WO2014147262A1 - Blank made of silicon, method for the production thereof and use thereof - Google Patents

Abstract

The invention relates to the use of quasi-monocrystalline silicon, which is preferably produced according to the method of directional solidification, for high ohmic optical components for use in the infrared spectral region or for high ohmic functional components in installations for semiconductor processing, in particular for showerheads. The blanks have a concentration of individual dislocations in the range of 10² to 10⁶ cm^-2 in both a dislocation cluster and in small-angular grain-boundary-free volumes, and a specific resistance of greater than 5 Ωcm, may contain foreign grains, twin grain boundaries or clusters of dislocations of small-angular grain boundaries, and can thus be produced more cheaply and efficiently than blanks made of monocrystalline silicon according to the Czochralski or float zone methods.

Description

 Blank of silicon, process for its preparation and use thereof

The present application claims the benefit of the following patent applications, the entire contents of which are hereby incorporated by reference: German Patent Application No. 10 2013 102 983.3, "Blank made of silicon, method for its production and use thereof", filed on 22.03.2013 German Patent Application No. 10 2013 107 189.9, "Blank of silicon, process for its preparation and use thereof", filed on 08.07.2013; German Patent Application No. 10 2014 101 222.4 "Method for producing a silicon ingot and a seed layer therefor", filed on 31.01.2014; International Patent Application No. PCT / EP2014 / 055453 "Blank made of silicon, method for its production and use thereof" , registered on 18.03.2014.

Field of the invention

The invention relates generally to a blank made of quasi-monocrystalline silicon, as defined below, and to the production and use thereof, and in particular to the use of quasi-monocrystalline silicon, which is preferably produced by the directional solidification method, for optical components for applications in the infrared spectral range, in particular in the range of 1.4 μιη to 10 μιη, more preferably in the range of 1.4 μιη to 8 μιη, and for functional components in systems for semiconductor processing such as for showerheads in eg Etching, cleaning or coating equipment.

PRIOR ART Monocrystalline high-resistance optical components made of silicon and monocrystalline high-resistance functional components in silicon semiconductor processing plants are currently obtained without exception from ingots produced by the Czochralski (CZ) method or the float zone (FZ) method. Here, high resistance refers to the range of resistivity greater than 5 Ωcm, and the used range of these ingots is cylindrical. The diameter corresponds to that of ingots which are used for the production of semiconductor wafers which are used for the production of integrated circuits. The diameters of such ingots are typically limited to 150 mm, 200 mm, 300 mm or 450 mm due to the limitations of the CZ method or FZ method. At the maximum, the CZ process has so far reached diameters of up to 550 mm.

The same geometric limitations also apply to the production of high-resistance monocrystalline showerheads, such as those used in plasma etching plants, plasma cleaning plants or coating plants. Because of this limitation, semi-finished products (also called blanks or preforms) for monocrystalline optical components or monocrystalline silicon functional components in semiconductor processing plants are always specified in such a way that they can be cut out of the cylindrical silicon ingots, from which silicon is also standard - Semiconductor Wafers are produced. The cutting takes place in the radial direction, resulting in round discs whose end faces correspond crystallographically the growth direction of the crystal (typically (100) or (111)). The maximum possible semifinished diameter corresponds to the ingot diameter. However, it is also possible to cut out semi-finished products for optical elements in the longitudinal direction from the ingot. The length is limited by the Ingotlänge and the width and the thickness by the ingot diameter. The majority of all ingots are grown in the (100) direction. The restriction to the production methods CZ and FZ therefore limits the length and width of (100) -oriented semi-finished products, in particular. Since the production of CZ ingots larger than 320mm raw diameter is extremely expensive and hardly available, this material is used for such functional components only in exceptional cases. Often it is necessary to resort to inexpensive multicrystalline material, which, however, has disadvantages in the application. The quality requirements imposed on blanks for such optical functional components or components have hitherto always involved an intrinsic material quality requirement with regard to the content of metallic impurities, nitrogen, oxygen, carbon, the specific electrical resistance and the transmittance at certain wavelengths of electromagnetic radiation. Depending on the application, a doping (p-type or n-type), a specific crystallographic orientation in connection with the format (round or square) and a certain purity (eg 7N) is required.

Since no other monocrystalline material of corresponding doping is available on the market in addition to CZ or FZ material, it was never questioned whether the high quality of the expensive CZ material or even more expensive FZ material for use as monocrystalline high-impedance optical components Silicon or for high-impedance functional components in systems for semiconductor processing of silicon really is absolutely necessary. Silicon as the transparent material in the infrared wavelength range is preferably used in the wavelength range 1.4 to 8μιη, with a main application area in the middle infrared wavelength range of 2 μιη-5 μιη. The much more expensive germanium is mainly used when the field of application has to be extended to 14 μιη. In the patent DE 11 2009 004 379 T5 (corresponding to US 2001/0243162 AI), an approach is described, which corresponds to the goal of a low-cost product for infrared optical elements, in particular those used in devices for detecting the emanating from living organisms thermal radiation , This solution uses high-purity polycrystalline material, which comes directly from the Siemens process and therefore has no specific crystal orientation. Also it achieves in particular in the SWIR (short wavelength IR) 1.4 μηι - 3.3 μηι but also the MWIR (mid-wavelength IR) 3.0 μιη - 8 μιη not achievable for monocrystalline material high transmission. Furthermore, it is limited in terms of its geometric dimensions to the typical for polycrystalline rod diameter size of about 200 mm.

In US2012 / 0176668 an infrared optical system is described, however, which uses monocrystalline silicon, which is produced by the CZ method, the MCZ method (CZ method using a magnetic field) or the FZ method. This procedure does not overcome the disadvantage of the high material price of the monocrystalline silicon and the limited geometric dimensions for usable semi-finished products for the production of the lens elements.

For components used in the semiconductor industry for processing semiconductors, in particular electrode plates for the distribution of gases (also referred to as showerheads or electrode plates or gas distribution plates) in plasma etching plants or plates for the distribution of gases in CVD (Chemical Vapor Deposition) plants the limitations, above all, of a geometric nature. Electrode plates must on the one hand have a larger diameter than the processed wafer and on the other hand have a high radial homogeneity of the specific resistance.

JP 2000-144457 describes a way of allowing dry etching of a wafer with max. 300 mm diameter, use an electrode plate of at least 20% larger diameter (according to concrete description in this application: diameter 365 mm x thickness 11.2 mm). Since at that time no CZ ingots larger than 300 mm were available, a multicrystalline ingot was used. However, since it was already known at that time that the many grain boundaries of a multicrystalline ingot are detrimental, this disadvantage was mitigated by cutting the wafer vertically out of the ingot (i.e., in the growth direction of the ingot) to make the electrode plate according to this invention.

At present, the maximum size of wafers is 450 mm, which requires correspondingly larger electrode plates or gas distribution plates during their processing in plasma etching systems or CVD systems. Regardless of the diameter of the wafer to be etched, electrode plates are used which have as large a diameter even up to twice as the wafer to be etched.

The mitigation of the negative influence of the grain boundaries (particle generation) is, however, bought in JP 2000-144457, that due to the segregation during the breeding process, a change of the resistivity in the growth direction of the ingot and thus occurs across the electrode plate.

JP2007-158007 describes the variation of resistivity as disadvantageous even when Electrode Plates are made from CZ ingots. Due to the method, these have a radial change in the specific resistance. JP 2007-158007 teaches that the radial variation in the resistivity of an electrode Plate may not exceed 5%. According to the embodiments (in particular Example 4), this limit is only exceeded if, for example, in Example 4 of a CZ ingot with 380 mm diameter, an edge region of 40 mm is removed. The final electrode plate thus has a diameter of only 300 mm.

Taking into account JP 2000-144457 and JP 2007-158007, the production of showerheads for dry etching systems for the processing of 450 mm wafers from ingots from the CZ process, which currently only ingots up to max. 550 mm diameter, without exceeding the fluctuation range for the spec. Resistance of 5% not possible.

According to the state of the art, a method of directional solidification of ingots was developed for the photovoltaic industry in order to be able to produce uniform (100) -oriented solar cells without grain boundaries and thus more powerful solar cells (cf EP 2028292, WO 2007/084934, WO 2009/014957 ). In this method, arranged on the bottom of the crucible monocrystalline seeds use that initiate directional solidification and should lead. However, the applicants of these patents have not further developed the described production methods and the methods of material evaluation for other applications, since they have neither considered nor considered these applications possible. These are exemplified by the statements in [0002] EP 2028292 or [003] in WO 2007/084934 or [002] in WO 2009/014957.

In these crystallization processes developed especially for the photovoltaic industry, the entire ingot is further processed into solar wafers after the separation of an edge region, which inevitably results from the standardized crucible size, and the separation of a lid and bottom region.

In the photovoltaic industry, the electrical efficiency of the solar cells used is the most important quality criterion for their use. It is known that with monocrystalline solar wafers, which are produced by the CZ process, higher efficiencies can be achieved than with multicrystalline solar wafers. This is due, in particular, to recombinations of charge carriers generated under solar irradiation at the large number of grain boundaries present in the material. SUMMARY OF THE INVENTION

The invention aims to provide a quasi-monocrystalline silicon crystal material, as defined below, for high-resistance quasi-monocrystalline silicon optical components for high-frequency spectral applications or high-resistance quasi-monocrystalline silicon functional components for use in semiconductor processing equipment, which is significantly less expensive than conventional monocrystalline material , which originates from the CZ or FZ process, is produced and meets only such technical requirements for these optical components or functional components such as lens blanks, mirrors, showerheads, which are actually mandatory for the application. This object is achieved by a method according to claim 1, by a blank (blank) according to claim 14 and by a use according to claim 24. Further advantageous embodiments are the subject of the related subclaims.

The term "quasi-monocrystalline" first appears as a technical term in solar cell production in about 2010. In the context of the present application, this term is intended to denote a silicon material which is directionally solidified in a crucible or the like, wherein the bottom of the crucible is integrally formed with one or multi-piece monocrystalline or quasi-monocrystalline seed layer, which originates in particular from a CZ or FZ process is designed, the seed layer are melted and this crystal growth or the directional solidification of the silicon melt their crystallographic orientation on freshly crystallized material The seed plate (s) of the seed layer may also be made of silicon material that already originates from a process as shown in the previous sentence and thus already consists of quasi-monocrystalline material For the purposes of the present invention, quasi-monocrystalline silicon material in the sense of the present invention consists of a single grain, which may, however, contain dislocation clusters and small-angle grain boundaries with a certain proportion, wherein the grain contains twins contained therein as well Foreign grains, if present, should completely enclose. For use in photovoltaics, the term "quasi-monocrystalline" also includes synonymous designations, such as, for example, mono-like, near-monocrystalline silicone, M-grades, U-grades and others, which, however, are specific to the manufacturer only for solar wafer material in this field are known.

The term "blank", as used in the present application, refers to a blank or a semi-finished product which is produced from a crude crystal produced after a cultivation in order to be able to evaluate the material quality of this blank on its surface and / or in its interior This blank is a round or angular disk with a thickness greater than 0.5 mm and less than 30 mm A blank rated as suitable for the intended application is further processed after its evaluation in order to adjust the final geometry and surface quality of the component The present application, as used, for example, in plasma etching systems, are also often synonymously referred to as electrode plates or shower head electrodes or gas distribution plates.

