Recovery Ultra-purified Silicon from Saw Slurry Experiment Report
Duration: From 06-01-2008 to 12-31 2008 Pengfei Xing，Jingjiang Li and Yaobin Wang etc. ，
/>Based on the past six months’ endeavor, the attempts to recover silicon fines from the slurry, which consists of ultra-purified silicon, silicon carbide, polyethylene glycol and traces of other impurities, have reached an encouraging stage and obtained very good results. The detailed procedure of experiments will be described as follows:
1. Test and analysis of the saw slurry
1.1 X-ray fluorescence analysis test (XRF) In order to get the components of saw powder, the slurry was analyzed with XRF E-30T, by which quantitative analysis of chemical components can be obtained. The results are given in table 1. It can be seen the amount of every chemical element in slurry is shown in Table1: the main components include: Si 56wt%, Fe 4.17wt%, O 16.7wt%(here the oxygen value is not accurate because this method and instrument can not be used as the oxygen content detection)，C 22wt%, here Carbon (C) is contributed by SiC and polyethylene glycol organic agent and Oxygen element is from SiO2 and polyethylene glycol organic agent. After calculation, the slurry is mainly composed of 30.5250% Si, which is physically dispersed in the state of simple substance; Si 34.9966%; ferric oxide 5.9570%; H2O and polyethylene glycol 16.7% in total. In addition, it also contains traces of other elements, which can be found from the original data. However, the oxygen element content was somewhat higher than expected. The higher oxygen element could come from oxidation of the silicon, inappropriate transport method or iron’s oxidation, which need to be averted.
Table 1.1 The original and calculated data of the scrap by XRF (a) The original data (b) The calculated data
(b) The calculated data Main components Si SiC Fe2O3 H2O and polyethylene glycol Percentage of components (W %) 30.5250 34.9966 5.9570 26.1260
1.2 Grain size analysis The grain size of slurry powder is tested with the instrument Mastersize 2000. Fig 1.2 showed the result of grain size analysis test. It illustrated that the slurry particles are small and their sizes mostly concentrated at the range of 1-100?m。Among the range of 1-100?m, 10%
of slurry powder is within 0-10?m, 50% of slurry powder is within 0-2.3?m and 90% of it is within 0-14.5?n. The average size of slurry powder is 2.319?m.
Fig. 1.2 The result of particle size analysis 1. 3 X-ray Diffraction (XRD) The crystalline form of all slurry is obtained by X-ray diffraction. From Fig. 1.3, it can be seen that the slurry contains simple substance of silicon; the crystalline index is (111), (220), (311). Diffraction peaks are as follows: 28.44°, 44.73°, 56.12°, and crystal structure is cubic. Silicon carbide has no less than two crystal structures, one is Hexagonal and the other is Rhombohedra. The silicon carbide in slurry is mianly Hexagonal. Moreover, the slurry contains a amount of silicon dioxide, iron oxide and some other impurities. Through the analysis, the physical and chemical properties of the slurry can be
understood comprehensively. It laid a solid foundation for later experiment.
Fig. 1.3 WAXD profiles of the saw powder
2. Hydrometallurgy Experiments
About the experiments and results of, the Si extraction from slurry with alloy by the pryometallurgical method have delivered by Dr. Li Jingjiang. Hence, here we would like just give the experiments and results with the method of Hydrometallurgy. 2.1 Sample preparation A series of experiments was carried out to the slurry by adopting hydrometallurgy. First of all, according to the physical and chemical properties of slurry, a kind of flotation agent was utilized, which is innoxious, non-polluting and cheaper. Then, the sample was refined adopting hydrometallurgy; finally, iron and other impurities were removed by appropriate chemical reagent. The experimental flow chart is shown in Fig 2.1:
The experiment flow chart with hydrometallurgy
2.2 X-ray diffraction (XRD) X-ray diffraction pattern was shown in Fig.2.2. It is obvious that the diffraction peaks of impurities decreased, but the diffraction peak of silicon increased drastically.
Fig. 2.2 WAXD profiles of the sample
From Fig.2.3, it can be seen that impurity peaks of sample reduced significantly compared to original slurry, and the relative intensity of silicon carbide diffraction peaks declined obviously compared with that of the slurry, that is, the content of SiC notably declined. It can also be observed that the silicon diffraction peaks of samples changed little. Based on the observation and discussion above, it can be found that the refinement methods of polycrystalline silicon might increase effectively without changing the properties and grade of its purity.
