern enthalpy (FPD). Interstitial aggregates are also known as large translocation (L/D) aggregates. For a discussion of the 矽 of single crystal 参考, refer to Chapter 2 of the aforementioned textbooks by Wulf and Tauber (the disclosures of which are incorporated herein by reference). It is known that many parameters need to be controlled, and then there is a small amount of defects (please read the back note and then fill out this page) I· -Installation · This paper scale applies to China National Standard (CNS) A4 specification (210 1251039 A7 B7

Ministry of Economic Affairs, Intellectual Property Bureau, employee consumption cooperative, printing

Still ingots. For example, it is known to control the rate of pull of the seed crystal and the temperature gradient of the hot section structure. According to the Fronkov theory, the ratio of 乂 to g (called V/G) determines the point 铸 degree of the cast bond, where v is the pull-up fan of the cast bond, and G is the cast (four) melt interface. Temperature gradient. The Fronkov theory is detailed in the "Shi Xixuan Manchu (four) into the mechanism,", in the author, Fronkov, Journal of Crystal Growth, Vol. 59, 1982, pp. 625-643. 曰/Fronkov Theory - The application can be found in the inventor's literature, entitled “The influence of the characteristics of the crystal contact device,” Proceedings of the 2nd International Symposium on Materials Science and Technology, 25-29 November 1996, p. SB. The fifteenth diagram of the document (herein reproduced in Fig. 2) shows a schematic illustration of the WG function as a function of the void concentration and the interstitial concentration. The Fronkov theory shows that the generation of the void/interstitial mixture of the 曰a circle is determined by V/G. In detail, the V/G ratio below a critical ratio creates a gap-filled prayer, while a WG ratio above a critical ratio creates a void-rich ingot. Although physicists, materials scientists, and others have conducted many theoretical studies and CZ puller manufacturers have done a lot of practical research, reducing the germanium density in single crystal germanium wafers is still a continuing need. The ultimate need is for pure stone wafers that do not contain void agglomerates and interstitial aggregates. SUMMARY OF THE INVENTION The present invention is directed to a method of making a bismuth ingot in a hot zone furnace by plotting a draw rate of an ingot derived from a bismuth melt in the hot zone furnace From each of the ingots in the hot zone furnace, the ingot is drawn along the axial direction, the pulling rate is sufficiently high to avoid interstitial aggregates, but also low enough to agglomerate the voids Body is limited to the ingot

This paper scale applies to the Chinese National Standard (CNS) A4 specification (210 X 297 mm)