The quasi-monocrystalline silicon crystal material according to the invention for high-resistance optical components for applications in the infrared spectral range or for high-resistance functional components for use in systems for semiconductor processing is only as good as really required and can thus be produced more cost-effectively than conventionally. In other words, according to the invention, cost advantages can be realized in that only those properties which are absolutely necessary for a satisfactory component are maintained. It has surprisingly been found that certain defects or the degree of their expression, which do not or can not be used in silicon material which can be used for semiconductor wafers are not allowed to occur to the same extent, the functionality of optical components or even functional components for use in systems for semiconductor processing not or not significantly affect. For components of a size that are not or only extremely expensive to produce with the conventional crystallization process for monocrystalline silicon (CZ, FZ), even the first time there is the possibility to monolithically produce this from the relatively defect-containing quasi-monocrystalline silicon according to the present invention. Until now, these large functional components either had to be manufactured as a whole part from multicrystalline material or assembled from monocrystalline or multicrystalline material segments. However, functional components made of multicrystalline material have far shorter service lives. Segmented functional components have a high production cost.

Surprisingly, it has also been found by the inventors that planar determination of carrier lifetime (or physical quantity correlating therewith) with silicon and other semiconductor materials such as germanium, gallium arsenide and other so-called compound semiconductors is a very simple, fast and nondestructive method to improve tool life To roughly evaluate the wear (removal of material by etching gas or generation of interfering particles) or the polishability of blanks for use for optical components or also for functional components for systems for semiconductor processing under operating conditions. Areas with only isolated dislocations (optically uncritical) can thus be distinguished from those with dislocation clusters (optically critical if the tilting of the subgrains reaches more than 20 ° and critical in certain electrical applications). The measurement technique, actually developed for evaluating the electrical quality of solar silicon or for determining the achievable efficiency of solar cells, can be used with the method described here for assessing the optical quality, the structural quality or chemical stability of quasi-monocrystalline silicon. Thus, simple, inexpensive and quick to perform evaluation procedures are available. In this case, the surface of the workpiece to be examined is examined with the following methods, for example: μ-PCD (microwave ve-detected), MWC (Microwave Detected Photoconductivity), PL (Photoluminescence), or similar scanning or imaging techniques. All these measurement techniques determine a value which is proportional to the charge carrier lifetime in the area of the surface. Here, the spatial resolution of the respective measurement techniques is different; It can range from a few μιη to a few mm. In the vicinity of closely spaced dislocation lines - ie in the region of dislocation clusters - the carrier lifetime decreases greatly, since the dislocation lines or small angle grain boundaries represent places of very high charge carrier recombination and the distances of the dislocation lines go below the diffusion length of the charge carriers. The only condition for the material is that no other recombination mechanisms are allowed to superimpose the recombination on the dislocation lines. This means that a massive contamination, for example, with evenly distributed in the volume of metallic impurities prevents them Type of detection of dislocation clusters. In particular, transition metals such as Fe, Cr, Co, Ni, Ti and the like should not exceed a concentration of 0.1 ppm.

A known grain-based method of directional solidification of silicon for the production of quasi-monocrystalline Si ingots for the photovoltaic industry is modified according to the present invention, in particular with regard to the possibility of targeted adjustment of the concentration of free charge carriers at a certain level less than 5 * 10 ¹⁵ / cm ³ and adds an allocation and evaluation method to define and evaluate a volume fraction of the resulting ingot so that it can be used to fabricate high-resistance quasi-monocrystalline optical components for infrared spectral applications as well as high-resistance quasi-monocrystalline functional components for use in processing equipment of semiconductors can be used. However, the allotment and evaluation method can be used expressly for differently manufactured silicon crystal material. These include silicon material which was produced by the known Czochralski method and contains unwanted twins and thus can not be used for semiconductor wafers, or silicon material which is produced by a method based on the Kyropoulos method for silicon without crucible contact (cf. Nakajima et al, J. Cryst. Growth 372 (2013) 121-128).

In the case of using a germ-based culture method in a crucible according to the method of directional solidification for the production according to the invention of mono or quasi-monocrystalline silicon in a Bridgman-type cultivating plant or a Vertical Gradient Freeze (VGF) plant, the following process steps are carried out in principle:

 Producing a quasi-monocrystalline ingot by directional solidification in a crucible;

 Bottom cut, mantle cut and lid cut on the ingot after its cooling and demolding from the crucible to obtain an ingot core;

 - Measurement of the resistivity as a function of the height coordinate of the ingot core on one of its outer sides for determining the height range useful for blanks;

 Determining a quasi-monocrystalline test surface on the surface of the ingot core or a slice separated horizontally thereof which is the same thickness or thicker than the later blank;

 - Testing and evaluation of the test surface with regard to content, distribution and / or quality relevance of at least one of the following: foreign grains, twin boundaries or clusters of dislocations or small-angle grain boundaries; and

Separating the blank from the quasi-monocrystalline region of the disc such that the blank removed in areas on its top or bottom containing no clusters of dislocations or small-angle grain boundaries has a concentration of discrete dislocations in the range greater than one x 10 ² to smaller lx l0 ⁶ cm ^"2 and wherein the separated blank has a specific resistance greater than 5 Ωαη. This process consists of further detailed steps:

It is the crucible bottom of a crucible with one or more mono- or quasi-monocrystalline silicon nuclei occupied, preferably with undoped silicon nuclei, so that forms a seed layer.

In principle, in the method, the seed layer can be formed in one piece and completely cover the bottom of the crucible or container. However, according to a further embodiment, the seed layer can also be formed from a plurality of seed plates, which are arranged directly adjacent to each other (forming as narrow gaps as possible between them) on the bottom of the crucible in order to completely cover it. In a method for producing the aforementioned seed layer from a plurality of seed plates, they can in principle also be produced by a Czochralski method or by a floating zone method.

According to a further embodiment, the plurality of germination plates is arranged on a planar seedbed, which is flat by a sawing process or even ground flat.

In this case, in particular a plurality of germ plates are arranged on the planar seed pad, the abutting surfaces or additionally at least one further surface are ground. In this case, the abutment surfaces of adjacent germination plates are ground at right angles so that, when using three or more germ plates, the width of the resulting column (viewed from above) is as small as possible. It is of equal importance that even during the melting of the germ plates no column arise. This means that even with a lateral (horizontal) view of the germ plate, the ground abutment surfaces have a right angle. After grinding, the ground surfaces (in particular the edge surfaces) have a roughness of Rz according to DIN 4762 of less than ΙΟΟμιη, more preferably less than ΙΟμιη and particularly preferably less than 5μιη and their angularity is considered both vertically and horizontally so good that gaps formed during the laying out of the germination plates and during the melting of the germination plates between immediately adjacent germinal plates are smaller than 1 mm, more preferably less than 0.1 mm, and more preferably less than 0.01 mm, at each point. As a result of this predetermined accuracy of fit, the silicon material according to the invention as a whole has no twin grains (shock twins) which are formed on germ buds. The crucible is then further filled with silicon raw material with the addition of substantially more dopant than required for the production of solar cells, ie, for example for the above-mentioned range of resistivity from 0.001 to 0.2 and the use of one or more already suitably doped nuclei a large amount of dopant is added, resulting in an initial concentration in the melt of about 1.2 * 10 ²⁰ atoms / cm ³ to 1.0 * 10 ¹⁷ atoms / cm ^{3 of} the dopant boron (B) or about 7.4 * 10 ¹⁹ atoms / cm ³ to 3.1 * 10 ¹⁶ atoms / cm ^{3 of} the dopant phosphorus (P) leads.

For directed solidification, the filled crucible is placed in a technically modified system for directional solidification of silicon, in particular a VGF plant.

This is followed by melting of the raw materials, melting of the germ or germs without keying the phase boundary, followed by a directional solidification of the liquid silicon to form an ingot.

After cooling, removal and demolding of the ingot, the cutting of the ingot soil from the ingot can be followed for preferential reuse as seed material. The quality of the cut makes it possible, if necessary, to visually recognize multicrystalline edge regions on both opposite cut surfaces and, if appropriate, separate them from the germplate. Subsequently, the cutting off of the cladding layer from the ingot can take place in a predetermined thickness or a thickness which is visually recognizable as non-monocrystalline at the cut surface. Subsequently, the remaining ingot can be cut into slices corresponding to the desired thickness in order to be able to produce blanks for high-resistance optical components or blanks for use in high-impedance functional components in systems for semiconductor processing, in particular for showerheads. Possibly. For example, a surface treatment of the wafer surfaces may be carried out to better ensure the visual distinctness of multicrystalline and quasi-monocrystalline regions of the processed surfaces and to mark the quasi-monocrystalline surface region. Subsequently, the marked quasi-monocrystalline surface region (test surface) on the disk or on the ingot can be checked for the presence and the location of detected foreign grains or detected twin boundaries.

Examination of a selected quasi-monocrystalline surface area (test surface) of the disk can be made by one of three preferred electrical methods for quantifying the content of this test surface to clusters of dislocations or small angle grain boundaries. In the process, it is detected whether and at which location on the test surface pixels with clusters of dislocations or small-angle grain boundaries are present. Subsequently, a determination is made of the percentage of clusters of dislocations or small-angle grain boundaries of adhered pixels on the test surface, as well as a specification of one or more new test surfaces of the disk, in terms of their content with respect to the percentage of clusters of dislocations or small angle grain boundaries Foreign grains and, in terms of their content of twin boundaries, the geometry and specification requirements for one or more blanks. Subsequently, a separation of the new test surface can be made, which is identical to the surface of the blank.

Quasi monocrystalline material purposes of this invention contains an average concentration of isolated dislocations of between 10 ² cm ^"2 to 10 ^{6 cm" 2,} typically between 10 ³ cm ^"2 and 10 ^{5 cm" 2.} Here, the quasi-monocrystalline silicon material according to the invention differs from the conventional monocrystalline silicon, produced for example with the Czochralski or the floating zone process with the target of use as a semiconductor wafer for the production of microelectronic components. This use requires a concentration of isolated dislocations usually below 10 ² cm ^-2 and is typically even wholly or nearly free of dislocations However, monocrystalline silicon from the Czochralski or the floating zone process can also have very high concentrations of dislocations, if it is However, such material is atypical and, in particular, because of the very high mechanical stresses associated therewith, also difficult to process (Czochralski and floating zone crystals with locally high dislocation densities tear easily).

In contrast to the method according to the invention, the microorganisms used in the methods according to the prior art for the production of p-doped quasi-monocrystalline solar cells have a comparatively very high resistance compared to a high-resistance mono- or quasi-monocrystalline silicon material according to the invention (specific electrical resistance greater than 5 Ωαη). typically 1 Ωαη to 3 Ωαη at B-doping). In contrast, for the production of high-resistance Si crystals according to the present invention, it is necessary to use seeds which have a lower, or even significantly lower, doping content or are not doped at all. That These germs have a higher or much higher resistance than in the prior art.

