METHOD FOR DETERMINING TRACE METALS IN SILICON

Abstract

A method for determining an amount of metallic impurities within silicon. The method includes the steps of (a) providing a rodlike silicon sample and a rodlike seed crystal in a zone melting apparatus, (b) zone melting to form a single silicon crystal having a conical end region with a droplike melt forming at the end of the single silicon crystal in a separation step, (c) cooling of the droplike melt to form a solidified silicon drop, (d) partial or complete dissolution of the silicon drop in an acid, and analyzing the solution obtained in step (d) by a trace analysis technique. Wherein the separation step further includes a remelting step for the silicon sample to reduce its diameter, forming a droplike melting zone, and separation of the seed crystal and the silicon sample by moving the seed crystal and the silicon sample apart from one another.

Claims

1-15. (canceled)

16. A method for determining an amount of metallic impurities within silicon, comprising the steps of: a) providing a rodlike silicon sample and a rodlike seed crystal in a zone melting apparatus; b) zone melting to form a single silicon crystal having a conical end region, with a droplike melt forming at the end of the single silicon crystal in a separation step; c) cooling of the droplike melt to form a solidified silicon drop; d) partial or complete dissolution of the silicon drop in an acid; and e) analyzing the solution obtained within step d) by a trace analysis technique; wherein the separation step further comprises the steps of: (i) remelting of the silicon sample to reduce its diameter, where for a first time interval the direction of movement of the silicon sample and of the seed crystal is reversed relative to its previous direction of movement, to form the conical end region; (ii) forming a droplike melting zone, where for a second time interval the movement of the seed crystal is halted and the direction of movement of the silicon sample is reversed again; and (iii) separation of seed crystal and silicon sample, where the direction of movement of the silicon sample is reversed and said sample for a duration of 5 to 20 s has a speed of movement of 150 to 400 mm/min.

17. The method of claim 16, wherein after the remelting, the silicon sample in an end region of length l has a diameter which is less than or equal to the diameter of the single crystal at its contact face with the melt.

18. The method of claim 17, wherein the diameter of the contact face of the single crystal with the melt is 3 to 8 mm, preferably 4 to 6 mm.

19. The method of claim 17, wherein the diameter of the silicon sample in its end region of length l is 2 to 8 mm, preferably 3 to 6 mm.

20. The method of claim 17, wherein the length l of the end region of the silicon sample corresponds to one to three times its diameter.

21. The method of claim 16, wherein the silicon sample during remelting is moved at a higher speed of movement than the single crystal.

22. The method of claim 21, wherein the speed of movement of the silicon sample is 5 to 15 mm/min, preferably 7 to 13 mm/min, more preferably 9 to 11 mm/min.

23. The method of claim 21, wherein the speed of movement of the single crystal is 2 to 10 mm/min, preferably 3 to 8 mm/min, more preferably 4 to 6 mm/min.

24. The method of claim 16, wherein the first time interval during remelting lasts 30 to 300 s, preferably 90 to 240 s, more preferably 60 to 120 s.

25. The method of claim 16, wherein the formation of the droplike melting zone and a speed of movement of the silicon sample is 1 to 5 mm/min, preferably 2 to 4 mm/min.

26. The method of claim 16, wherein the second time interval lasts 1 to 4 s, preferably 2 to 3 s.

27. The method of claim 16, wherein the separation of seed crystal and silicon sample and a speed of movement of the silicon sample is 250 to 350 mm/min.

28. The method of claim 16, wherein the cooling of the droplike melt, the movement of the silicon sample is halted and the seed crystal is removed in its original direction of movement with a speed of movement of 150 to 400 mm/min, preferably 250 to 350 mm/min, from the silicon sample.

29. The method of claim 16, wherein the silicon drop is dissolved partially by immersion in the acid for a duration of 3 to 15 min, preferably 5 to 10 min.

30. The method of claim 29, wherein the acid comprises a mixture of concentrated nitric acid and hydrofluoric acid in a ratio of 4:1 to 3:1, preferably 2:1 to 1:1.

Description

[0058] FIG. 1 shows the zone melting before the separation step according to the prior art.

