WO2025051364A1 - Continuously melting silicon for downstream processes - Google Patents

Abstract

The invention relates to a method for continuously providing a silicon melt and further processing the silicon melt immediately thereafter, having the steps of (A) providing solid silicon in a melt container, (B) controlling the temperature of the melt container, thereby forming a silicon melt, (C) supplying solid silicon, wherein the supplied solid silicon is melted, and (D) immediately supplying the obtained silicon melt to a processing unit for melted silicon. The invention additionally relates to processing products of the silicon and to the use thereof.

Description

Continuous melting of silicon for downstream processes

The present invention relates to a process for the continuous preparation of silicon melt with immediate subsequent further processing, comprising the steps of (A) providing solid silicon in a melting container, (B) tempering the melting container to form a silicon melt, (C) feeding solid silicon, wherein the fed solid silicon is melted, and (D) directly feeding the resulting silicon melt to a processing unit for molten silicon. The invention is also directed to processed products of silicon and their use.

Silicon melt can be further processed in a variety of ways, for example by atomization, by casting, by stirring to form an alloy or by granulation.

Processed products from silicon melt, such as silicon particles, which are accessible by atomization of silicon melt, have a broad application potential. However, the melting and processing methods described in the prior art are generally batch processes and require very complex process control, as summarized below.

Batch processes have several disadvantages, such as low space-time yield and high material wear due to regular heating and cooling processes. Therefore, the large-scale industrial use of the processed products from silicon melt requires a continuous production process in order to achieve the required

To be able to provide quantities cost-effectively. Although EP 3841225 A1 describes a process for the continuous preparation and processing (here: atomization) of molten metal, it is necessary to transfer the liquid melt through several containers before it can be further processed by atomization. This complex transfer of melt to further processing leads to high operating costs and complex plant construction, while also being highly error-prone.

EP 0372918 B1 is also directed to a process for the continuous production of silicon powder, wherein silicon is melted in a melting furnace, poured off at the top of the melting furnace, refined and alloyed in a further reaction unit, temporarily stored in a holding furnace and then transferred to a processing unit - in this case an atomization device. The process described here requires a very complex process control in which silicon melt, which is already difficult to handle due to its high temperature, has to be transported across several reactor compartments until the melt is processed. Because the molten silicon is poured off at the top of the melting furnace together with any slag that may be formed and which floats on top of the melt, a further purification step for silicon is necessary. Overall, the process described in EP 0372918 B1 is very complex, time-consuming and therefore prone to errors.

In particular, there is extensive state of the art for further processing of molten metal by atomizing or atomizing the molten metal, which will be further supplemented as follows: For example, DE 3533964 C1 already describes a process in 1987 for producing ultrafine metallic powder in spherical form with a diameter of between 5 and 30 pm. It is proposed to produce metal melts in a crucible and to atomise a melt jet emerging from the crucible in a Laval nozzle system through which a propellant gas flows. The metal melt can be produced batchwise by placing metal ingots in a crucible, melting them there and keeping them in the melt. Alternatively, a continuous process is described in which a metal ingot is pushed into the crucible from above by a complex pushing device so that the ingot is always in contact with the melt and melts on the surface. This document thus discloses a highly complex and error-prone continuous process, in which the metal to be melted must be fed directly to the melt or directly to the Laval nozzle. Furthermore, it is not described that the claimed process is also suitable for silicon.

DE 3311343 A1, which originates from the same inventor and applicant as the above-mentioned document, emphasizes that the metal particles obtainable by melting and atomizing molten metal have a smooth surface. However, such particles are only inadequately suited for the demanding applications of the invention, such as, for example, as a thermal conductor, as will be explained in detail later.

DE 3737130 A1 represents a further development of the system from DE 3533964 C1, whereby the melt is generated in a melting device spatially separated from the Laval nozzle system and is transported to the Laval nozzle system via a U-tube. DE 3737130 A1 thus pursues the objective of ensuring that the melt jet to be atomized has a particularly uniform material flow. In this way, metal particles can be particularly can be produced reproducibly. However, even the process described requires a very complex reaction procedure and is not described in connection with silicon.

The process just described was further improved in EP 1042093 B1. An optimized Laval nozzle with specially arranged gas nozzles reduced gas consumption and increased production output. Only solder melts, Cr-Mo steel, silver, and aluminum melts are mentioned as atomizable metal melts. There is no mention of silicon melts. The resulting particles are spherical and very small, with grain sizes of 8.5–10.1 pm.