Also, in the prior art method, a higher or much higher amount of dopant boron is additionally added to the silicon raw material disposed above the seed (s) than in the present invention. This higher doping is responsible for setting the specific electrical resistance of ingots commonly used in photovoltaics in the range from 1 Ωαη to 3 Ωαη.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The aforementioned process steps using the directional solidification method for producing the quasi-monocrystalline silicon will be described in more detail with reference to the accompanying drawings. The following descriptions and figures refer to the VGF process as one of the methods of directional solidification. Silicon material produced by this process is referred to as VGF mono-silicon. Show it:

 Figure 1 is a vertical section through a generation 4 Si ingot made by a crystallization process according to the present invention, with exemplary distribution of the material in quasi-monocrystalline and multicrystalline and a possible division of, for example, horizontally excitable slices to obtain blanks for

Functional components, eg the use of the high-impedance optical components indicated in the figure or the high-resistance functional components for systems for semiconductor processing; Fig. 2 is another vertical section through a generation 5 Si ingot made by a crystallization process according to the present invention, with exemplary distribution of the material in quasi-monocrystalline and multicrystalline and a possible dicing of, for example, horizontally excisable slices for obtaining

Blanks for functional components, e.g. the use of the high-impedance optical components or high-impedance functional components for semiconductor processing plants indicated in the figure;

 Fig. 3 is a photograph of one of such ingot in the middle

 Height cut out slice with quasi-monocrystalline region A and multicrystalline region B with exemplarily blanks for e.g. high-resistance optical components or high-resistance functional components for semiconductor processing equipment;

 4a and 4b show PL (left) and μPCD images (right) of a rough-cut wafer surface of monocrystalline dislocation-free silicon material, which after a CZ

Method was made (annular structures are visible);

 Fig. 4c and 4d corresponding PL (left) and μPCD images (right) of a sawn

 Wafer surface of dislocation-containing quasi-monocrystalline silicon material prepared by a VGF method according to the present invention (no ring-shaped structures visible, no clusters of

Dislocations, not foreign or twin);

 Figures 5a and 5b show respective PL and μPCD images of a rough-sawn wafer surface of quasi-monocrystalline silicon material derived from a bottom-near region of an ingot made by a VGF process according to the present invention, wherein clusters of dislocations or small-angle grain boundaries are visible ;

 FIGS. 6a and 6b show respective PL and μPCD images of a rough-sawn wafer surface of quasi-monocrystalline silicon material originating from a capped region of an ingot made by a VGF process according to the present invention, wherein clusters of higher pitch or small angle grain boundaries Area fraction are visible as in Figures 5 a and 5b;

 7 shows a μPCD image of a VGF monocrystalline test surface from the

 Area A of an ingot;

Fig. 8a, the arrangement of a thermocouple off-center at the bottom of a

 Crucible mounting plate on which a crucible stands, for an exemplary method according to the present invention when using a one-piece seed layer;

Fig. 8b, the arrangement of a thermocouple off-center on the underside of a

Crucible standing plate on which a crucible stands, for an exemplary method according to the present invention using a seed layer formed of a plurality of seed plates whose edges are ground and worked, and in which the gaps between individual seeds are not drawn to scale; 9 shows an exemplary temperature profile in a method according to the present invention.

DETAILED DESCRIPTION

As a crucible according to the present invention, in principle, a ready-to-use crucible inside coated by the crucible manufacturer or a crucible coated by the user himself (eg a quartz or fused silica crucible) or a graphite crucible or a Si ₃ N 4 crucible can be used. The crucible base area depends on the size of the high-resistance optical components to be produced therein or the high-resistance functional components for systems for semiconductor processing. It can have square crucibles of size G4 (720 mm x 720 mm), G5 (880 mm x 880 mm), G6 (1050 mm x 1050 mm) or currently not in use larger crucibles up to 3000 mm x 3000 mm, round crucibles a diameter greater than 450 mm or rectangular crucibles of minimum base surface edge length of 450 mm and maximum base edge length of 3000 mm are used. A crucible height below 250 mm is just as meaningless as a crucible height (possibly including crucible attachment) of greater than 800 mm. Total crucible heights of 450 mm, 550 mm, 650 mm, 780 mm or sizes in between have proved to be very practical.

Hereinafter, a method for producing a Si ingot according to the present invention will first be described with reference to FIGS. 8a to 9. FIG. 8 a shows a schematic cross section through a crucible 2 of a plant 1 for the production of silicon ingots. The container 2, usually a quartz crucible, has a bottom 3, which extends perpendicular to the longitudinal direction 5, and at least one side wall 4, which extends in the longitudinal direction 5 and can be formed circumferentially, in particular of four rectangular surfaces can be formed specify an overall rectangular or preferably square base of the container 2. The bottom 3 may according to further embodiments also have a different cross section, for example, an octagonal, circular or oval cross-section.

According to FIG. 8a, a seed layer 10 made of silicon is arranged on the bottom 3 of the container 2 as a seed template, which is formed in one piece and completely covers the bottom 3. The seed layer 10 preferably has the same material properties as the silicon ingot to be formed. To compensate for ripples and bumps of the bottom 3 of the container 2, which could cause dislocations and other lattice defects in undesirable high number and concentration in the produced Si ingot, a planar seed pad 13 is provided on the bottom 3, of which at least the top of a sufficient planarity having. This corresponds at least to the quality of a standard saw cut, the top can also be ground flat. In this way, an orientation of the seed layer is achieved exactly perpendicular to the perpendicular bisector 5 on the bottom 3.

The container 2 is shown in FIG. 8a on a crucible setting plate 14 made of a good heat-conducting material, preferably of graphite. This leaves, as well as the underneath located cooling plate 15, heat of the underlying meandering bottom heater 17 very well through to heat the bottom 3. The crucible mounting plate 14 is designed for an exactly vertical arrangement of the container 2. By means of the cooling plate 15, which can be traversed by a gas as the cooling medium, a predetermined cooling power for cooling the bottom 3 can be set. Referring to Figure 8a, a thermocouple 16 having a protective tube is provided off-center in the cooling plate 15 to monitor the temperature during directional solidification. An elaborate and disturbing monitoring of the temperature by means of a measuring tube immersed in the Si melt can thus not be dispensed with according to the invention. More specifically, the thermocouple 15 is arranged at a radial distance from the center of the bottom 3, which corresponds to half the diameter of the container 2, if this is round, or half of an edge length of the container 2, if this has a square base , in particular with a tolerance of + 30mm and -100mm. As explained in more detail below with reference to FIG. 9, in the method, the temperature signal of the thermocouple 15 is controlled.

FIG. 8b shows, in a schematic cross-section through a crucible, an alternative embodiment for the production of silicon ingots, in which a seed layer is used, which is formed from a plurality of seed plates 11 whose edges or joints form as narrow gaps as possible 12 directly abut each other and completely cover the bottom 3 of the container 2. The germ plates 11 do not necessarily have a uniform cross section, in particular a rectangular or square cross section, so that the bottom 3 can be completely covered. The germ plates 11 may be arranged in a mirror-symmetrical arrangement with respect to a center plane (not shown) extending perpendicularly to the bottom 3 of the container 2. The shape of the germ plates 11 may also be suitably adapted to the shape of the bottom 3 in other ways. The thicknesses of the seed plates 11 are not necessarily the same, so that a substantially planar surface of the seed layer is formed, however, the bottom and the edges are processed so that all gaps between the germs at each point are extremely low.

In order to prevent tilting of the germ plates 11, which could cause displacements and other lattice defects in undesirably high numbers and concentrations in the Si ingot to be produced, a planar seed pad 13 is provided in this case on the bottom 3, of which at least the upper side has sufficient planarity having. This corresponds at least to the quality of a standard saw cut, the top can also be ground flat. The germ plates 11 can thus be arranged with exactly uniform orientation on the bottom 3 of the container 2.

In order to prevent tilting of the germ plates 11 and the formation of unnecessarily wide gaps 12 between them, the surfaces of the germ plates 11 are preferably completely or partially ground. The roughness Rz in accordance with DIN 4762 should be less than ΙΟΟμιη, more preferably less than ΙΟμιη and more preferably less than 5μιη. When using multiple seed plates in the seed layer is the distance between the individual Germplates preferably small, regardless of whether this is caused by tilting of the germ plates or non-plane-parallel edges. This distance is less than 1 mm, more preferably less than 0.1 mm and more preferably less than 0.01 mm, so that resulting gaps 12 between immediately adjacent seed plates 11 at any point are of this order of magnitude.

After the arrangement of the above-described seed layer 10, 12 at the bottom 3 of the container 2, a silicon melt 6 is provided in the container 2. For this purpose, lumped silicon can be introduced into the container 2 and melted, for example as disclosed in the applicant's EP 2028292 A2. The silicon melt 6 can also initially be provided in another container (not shown) and transferred into the container 2 in liquid form. The process is carried out in such a way that the seed layer 10, 12, which acts as a seed receiver, is merely melted on, that is to say only partially but not completely melted. Subsequently, the silicon melt 6 is directionally solidified in a conventional manner.

After solidification of the silicon melt 6 to a silicon ingot it is removed from the container 2 and further processed. For this purpose, this is suitably trimmed, for example by cuts along the longitudinal direction 5 of the silicon ingot. Furthermore, a bottom and a cap of the silicon ingot are removed by cuts perpendicular to the longitudinal direction 5 of the silicon ingot. Depending on the process conditions, a portion of the machined core of the silicon ingot on the periphery may not have the desired material properties. As a germ, a surface germ filling almost the entire inner surface of the crucible is particularly useful. This can be obtained by a bottom cut from a previous ingot. For reasons of effectiveness, the germ or the germ layer should not be unnecessarily thick, but also not too thin, so that there is no danger of it completely melting in the later germination phase. Seed thicknesses of 15 mm, 30 mm, 45 mm, 50 mm, 70 mm, 80 mm or intermediate sizes have proven useful, with 40 mm or 45 mm being preferred.

For the production of high-impedance optical components or high-impedance functional components for systems for semiconductor processing, the orientation of the microbes is rather of secondary importance. Although the (100) orientation of the microorganisms desired in photovoltaics is not mandatory for the production of high-resistance optical components or high-resistance components for functional components in semiconductor processing plants, it is the preferred microbial orientation for breeding-related considerations. Depending on the objective, however, it can also be bred based on (111), (HO), (211) - or even differently oriented germs.

However, the above-mentioned and preferred area germ does not limit the use of germs. If the bottom of the crucible of smaller germs over the entire surface or to a small edge area is to be designed over the entire surface, it is recommended that the individual germs as possible without joints abut each other and also to leave no gaps. Germs can be worked out quadratically from parts of a surface germ, a round crystal produced according to a CZ method, in the above seed thickness. Germs may also be obtained from longitudinal sections of the desired thickness mentioned above from a crystal made by a CZ method. Germs can also be cut out vertically from a previously ingame-based directionally solidified ingot as vertical boards or ingots in the desired thickness mentioned above.

It is important for all types of seed production that the seed already has a high purity and little or no doping. With knowledge of the germ quality, the addition amount of dopant to the pure Si raw material is chosen so that the target value for the resistivity within the desired range of greater than 5 Ωαη is achieved. After placement of the germs so the remaining crucible volume is still filled with virgin or recycled Si raw material (at least in the purity solar grade) and possibly dopant. When filling the crucible or provided with a crucible top crucible, the resulting total height is used as possible. A filling height of less than 250 mm is just as meaningless as a filling height (possibly including crucible attachment) of more than 800 mm. Total crucible heights of 350 mm, 450 mm, 550 mm, 650 mm and 780 mm or intermediate sizes have proved to be very practical.