[0059] FIG. 2 shows a comparison between pin-nub and freeze-nub methods.

[0060] FIG. 3 shows a pin-nub on a single crystal.

[0061] FIG. 4 shows a pin-nub in an acid bath.

[0062] FIG. 5 shows the concentration of iron in a step etch.

[0063] FIG. 6 shows the concentration of copper in a step etch.

[0064] FIG. 1 shows schematically a zone melting apparatus 10 with a silicon sample 12 and a single crystal 14, which are present shortly before the separation step. The silicon sample 12 and the single crystal 14 are each clamped into a rotating shaft, which for reasons of clarity is not illustrated. The direction of rotation is indicated by the arrows 11a, 11b, the direction of movement by the arrows 13a, 13b. The silicon sample 12 and the single crystal 14 are joined to one another via a melting zone 20, which is heated by an induction coil 21. The silicon sample 12 is a rodlike polysilicon sample having a cylindrical region 16 and a conical region 18 in the melting zone 20. The original length of the polysilicon sample was about 15 cm. The diameter of the region 16 is about 20 mm. The single crystal 14 recrystallized on a single seed crystal 15 consists of an initial cone 22, a cylindrical section 24, and a conical section 26 in the melting zone 20. The length of the single crystal 14 is around 110 mm, and the diameter in the cylindrical section 24 is about 14 mm.

[0065] FIG. 2 shows a pin nub 30 on a conical end region 27A of a single crystal 14A which has been produced according to the method of the invention. Illustrated in comparison to this is a freeze nub 32 (according to DE 10 2010 039 755 A1) on a conical end region 27B of a single crystal 14B. The end regions 27A, 27B may have a length of 3 to 6 cm. The numbering of elements already identified in FIG. 1 is retained, with elements from the pin nub method being distinguished by an “A” and elements from the freeze nub method by a “B”. The corresponding silicon samples 12A and 12B, with their end regions 34 and 36 resulting from the methods, are likewise illustrated. There are no substantial differences in the shape of the single crystals 24A, 24B. There is not necessarily any difference in the diameter of the cylindrical sections 24A, 24B. The essential visible difference after the implementation of both methods lies in the end regions 34, 36 of the single crystals 12A, 12B. In the case of implementation of the freeze-nub method, the part of the silicon sample 12B that is located in the melt during separation forms an end region 36 which is like a round head. This end region 36 corresponds in general to 50-80% of the melt. Hence only 20-50% remains in the form of the freeze nub 32 at the conical end region 27B of the single crystal 24B. Conversely, in the case of the pin-nub method, as a result of the remelting, an end region 34 is formed with a partial section 38 having a length l, which in terms of diameter corresponds at most to the diameter of a contact face 39 between the conical end region 27A and pin-nub 30. In this way, the part of the silicon sample 12A that is located in the melt during separation is reduced to a small fraction 37. This fraction 37 corresponds customarily to only 5-10% of the melt during the separation procedure. Commonly, thus, only 5-10% of silicon melt is lost to the analysis.

[0066] FIG. 3 shows a pin-nub 30 which has formed, after the cooling, on the conically tapering end region 27A of the single crystal 14A (cf. FIG. 2). An outer layer 40 which has been marked is intended to illustrate approximately the etch removal, which is provided for the subsequent analysis on partial dissolution of the silicon drop in an acid. The metallic impurities are contained in this outer layer 40. The diameter of the contact face 39 with the pin-nub 30 is about 5 mm. This diameter also corresponds essentially to the diameter of the contact face of the single crystal 14A in its conical end region 27A with the melt at the moment of separation.

[0067] FIG. 4 shows a conical, acid-filled vessel 50 in a mount 52. The single crystal 14A with the pin-nub is immersed in the vessel 50, where it is secured by a holding bracket 54. As a result of the conically tapering vessel, the amount of acid used for the partial dissolution can be reduced.

EXAMPLE 1

[0068] 6 drill cores (silicon samples) all having a diameter of 22 mm and a length of around 8 cm were taken from a silicon rod pulled using the Czochralski process and contaminated with a known contamination with the metals Fe, Cr, Ni, Cu, Zn and Sn. The samples were etched with a cleaning etch for about 15 min in an acid mixture of HF (45%) and HNO.sub.3 (65%) in a ratio of 1:6, and were rinsed with ultra-pure water. Preparation for this purpose took place under clean-room conditions (class 10).