From patent DE 10340606 B4, it is known to melt metal in a crucible and atomize it into metal powder using a Laval nozzle. It is crucial to thermally shield the melt nipple, which is used to introduce the molten metal in the crucible into the nozzle, from the cold atomizing gas. Otherwise, the melt cools excessively, which significantly impairs the quality of the produced powder (particle shape, particle size, particle size distribution) or makes atomization impossible. Therefore, a corresponding shield was proposed in DE 10340606 B4. One challenge was to design the shield in such a way that it does not adversely affect the flow profile of the atomizing gas before entering the nozzle, because this flow profile also has a significant impact on the quality of the produced powder. The described process is intended for discontinuous operation, with a stopper rod initially sealing the crucible at the bottom and then, after the melt is generated, pulling it upward to release the crucible's outflow toward the atomization. Continuous operation is only recommended for low-melting metals (Pb, melting point approximately 327 °C). Silicon is not mentioned. EP 3083107 B1 describes a crucible-free melting and atomization process for metals, in which a specially prepared metal rod is melted within an induction coil to generate a molten stream, which is then introduced directly into the atomization nozzle. The process requires a complex lifting and lowering device for holding, raising and lowering the metal rod in order to constantly adjust the position of the rod within the induction coil. An advantage of this process is that no crucible is necessary, which could potentially contaminate the metal to be melted therein. However, the process is comparatively complex and requires the production of corresponding metal rods. These rods must also be produced with the required purity, since purification of the metal to be atomized is no longer possible during the process. This document discloses, for example, aluminum, stainless steel and titanium as suitable metals; Silicon is not disclosed.

As explained above, the methods known to date for melting metals and subsequently processing the melt have in common that they require a very complex and laborious—and thus error-prone—process control. If purification of the silicon used prior to further processing is planned at all, this is carried out in a separate process step. Furthermore, no efficient approaches to continuous process control are known, especially when silicon is used as the starting material.

It would therefore be desirable to provide a simple and cost-effective process for the preparation and further processing of silicon melt with simultaneous possible purification of the silicon, which overcomes the above-mentioned disadvantages overcomes and can be operated particularly cost-efficiently and robustly.

The intended process should also be capable of providing high-performance silicon processing products that are used in demanding applications and are characterized, for example, by high thermal conductivity.

The object underlying the invention is achieved by the first aspect of the invention, directed to a process for the continuous preparation of silicon melt with immediate subsequent further processing, comprising the following steps in the given order:

(A) providing solid silicon (initial silicon) in a melting vessel;

(B) tempering the melting container to a process temperature using a suitable tempering device, whereby the solid silicon is melted to form a silicon melt;

(C) feeding solid silicon into the melting vessel, whereby the fed solid silicon is melted; and

(D) directly feeding the silicon melt obtained after step (C) via a melting vessel outlet located at a bottom of the melting vessel to a molten silicon processing unit to further process the molten silicon.

"Directly" in the sense of the invention means that the silicon melt obtained in step (C) is fed directly to the processing unit, without, for example, being fed into a The silicon slurry is then transferred to an intermediate container. The feed can be effected, for example, contactlessly through a free outlet or via lines made of a suitable material. Suitable materials that are resistant to the silicon melt are known to those skilled in the art.

The method according to the invention has some decisive advantages compared to the prior art methods:

Using the method according to the invention, liquid silicon can be continuously provided for downstream processes, such as granulation or atomization (so-called atomization). This allows the processing of liquid silicon to be carried out without interruption over a long period of time, thereby increasing, for example, process efficiency, product homogeneity, and product quality, and significantly reducing susceptibility to failure.

Compared to standard melting processes, the continuous melting of silicon on a small scale also avoids the complex and dangerous handling and transfer of large volumes of melt. This significantly increases process reliability.

The continuous process also increases throughput, as downstream processes such as granulation or atomization can also be operated continuously. Since neither warm-up nor cooling times are required, the costs for the particulate end product are lower.

Continuous processes also require less technical effort. The process according to the invention also offers the advantage that the silicon can be processed and purified simultaneously in a single process step by removing the impurities in the starting material by separating any slag that may be formed and which comprises the impurities.