-· Μ m / ϋ I nr Please read the back 2 Ϊ 填写 填写 填写 i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i Inside. Subsequently, the thus-drawn ingot is cut into a plurality of semi-perfect wafers having a void region including a void agglomerate at a center thereof, and a gap region between the void-rich region and The pure shell d ' between the edges of the wafer does not contain empty p-aggregates and interstitial aggregates. The present invention is based on the understanding that an agglomerate is formed from a point enthalpy when the concentration of a certain enthalpy exceeds a certain critical concentration. If the concentration of the plague at the (void or interstitial) point can be maintained below this critical concentration, the aggregate will not form when the prayer is pulled. In order to maintain the point peek concentration below the critical point enthalpy concentration, the ratio of the pulling rate to the temperature gradient (V/G) at the ingot-melt interface is limited to be higher than the ingot_melt. a first critical ratio of the pull rate to the temperature gradient at the interface, the first critical ratio needs to be maintained to avoid interstitial aggregates; and (2) the pull rate below the ingot-melt interface For the second critical ratio of the temperature gradient, the second critical ratio is not excessively large, and the void agglomerates are confined to one of the void-rich regions at the center of the ingot. Therefore, when the ingot is pulled out of the crucible melt of the hot zone furnace, the graph of the pull rate is adjusted to maintain the pull rate versus the temperature gradient above the first critical ratio and below Two critical ratio. According to another aspect of the present invention, it is provided by one of the hot zone furnaces by a drawing rate of the ingot derived from the melt of one of the hot zone furnaces. In the crucible melt, the ingot is pulled out, the pulling rate is sufficiently high to avoid interstitial aggregates, but also low enough to avoid void collection. Thus, when the ingot is cut into wafers, the wafer is a pure tantalum wafer that includes a bit of fatigue but does not contain void aggregates and interstitial agglomerates. According to this aspect of the invention, it has been determined that if the V/G ratio is limited to the paper size, the Fia standard (CN) A4 specification (21Q 297 297 cc) is applied (please read the following notes: This page) I· Armored J. Description of invention (6) In a narrower range, both the point gap concentration and the point void concentration can be maintained below the critical point concentration of the agglomerate. Therefore, the whole ingot may not contain agglomerates. To 幵 y into pure stone eve, it is necessary to maintain the first criticality of the first critical ratio of the temperature gradient of the pulling rate change diagram of the ingot smelting interface The ratio of the pull rate change diagram of the ingot-melt interface to the second critical ratio of the temperature gradient cannot exceed the second critical ratio to prevent the void aggregate. Then the pull rate change is determined. The graph can be maintained when the prayer key is pulled by the crucible melt of the hot section furnace, and the temperature gradient ratio is higher than the first critical ratio and lower than the second critical ratio. - The pull rate of the melt interface is compared with the temperature gradient The radial temperature gradient and the axial temperature gradient are considered between the two critical ratios. In the radial direction, the temperature gradient across the wafer is usually changed because the wafer is in contact with different thermal environments. The temperature gradient at the edge is usually higher than the center of the wafer, which is caused by thermal characteristics. The pull rate is usually constant across the wafer. Therefore, the V/G ratio is usually reduced from the center of the wafer toward the edge of the wafer. The rate and heat zone furnace is designed to be within the diffusion length from the center of the wafer to the edge of the wafer, maintaining the V/G ratio below the critical point of the agglomerate, ie, the _th critical ratio and the second critical Compared with the room, it can be applied to the direction of the vehicle. In the axial direction, due to the thermal mass of the ingot, the gradient of the dish decreases with the pulling of the casting. So when the prayer is pulled, the pulling rate must be reduced. Slow to maintain the V/G ratio between the first and second critical ratios. Therefore, by controlling the pull rate graph, the V/G is maintained at the Intellectual Property Office of the Ministry of Economic Affairs.

A7 B7

1251039 Between the two critical concentrations, a semi-pure beijing circle with a void-rich region at the center and a pure region between the void-rich region and the edge of the wafer can be generated. The void-rich zone may include void spots and aggregate enthalpy, while the pure zone does not contain any void aggregates and interstitial aggregates. Preferably, a pure wafer can be formed which includes a little fatigue but does not contain void aggregates and interstitial aggregates. The first and second critical ratios can be determined by experiment or simulation. The ratios can be empirically measured by cutting a reference ingot into a plurality of wafers or axially cutting the reference ingot. Experimental and simulation techniques can also be applied in combination. In particular, the first and second critical ratios can be experimentally extracted from the crucible melt of the hot zone furnace by pulling at a pulling rate in a range of a pulling rate. Sexually measured. Subsequently, the reference ingot is cut into a plurality of wafers. For semi-pure wafers, a wafer with a pre-size of g-containing void regions and no interstitial agglomerates can be identified. Preferably, a crystal circle having a minimum void-rich region and no interstitial aggregate is identified. The first and second critical ratios of the semi-pure wafer are calculated from the extracted wafer's pull rate and the location of the identified wafer in the ingot. # To determine the first and second critical ratios of a pure wafer, the reference wafer can be pulled into several wafers and the wafers containing no void agglomerates and gap agglomerates can be identified. The first and second critical ratios of the pure wafer are calculated from the lift rate of the identified wafer and the location of the identified wafer in the ingot. The reference ingot is preferably at a pull rate that varies in a certain pull rate range

Bian's Zhang scale applies to Chinese national standards (CNSU4 specifications (10)