[0069] The silicon samples were each installed into an FZ apparatus (upper pulling shaft: silicon sample; lower pulling shaft: seed). The seed used was a single silicon crystal. The zone melting operation is set out in Table 1, with reference also being made to the figures. Speeds v of movement that are labelled with a negative sign indicate the usual pulling direction when the melting zone is moved through the silicon sample. Referring to FIGS. 1 and 2, a negative speed of movement denotes a downward movement of the single crystal 14A and/or of the silicon sample 12A in relation to the induction coil 21. A positive sign indicates an upward movement.

TABLE-US-00001 TABLE 1 Step Pin-nub method Preheating The Si sample 12A is heated using a preheater Generator power (P): 36 kW (70% power) Coupling-in Induction power couples in Melt drop is formed on the Si sample 12A Attachment Melt drop is joined to the seed at the level of the induction coil 21 (the two pulling shafts are moved toward one another manually) Formation of Parameters for cylindrical section 24 initial cone 22 diameter of 14 mm: P = 14 kW (40% power) v top pulling shaft: −2.4 mm/min v bottom pulling shaft: −5.8 mm/min Pulling length (travel of bottom pulling shaft): 5 mm Cylindrical P = 14.2 kW (40.5% power) section 24A v top pulling shaft: −2.4 mm/min v bottom pulling shaft: −6.0 mm/min Length: (travel of bottom pulling shaft): 70 mm Diameter: 14 mm Transition to Step 1: region 27A P Generator: 13.7 kW (39.0% power) v top pulling shaft: −1.8 mm/min v bottom pulling shaft: −10.0 mm/min Pulling length (travel of bottom pulling shaft): 10 mm Step 2: P = 13.5 kW (38.5% power) v top pulling shaft: −1.2 mm/min v bottom pulling shaft: −10.0 mm/min Pulling length (bottom): 10 mm Region 27A P = 13.3 kW (38% power) v top pulling shaft: −1.0 mm/min v bottom pulling shaft: −14.5 mm/min Pulling length (travel of bottom pulling shaft): 22 mm Diameter of contact face 39: 5-6 mm Remelting P = 13.3 kW (38% power) v top pulling shaft: +10 mm/min v bottom pulling shaft: +6 mm/min Travel of bottom pulling shaft: 8 mm Diameter of partial section 38: 5-6 mm Formation of P = 13.7 kW (39% power) pin-nub 30 v top pulling shaft: −2.0 mm/min v bottom pulling shaft: 0 mm/min Pulling length (travel of bottom pulling shaft): 0 mm Separation P = 0 kW v top pulling shaft: +350 mm/min v bottom pulling shaft: 0 mm/min Travel of top pulling shaft: 40 mm Travel of bottom pulling shaft: 0 mm Cooling Step 2: P = 0 kW v top pulling shaft: 0 mm/min v bottom pulling shaft: −350 mm/min Travel of top pulling shaft: 0 mm Travel of bottom pulling shaft: 300 mm

[0070] The rotary speed of the single crystal 14A was typically from 15 to 30 rpm, and that of the silicon sample 12A was typically from 3 to 10 rpm in the opposed direction of rotation (cf. FIG. 1, arrows 11A, 11B). After separation and cooling, the largest diameter of the pin-nub was around 8 mm.

[0071] After dismounting, the pin-nub 30 was transferred with a holding bracket 54 into an acid-filled conical vessel 50 of perfluoroalkoxyalkane (PFA) under clean-room conditions (class 10) and partially dissolved (partially etched). The acid consisted of a mixture of HNO.sub.3 (69 wt %) and HF (40 wt %) in a ratio of 1:1. The vessel 50 was filled with around 6 ml of acid. The fully immersed pin-nub was exposed to the acid for 6 min. The pin-nub was subsequently washed with 1 ml of fresh acid. The etched solution obtained was then concentrated at a temperature of 250° C. for around 30 min. The residue obtained was dissolved with a mixture of 25 μl of HF (40 wt %), 25 μl of HNO.sub.3 (65 wt %) and 1450 ml of ultra-pure water to give a measurement solution. This treatment was carried out with each of the six pin-nubs obtained.