In order to obtain in the desired range of the electrical resistivity greater than 5 Ωαη, the concentration of free charge carriers with respect to acceptors, formed as the difference of the acceptor concentration minus the donor concentration, a concentration of 2.7 * 10 ¹⁵ / cm ³ or Donors, formed as the difference of the donor concentration minus the acceptor concentration, should not exceed a concentration of 9.05 * 10 ¹⁴ / cm ³ . For higher specific electrical resistances, the permissible concentrations are correspondingly lower. In order to fulfill this condition, special measures are necessary. This can be achieved, for example, by one or more of the following measures:

 the compensation of the introduced impurities with acceptors such. Boron, aluminum, gallium or donors, e.g. Phosphorus, arsenic, antimony or combinations of different acceptors and donors depending on the nature of the impurity;

the use of particularly pure crucible coating material as an impurity source with respect to acceptors or donors, eg use of Si ₃ N ₄ crucible coating material with a boron content of less than 5 ppm (preferably less than 1 ppm), an aluminum content of less than 100 ppm (preferably less than 20 ppm) a phosphorus content of less than 4 ppm (preferably less than 0.8 ppm);

the use of particularly pure crucibles as a further source of contamination with respect to acceptors or donors, eg quartz crucibles having a boron content of less than 0.1 ppm (preferably less than 0.04 ppm), an aluminum content of less than 160 ppm (preferably less than 15 ppm) Phosphorus content less than 2 ppm (preferably less than 0.1 ppm); the in situ sampling of the melt and targeted compensation with the missing acceptors (eg boron, aluminum, gallium or combinations of different acceptors) or donors (eg phosphorus, arsenic or combinations of different donors);

 - the choice of a breeding course that is short enough to the diffusion of the

To keep foreign matter in the melt much smaller than 1 x 10 ¹⁵ / cm ³ ;

- To keep the height of the molten phase during the process over, for example, a Nachchargierung so small that the entry of impurities from the coating in the melt remains much smaller than 1 x 10 ¹⁵ / cm ³ ;

 - The use of a diffusion barrier for acceptors or donors from very pure

Material between crucible and melt according to the principle of DE 10 2012 100 147 AI, the content of which is hereby expressly incorporated by reference; the use of a cold crucible wall to create a bulk material crucible as used in the skull melting process; - The use of high purity crucible made of SiC or S13N4 the one coating with

Materials that serve as a source of pollution can be unnecessary

The filled crucible is then placed in a Bridgman-type crystal growing facility or in a VGF facility. The latter is equipped with different heater configurations depending on the system type. Commonly used are systems with only ceiling heaters, systems with ceiling and floor heaters, systems only with jacket heaters, systems with ceiling and jacket heaters or systems with ceiling, jacket and bottom heaters. According to the invention, the raw material is melted from above. By a suitable temperature control of the heater with simultaneous cooling of the seed layer is ensured both that the added raw material is completely melted, the germ or the germs are not completely melted, but necessarily melted.

In order to be able to represent the melting of the seed layer in an understandable manner, each individual area fraction of the seed layer is defined with a size of approximately 1 cm ² . Each of these individual surface portions should be melted above a crucible bottom up to a certain height. This height coordinate, up to which the melting of each area fraction takes place, is in the range of 20% to 90% of the thickness of the inserted seed layer. Depending on the size of the pot, the type of plant and the seed thickness, areas are found that are preferred. For example, with a seed thickness of 40 mm, an area of the height coordinate of 35% to 75% of the original seed thickness or, with a seed thickness of 45 mm, a range of the height coordinate of 30%> to 80%> of the original seed thickness is preferred. By means of a suitable temperature / time control of the heaters, it is achieved that a horizontal phase boundary forms in the flatness in the interior of the crucible, which makes it possible to melt the seed layer without a key for determining the position of the phase boundary. Advantageously, the measurement of the temperature is not made in the center of symmetry on the underside of the plate, which is referred to as a crucible mounting plate. The measurement of the temperature is made by placing a thermocouple in the graphite crucible setting plate at a radial distance from the center of the crucible bottom which is half the edge length or diameter of the crucible. The location of measurement may be within a ring defined by a positive Tolerance deviation of 30mm and a negative Tolerance deviation of 100mm of a radius range, which corresponds nominally half of the crucible edge length. The measurement thus takes place, as shown in FIGS. 8a and 8b, at a point further out which is close to the edge of the crucible and provides representative and accurate measurement results by this layer in order to prevent complete melting of the seed layer near the crucible wall. This danger is due to the fact that the system tends to form heat flows in the direction of the center by means of germ cooling, so that the temperature in the seed layer center is colder than at the edge of the seed layer. Important parameters are cooling capacity and geometry of the cooling arrangement. The cooling capacity must be adjusted to the heater temperatures so that a flat phase boundary is established. The geometry, or rather the cooled surface, must be approximately equal to the area of the crucible. The position of the temperature measuring point must therefore be chosen so that it detects the edge area under the germ, but is not distorted by the cooling. When the melting of the seed layer is completed, by further increasing the heat dissipation through the crucible mounting plate, further melting of the seed is stopped and directional solidification is initiated.

As shown in the exemplary embodiments, the melting of the seed layer takes place without mechanically touching the position of the phase boundary. Instead, the procedure is such that the temperature of the underside of the crucible mounting plate is measured at a defined distance from the center of symmetry of the crucible bottom, specifically at the periphery of the crucible mounting plate of good heat-conducting graphite, by means of a pyrometer or thermocouple. 9 shows a typical time profile of the temperature profile, as measured by the thermocouple 16 shown in FIGS. 8a and 8b. A bottom heater is used (see reference numeral 17 in Figures 8a and 8b), which operates from the beginning of the heating of the silicon until the melting temperature of the silicon is reached. The shutdown of the floor heater takes place on reaching the melting temperature of the silicon at the time ti in FIG. 9 and there is a simultaneous connection of a floor cooler with a cooling capacity of a maximum of lW / cm ² , with a maximum of 0.65W / cm ^{2 are} preferred. Despite the switching off of the bottom heater and the connection of the cooling, the temperature measured at the location of the thermocouple increases even after the time ti until it finally falls. This temperature drop is shown in FIG. 9. In this phase, further heating of the container by means of a not shown in Figures 8a and 8b lid and jacket heater, which are designed in the usual way. Due to the gradual melting of the silicon material in the container, the temperature drop according to FIG. 9 is gradually decelerated. The temperature eventually passes through a minium and then begins to rise again due to the heating by the lid and jacket heater. The crystallization phase is shown in FIG. 9 according to the invention at a time t ₂ initiated by increasing the cooling capacity of up to a maximum of lW / cm ^2, preferably of at most 0,65W / cm ² to a minimum of 2W / cm ^2, preferably minimally l, 5W / cm ² , as soon as the temperature measured on the thermocouple at the crucible mounting plate has passed a predetermined amount after passing through the bottom heater cut-off and cooling power switch-on minimum (as described above) Temperature difference (ΔΤ) of preferably 5 K to 25 K has risen again above this minimum. In the case of the invention, it is thus preferably controlled to the temperature signal of the thermocouple located in the crucible mounting plate, so that it is not necessary to immerse a temperature sensor in the Si melt, which is conventionally complicated and causes a great variety of defects in the Si ingot. As will be apparent to one skilled in the art upon reading the foregoing description, the thermocouple, as an example of a more common temperature sensor, may also be located elsewhere near the crucible of the apparatus of Figures 8a and 8b so long as there is sufficient correlation between the temperature signal of such a temperature sensor and the actual temperature prevailing in the crucible. This correlation can for example be determined in advance by means of calibration or reference measurements and stored in the memory of a control device, for example a processor, in order to achieve a suitable temperature control. It is advantageous to make the phase boundary solid-liquid convex in the crystallization phase, in order to curb the propagation of crystal grains of other orientation, which always occur due to foreign nucleation on the crucible wall in the direction of Ingotmitte. At the same time, the convexity of the phase boundary must not be too large in order not to excite the dislocation formation caused by thermal stress. The crystallization of an ingot takes place according to the invention:

 - with the directed solidification from bottom to top;

 - With arranged on a germination plate germs whose germinal surfaces and especially the seed edges of the individual germs are ground as described earlier, so that the abutting edges have a low roughness and as possible without joints abut and thus

Prevent bump twins or alternatively the

 - with sprouts that overlap and are not perpendicular to the phase boundary to reduce the twin frequency

 - With a thermal regime, which avoids any jumps in the temperature, pressure and position control to reduce the twin frequency

 - With a convex phase boundary, which prevents the ingrowth of foreign nuclei formed on the crucible wall, in the Ingotmitte

 with a convex phase boundary whose deflection is low enough to keep the thermal stress induced by this deflection lower than the critical shear stress of 1.6 MPa in the crystal so as not to produce dislocation multiplication.

After the finished cultured ingot has been cooled to a temperature close to room temperature according to a standard program, it is taken out of the crystal growing plant together with the crucible and demoulded. If one cuts a round ingot along the axis of symmetry or a cuboid ingot along a parallel to an edge extending and extending through the center of symmetry of the ingot line virtually, we obtain a fictional sectional surface, is illustrated at where the unmelted seed layer is located and where multi- and quasi-monocrystalline Areas are present, as they are suitable for the production of high-impedance optical components or high-resistance components for semiconductor processing equipment with the sizes or dimensions exemplified in Figures 1 and 2. Figures 1 and 2 show how such an ingot can be cut horizontally, for example, in order to obtain blanks for various end products. The photograph of such a horizontal section is shown in FIG. 3, which will be explained in more detail below.

Such a basic cross-section of a generation 4 ingot (630 mm × 630 mm base and height 430 mm) is shown in FIG. 1 and a generation 5 ingot (780 mm × 780 mm base and height 430 mm) is shown in FIG. 2. In both figures, areas are marked which are exemplary of two frequently required dimensions of high-resistance optical components or high-resistance components for semiconductor processing equipment (0 600 mm × 10 mm or 0 540 mm × 10 mm or 0 300 mm × 6 mm or 0 250 mm x 10 mm) can be used. Basically, there are a variety of other products with different dimensions; These possibilities are referred to by way of example in FIGS. 1 and 2 as blank.

The further processing of an ingot takes place in the manner described below by way of example:

First is

a) sawed off a bottom layer, which usually takes place with a band saw. At the sawn cut surfaces it is possible to see visually exactly under suitable incidence of light and different viewing directions where the quasi-monocrystalline center region ends and a multicrystalline edge region begins. After the distance of the vertical section of each Ingotrand, which may be, for example, between 2 cm and 10 cm, is defined concrete, takes place

b) the sawing of the four side surfaces or a cladding layer. Alternatively, the four side surfaces or the cladding layer can also be sawed off in a predefined thickness first. Which thickness is necessary results from experience, i. from previous processes, or from whether the crucible bottom was laid over the entire surface or with the omission of a narrow edge area with germs. The soil layer to be sawn off after this step is slightly smaller in this procedure than in the procedure according to a). Thus, in the case of reuse of the soil layer as a germ in the subsequent process of this does not fill the crucible bottom to 100%.

As a next step, horizontal slices are cut out of the remaining ingot volume, which have a thickness which is still an allowance for the further processing steps for the production of e.g. have high-impedance optical components or high-resistance components for semiconductor processing equipment or other blanks.