When using silicon as a starting material, it has proven surprisingly advantageous to drain the molten metal at the bottom of the melting vessel. Solid silicon added to the melt does not sink to the bottom of the melt until it has completely melted, but floats on top of the melt and is itself melted in the process. Any slag that may have formed, which contains impurities from the silicon, also floats on the melt pool. If the silicon melt is now drained at the bottom of the melting vessel, this occurs with spatial separation of the silicon that has not yet melted and the slag. The drained melt is therefore characterized by increased purity. The process according to the invention is therefore surprisingly capable of carrying out purification directly during the processing of the silicon. The high-purity silicon processing products are particularly well suited for the high-performance applications according to the invention, such as heat transfer technologies.

In order to avoid making the number of pages of the description of the present invention too extensive, only the preferred embodiments of the individual features are listed below.

However, the expert reader should understand this type of disclosure to mean that every combination of different levels of preference is explicitly disclosed and explicitly desired. The melting vessel preferably has a top and bottom side, the spacing of which defines a melting vessel height, wherein the top side preferably has an opening through which the solid silicon supplied in step (C) is supplied.

In a preferred embodiment, during melting in step (B), a silicon melt is formed with a slag floating thereon, which in particular comprises impurities of the silicon used.

As already explained, the continuous melting process according to the invention enables the simultaneous purification of the starting material without complex, additional process steps. When using raw silicon with low purity or even pre-purified silicon with at least 99% purity, the slag, which particularly comprises impurities from the silicon used, floats on the surface of the melt and is thereby separated from the purified silicon flowing out at the bottom. This is a significant advantage of the process. A complex separation of the impurities by decanting or reactive gas atmosphere, as often used in the field, is not required.

The suitable tempering device is, for example, an induction heater.

Preferably, the process temperature is at least 10 °C, preferably at least 20 °C, particularly preferably at least 30 °C higher than the melting temperature of the solid silicon used.

In a preferred embodiment, the method is designed to readjust the process temperature, which is set in step (B), during step (C). This is This is particularly necessary when the temperature of the mixture of silicon melt and solid silicon fed into the melting vessel falls below the desired process temperature. Step (C) preferably starts as soon as a large portion of the initial silicon has melted, for example, as soon as at least 50 wt.% of the initial silicon is in molten form.

The addition of solid silicon in step (C) can be carried out continuously or in portions, preferably continuously. Particularly preferably, the addition of solid silicon in step (C) is arranged such that as much solid silicon is added as molten silicon leaves the melting vessel via the melting vessel outlet.

The feeding in step (C) can be carried out by means of a suitable device, for example by gravimetric dosing with a rotary valve, directly onto the melt or below the level into the melt.

The initial silicon and/or the solid silicon supplied in step (C) is preferably metallurgical silicon.

Preferably, the initial silicon and/or the solid silicon added in step (C) has an average grain size in the range 0.5 mm to 20 cm, preferably 1 mm to 20 mm.

The initial silicon and/or the solid silicon added in step (C) is, for example, commercially available as pre-crushed material or via a grinding process with protective screening with a suitable sieve mesh size, for example, not more than 2 cm. Surprisingly, the grain size according to the invention leads to particular advantages in process control: Silicon particles with a small grain size float on the silicon melt and do not melt quickly enough. Silicon particles with a larger grain size than that according to the invention also do not melt quickly enough. Only the particles with the grain size according to the invention exhibit a melting profile suitable for the process according to the invention.

In a preferred process, the solid silicon added in step (C) has a grain size of less than 10 wt.% with < 150 pm, preferably less than 5 wt.% with < 100 pm, more preferably less than 3 wt.% with < 80 pm, determined by sieve analysis according to DIN 66165.

Maintaining the low levels of microparticles contributes particularly to occupational safety. If the particle size distribution according to the invention is maintained, no complex dust protection measures, such as dust extraction devices on the melting vessel, are necessary.

The solid silicon added in step (C) is in particular added in such a way that it has direct contact with the molten silicon in the melting vessel.

This is particularly advantageous because it allows for better heat transfer between the already melted silicon and the subsequently added solid silicon and high melting rates can be achieved, which promote a continuous process flow.

Preferably, the purity of the initial silicon and/or the solid silicon supplied in step (C) is > 96%, preferably > 98%, more preferably > 99% purity. A preferred method is configured so that the filling level of the melt within the melting vessel is sufficiently high to avoid clogging of the outlet and/or discharge of the slag, in particular at least 10 cm high.

Alternatively, a preferred method is configured such that the fill level of the melt within the melting vessel is at least 10% of the height of the melting vessel, particularly preferably 20-80%, in particular 30-70%.