1251039 A7 V. INSTRUCTIONS (8) The Ministry of Economic Affairs, Intellectual Property Office, Staff Consumer Cooperative, printed by the 矽 melt of the hot zone furnace, the second pull rate of the first pull rate lower than the first pull rate a third pull rate that is higher than the second pull rate but lower or higher than the first pull rate. The first, second and second pull rates are preferably based on the desired ingot diameter and the desired V/G The linear change in the pull rate is preferably used to determine the first and second critical ratios. In another experimental technique, the reference ingot is heated at a pull rate that varies over a range of pull rates. The zone furnace is pulled out of the melt. Subsequently, the reference ingot is axially cut. For semi-pure wafers, at least one of the ingots that have been axially cut is identified as having the least The void-containing region does not contain the axial position of the interstitial aggregate. Next, the first and second of the semi-pure wafer are calculated from the corresponding pull rate of the axial position identified in the axial cut casting: Critical ratio. In order to produce a perfect wafer, in the axially cut ingot, At least one axial position without the void agglomerate and the interstitial aggregate. The tensile rate from the identified axial position and the position of the axial position are different from the first wafer of the perfect wafer. The second critical ratio. The first and second critical ratios can also be determined using the analog mode theory. The first and second critical ratios can be identified by the Fronkov theory. The pull rate change graph can be used to simulate the radial temperature. The operation of the specific hot section of the pulling process is determined. The pulling rate change graph determines the axial temperature can be determined by the operation of the hot section furnace when the simulated ingot is pulled. The pulling rate can be maintained to be higher than the ingot temperature gradient ratio. - the critical ratio, and lower than the second critical ratio of the pull rate change curve 11, which can then be determined by the analog pull rate audible temperature map and the mode pull rate versus the axial temperature map. First read the note on the back: fill out this page) Loading ·- *k_ 11 1251039 A7 B7 V. Description of invention (9) It is also necessary to understand that the first and second critical ratios of pure stone eve can be identified by a two-step method. By inspection and / or theory The first and first critical ratios of the semi-pure enthalpy are determined. The thermal segment structure is then modified until the first and second critical ratios are determined experimentally and/or theoretically. The present invention provides a plurality of similar halves made from tantalum ingots. Pure single crystal germanium wafers. Each semi-pure quartz wafer has a void-rich region at the center, which contains void collectors and a pure region between the void-rich region and the edge of the wafer, which does not contain void agglomerates and interstitials. Aggregates. The gap-rich areas of each wafer are generally straight and equal. Fortunately, the pure shell area accounts for at least 36% of the wafer area. The better pure area accounts for at least 60% of the wafer area. If the V/G is tighter. In contrast, the present invention can produce a plurality of pure single crystal germanium wafers fabricated from a single germanium slave, wherein each pure germanium wafer comprises void aggregates and interstitial agglomerates. Thus, by maintaining the pulling rate against the ingot_melt The temperature gradient of the interface is between the upper and lower limits, which can limit the aggregates to the void-rich areas in the center of the wafer or eliminate the agglomerates to produce pure germanium wafers.简单 Simple description of the figure Fig. 1 is a representative diagram of the sound of the Chai Kelasji puller for growing single crystal stone ingots. ~ Figure 2 illustrates the Fronkov theory.苐3 A-3 E illustrates an overview of the fabrication of wafers with a void-rich region in the center and a pure region between the voided region and the edge of the wafer. Fields 4A-4E illustrate an overview of wafer fabrication without agglomerates. 7 paper scale applies to China National Standard (CNS) A4 specification (210 X 297 mm) (please read the note on the back: fill out this page), loading wheel · Ministry of Economic Affairs Intellectual Property Bureau employee consumption cooperative printing 12 1251039 5 OBJECT OF THE INVENTION Figure 5 illustrates an example of a money change in accordance with the present hair rate. Theory ^Decision Figure 6 illustrates a graph of the use of an axial sag pull rate as a function of the present invention. , 'X-type Figure 7 illustrates an example of a graph of the pull rate change according to the present invention. Ming Department by wafer identification experimental method Ministry of Economic Affairs Intellectual Property Bureau employee consumption cooperative printing Figure 8 illustrates the combination of simulation, chip and wafer identification for manufacturing ingots according to the present invention. Fig. 9 is a schematic representation of a perfect Shih's Chaucer puller according to the modification of the present invention. Figure 10 illustrates an example of changing the pull rate to determine the preferred pull rate in accordance with the present invention. Figure 11 is a representative view of an X-ray tomographic image illustrating the void-rich region of the first reference ingot according to the present invention, rich in interstitial regions and perfect regions. Figure 12 is a representative image of an X-ray tomography image, illustrating an example of a void-rich region of the first reference spine according to the present invention, rich in interstitial regions and a perfect region. Figures 13 and 14 illustrate plots of pull rate changes for the addition of void-filled wafers and perfect wafers in accordance with the present invention. Component label comparison table Woodcoras (CZ) puller 106 internal disaster 102' 104 heating element 1〇8 External collapse + paper scale applicable Chinese National Standard (CNS) A4 specification (210 X 297 mm) X λ L; 轴向 Axial cutting (please read the note on the back: fill out this page) ·!· -