The determination of the specific resistance is expediently carried out in two steps: First, the resistance at the inner cut surface of a side surface along a vertical line is determined pointwise. This is used to check at which height coordinates the Ingot that achieves the required resistivity for one or more types of final products. After cutting out horizontal slices from the ingot, specimens cut adjacent to the area selected for the final product can additionally determine the resistivity in higher accuracy in the form of a surface scan across the specimen. However, it is also possible to perform a planar raster measurement of the specific resistance over the entire interesting surface area of the slice, in order to specify the mean value and also the fluctuation range of the measured values within the slice. Similar to the latter case, this is also possible directly on the blank to be delivered. In most cases, this effort is not required.

Essentially, the methods or devices that can be used for the measurement differ only in what range of resistance is expected and whether it is necessary to record the measured values over a wide area. Thus, the measurement may be e.g. offline and punctiform or scanning in the range up to 30 Ωαη done with an eddy current measuring method (eg with the RT 100 from the company. S emilab / Hungary) or pointwise or raster with a 4-tip measuring method done (eg the Automatic Four-Point Sample, Model 280SI Series from Four Dimensions Inc./ Hayward, CA, USA for measuring previously defined suitable sample geometries or the 4PP system for ingots of the same company).

FIG. 3 shows a photograph of the plan view of a slice cut from the ingot in the middle ingot height. Although in this case the side surfaces have already been separated, one can see a multicrystalline region (B), which has grown in from the edge in the direction of the quasi-monocrystalline center (A) and is separated in this image by a hand-drawn line from the quasi-monocrystalline material region. In the quasi-monocrystalline center (A), three round blanks of different diameters are shown by way of example. The sample shown in FIG. 3 is representative of different sample geometries and sample positions. This can be for the o.g. Resistance measurements, transmission measurements, FTIR measurements or for contamination determination (AAS, GDMS, ICP-MS, ...). Within the visual range quasi monocrystalline good range, an attempt is now being made to utilize the largest possible volume for end products. For this purpose, the quasi-monocrystalline material region must be investigated in more detail whether it is not the position and above all local strong concentrations of clusters of dislocations and small-angle grain boundaries, foreign grains or twin boundaries rendering certain areas unusable for the intended use. As a result of these tests described below, the size and position of the blanks obtainable from a disk must be finally determined.

Such disks - in particular their quasi-monocrystalline region - are now in terms of content, distribution and quality relevance of a first group of Crystal defects, known as "dislocation clusters", "dislocation agglomerations", "small angle grain boundary clusters" or the like.) This group has been referred to in the preceding text and will be referred to hereinafter as "clusters of dislocations or small angle grain boundaries" or simply as dislocation clusters or clusters. The assessment of this first group of crystal defects is based on the following relationships and should be described as follows:

VGF-mono-silicon material, which is produced as a result of the directional solidification containing an average concentration of isolated dislocations between greater than 1 x 10 ² cm ^"2 and less than 1 x 10 ^cm", typically between 10 cm ^"and 10 ^cm" , Here VGF mono-silicon differs from the ideal monocrystalline silicon, manufactured for example with the Czochralski or the floating zone method. Silicon produced by the latter method is usually below 10 ² cm ^-2 and is typically even wholly or nearly free of dislocations However, monocrystalline silicon from the Czochralski or the floating zone method may also have very high concentrations of dislocation multiplication offsets. However, such material is atypical, it does not meet the specification requirements of the uses of material produced by this process, is typically recycled, and is also difficult to process, especially because of the high mechanical stresses involved (Czochralski and Floating zone crystals with locally high dislocation densities break easily).

As long as the dislocations are isolated, i. unless they are partially aligned to dislocation lines, small angle grain boundaries, and clusters thereof, the local orientation of the crystal lattice is not macroscopically affected by the presence of the dislocations. The optical appearance of a machined, ground or polished workpiece surface of quasi-monocrystalline silicon workpieces of different dislocation concentration does not differ. Transmission and reflection behavior are identical.

This is different if dislocations during the crystallization or even during the cooling to arrange dislocation lines and accumulations of dislocation lines, the aforementioned dislocation clusters. The linear dislocations may also extend to small angle grain boundaries. The latter then enclose crystal regions which, compared to the rest of the matrix of the workpiece, can have a significant tilt of a few arc minutes to, in the extreme case, a few degrees, so-called subgrains. Typically, dislocation clusters of this type have many such small and minute sub-grains tilted adjacent to each other, with distances and dimensions ranging from μιη to cm.

Areas of greater tilt (sub-grains with a tilt angle of greater than about 20 °) are comparable in appearance to foreign grains and can be visually recognized with unaided eye and under good lighting on the mechanically machined surface, if they are sufficiently large. They represent a second group of crystal defects in VGF mono-silicon. Also the third group of Crystal defects in VGF mono-silicon, the twin boundaries, are visually recognizable, as is the second group.

It is important in visual detection that the angle of incidence of the illumination to the surface to be assessed covers the angle range from 10 ° to 75 ° and the light source rotates azimuthally in 10 ° increments over 360 °. The viewing direction of the area to be assessed by the appraiser must be from the direction of incidence and the same incidence angle range from 10 ° to 75 °. In addition to the viewing direction of exactly opposite the direction of incidence of the light source, an azimuthal viewing angle range of -90 ° to + 90 ° to the viewing direction must be covered. Sanded surfaces are suitable for this, better are sandblasted surfaces. However, it has been found that a roughly ground surface quality typically achieved by wire-cutting lapping or on a band saw is also sufficient for the evaluation. However, areas of greater tilting can also be very small and thus avoid visual recognition. These very small defects are not included in the corresponding material classes defined below. These small undetectable and the larger detectable foreign grains, twin boundaries or regions of greater tilt are always in a quasi-monocrystalline matrix. Quasi-monocrystalline matrix is defined as a quasi-monocrystalline region that embeds a foreign metal all around.

As a result of the visual evaluation, the material classification for foreign grains takes place in 5 classes:

Class FK1: Silicon material with discrete dislocations in the range of greater than 1 x 10 ² cm ^"2 to less than 1 x 10 ⁶ cm ^{" 2} in its quasi-monocrystalline, dislocation cluster or small angle grain boundary-free volume which has no foreign grains on a cut surface through the material.

 Determination criterion: Visibility in the test surface. The test surface is either the disk surface or the blank surface, which is at Final components of small thickness in the direction of the load or the chemical

Attack indicates (functional area).

 In this context, a small thickness means that the blank or the slice cut out of the ingot has a thickness of at least 5 times less than its largest surface dimension.

Class FK2: silicon material with discrete dislocations in the range of greater than 1 x 10 ² cm ^"2 to less than 1 x 10 ⁶ cm ^{" 2} in its quasi-monocrystalline, dislocation cluster or small angle grain boundary-free volume which has a maximum of 1 foreign particle per dm on a cut surface through the material ² in a size smaller than 50 mm ² has.

Determination criterion: Visibility in the test surface. The test surface is either the disk surface or the blank surface, which points to the final component in the direction of stress or chemical attack (functional surface). The direction of stress or chemical attack present at the final component means the side of the component which exposed to stress, such as the area near a plasma or exposed to reactive gases.

The quantity "number of foreign particles per dm ² means the area density of foreign grains on the evaluated surface.

Class FK3: silicon material with discrete dislocations in the range of greater than 1 x 10 ² cm ² to less than 1 x 10 ⁶ cm ² in its quasi-monocrystalline, dislocation cluster or small angle grain boundary-free volume which has a maximum of 2 foreign grains per dm ² having a size smaller than 50 mm ² ,

 Determination criterion: Visibility in the test surface. The test surface is either the disk surface or blank surface accessible to the evaluation.

 The blank surface accessible for evaluation is the outer surface of a blank, through which suitable measuring methods provide information on a blank

Volume representative surface quality or the surface quality achievable after final processing.

Class FK4: silicon material with discrete dislocations in the range of greater than 1 x 10 ² cm ² to less than 1 x 10 ⁶ cm ² in its quasi-monocrystalline, dislocation cluster or small angle grain boundary-free volume which has a maximum of 5 foreign grains per dm ² in a size smaller than 50 mm ² has.

 Determination criterion: Visibility in the test surface. The test surface is either the disk surface or in one of the evaluation accessible

Blank surface.

Class FK5: Silicon material with discrete dislocations in the range of greater than 1 x 10 ² cm ^"2 to less than 1 x 10 ⁶ cm ^{" 2} in its quasi-monocrystalline, dislocation cluster or small-angle grain boundary-free volume, on a sectional surface through the material foreign matter but without numerical limitation having a size of less than 50 mm ² each.

 Determination criterion: Visibility in the test surface. The test surface is either the disk surface or blank surface accessible to the evaluation.

The method described above for the determination of foreign grains on the respective test surface is also suitable for the determination of twin limits. A twin boundary is defined as a coherent grain boundary, preferably as a Σ3 grain boundary.

As a result of the visual evaluation, the material classification with respect to twin boundaries takes place in 5 classes:

The above-mentioned determination criteria for foreign grains in each class 1-5 apply analogously to twin limits of classes 1-5. Class ZI: Silicon material with discrete dislocations in the range of greater than 1 x 10 ² cm ² to smaller l x 10 ⁶ cm ² in its quasi-monocrystalline, dislocation cluster or small angle grain boundary-free volume which has no twin boundaries on a cut surface through the material

Class Z2: silicon material with individual dislocations in the range of greater than 1 x 10 ² cm ^"2 to less than lx l0 ^{6 cm" 2} in its quasi monocrystalline versetzungscluster- or small-angle grain boundaries free volume which to a sectional area through the material of a maximum of 4 pieces of twin boundaries of a total length of maximum. 0.5 m per dm ² .

Class Z3: silicon material with individual dislocations in the range of greater than 1 x 10 ² cm ^"2 to less than lx l0 ^{6 cm" 2} in its quasi monocrystalline versetzungscluster- or small-angle grain boundaries free volume which to a sectional area through the material of a maximum of 10 twin boundaries having a Total length of 1.4 m per dm ² has.

Class Z4: silicon material with individual dislocations in the range of greater than 1 x 10 ² cm ^"2 to less than lx l0 ^{6 cm" 2} in its quasi monocrystalline versetzungscluster- or small-angle grain boundaries free volume which to a sectional area through the material of a maximum of 100 twin boundaries having a Total length of 14 m per dm ² has.

Class Z5: Single-dislocation silicon material in the range of greater than 1 x 10 ² cm ² to smaller l x 10 ⁶ cm ² in its quasi-monocrystalline, dislocation cluster or small angle grain boundary-free volume which has more than 100 twinned boundaries per section through the material dm ² has. Visually not at all conspicuous are all dislocation clusters with sub-grains or areas of only slight tilt, regardless of their size. However, they represent structural inhomogeneities, which affect their function in reflective or transmissive optical components at too high a concentration. Also in other products made of VGF mono-silicon material, such as electrode plates (also referred to as showerhead or electrode plate or Gas Distribution Plate) in plasma etching or plates for distribution of gases in CVD (Chemical Vapor Deposition) and other in In the semiconductor industry for the processing of semiconductors or semiconductor devices required components, these clusters may be undesirable and are to be assessed in terms of their area ratio of the total area of the component. It is therefore necessary to detect the content of such clusters of dislocations or small-angle grain boundaries already on the partially machined workpiece or blank in order to be able to carry out a suitable allocation on the basis of this detection.