Maintaining the melt level within the melting vessel at the heights according to the invention is particularly advantageous because this prevents undesired clogging of the outlet with slag and/or discharge of the slag either via the melting vessel outlet (if the melt level is too low) or the top of the melting vessel (if the melt level is exceeded according to the invention). Furthermore, it can be prevented that the melting vessel outlet becomes clogged by silicon that has not yet melted if the melting vessel is too low.

The slag possibly formed in step (C) can be collected on the surface of the melt and/or separated from the surface in portions or continuously, for example by decanting into a slag collecting container.

In a special design, the separation of the slag floating on the surface of the melt is designed in such a way that the filling level of the melt is increased, for example by adding solid silicon, until the filling level reaches the top of the melting vessel. In this case, the top of the melting vessel is open. designed so that the slag, possibly together with a small portion of the melt, overflows and is thus removed from the melting vessel. The overflowing slag can then be collected, for example, in an overflow channel surrounding the melting vessel and located near the top of the melting vessel.

A preferred embodiment of the process is carried out at a melting rate of at least 45 kg of silicon per hour, preferably 60 kg of silicon per hour, particularly preferably 75 kg per hour.

The processing unit can be selected from an atomization unit, a casting unit, a stirring unit, a granulation unit or a purification unit, in particular an atomization unit.

In the case of a casting unit, the molten silicon is poured from the melting vessel outlet directly into a mold for shaping, in which the silicon solidifies.

In the case of a stirring unit, the molten silicon is poured from the melting vessel outlet directly into a stirring vessel, in which the silicon melt is then added, for example, to a metal melt in the stirring vessel to form an alloy.

In the case of a granulation unit, the molten silicon is transferred from the melting vessel outlet directly into a quenching vessel filled with water, where the molten silicon solidifies in the water in the form of granules.

In the case of a purification unit, the molten silicon is fed, for example, to the Czochralski process, in which in which high-purity single crystals are obtained from the silicon melt with the help of a silicon crystallization nucleus.

In a particularly preferred embodiment, the immediately subsequent further processing of the silicon melt represents a production of silicon particles.

The processing unit then particularly preferably represents an atomization unit for molten silicon and the method is configured such that the molten silicon is atomized in the form of a spray jet immediately after melting and solidifies to form silicon particles, preferably predominantly round, spattered and/or nodular silicon particles with irregular surface morphology.

The molten silicon is fed directly to the atomization unit, i.e. the melt is fed directly to the atomization unit without being transferred to an intermediate container.

The atomization in step (D) can be carried out, for example, together with an inert gas as atomization gas selected from nitrogen and argon, in particular nitrogen.

The atomization unit preferably comprises at least one atomization unit. Gas atomization techniques are widely used industrially for atomizing metal melts. A wide variety of nozzle designs are used, all of which have in common that a pressurized atomization gas escapes from one or more gas nozzles and approaches a melt, thereby atomizing it. The atomization unit can therefore be a closed coupled unit or a Laval nozzle, in particular a Laval nozzle.

In a closed-coupled unit, the silicon melt emerging from the melting vessel via the melting vessel outlet is immediately exposed to the atomizing gas, which is released from a plurality of gas nozzles arranged near the melting vessel outlet, in order to achieve shearing of the melt and thus atomization.

When using a Laval nozzle, the melt flows through the nozzle, surrounded by a coaxial flow of atomizing gas in the same direction. The melt is subjected to shear stress and expands, thus achieving atomization.

Laval nozzles for atomizing a metal melt are well known in the field and are described, for example, in US6481638B1.

Preferably, the atomization is carried out at a throughput rate of at least 45 kg per hour, preferably 60 kg per hour, particularly preferably 75 kg per hour.

If the sputtering unit is surrounded by a housing, the sputtering in step (D) is preferably arranged so that the molten silicon solidifies substantially completely before it hits a housing part of the sputtering unit.

The spray jet from the expansion nozzle, which contains the solidifying predominantly round, spattered and/or nodular silicon particles, is preferably maintained continuously for at least 60 minutes, more preferably for at least 90 minutes, particularly preferably for more than 120 minutes. Predominantly round, spiky and/or nodular silicon particles with particularly advantageous properties, which arise from a special surface morphology of the particles, are obtainable in particular via the process according to the invention for the continuous preparation of silicon melt with immediately subsequent further processing in the form of atomization.

The object of the invention is also achieved by a further aspect of the present invention, directed to silicon particles having a predominantly round, spiky and/or nodular particle shape with an average particle size d50 in the range from 20 to 200 pm, in particular 30 to 150 pm, wherein the particles have an irregular surface morphology which is characterized by a raspberry-shaped surface structure.