-n / I . _轮· 13 1251039 A7 Ministry of Economic Affairs Intellectual Property Bureau Employees Consumption Cooperatives Printing V. Inventions (11) 128 Ingots 130 Chamber enclosures, interface 132 Cooling ports 502-514, 602-612, 702-712, 802 DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT OF THE INVENTION The present invention will be described in more detail hereinafter with reference to the accompanying drawings in which FIG. The present invention may be embodied in a variety of different forms and is not to be construed as limited to the details of the details. Similar numbers throughout the text indicate similar elements. Summary ··Enriched voids and perfect wafers Now refer to Figure 3A-3E, which illustrates the fabrication of semi-pure wafers. The wafer has (1) a void-rich region at its center containing void aggregates, and (7) The pure region rich in the void region and the edge of the wafer does not contain void aggregates and interstitial aggregates. As shown in Figure 3A, the fabrication of void-rich wafers can begin with a review of the Fronkov theory. The Fronkov theory is illustrated in the figure. As shown by the straight line starting at the edge (E) and ending at the center (C), it is found that the temperature gradient ratio (referred to as V/G) of the right draw rate to the ingot melt surface according to the present invention is maintained greater than by the edge E (marked For point a) diffusion length (v/g)i and small 110 rotation axis 112 first direction 114 heat shield 120 crystal pulling shaft 121 reverse 124 seed crystal 126 fire valley melt polycrystalline crush feed ----- -------装--- (Please read the note on the back: fill out this page) . - : The paper size applies to the Chinese National Standard (CNS) A4 specification (210 x 297 mm) 14 1251039 Ministry of Economic Affairs Intellectual Property Bureau Staff Consumer Cooperative Printed A7 B7 V. Invention Description (12) At Center C (V/G) 2, a semi-pure wafer can be fabricated which has a void-rich zone and a void-rich zone at the center. There is a pure zone between the edge of the wafer. In particular, V/G can be changed in the radial direction of the ingot across the wafer. V/G is usually reduced from the center of the wafer toward the edge of the wafer due to differences in thermal characteristics at the center and edge of the wafer. Such a specific wafer appears from the center (C) to the edge (E) in the radial V / G range shown in Fig. 3A. Tantalum ingots and wafer fabrication are of considerable concern for the formation of agglomerates within the wafer, including voids or interstitial aggregates. It is known that an agglomeration system is formed by coalescing of dots formed when an ingot is originally produced from a molten body. The point concentration is usually determined by the interface conditions between Shixi ingot and Shishi melt. Then, as the casting spin is further pulled, the diffusion and cooling determine the point of fatigue and coalescing to form an agglomerate. As shown in Fig. 3B, it was found that according to the present invention, there is a critical void point [ concentration [V]* and a critical interstitial point 瑕疵 concentration [Q], below which the point 瑕疵 does not coalesce to form an aggregate. It has been found that, according to the present invention, if the concentration of the spot is maintained below the critical concentration of the peripheral region of the wafer, a void-rich region is formed at the center of the wafer, but a pure region is formed between the wafer-rich region and the edge of the wafer. Thus, as shown in Fig. 3B, the gap density across the wafer is maintained below the critical void concentration [V]*, except for the center c. Thus, as shown in the redundancy diagram, 'the void-rich region [v] is formed in the central region but the void-rich region [V] does not contain void aggregates in the outer edge of the wafer edge and is therefore labeled as [p] (pure or perfect) . Referring again to Figure 3B for interstitial, the interstitial concentration is maintained below the critical interstitial concentration [1]* from the center of the wafer C to the diffusion length L from the edge E of the wafer corresponding to point A! The diffusion length Li between the wafer and the edge B, even if the interstitial concentration is initially higher than the critical concentration (1)* of the cast bond's melt interface, expansion (please read the note on the back: fill this page) - install 15 1251039 V. INSTRUCTIONS (13) The interstitial voids can be diffused from the ingot, and the moon will emerge without crystals during the crystal growth process. The diffusion length L 丨 is 8 吋 yen and _, s ^ is from a 曰曰 曰曰 0 and each is usually about 2.5 to 3 cm. Thus, as shown in Fig. 3C, the shape of the 士士/丄α _ " ° / semi-pure Japanese yen has a rich gap area [V] at its center and between the rich margins of the field Wu District [p]. The pure area [P] accounts for at least 36 〇 / 口 of the wafer area and preferably at least 60% of the wafer. Resolving the wafer forming the 3C pattern, V / G The dimension a must be greater than (V/G), and the middle to C is less than or equal to (v/g) 2. To maintain the μ ratio between the two critical values, two thermal conditions must be considered. The radial temperature gradient G at the wafer center C to the wafer diffusion length a must be maintained within this equivalent range. Thus, the center V/G must be close to (v/g) 2俾 to limit the void agglomerates in the void-rich In addition, the WG at the distance from the edge diffusion length L1 must be maintained greater than (V/G)| to prevent interstitial agglomerates. Such a hot segment furnace is preferably designed to maintain the change in G from wafer center to wafer diffusion length. V/G can be maintained between (¥/(})2 and (¥/(5)1. The second consideration of the Ministry of Economic Affairs Intellectual Property Bureau employee consumption cooperatives is that it starts at the seed crystal and ends at the tail. Round pulled by molten body G / Mouth from the change to Ai. The test temple does not increase the heat of the casting spin, > salt less melt heat and /, his heat test can make G fall when the ingot is pulled by the melt. So to maintain V / G Between the first and second critical ratios, it is necessary to adjust the graph of the pulling rate change when the prayer is pulled by the crucible melt of the hot section furnace. The void aggregate can be obtained by controlling V/G when the ingot is pulled up. Restricted to the void-rich region [V] close to the ingot axis A, as shown in Fig. 3D. No interstitial aggregate is formed, so the ingot region outside the void-rich region [v] is indicated by [p],屯, 凡 or 凡美. Also as shown in Figure 3D, obtain a number of semi-pure crystals at the midday paper scale for the China National Standard (CNS) A4 specification (21Q X 297 mm) 1251039 V. Ministry of Economic Affairs Intellectual Property Bureau Employees' Consumer Cooperatives Printed Inventions (14), with a void-rich zone containing void agglomerates and a gap between the edges of the wafer. f(4) Tail Touch, Field 3 Gap Fields, Occupational Fruits, and Interstitial Gathering The pure area of the body. Μ--- (Please read the note on the back and fill in this page) The diameter of the gap is m. Equivalent. A plurality of wafers formed by a single _ may be represented by an id number, and the 3D icon is shown as m, which is usually a text number marked on each wafer. The 18-character domain identification wafers are all from a single ingot. Figure 3E illustrates a plot of the pull rate change used to maintain V/G between two critical ratios when the ingot is pulled from the melt.