As a direct, visual detection of the clusters of dislocations or small-angle grain boundaries within the VGF monocrystalline crystal volume such as described is certainly not possible, the detection must be done indirectly. It has surprisingly been found that a planar determination of the carrier lifetime (or a physical quantity correlated therewith) in the case of silicon and other semiconductor materials such as germanium, gallium arsenide or other so-called compound semiconductors is a very simple, fast and non-destructive method in order to predict a predicted service life or wear (Material removal by Ätzgaseinwirkung or generation of interfering particles) of the functional components under operating conditions rough estimate. Areas with only isolated dislocations (optically uncritical) can be distinguished from those with dislocation clusters (optically critical and critical for certain electrical applications). The measurement technique, actually developed for assessing the electrical quality of solar silicon or for determining the achievable efficiency of solar cells, can be used with the method described here for assessing the structural quality of mono or quasi-monocrystalline silicon. Thus, simple, inexpensive and quick to perform evaluation procedures are available.

In this case, the surface of the workpiece to be examined is investigated with the following methods, for example: μ-PCD (mierowa ve-detected Photo-Conductance Decay Measurement), MWT (Microwave Detected Photoconductivity), PL (Photoluminescence), or similar scanning or imaging measurement techniques. For all these measuring methods, there are already manufacturers of commercial measuring instruments, such as the company Semilab / Hungary, the company Freiberg Instruments / Germany or the company Hennecke / Germany. All these measurement techniques determine a value which is proportional to the charge carrier lifetime in the area of the surface. Here, the spatial resolution of the respective measurement techniques is different; It can range from a few μιη to a few mm. Near closely spaced dislocation lines - i. in the range of dislocation clusters - the carrier lifetime decreases strongly, since the dislocation lines or the small angle grain boundaries represent places of very high charge carrier recombination and the distances of the dislocation lines go below the diffusion length of the charge carriers. The only condition for the material is that no other recombination mechanisms are allowed to superimpose the recombination on the dislocation lines. That is, massive contamination, e.g. with uniformly distributed in volume metallic impurities prevents this type of detection of Versatzungsclustem. In particular, transition metals such as Fe, Cr, Co, Ni, Ti and the like should not exceed a concentration of 0.1 ppm.

Furthermore, efficient use of the method benefits that dislocation clusters once present in the VGF monosilicon material continue to propagate and spread in the course of directional solidification, but never disappear or dissolve. So it is sufficient for a simple classification to examine the side of the workpiece, which is finally solidified. This corresponds to the measurement of the total area shown in FIG. 3 or the minimum of the quasi-monocrystalline area A marked in this image. The evaluation or classification of the examination-accessible side of a cuboidal, round, annular or other shaped blank takes place according to the area fraction of dislocation clusters found. The determination of this area proportion takes place by means of image evaluation of the areal images (raster measurements or camera shots, depending on the measuring technique) of individual sides of the workpiece (possibly only the last solidified side). The measured values for the carrier lifetime in the region of the dislocation clusters are clearly below the mean of the measured values outside the dislocation clusters. Although the absolute values and the resolution depend on the chosen measuring technique, the type of mechanical surface treatment, the electrical conductivity of the material and the content of metallic impurities, the areas of dislocation clusters are always distinguished from the unloaded areas by significantly lower measured values. By way of example, the definition of a threshold value for the charge carrier lifetime or a measured variable correlated with this charge carrier lifetime can then be used to separate area proportions of areas with dislocation clusters from such areas without dislocation clusters or to quantify them in terms of area.

 The following description of the basic procedure is based on the μPCD method. FIGS. 4a and 4c each show PL recordings, and FIGS. 4b and 4d respectively show the μ-PCD mapping of a respective rough-cut wafer surface of quasi-monocrystalline dislocation-free CZ material (FIGS. 4a and 4b) and VGF monosterization-related Silicon (Figures 4c and 4d). In both cases, there are no dislocation clusters (defect area fraction 0%). The materials are very similar.

The color gradient in the μPCD images is due to thickness variations of the wire-trimmed wafer and has nothing to do with quality differences. These pictures are merely intended to demonstrate that the defect type individual dislocation can not be determined with the named measuring methods and this is also not necessary for the quality assessment according to the present invention.

FIG. 5 a shows a PL recording and FIG. 5 b shows a μ-PCD mapping of a rough-cut surface of VGF monosilicon material in the geometric shape of a wafer. The sample contains dislocation clusters, which can be evaluated by means of image evaluation with respect to the area fraction. The area fraction interspersed with dislocation clusters and small-angle grain boundaries is low in this case.

FIG. 6a shows a PL recording and FIG. 6b shows a μ-PCD mapping of a VGF monosilicon material in the geometric shape of a wafer. The area fraction interspersed with dislocation clusters and small-angle grain boundaries is high in this case.

FIG. 7 shows a surface element (test surface) of a larger blank as a μ-PCD surface image, a partial surface of an IR blank (blank for use for optical components in the IR spectral range in the sense of the present application), which was produced from the VGF monocrystalline region A of an ingot.

The entire area of the quasi-monocrystalline region A or blank to be evaluated is either taken as a frame and evaluated or composed of several frames (e.g., area elements as shown in Fig. 7) and evaluated. If a statement is to be made about a specific blank geometry in the sense of a quality classification in the five classes C1 to C5 mentioned below, a contour (new test surface) is inserted into the evaluated surface or into the composite image, within which the area fraction in the total surface determined by dislocation clusters and small-angle grain boundaries.

In DE 102011056404, the content of which is expressly incorporated herein by reference, the procedure of determining the area fraction of areas with dislocation clusters on a total area is described on the basis of the measuring method MDP. This procedure is suitable when using the method μ-PCD or PL is in the same way.

Specifically, the charge carrier lifetime or a value correlated with the charge carrier lifetime of the material is always determined for a pixel of the test surface by means of the aforementioned rastering or imaging measurement techniques (in the case of high-resistance material preferably PL). This is also done for all adjacent pixels in a subarea of the total area to be evaluated. This subarea may be square, rectangular, circular or elliptical. From the pixels completely enclosed in this area with a center pixel (border pixels are not allowed), an averaged image is generated. This can e.g. from an area of 11 x 11 pixels, i. 121 pixels, or an arbitrarily defined number of pixels, e.g. greater than 50 pixels or greater than 100 pixels are obtained. Since the charge carrier lifetime is significantly lower in a dislocation cluster or pixels containing small-angle grain boundaries than in dislocation cluster-free or small-angle grain boundary-free good domains, a difference between the average value is formed from the defined number of neighboring pixels and the center pixel measurement value. If this difference exceeds a previously defined amount, the center pixel is evaluated as a bad pixel. This threshold value has to be specified according to the measuring method, the range of the specimen resistivity and the doping (p- or n-type). For medium-resistance material and the measurement method MDP, e.g. a threshold of 0.22 is an appropriate value.

According to the procedure described in [0059] to [0062] of the referenced DE 102011056404, the cluster content in% is determined as follows:

Number of pixels

 Cluster level = 100% - -; ; -;

 Anza n i _ S cn ι e cn tp? x e ι- A nzanijG tpi x e t

21 The cluster content in% in the total area to be evaluated or in the stored contour (eg the blank surface or new test surface) is used to make a quality rating.

The lower the range interspersed by dislocation clusters and small angle grain boundaries, the lower the probability of the appearance of the smallest, optically prominent grains or subgrains of differing crystallographic orientation.

High-resistance quasi-monocrystalline VGF silicon material for optical components for applications in the infrared spectral range, in particular in the range of 1.4 μιη to 10 μιη, more preferably in the range of 1.4 μιη to 8 μιη, particularly preferred range 2 μιη to 5 μιη, or for functional components in systems for semiconductor processing such as showerhead is divided into the following five quality classes: class Cl: silicon material with individual dislocations in the range of greater than 1 x 10 ² cm ^"2 to less than lx l0 ^{6 cm" 2} in its quasi monocrystalline versetzungscluster- or small-angle grain boundaries free volume which has an area fraction of 0% clusters interspersed by dislocation clusters or small-angle grain boundaries on a cut surface through the material.

Class C2: silicon material with individual dislocations in the range of greater than 1 x 10 ² cm ^"2 to less than lx l0 ^{6 cm" 2} in its quasi monocrystalline versetzungscluster- or small-angle grain boundaries free volume which to a sectional area through the material of a surface portion of less than or equal 25% interspersed with dislocation clusters and small angle grain boundaries.

Class C3: Silicon material with discrete dislocations in the range of greater than 1 x 10 ² cm ^"2 to smaller lx 10 ⁶ cm ^{" 2} in its quasi-monocrystalline, dislocation cluster or small angle grain boundary-free volume which has an areal fraction of less than or equal to a sectional surface through the material 50% interspersed with dislocation clusters and small angle grain boundaries.

Class C4: silicon material with individual dislocations in the range of greater than 1 x 10 ² cm ^"2 to less than lx l0 ^{6 cm" 2} in its quasi monocrystalline versetzungscluster- or small-angle grain boundaries free volume which to a sectional area through the material of a surface portion of less than or equal 80% interspersed with dislocation clusters and small angle grain boundaries.

Class C5: Silicon material that passes through the material on a cut surface

 Area fraction of less than or equal to 100% interspersed with dislocation clusters and small-angle grain boundaries.

When classifying mono or quasi-monocrystalline silicon for applications, the classes FK1 to FK5, ZI to Z5 and Cl to C5 shall be used for the assessment.

Depending on the application, different material behavior under operating conditions can be observed and therefore material of a particular class must be selected. An important material behavior is for example:

 the flat etching removal per unit time (in particular, for example, in plasma etching plants) the homogeneity of the etching removal (in particular, for example, in plasma etching plants)

 the release of particles under conditions of use (in particular, for example, in plasma etching plants)

 - the service life of the component

With regard to the strength of the influencing of the aforementioned material behavior, the defined classes have different strengths:

Classes Z: comparatively weak negative effect

Classes C: relatively moderately negative effect

Classes FK: relatively strong negative effect

Regardless of whether a high-resistance silicon material evaluated in accordance with the present invention has been categorized into one of Z, C or FK classes and for fabrication of functional components in etching equipment (whether with or without plasma support), CVD equipment (with or without plasma support) or other equipment for processing semiconductor devices, it has always had better performance characteristics than traditional high resistance multicrystalline material.

High-resistance quasi-monocrystalline silicon material from the CZ or FZ process which is suitable for semiconductor production is only slightly superior to the high-resistance silicon material according to the present invention with regard to its application characteristics.