The d50 value can be determined using methods known to those skilled in the art, e.g., laser diffraction. In laser granulometry, the particles are dispersed in a liquid or gaseous medium. The particle size distribution is then determined based on the diffraction of a laser beam directed onto the sample medium. The d50 value means that 50% of the particles are smaller than the specified value.

The silicon particles produced by the melting and atomization process preferred according to the invention, having a predominantly round, spiky and/or nodular particle shape, thus have a raspberry-shaped surface structure. This means that the silicon particles are composed of predominantly round, spiky and/or nodular primary particles, preferably nodular or predominantly round primary particles, particularly preferably predominantly round Primary particles on which a large number of secondary particles are located. The average diameter of the primary particles is greater than the average diameter of the secondary particles, the average diameter of the primary particles preferably being at least 2 times larger, more preferably at least 3 times larger, than the average diameter of the secondary particles. These secondary particles can also be spherical or protrude from the surface of the primary particle in the form of prisms, prism-like structures or cones. Preferably at least 5% of the surface of the primary particles is covered with secondary particles, preferably at least 10%, particularly preferably at least 20%, as can be confirmed, for example, by scanning electron micrographs. Completely surprisingly, in the silicon particles with a raspberry-shaped surface which are preferred according to the invention, the secondary particles are firmly fused to the primary particles and are partially immersed in the secondary particles, preferably on average to at least 10% of the diameter thereof, as can be seen, for example, in electron micrographs. This distinguishes the silicon particles with a raspberry-shaped surface preferred according to the invention from agglomerates of primary particles and secondary particles, in which the secondary particles adhere to the surface of the primary particles only through physical interactions and can therefore be easily detached from them.

The term "predominantly rounded" (see Figures 1A and 1B) indicates that minor deviations from the spherical shape, such as slight oval shapes, are included.

"Sparkling" is a term well known in the field of

Particle technology and means, for example, that the primary particles have jagged and possibly jagged surfaces. The spattered primary particles are preferably composed of agglomerates of ellipsoidal and/or spherical individual particles.

"Nodular" (see Figures 2A and 2B) is also a term known in the field of particle technology and means, for example, that the primary particles are made up of nodular agglomerates of irregularly arranged individual particles.

Completely surprisingly, it has been found that the silicon particles produced by the preferred process according to the invention exhibit the advantageous properties according to the invention, in particular a comparatively high thermal conductivity. The high thermal conductivity of the particles according to the invention is particularly useful, for example, in the production of thermally conductive plastics in which the particles according to the invention are embedded in a plastic matrix, for example a silicone matrix.

Figure 3A shows a cross-section of the interface between a particle according to the invention with irregular surface morphology (bottom) and the plastic matrix (top). Figure 3B shows that the heat transfer from the particle into the plastic surprisingly occurs via the secondary particles (i.e., the raspberry-shaped surface structure) into the plastic matrix. The particles according to the invention are thus surprisingly characterized by excellent

Heat transport properties from .

Due to their low heat transfer resistance, the particles according to the invention are particularly suitable as thermally conductive fillers for various matrices, such as plastics. The Si particles used as fillers in the prior art are mostly obtained by grinding processes and are prism-shaped (see Figure 4) and have an angular surface.

The disadvantage is that such particles have a large surface area and, as a composite with polymer, bind a large amount of polymer. This significantly increases the viscosity of the plastic composition. Only mixtures with comparatively low filler levels and low thermal conductivity can be produced. At higher filler levels, the composition becomes very stiff and can no longer be processed using conventional methods, such as dispensing. It has also been shown that plastic compositions containing ground silicon particles with angular surfaces exhibit comparatively strong abrasion during processing.

Completely surprisingly, experiments have shown that the plastics provided with the silicon particles according to the invention have a significantly increased thermal conductivity.

The irregular surface morphology of the particles significantly improves the thermal conductivity properties.

The particles according to the invention preferably have an aspect ratio b/1 of at least 0.6, preferably at least 0.7.