Ik耆 ingots are reduced by the pulling of the melt, usually slowing down the pulling rate v

Holding WG is between two critical ratios. In order to obtain the expected processing changes, the V/G rut is in the middle of the first and second critical ratios. It is thus preferable to maintain the guard band to allow for processing variations. Summary: Pure germanium wafers Figures 4A-4E correspond to the 3A-3E diagram illustrating the control of the pull rate: pure stone casting and wafer. As shown in Fig. 4A, if v/g is maintained within a tight tolerance between the center C of the circle and the diffusion length a from the edge E of the wafer, void formation and interstitial accumulation can be prevented throughout the wafer. ^ As shown in Fig. 4B, at the wafer center (ingot axis A), the v/g ratio is maintained below the critical ratio (V/G) 2 at which the void collector can be formed. Similarly, V/G maintains a critical ratio (v/g)i that can form a gap-filled agglomerate. Thus, the formation of the figure, 屯石夕 [P] does not contain interstitial aggregates and void agglomerates. The pure bond is shown in Figure 4D along with a pure wafer. The graph of the pull rate change of pure Shi Xi is shown in Fig. 4E. X 297 mm) t paper scale China] 17 051039

The determination of the pull rate change graph according to the present invention is to form a semi-pure wafer having a void-rich region at the center and a pure region between the void-rich region and the edge of the wafer, and determining a pull rate change graph ( Fig. 3E) The temperature gradient ratio of the pulling rate to the ingot is maintained to be higher than the first critical ratio and lower than the second critical ratio when the ingot is melted by the hot zone furnace. Similarly, in order to form a pure crucible containing a point enthalpy but no void agglomerates and interstitial aggregates, the graph of the pulling rate change (Fig. 4E) is determined when the ingot is pulled by the crucible melt of the hot section furnace. The pull rate to temperature gradient ratio is maintained above the first critical ratio and below the second critical ratio. The decision of the pull rate change graph will now be described. The pull rate change profile can be determined theoretically by simulation, experimentally determined via axial slice reference ingots, experimentally determined by slicing the reference spins into wafers, or determined by a combination of such techniques. In addition, the curve of the pull rate of the pure stone can be determined by first determining the pull rate curve of the semi-pure stone and then modifying the heat segment structure to obtain the pure pull rate curve. These techniques are now described. (Please read the note on the back first) Fill in this page)