Embodiment 1

A crucible with crucible attachment with a total height of 760 mm is placed in a G4 furnace. In the crucible are a monocrystalline or quasi-monocrystalline germ plate with a length x width of 600 mm x 600 mm and a height of 40 mm, crystalline silicon raw material with a total weight of 450 kg. The germination plate is positioned on the crucible bottom in such a way that a uniform gap remains between the germination plate and the crucible wall. The furnace system has three heating zones: ceiling heater, jacket heater and bottom heater. In the lower part of the system is an active cooling arrangement, which consists of water-cooled copper plate and an overlying highly heat-conductive graphite block identical geometry shape. The cooling arrangement can be moved vertically by means of a lifting mechanism and contacts the crucible installation plate via the graphite block. The water-cooled copper plate and the graphite each have a hole at the edge through which a pyrometer can see directly from below on the crucible mounting plate. This pyrometer is used to control the germplasm temperature. At the beginning of the melting phase, the cooling plate is in the lower position and all heaters are active. The crystalline silicon is melted from above. From a certain temperature, which corresponds to the melting point of the silicon, the cooling plate is partially moved up and reduces the performance of the bottom heater. At this time, the cooling plate still has no contact with the overlying crucible mounting plate. Of the Bottom heater encloses the graphite cylinder in the completely upwardly driven state. During the melting phase, the soil is simultaneously cooled and heated to not completely melt the seed but to minimize heat losses at the edge of the crucible. The cooling capacity must be adjusted to the heater temperatures in such a way that, especially during the seeding process, a flat phase boundary is established. The temperature determined by the pyrometer at the measuring location below the crucible setting plate goes through a minimum in each process, which is determined and stored as an absolute value. The temperature difference between the current temperature value after passing through the minimum and the previously determined minimum is used to determine the height of the seeding point. This corresponds to a temperature difference between measured value and temperature minimum of 20 K. If this difference is reached, the crystallization is initiated. The Ankeimstelle lies with a germ with a height of 40 mm and then in the range of 15-25 mm above the crucible bottom. To start the crystallization, the cooling plate is fully moved up to the crucible mounting plate. The contact with the crucible mounting plate increases the heat dissipation downward as a result of an increase in the maximum cooling capacity. 0,65W / cm ² on minimal l, 5W / cm ^2, and the phase boundary moves upwards. During crystallization, all heater temperatures are additionally reduced according to a temperature-time profile. When crystallization is complete, the cooling phase is initiated. The cooling rates are a maximum of 100 K / h.

Thereafter, the ingot is removed from the crystallization unit and removed from the crucible.

On a band saw the ingot soil is separated in a thickness of 45 mm. In this way, a plate is recovered, which is used again as a germ after a sandblasting process with suitable material and subsequent cleaning. From the remaining ingot side parts are now sawed off so that the resulting ingot base area is reduced to 630 mm x 630 mm.

By measuring the resistivity profile according to the 4-peak measurement method on a sample from a side part along the growth direction, it is confirmed that the course predicted by the doping has actually been achieved.

The ingot is placed on a side surface, and there is a lid cut, which removes the experience according to segregation contaminated with impurities and therefore useless material of appropriate thickness. This shows how large the quasi-monocrystalline region is in the center of the ingot both on the bottom and on the lid.

Based on this result and according to the specific geometric requirements, slices with a corresponding thickness allowance are cut out of the ingot, which are sufficient to cut out blanks for functional components.

Under obliquely incident light of a source of illumination is a marking of the quasi-monocrystalline region (center region) of the disc. The outer edge region is therefore the multicrystalline region of the disc. As already described above, the quasi-monocrystalline center region is now being examined in more detail in which number or at what location within the quasi-monocrystalline matrix foreign or twin boundaries could be present and in which class FK 1 to FK 5 or ZI to Z 5 this center region or correspondingly predetermined Areas thereof, which are suitable for the production of high-impedance optical components or high-impedance functional components for systems for semiconductor processing, can be grouped. Such high-resistance optical components or high-impedance functional components for systems for semiconductor processing are, for example, blanks for optical components (lenses, mirrors) or for showerheads or other functional components for systems for semiconductor processing. Dimensions of such blanks are, for example, 0 600 mm × 10 mm or 0 540 mm × 10 mm or 0 300 × 6 mm or 0 250 × 10 mm. Furthermore, this center area or corresponding predetermined areas thereof, which appear suitable for the production of functional components, examined by means of scanning or imaging measurement techniques (such as MDP, μ-PCD or PL) and divided into the classes Cl to C5.

If necessary, a sample is also cut from the quasi-monocrystalline center region to control the resistivity by means of a 4-peak measuring site, i. to confirm that he e.g. is within the required range of greater than 5 Ωαη.

According to the requirements of the blank and based on the measured resistivity, the selection of a range of dimensions and classification (FK1-FK5, Z1-Z5 and C1-C5) suitable range of the quasi-monocrystalline center region A (see example in Fig. 3) for the blank Now work out the blanks from the disc. The blank surface now corresponds to a newly defined test surface for which the classification applies. If blanks are required for a plurality of identical or different components, an optimization based on the geometrical dimensions and the locally different classification results of the quasi-monocrystalline regions of removed slices is carried out according to FIGS. 1, 2 and 3. Embodiment 2

In a G5-VGF furnace plant, a crucible with crucible attachment with a total height of 780 mm is introduced. In the crucible, a planar seed pad made of silicon is introduced, on which a plurality of silicon seed plates are arranged in a layer. The thickness of the seed pad is 5 to 20mm, with 10 to 20mm being preferred. The germplates are cut horizontally from quasi-monocrystalline areas of previous ingots to a thickness of 45mm. The abutting surfaces of the germ plates or additionally at least one further surface are ground. The abutment surfaces of adjacent germination plates are ground at right angles so that when using three or three more germ plates the width of the resulting column (viewed from above) is as low as possible. It is of equal importance that even during the melting of the germ plates no column arise. This means that even with a lateral (horizontal) view of the germ plate, the ground abutment surfaces have a right angle. After grinding, the ground surfaces (in particular the edge surfaces) have a roughness of Rz according to DIN 4762 of less than ΙΟΟμιη, more preferably less than ΙΟμιη and particularly preferably less than 5μιη and their angularity is considered both vertically and horizontally so good that gaps formed during the laying out of the germination plates and during the melting of the germination plates between immediately adjacent germinal plates are smaller than 1 mm, more preferably less than 0.1 mm, and more preferably less than 0.01 mm, at each point. Due to this predetermined accuracy of fit, the silicon material to be grown according to the invention does not have any twin grains (shock twins) which are formed on germ buds.

On the seed layer, crystalline silicon raw material of a total weight of 700 kg is filled. The seed layer is dimensioned and positioned on the microplate so that there is a gap of approx. 20 mm all around. Their upwards or, if necessary, additionally downwards pointing surface is roughened in a sandblasting process with suitable material. Above and into the gap between germ plate and crucible wall polycrystalline silicon raw material is filled. The Si raw material has at least 7N purity. In order not to fall below the required content of a maximum of 2.5 * 10 ¹⁵ / cm ³ for donors and still receive n-conducting material, the amount of introduced by pollution acceptors in the amount of 1 * 10 ¹⁵ / cm ³ by the Addition of 38 mg of phosphorus to 700 kg of Si can be compensated.

The kiln plant is a multi-zone kiln plant with a total of four temperature-controlled heating zones: ceiling heaters, overhead heaters, bottom heaters and bottom heaters. Under the crucible mounting plate is an active cooling device. The cooling medium used is gaseous nitrogen. The dimensions of the cooling device (length, width) correspond at least to the dimensions of the crucible mounting plate. There is a thermocouple to control the temperature of the seedplate below the crucible shelf near the edge of the crucible. This thermocouple is guided from below in a protective tube in the cooling device, which consists of a good thermal conductivity graphite. The protective tube abuts the overlying graphite plate from below. On top of this graphite plate are the crucible mounting plate made of graphite and above it the crucible.

The melting process, as shown in FIG. 9, is designed such that the silicon raw material is melted from above and the silicon germination plate is only partially melted. For this purpose, a temperature profile is set, which typically has a higher temperature on the ceiling heater than on the bottom heater. During the melting phase, the bottom heater is in operation only in the heating phase and is switched off at the time t _ls reaching the melting temperature of the silicon, in order to prevent melting of the quasi-monocrystalline nucleus. Previously, the gas cooling is activated from reaching about 1400 C on the side heater. The cooling capacity must be the heater temperatures so be adapted that sets in particular during Ankeimprozess a flat phase boundary. Even after activation of the gas cooling, the crystalline silicon raw material continues to melt above the germplate as a result of the ceiling and side heaters, which continue to be held at a predefined temperature value. The temperature at the measuring point below the crucible setting plate goes through a recognizable minimum in each process in Fig. 9, which is determined and stored as an absolute value. The total heating power inevitably decreases continuously in the same period of time. At the temperature of the thermocouple or more precisely by the temperature difference between the current measured value and the previously stored temperature minimum, the height of the Ankeimstelle can be determined. In the case of a germ of 45 mm height, the germination site is ideally in the range of 25 mm to 35 mm above the crucible bottom. This corresponds to a temperature difference ΔΤ between measured value and temperature minimum of 8 K-12 K. If this difference is reached at time t ₂ (see FIG. 9), the crystallization is initiated. It can be several hours between temperature minimum and start of crystallization. The crystallization is initiated on the one hand by increasing the cooling capacity of the gas cooling and on the other hand via a controlled temperature-time profile of the active heater. In this case, the cooling capacity is rapidly increased, for example, from a maximum of 5 kW (specifically, a maximum of 0.65W / cm ² ) to at least 15 kW (specifically, a maximum of 1.5 W / cm ^{2 considered} ) to prevent melting of the germ. The temperatures of the heating zones are slowly reduced. The cooling rates are in the range of -0.4 K / h to - 15 K / h. By appropriate selection of the heater temperatures and the cooling capacity, a convex phase boundary is set in the center, which pushes the polycrystalline edge region further outward or supports a vertical columnar growth there. When the ingot has finished crystallizing, the cooling phase begins. During the cooling phase, the heaters are controlled by a further temperature-time profile. The cooling rates are -10 K / h to -80 K / h.

Thereafter, the ingot is removed from the crystallization unit and removed from the crucible. On a band saw the ingot soil is separated in a thickness of 45 mm. In this way, a plate is obtained, which can be used again as a germ after a sandblasting process with suitable material and subsequent cleaning. From the remaining ingot side parts are now sawed off so that the resulting ingot base area is reduced to 780 mm x 780 mm.

By measuring the profile of the resistivity at a side part along the growth direction, it is confirmed that the course predicted by the doping has actually been achieved. The ingot is placed on a side surface, and there is a lid cut, which removes the experience according to segregation contaminated with impurities and therefore useless material of appropriate thickness. This shows how large the quasi-monocrystalline region is in the center of the ingot both on the bottom and on the lid.

Based on this result and according to the specific geometric requirements, slices with a corresponding thickness allowance are removed from the ingot cut out from their quasi-monocrystalline volume blanks for functional components for equipment for semiconductor processing (especially for the production of so-called showerheads) cut out. Under obliquely incident light of a source of illumination is a marking of the quasi-monocrystalline region (center region) of the disc. The outer edge region is therefore the multicrystalline region of the disc.

As already described above, the quasi-monocrystalline center region is now being examined in more detail in which number or at what location within the quasi-monocrystalline matrix foreign or twin boundaries could be present and in which class FK 1 to FK 5 or ZI to Z 5 this center region or correspondingly predetermined Areas thereof, which are suitable for the production of high-impedance optical components or high-impedance functional components for systems for semiconductor processing, can be grouped. Such high-resistance optical components or high-resistance functional components for systems for semiconductor processing are, for example, blanks for round or high-impedance optical components (eg mirrors, lenses) or high-resistance functional components for systems for semiconductor processing (in particular showerheads) of the dimensions 0 600 mm × 10 mm or 0 540 mm x 10 mm or 0 300 mm x 6 mm or 0 250 mm x 10 mm.

Furthermore, this center area or corresponding predetermined areas thereof, which appear suitable for the production of functional components, examined by means of scanning or imaging measurement techniques (such as MDP, μ-PCD or PL) and divided into the classes Cl to C5.