The aspect ratio is used as a key figure to describe the particle shape. In older technology, the aspect ratio is often described by the ratio of length to width (1/b). This results in values greater than or equal to 1. In more recent literature, for example according to ISO 9276-6, the aspect ratio is calculated from the inverse ratio of width to length (b/1). This results in Values of less than or equal to 1 result. Both parameters can be converted into one another by taking the inverse. For the purposes of this invention, the aspect ratio is defined as the ratio of the width to the length (b/1) of the particle. The particle width is given by x _c min , the smallest of all measured maximum chords of the particle projection, and the particle length is given by x je max , the longest Feret diameter of all measured Feret diameters of a particle. More detailed information can be found, for example, in "Operating Instructions / Manual Particle Size Analysis System CAMSIZER®", Retsch Technology GmbH, 42781 Haan; Doc. No. CAMSIZER V0115. This results in the following formula for the aspect ratio: b/1 = x c min / X fe max

In a preferred embodiment, the particles according to the invention have a sphericity index SPHT of at least 0.68, preferably at least 0.74.

The sphericity SPHT is calculated from the projection area A of the measured particle in relation to the area of a circle with the same circumference P of the projected particle according to the following formula (more detailed information can be found, for example, in "Operating Instructions / Manual Particle Size analysis System CAMSIZER®", Retsch Technology GmbH, 42781 Haan; Doc. No. CAMSIZER V0115):

SPHT = 4%A / P ²

The characteristic value SPHT corresponds to the square of the circularity C according to ISO 9276-6.

The particles according to the invention preferably have a particle size distribution with a distribution width SPAN ((dgo-dio) /dso) of at least 0.4, preferably at least 0.6, particularly preferably at least 0.8. In a preferred

In this embodiment, the SPAN is between 0.6 and 3.0, in particular between 0.8 and 2.5.

Those skilled in the art know that the standard deviation is not standardized and is only a useful characteristic for evaluating the particle size distribution of different samples if the mean particle sizes of the comparison samples are approximately the same. Therefore, to describe the relative width of the particle size distribution within the scope of this invention, the width of the particle size distribution weighted by the mean particle size dso is used, the dimensionless distribution width SPAN (span), which is defined as:

SPAN ( dgo-dio ) / dso •

For the purposes of this invention, all parameters describing particle size (parameter: mean diameter d50), particle size distribution (parameter: standard deviation sigma and distribution width SPAN), or particle shape (parameter: aspect ratio b/1 and sphericity SPHT) refer to a volume-related distribution. These parameters can be determined, for example, using dynamic image analysis according to ISO 13322-2 and ISO 9276-6, for example, using a Camsizer X2 from Retsch Technology.

The particles according to the invention preferably contain at most 1.5 wt.% silicon particles smaller than 2 μm, preferably at most 1 wt.%, particularly preferably at most 0.5 wt.%, in each case based on the total amount of silicon particles. Particularly preferred silicon particles are essentially free of particle fractions smaller than 2 μm. Essentially free of means that the presence of such particles is tolerated to the extent of a "contamination" of the particles according to the invention and does not interfere with their inventive effect. The silicon particles according to the invention preferably contain less than 15% by weight, more preferably less than 10% by weight, particularly preferably less than 5% by weight of a particle fraction having a diameter of less than or equal to 10 pm, in each case based on the total amount of silicon particles.

The silicon particles according to the invention preferably contain less than 20% by weight, more preferably less than 15% by weight, particularly preferably less than 10% by weight of a particle fraction having a diameter of less than or equal to 20 pm, in each case based on the total amount of silicon particles.

The person skilled in the art is also aware that metallic silicon particles are flammable under certain conditions and that the dusts are explosive. The person skilled in the art also knows that the risk of dust formation, flammability and explosiveness of metal powders increases sharply with decreasing particle size. For this reason, high proportions of very small silicon particles below 20 pm are unsuitable for many applications. Such particles are dangerous to handle due to their low minimum ignition energy and require complex and costly safety precautions during industrial processing.

Silicon particles with an average particle size of over 200 pm are not suitable for many applications of thermally conductive plastic compositions, since such large-grained silicon particles often do not fit into the fine gaps that have to be filled, for example, with gap filler.

A further aspect of the present invention relates to the use of the predominantly round, spiky and/or nodular silicon particles with raspberry-shaped surface morphology according to the invention as a thermally conductive filler. Predominantly round, spiky, and/or nodular silicon particles with a raspberry-shaped surface morphology possess significant application potential, particularly as thermally conductive fillers for gap fillers or potting compounds in the field of e-mobility. Such silicon particles offer significantly higher thermal conductivity while simultaneously reducing sealing and manufacturing costs compared to conventional state-of-the-art thermally conductive fillers.

Figures 1A and B show electron micrographs of the predominantly round particles with a raspberry-shaped surface structure.

Figures 2A and B show electron micrographs of the nodular particles with raspberry-shaped surface structure.