I 装 _ Ministry of Economic Affairs Intellectual Property Bureau employee consumption cooperative printing

The graph of the pull rate change by simulation is shown in Fig. 5. Now, the graph of the pull rate change is determined by the simulation theory. As shown in Fig. 5, the commercial simulation software can be used to simulate the V/G variation of block 502 (referred to as. Then, at block 5〇4, it is determined whether the V/G variation from the center to the edge diffusion length L1 is small enough. Can meet the criteria for forming semi-pure wafers or pure wafers. Especially for the example of Figure 3C, this paper scale applies to China National Standard (CNS) A4 specification (210 X 297 mil) 18 1251039 A7 V. Invention Description (16 In the case of a region rich in voids, the Δ between the wafer radius d and the radius a (ν/〇 must be between (V/G) 丨 and (V/G) 2. In other words, fill The 瑕疵 point concentration must be less than [Q] * for the wafer between the centers c and a, and the groove radius should be less than [V] * for the wafer radius greater than d. Similarly, as shown in Figure 4. As shown, to form pure enthalpy, Δ(ν/(}) must be less than or equal to (v/(7)^ minus (V/G) 1俾 maintain [V] below the critical concentration [V]*, from center c to The full extent of the expansion distance a must also be maintained (1) below the critical concentration [〗]*. Follow the 5th picture of the 5th, at block 5〇4, if the V/G radial is determined at block 5〇2. Change too large to meet semi-pure wafers or The condition of the wafer (Fig. 4D), the hot section can be modified at block 5〇6 and re-simulated until the gradient is small enough to meet the desired condition. Particularly as shown in Fig. 9, the hot section can be as follows Modify, add cover 914 to heat 4 114 to fill the space between cover 914 and hot section screen 114 with a heat retaining material such as ferrous carbide. Other hot section modifications may be made to reduce the temperature gradient as needed. In the figure, the V/G axial change mode is performed at block 5〇8 to determine the Δ((change) when the ingot is pulled. Again, at block 51, it is understood whether the change is small enough to maintain the desired characteristics as the wafer grows. If not, then the hot section is modified at block 506. Then at block 512, the pull rate profile is determined to maintain the critical v/(} as shown in the figure or the map. Then use this trick at block 514. Ingots. The preferred pull rate curve is used in block 512 to maintain V/G midway between the two critical ratios' thus maintaining the guard band and considering typical process variations. Lifting rate change graph g paper ruler money and money standard (cns) A4 19 V. INSTRUCTIONS (Π) Now refer to Fig. 6 'Description of the use of the shaft, ", as shown in the figure. Reference to the heart =: rate pull. To determine the preferred pull rate, such as the first. The use = rate range. As shown in Figure 10, the pull rate is adjusted from a high pull rate (4) ^•2 milliseconds to a low pull rate (4) Q5 mm/min and a pullback pull rate. Low to 0.4 mm/min or less. The change of the pulling rate (b) and (4) is preferably linear. The cast m1 and 12 drawings with the profiles shown in the first and the 12th are respectively illustrated as examples of the enrichment of the prayer key. The void region is rich in interstitial and perfect regions [v], m and [p]. Industry insiders understand the concentration of various oligomers in these areas that are not shown in the line diagrams in Figures n and 12. ^ Referring back to Figure 6, the ingot is sliced in the axial direction of the block 6〇4. Thus, referring to Fig. 11, for semi-pure sputum, the ingot is sliced in the axial direction, and the axial section is measured using conventional techniques such as copper decoration, Secc〇-etching, X-ray tomography analysis, life measurement, and other conventional techniques. Aggregate concentration. Preferably, the agglomerates are cut in the axial direction, mirror-etched and measured at 8 Torr under nitrogen atmosphere (arc annealing for 4 hours and annealing at 1000 ° C for 16 hours). As shown in Fig. u, the axial position P! It has a large void-rich zone and a relatively small perfect zone. The axial position P2 has a small void-rich zone and a large perfect zone. The axial position P3 has a small gap-rich zone and the largest possible Perfect zone, but no rich interstitial agglomerate zone is introduced. The axial position p4 has a very small void-rich zone but a large rich interstitial zone, which is undesirable. Thus at block 608, based on the eleventh The axial position of the figure determines the V/G for the axial position p3. At block 610, the position p3 is determined to be satisfied when the wafer is pulled. -20 - The paper size applies to the Chinese National Standard (CNS) A4 specification (210 X 297) Gong) Printed by the Ministry of Economic Affairs Intellectual Property Bureau employee consumption cooperative