If necessary, a sample is also cut out of the quasi-monocrystalline center region to control the resistivity by means of a 4-peak measuring site, i. to confirm that he e.g. is within the required range of greater than 5 Ωαη.

According to the requirements of the blank and based on the measured resistivity, the selection of a range of dimensions and classification (FK1-FK5, Z1-Z5 and C1-C5) suitable range of the quasi-monocrystalline center region A (see example in Fig. 3) for the blank Now work out the blanks from the disc. The blank surface now corresponds to a newly defined test surface for which the classification applies.

If blanks are required for a plurality of identical or different components, an optimization based on the geometrical dimensions and the locally different classification results of the quasi-monocrystalline regions of removed slices is carried out according to FIGS. 1, 2 and 3. The quasi-monocrystalline Si material prepared and characterized in the geometric shape of a blank as described above according to the method of the invention is particularly suitable for

 a) as a blank for optical components or components for applications in the infrared spectral range, in particular in the range 1.4 to 10 μιη, more preferably in the range 1.4 μιη to 8 μιη, particularly preferred range 2 μιη to 5 μιη. Such components or components may be:

 Imaging components, such as lenses, mirrors, prisms or disks, which may also be provided with reflective or diffractive structures or coatings.

 b) as blank for components for use in semiconductor processing equipment.

 An example of such a component is an electrode plate (also referred to as a showerhead or electrode plate or gas distribution plate), as used in semiconductor processing, for example, for the distribution of gases in plasma etching, plasma cleaning or CVD (Chemical Vapor Deposition) systems, as disclosed, for example, in US 2005/0173569 A1 or US 2001/0076401 Al, the entire contents of which are hereby incorporated by reference.

In summary, a blank, as described above, a method for its production, as described above, and its use for optical components of quasi-monocrystalline silicon for applications in the infrared spectral range, in particular in the range of 1.4 μιη to 10 μιη, more preferably in the range , 4 μιη to 8 μιη, particularly preferred range 2 μιη to 5 μιη and disclosed for components of quasi-monocrystalline silicon in plants for processing of semiconductors, as described above.

LIST OF REFERENCE NUMBERS

1 Device for directed solidification

 2 containers

 3 floor

 4 side wall

 5 longitudinal direction

 6 silicon melt

 10 germ layer

 11 germ plate

 12 gap

 13 germ support

 14 crucible mounting plate

 15 cooling device / gas cooler

 16 thermocouple with protective tube

 17 floor heaters

Distance between the location of the thermocouple 16 and the center of the container 2

Claims

Method for producing a quasi-monocrystalline silicon blank having a predetermined thickness from a silicon ingot, for optical components for applications in the infrared spectral range or for functional components in semiconductor processing systems, in particular for showerheads, comprising the following steps:

 Providing a container (2) for receiving a silicon melt (6) having a bottom (3) and at least one side wall (4);

 Placing a seed layer (10) at the bottom of the container;

 Providing the silicon melt (6) in the container while melting the seed layer; and

 directionally solidifying the silicon melt in the container to the silicon ingot; in which

 after cooling the silicon ingot and removing it from the crucible, a bottom portion, skirt portion and upper portion of the silicon ingot are separated to obtain an ingot core;

 the specific resistance of the ingot core is measured in dependence on its height coordinate on one of its outer sides in order to determine a usable height range for cutting out a blank;

 setting a monocrystalline test surface on the surface of the ingot core or a slice separated therefrom which is equal to or thicker than the predetermined thickness of the blank to be produced;

 the test surface is examined and evaluated in terms of content, distribution and / or quality relevance of at least one of the following: foreign grains, twin boundaries or clusters of dislocations or small-angle grain boundaries;

the blank or blanks-suitable volume is separated from the quasi-monocrystalline region of the disc so that the blank removed or the volumes suitable for blanks in areas on its top or bottom containing no clusters of dislocations or small-angle grain boundaries, a concentration of discrete dislocations in the range of greater than 1 x 10 ² to less than 1 x 10 ⁶ cm ² and wherein the cut-out blank has a specific resistance of greater than 5 Ωcm.

The method of claim 1, wherein the blank or volumes suitable for blanks are separated from the quasi-monocrystalline region of the disk such that regions are still included on its top or bottom in addition to the regions which do not contain clusters of dislocations or small angle grain boundaries. Containing clusters of dislocations or small angle grain boundaries and containing neither visually identifiable twin boundaries (Class ZI) nor visually identifiable foreign grains (Class FK1). A method according to claim 1 or 2, wherein on the top or bottom of the blank or the volume suitable for blanks foreign grains of a size less than 50 mm ² and / or twin boundaries are included.

Method according to one of the preceding claims, wherein on the top or bottom of the blanks or the volume suitable for blanks a combination of the class of the error clusters of dislocations or small-angle grain boundaries (class Cl to class C5) with a class of foreign matter defects (class FK2 to Class FK5) or a class of error Twin Limits (Class Z2 to Class Z5) are provided.

Method according to one of the preceding claims, wherein combinations of all three classes of defects on the top or bottom of the blank or the blank suitable for blanks from clusters of dislocations or small-angle grain boundaries (class Cl to class C5), foreign grains (class FK1 to class FK5) and twin borders (Class ZI to Class Z5).

Method according to one of the preceding claims, further comprising the steps:

 Placing a planar seed pad (13) on a bottom (3) of the container (2) to completely cover it; and

 Arranging the seed layer (10) on an upper surface of the seed pad (13); wherein at least the upper surface of the planar seed pad (13) is machined, in particular of the quality of a saw cut, or ground flat to form a planar surface on which the seed layer (10) is placed without the formation of any hollow spaces.

The method of claim 6, wherein the seed layer (10) is formed as a one-piece seed plate substantially completely covering the seed pad (13) on the bottom (3) of the container (2), or wherein the seed layer (10) is made up of a plurality of Germial plates (11) are formed, which are arranged directly adjacent to each other on the germ support (13) on the bottom (3) of the container (2) to the germ support (13) on the bottom (3) of the container (2) substantially completely cover.

The method of claim 7, wherein the plurality of seed plates (11) are separated from a silicon ingot produced in a previous process by directional solidification of a silicon melt by a process according to any one of claims 1 to 6.

The method of claim 7 or 8, wherein the surfaces of the plurality of seed plates (11) arranged on the planar seed support are completely or partially ground and have a roughness of Rz according to DIN 4762 of less than ΙΟΟμιη, more preferably less than l0μm and more preferably less than 5μιη and wherein their angularity is so good that the resulting gap (12) between immediately adjacent Germplates at each location are less than 1 mm, more preferably less than 0.1 mm, and more preferably less than 0.01 mm.

10. The method according to any one of the preceding claims, wherein the temperature of the container (2) or received therein silicon melt (6) by means of a

Temperature sensor (16), in particular by means of a thermocouple is monitored and controlled, which is arranged off-center in a crucible setting plate (14) on which the container (2) is arranged. 11. The method of claim 10, wherein the crucible setting plate (14) is formed of graphite and wherein the temperature sensor (16) at a radial distance from the center of the crucible bottom, which is half the diameter of the container (2), if this formed round is, or which corresponds to half an edge length of the container (2), provided that it has a square base, in particular with a tolerance of + 30mm and - 100mm.

12. The method according to claim 9 or 10, wherein the container (2) by means of a bottom heater for heating the bottom (3) of the container (2), by means of at least one jacket heater for heating side surfaces of the container (2) and a lid heater for heating a is heated at the upper end of the container (2) and wherein the temperature of the container (2) or the silicon melt (6) received therein is controlled by

 the bottom heater is operated until reaching the melting temperature of the silicon,

the bottom heater is switched off when the melting temperature of the silicon is reached, while the jacket heater and lid heater continue to operate and at the same time a bottom cooler with a maximum cooling power of 0.65W / cm ^{2 is} switched on, and

a crystallization phase of the molten silicon melt melted in the container (6) by increasing the cooling power of a maximum of 0.65 W / cm ² to a minimum of l, 5W / cm ^{2 is} initiated as soon as the on the thermocouple (16) on the crucible mounting plate (14) measured temperature has risen again by a minimum temperature difference (ΔΤ) of 5 K to 25 K after passing through a by the switching off of the bottom heater and the connection of the bottom cooler minimum.

13. The method according to any one of the preceding claims, wherein for a visual detection of twin boundaries and foreign grains on the herausgetrennten blank or the volume suitable for blanks an angle of incidence of a light source to the surface to be assessed covers an angle range of 10 ° to 75 ° and the light source azimuthally in 10 ° increments the surface to be assessed rotates through 360 ° and a

Viewing direction for viewing the area to be assessed from the direction opposite to the Einstrahrichtung and the same Einstrahlwinkelbereich from 10 ° to 75 ° sweeps over and in addition to the viewing direction of exactly opposite the direction of the light source an azimuthal viewing angle range from -90 ° to + 90 ° is swept to the viewing direction.

14. Blank of quasi-monocrystalline silicon material for use for optical components for applications in the infrared spectral range or for functional components in systems for semiconductor processing, in particular for showerheads, wherein

the blank in areas on its upper or lower side, which do not contain clusters of dislocations or small-angle grain boundaries, has a concentration of individual dislocations in the range of greater than 1 x 10 ² to less than 1 × 10 ⁶ cm ^-2 and

 the blank has a specific resistance greater than 5 Ωαη.

Blank according to claim 14, further comprising areas containing clusters of dislocations or small-angle grain boundaries on its top or bottom in addition to the areas which do not contain clusters of dislocations or small-angle grain boundaries.

16. blank according to claim 14 or 15, which has foreign grains of a size smaller than 50 mm ² on its upper or lower side.

17. Blank according to claim 16, which has on its upper or lower side a predetermined number of foreign grains of a size smaller than 50 mm ²

Blank according to one of claims 14 to 17, which has twin boundaries on its top or bottom.

Blank according to claim 18, having a predetermined number of twin boundaries on its top or bottom.

Blank according to one of claims 14 to 19, which has on its top or bottom side a combination of classes of defects clusters of dislocations or small angle grain boundaries, foreign grains and twin boundaries.

Blank according to claim 20, comprising on its top or bottom side combinations of all three classes of defects from clusters of dislocations or small angle grain boundaries, foreign grains and twin boundaries.

22. Blank according to one of claims 14 to 21, which has a diameter or a diagonal of greater than 320mm or whose shorter edge length is greater than 320mm.

Blank according to one of claims 14 to 21, which has a diameter or a diagonal of greater than 470 mm or whose shorter edge length is greater than 470 mm.

24. Use of the blanc according to one of claims 14 to 23 for applications in the infrared spectral range, in particular in the range of 1.4 μιη to 10 μιη, more preferably in the range 1.4 μι Μβ 8 μιη, preferably as Linsenblanks or mirror.

25. Use according to claim 24, wherein the blank

a concentration of individual dislocations in the area in the range of greater than 1 x 10 ² to less than lx l0 ^{~ 6} cm ² in its versetzungscluster- and small-angle grain boundaries free volume and / or

 has a resistivity of greater than 5 Ωαη.

26. Use according to claim 24 or 25, wherein the blank on its top or bottom no visually identifiable twin boundaries (class ZI) and also no visually recognizable foreign particles (class FK1) contains.

27. Use according to claim 24 or 25, wherein the blank on its top or bottom visually identifiable twin boundaries (class Z2 to class Z5) and / or visually identifiable foreign grains (class FK2 to class FK5) contains.