Figures 3A and B show heat transfer profiles of a silicon particle according to the invention in a plastic matrix.

Figure 4 shows silicon particles obtainable by a conventional milling process.

The present invention also relates to the following embodiments described via items 1-28:

1. A process for the continuous preparation of silicon melt with immediate subsequent further processing, comprising the following steps in the specified order:

(A) providing solid silicon (initial silicon) in a melting vessel;

(B) tempering the melting container to a process temperature using a suitable tempering device, whereby the solid silicon is melted to form a silicon melt;

(C) feeding solid silicon into the melting vessel, whereby the fed solid silicon is melted; and

(D) directly feeding the silicon melt obtained after step (C) via a melting vessel outlet located at a bottom of the melting vessel to a molten silicon processing unit to further process the molten silicon.

2. The method according to item 1, wherein the melting vessel has a top and a bottom, the spacing of which defines a melting vessel height, wherein the top preferably has an opening through which the solid silicon supplied in step (C) is supplied.

3. A method according to any one of the preceding points, wherein the melting in step (B) and step (C) forms a silicon melt with a slag floating thereon. 4. Method according to one of the preceding points, wherein the suitable tempering device is an induction heater.

5. Process according to one of the preceding points, wherein the process temperature is at least 10 °C, preferably at least 20 °C, particularly preferably at least 30 °C higher than the melting temperature of the initial silicon or the solid silicon used.

6. The method according to any one of the preceding points, wherein step (C) starts as soon as a large part of the initial silicon has melted, preferably at least 50 wt.% of the initial silicon is in molten form.

7. Process according to one of the preceding points, wherein the supply of solid silicon in step (C) takes place continuously or in portions, preferably continuously.

8. The method according to any one of the preceding points, wherein the feeding in step (C) is carried out by means of a suitable device, for example by gravimetric dosing with a rotary valve, directly onto the melt or below the level in the melt.

9. The process according to any one of the preceding points, wherein the initial silicon and/or the solid silicon supplied in step (C) is metallurgical silicon.

10. Process according to one of the preceding points, wherein the initial silicon and/or the solid silicon supplied in step (C) has an average grain size in the range 0.5 mm to 20 cm, preferably 1 mm to 20 mm. 11. Process according to one of the preceding points, wherein the solid silicon added in step (C) has a grain size of less than 10 wt.% with < 150 gm, preferably less than 5 wt.% with < 100 gm, more preferably less than 3 wt.% with < 80 gm, determined by sieve analysis according to DIN 66165.

12. A method according to any one of the preceding points, wherein the solid silicon supplied in step (C) is supplied in such a way that it has direct contact with the molten silicon in the melting vessel.

13. The process according to any one of the preceding points, wherein the purity of the initial silicon and/or the solid silicon fed in step (C) is > 96%, preferably > 98%, more preferably > 99% purity.

14. Method according to one of the preceding points, which is configured such that the filling level of the melt within the melting vessel is sufficiently high to avoid clogging of the outlet and/or discharge of the slag, in particular is at least 10 cm high.

15. Method according to one of items 1-13, which is configured such that the fill level of the melt within the melting container is at least 10% of the height of the melting container, particularly preferably 20-80%, in particular 30-70%.

16. A process according to any one of the preceding points, wherein the slag possibly formed in step (C) is collected on the surface of the melt and/or is separated from the surface in portions or continuously, for example by decanting into a slag collecting container. 17. A process according to any one of the preceding points, which is carried out at a melting rate of at least 45 kg of silicon per hour, preferably 60 kg of silicon per hour, particularly preferably 75 kg per hour.

18. Method according to one of the preceding points, wherein the processing unit is selected from an atomization unit, a casting unit, a stirring unit, a granulation unit or a purification unit.

19. The method according to item 18, wherein the processing unit is an atomization unit for molten silicon and the method is configured such that the molten silicon is atomized in the form of a spray jet immediately after melting and solidifies to form silicon particles, preferably predominantly round, spattered and/or nodular silicon particles with irregular surface morphology.

20. The method according to item 18 or 19, wherein the atomization in step (D) is carried out together with an inert gas as atomization gas selected from nitrogen and argon, in particular nitrogen.

21. Method according to one of items 18-20, wherein the atomization unit comprises at least one atomization unit, for example selected from a closed-coupled unit and a Laval nozzle, in particular a Laval nozzle.