[0072] On measurement by means of ICP-MS, determinations were made of the elements Fe, Cr, Ni, Na, K, Sn, Zn, Al, Cu, Mo, La, Cs, Ce, Te, Sc, Se, Ti, Ta, Ge, W, Mg, Ag, Li, V, Mn, Zr, Pb, Y, Sr, Ba, Bi, Cd, Sn, As, Ru, Rb, U, Ga, In, Ca and Co. Blank values were generated as well. For this purpose, 4 conical vessels were filled with acid and processed analogously without samples. The amounts measured in the context of these blank values were averaged and subtracted in each case from the amounts measured for the samples. The calculated value was then expressed relative to the total weight of the single crystal.

[0073] This shows significantly higher recovery rates (RR) for selected metals in comparison to the freeze-nub method. The values in question are average values from all six silicon samples.

TABLE-US-00002 TABLE 2 RR in % Fe Cr Ni Cu Zn Sn Freeze-Nub 37 40 54 38 37 41 Pin-Nub 94 95 95 96 94 94

[0074] Table 3 shows reduced detection limits in comparison to the freeze-nub method.

TABLE-US-00003 TABLE 3 DL (pg/g) Fe Cr Ni Cu Zn Na Freeze-Nub ≤20 ≤20 ≤20 ≤10 ≤10 ≤20 Pin-Nub ≤2 ≤1 ≤1 ≤1 ≤1 ≤1

[0075] FIGS. 5 and 6 show the concentration profile of iron (C.sub.Fe, FIG. 5) and copper (C.sub.Cu, FIG. 6) in measurement solutions obtained after 5 min, 16 min, 29 min, 44 min and 79 min of partial etching of a pin-nub (step etch: after each step, the pin-nub was transferred to fresh etch solution). Corresponding curves were prepared for chromium and nickel, with each experiment being repeated once (samples 1 and 2 in each case). The results are set out in Table 4.

TABLE-US-00004 TABLE 4 Etch removal per step in % Fe Cr Ni Cu 1.sup.st etching sample 1 98 96 98 98 2.sup.nd etching sample 1 2 3 2 2 3.sup.rd etching sample 1 0 0 1 0 4.sup.th etching sample 1 0 0 0 0 5.sup.th etching sample 1 0 0 0 0 1.sup.st etching sample 2 98 97 98 98 2.sup.nd etching sample 2 2 2 1 1 3.sup.rd etching sample 2 0 1 0 0 4.sup.th etching sample 2 0 0 0 0 5.sup.th etching sample 2 0 0 0 0

[0076] After an etching time of just 5 min and a removal of only 0.06 g, at least 96% of the metallic impurities had been etched away in all cases. This result shows that by cleaning of a freeze tip (cf. EP 0 349 117 A2) with a cleaning etch, a portion of the metallic impurity is already lost to analysis.

EXAMPLE 2

[0077] Various Parameters were Tested in the Zone Melting:

[0078] 7 silicon samples (drill cores) with 19 and 22 mm were used. By adapting the corresponding parameters in the zone melting, different diameters (D) were obtained on the single crystal and on the end region of the silicon samples. Moreover, different travel lengths of the lower pulling shaft were tested during remelting, and different pin-nub sizes were obtained. In addition, samples with different single crystal weights were pulled and tested. The results are summarized in Table 5.

TABLE-US-00005 TABLE 5 (D) Single Sample Length of Length sample RR crystal section Remelted conical of end on Nub (averages (D) (D) silicon region remelts separation size Fe, Cr, Ni, Cu) # mm mm g mm mm mm g % 1 12 22 30 30 8 4 0.4 97 2 12 19 20 30 8 4 0.3 94 3 14 22 30 35 8 4 0.5 95 4 15 22 20 50 8 3 0.4 97 5 10 22 40 20 8 3 0.4 96 6 14 22 50 30 5 5 0.4 94 7 10 22 15 30 15 8 0.4 90

[0079] The recovery rate was always between 90% and 97%.