1251039 A7 B7 V. INSTRUCTION DESCRIPTION (18) Based on the graph of the pull rate change, then the casting spin is produced. The industry needs to understand that the axial position P3 is not used for real production because the process variation causes the interstitial zone to grow. So can I choose a location? The axial position between 2 and P3, regardless of process variations, may include acceptable small void-rich zones without the introduction of interstitial aggregates. The axial ingot slicing can also be carried out on ingots that are pulled up in the hot section, which is designed for use in pure crucibles. Such an ingot is shown in Figure 12. As shown in Figure U, it shows the void-rich zone, the perfect zone and the interstitial zone [v], m and [P]. As shown in Fig. 12, the car consists of a zone 5 in the rich position of the VP11-like position shown in the position 1VPio. The positions P? and P(7) contain a rich interstitial ring and the perfect middle position P6 and p9 are perfect because the center does not contain voids, and the gap is not filled. Thus at block 6〇6, the axial position corresponding to position h or P9 is selected and at block 6〇8 the v/G is determined for this axial position. The pull rate profile of such a V/G can be maintained at block 61, and the ingot is grown at block 612. The industry needs to know that pure 矽 can be produced by selecting the axial position range of the adjacent position & and P9. This allows for true v/g licensing process changes while maintaining perfect flaws. Determining the pull rate change curve by wafer identification Now, referring to Fig. 7, the pull rate change experiment and the graph are determined by the wafer identification experiment. As shown in Figure 7, the reference ingot is tied to the cube at a pull rate. To determine the preferred pull rate profile, a certain pull rate range is preferred as described in Figure (7). For example, the pull rate is adjusted from a high pull rate (4) (here 4 12 胄 m / min) to a low pull rate ^

$ paper ruler money relationship (21Qx297 public love 1251039

Home meter/minute and return high pull rate. The low pull rate can be as low as U mm / min or less. The change in the pull rate (b) and (4) is preferably linear. An ingot having a cross section as shown in Fig. 11 or Fig. 12 can be produced. Referring back to Figure 7, a plurality of wafers are produced by slicing in the warp direction at block 7〇4. Referring to Figure 11, for the semi-pure, the ingot is sliced to provide a representative wafer W1_W4. The aggregate concentration of the wafer, such as copper decoration, secco etching, lifetime measurement or other conventional techniques, is then measured using a known technique. As shown in the figure, the wafer W1 has a large void area and a relatively small perfect area. Wafer W2 has a smaller void-rich: and a larger perfect zone. Crystal K w3 has the smallest possible void-rich region and the most probable perfect region without introducing a region rich in interstitial aggregates.曰曰曰s % has a very small void-rich zone but has a large interstitial-rich zone. Thus at block 708, the WG of wafer 1 is determined based on the axial position of the wafer along the prayer key of FIG. At block 71G, a plot of the pull rate of the wafer W3 as the wafer is pulled is determined, and then the ingot is fabricated. Printed by the Consumer Intellectual Property Office of the Ministry of Economic Affairs. The industry insiders must understand that the axial position of the wafer W3 is not used for real production because process variations may cause the interstitial zone to grow. Thus, the axial position of the wafer between wafer W3 and wafer W2 can be selected to include acceptable small void-rich regions, but no interstitial aggregates are introduced and are independent of process variations. Wafer slicing can also be performed on ingots designed for use in the hot section of a perfect crucible. Such an ingot is shown in Figure 12. As shown in Figure u, it shows the void-rich zone, the perfect zone and the interstitial zone [V], [Q and [p]. As shown in Fig. 12, it is not possible to produce a plurality of wafers Ws and W10. Wafers W5 and Wi〇 are similar to those shown in the figure, including a void-rich central region. Wafer w? and Ws contain rich fill 22 1251039 A7 ---------- B7 _ five, invention description (2 〇 ) " 'gap ring and perfect center. However, the wafer WdW9 is perfect because the center does not contain voids and the edges do not contain gaps. Thus at block 706, the axial position corresponding to the crystal H W64 w9 is selected, and at block 7 〇 8 the V/G is determined for this axial position. At block 710, a pull rate change graph for maintaining such V/G is determined and the ingot is fabricated at block 712. Industry insiders need to know how to select a wafer location range in the field of W6 and w9 to produce Shi Xi. This selects true V/G and allows for multiple process changes while maintaining perfect 矽 characteristics. The thin 1 test and the mold technique are used to determine the pull rate. Referring now to Fig. 8, the ingot of the present invention is produced using a combination of simulation, axial slicing and wafer identification. As shown in Figure 8 and Figure 8 of Figure 8, the simulation can be used to determine a certain range of pull rates. At block 8〇4, multiple reference ingots can be grown. A number of ingots are sliced in the axial direction at block 8〇6, and a number of ingots are sliced into wafers at block 808. The optimal v/g is determined by the interaction of the axial slice, wafer identification and simulation results as determined by block 81. The pull rate is then determined at block 812 and the ingot is fabricated at block 814. This process is carried out twice to obtain a pure enthalpy, so that the heat section can be modified as needed to obtain a semi-pure enthalpy. Printed by the Intellectual Property Office of the Ministry of Economic Affairs, the Consumer Cooperatives. The true pull rate will vary depending on a number of variables, including but not limited to the diameter of the casting spin' using a specific hot zone furnace and the melt quality of the crucible. Figures 13 and 14 illustrate a plot of the pull rate change determined using a combination of simulation and experimental techniques (Figure 8). Figure 13 illustrates a plot of the pull rate of a growing ι 〇〇 cm, diameter 200 mm ingot to form a void-rich zone with a diameter of 12 PCT and a pure 矽 zone of 64%. A model Q41 hot section furnace manufactured by Mitsubishi Materials Corporation was used. Figure 14 illustrates the use of this paper scale for the Chinese National Standard (CNS) A4 specification (210 X 297 mm) 23 1251039 A7