22. Process according to any one of items 18-21, wherein the atomization is carried out at a throughput rate of at least 45 kg per hour, preferably at least 60 kg per hour. 23. Method according to one of the items 18-22, wherein the sputtering unit is surrounded by a housing and the sputtering in step (D) is preferably arranged such that the molten silicon solidifies substantially completely before it strikes a housing part of the sputtering unit.

24. Process according to any one of items 18-23, wherein the spray jet from the expansion nozzle, which contains the solidifying spherical silicon particles, is maintained continuously for at least 60 minutes, preferably for at least 90 minutes, particularly preferably for more than 120 minutes.

25. Silicon particles having a predominantly round, spiky and/or nodular particle shape with an average particle size d50 in the range from 20 to 200 pm, in particular 30 to 150 pm, wherein the particles have an irregular surface morphology which is characterized by a raspberry-shaped surface structure.

26. Silicon particles according to item 25, which are composed of predominantly round, spiky and/or nodular primary particles, preferably nodular or predominantly round primary particles, particularly preferably predominantly round primary particles, on which a plurality of secondary particles is located, wherein the primary particles are larger than the secondary particles and wherein the secondary particles are spherical or protrude from the surface of the primary particle in the form of prisms, prism-like structures or cones.

27. Silicon particles according to item 25 or 26, wherein at least

5% of the surface of the primary particles with secondary particles are covered, preferably at least 10%, particularly preferably at least 20%.

28. Use of the silicon particles according to any of items 25-27 as a thermally conductive filler.

Claims

Claims 1. A process for the continuous preparation of silicon melt with immediate subsequent further processing, comprising the following steps in the specified order: (A) providing solid silicon (initial silicon) in a melting vessel; (B) tempering the melting container to a process temperature using a suitable tempering device, whereby the solid silicon is melted to form a silicon melt; (C) feeding solid silicon into the melting vessel, whereby the fed solid silicon is melted; and (D) directly feeding the silicon melt obtained after step (C) via a melting vessel outlet located at a bottom of the melting vessel to a molten silicon processing unit to further process the molten silicon. 2. The method according to claim 1, wherein the melting in step (B) and step (C) forms a silicon melt with a slag floating thereon. 3. The method according to any one of the preceding claims, wherein the process temperature is at least 10 °C, preferably at least 20 °C, particularly preferably at least 30 °C higher than the melting temperature of the initial silicon or the solid silicon used. 4. The process according to any one of the preceding claims, wherein the supply of solid silicon in step (C) is carried out continuously or in portions, preferably continuously. 5. A method according to any one of the preceding claims, wherein the solid silicon supplied in step (C) is supplied in such a way that it has direct contact with the molten silicon in the melting vessel. 6. Method according to one of the preceding claims, which is configured such that the filling level of the melt within the melting vessel is sufficiently high to avoid clogging of the outlet and/or discharge of any slag present, in particular is at least 10 cm high. 7. Method according to one of claims 1-5, which is configured such that the filling level of the melt within the melting container is at least 10% of a height of the melting container, particularly preferably 20-80%, in particular 30-70%. 8. A process according to any one of the preceding claims, wherein the slag possibly formed in step (C) is collected on the surface of the melt and/or is separated from the surface in portions or continuously, for example by decanting into a slag collecting container. 9. The method according to any one of the preceding claims, wherein the processing unit is selected from an atomization unit, a casting unit, a stirring unit, a granulation unit or a purification unit. 10. The method according to claim 9, wherein the processing unit is a sputtering unit for molten silicon and the method is configured so that the molten silicon is atomized in the form of a spray jet immediately after melting and solidifies to form silicon particles . 11. Method according to claim 9 or 10, wherein the atomization unit comprises at least one atomization unit, preferably selected from a closed-coupled unit and a Laval nozzle, in particular a Laval nozzle. 12. Silicon particles having a predominantly round, spiky and/or nodular particle shape with an average particle size d50 in the range from 20 to 200 pm, in particular 30 to 150 pm, wherein the particles have an irregular surface morphology which is characterized by a raspberry-shaped surface structure. 13. Silicon particles according to claim 12, which are composed of predominantly round, spiky and/or nodular primary particles on which a plurality of secondary particles are located, wherein the primary particles are larger than the secondary particles and wherein the secondary particles are spherical or protrude from the surface of the primary particle in the form of prisms, prism-like structures or cones. 14. Silicon particles according to claim 12 or 13, wherein at least 5% of the surface of the primary particles is covered with secondary particles. 15. Use of the silicon particles according to any one of claims 12-14 as a thermally conductive filler.