First read the back of the note, m fill in the clothes page

Claims (1)

1251039 A8 B8 C8 D8 97 mm) Sixth, the scope of application for patents ι·: a kind of semi-pure single crystal weight, which is made up of a single semi-pure single crystal stone wafer at its center, the enriched void area of the aggregate A pure region between the void-rich region: the void-free agglomerate and the interstitial aggregate:: back, the rounded void-rich region has the same diameter.曰曰2·If you apply for the patent scope of the first half of the pure single crystal stone wafer, the wafer also contains the identification index, which is located in each, the identification index on the daily yen will be used to extract the wafers. Made from a single ingot casting. , ', 3 · For example, the semi-pure single crystal wafer of the i-th patent scope of the patent application, ... the wafer has an equal wafer area, and the pure region of the middle and the round is at least the wafer 36% of the area. 4_ : Apply for the semi-pure monocrystalline single crystal wafer of item 3 of the patent scope, in which the pure area of each Japanese yen accounts for at least 6〇% of the wafer area. 5·-a pure single crystal ^ round' is made of a single ingot, and the pure single crystal silicon wafer does not contain void aggregates and interstitial aggregates. • For example, the pure single crystal germanium wafers in the third paragraph of the patent application, each of which contains an identification index, and the identification index on each wafer means that each wafer is made of a single tantalum ingot. 7·=Growth growth single crystal cut casting bond of the Chailasaki (a. Such as (four) lifted as 'which contains: a bounding body; a beak inside the surrounding body; crystal in the body surrounded by β Hai a pulling vehicle having an axis and an axial end adjacent to the hanging horse; ^ Zhang scale suitable _ 1251039 Applying for a patented garden to pull the shaft axially away from the device; the surrounding heater is located in the enclosure and surrounds the addition: between the material and the crystal pulling shaft The heat shield has a crucible at one end and a heat retaining material inside the cover. δ·If the application of the scope of the patent scope item 7 is the Chaucer base puller: ★, wherein the heat screen is - between the (four) and the crystal pulling shaft, the soil / heat screen 'the heat screen m Extending from a first end adjacent to a second end remote from the hanger; and wherein the cover is located at a first end of the cylindrical heat shield. 9. A Chaucer base puller according to claim 8 wherein the heat retaining material is ferrous carbide. -26 This paper scale applies to Chinese national standard (〇jS) A4 specification (210X297 mm) ........................attack... (Please read the note on the back of the page first), stare at – line: 丨丨_