(a) The original slurry
(b) The sample after refining
Fig. 2.3 WAXD contrast profiles of the sample before and after refinements 2.3 X-ray fluorescence analysis test system (XRF) A lot of the attempt and experiments are carried out with the hydrometallurgy methods. Finally we get an ideal experimental result. The result is presented in Table 2.1. From Table 2.1, it can be seen that silicon content in the sample has reached the level of 94.8wt%, and carbon is about 4.5wt%. Compared to the result in Table 1.1, it is proved that the result is good enough for next stage experiments: Si in chemical quantization increase to 94.8wt% from 56.7wt%, meanwhile, the carbon decrease to 4.5wt% from 22.1wt%. Table 2.2 shows the comparison of the components between the original slurry and purified sample. It is obvious that silicon component in slurry notable increased and SiC of it significantly decreased. The amount of SiC can be calculated out from carbon chemical quantization the as follows: C 12 4.48wt% -------------- SiC 40 X
X(SiC) = 40*4.5%/12 = 14.6wt%
Table 2.2 XRF results for the purified slurry powder by hydrometallurgy method
Table 2.3 The comparison of the components between slurry and the purified sample Main Element Si SiC Fe2O3 H2O and polyethylene glycol Percentage Components (wt %) Original slurry 30.5250 34.9966 5.9570 26.1260 Purified sample 85.4 14.6 0.458 0
Recently, we have received analysis result of our samples A and B from REC/Alkem.
Compared with our analysis result done in China Northeastern Universtiy, it can be seen that the analysis result from REC/Alkem is much similar to ours (Sample B), for example, the content of C, Fe, Cu, Zn, it proves that our result are very reliable and convincible. Table 1: The comparison of the analysis results which from REC/Alkem and from NEU (China Northeastern University) Enhet % % % % % % % % % Parameter C Si-met Fe Cu Zn Al Ti Ca Mg REC/Alkem Sample A 4.22 40.6 0.21 0.027 0.0022 0.010 0.0019 0.048 0.018 Sample B 4.91 58.2 0.48 0.056 0.0058 0.015 0.0028 0.076 0.028 0.107 NEU(Northeastern University) Sample B 4.48 94.8 (elememt) 0.46 0.066 0.0057 0.025
The only difference in the two results is the content of Si (REC=58.2wt%, and NEU=94.8wt%). The reason is that: in our result of sample B, Sitotal(element)=Si(metal)+ Si(SiC) + Si (SiO2)=94.8wt%, but in REC/Alkem results, Sitotal = Si(metal). According to the content of Carbon, we have calculated out above that the SiC in Sample B is 14.6%, and the amount of the remainder is 100-Si(mat)-SiC=100-58.2-14.6=27.2wt%, what is the remainder 27.2%? We think that most possibly it is Si and/or SiO2 because the Sitotal(element)= Si(metal)+Si(SiC)+Si (SiO2) =94.8% in sample B. Why the metal Si become SiO2? It is since Si in slurry is immersed in water for a long time, about more than half a month before we receive it, thus part of Si metal in powder is oxidized into SiO2 by oxygen at the wet condition. But the oxidization of Si can be avoided very easily if only making the powder dry. After finishing saw the Si wafer, we just bake the slurry in time, remove water and make it dry, then package it in seal or filled the package inert gas for example Nitrogen or Argon. These procedure will easily protect the metal Si in slurry from oxidization. I can be seen that, if only the saw slurry were dried in time and held properly before sending to us, we definitely can refine more high Si content powder from the saw slurry, at least 80-83wt% Si using our hydrometallurgy process. This is an important breakthrough for
the technique to obtain 6N polycrystalline silicon. Now we are optimizing the parameters of our process by Ph.D and Master students. After optimization, it is very available to increase the content of Si more than 90％ in the powder. In fact, we can go to next step with separating SiC from Si by melting the powder as long as Si contents rising to 80%. After melted, Si of powder will gather into an ingot and SiC will be removed. If the content of Si in powder rising to more than 85%, it will be easier done and the purity of Si will be more high.
3. Melting the Si powder and Cast a Si ingot
Next step, in semi-industrial scale experiment, the samples obtained by
hydrometallurgical process will be melted in vacuum. The process is shown in Fig. 3.1. After melting, Si powder become a Si ingot and SiC will be removed. Through melting process, Si content of ingot would reach more than 99%. In order to make Si ingot becoming the multicrystal Si(6N), a subsequent treatment for Si ingot will be needed.
Fig. 3.1 The flow chart of making Si ingot
4. Unidirectional Solidification and Zone Refining
In semi-industrial scale experiment, the further purification of Si ingot can be done with unidirectional solidification and zone refining. Fig. 4.1 gives the process flow chart of unidirectional solidification and Fig.4.2 illustrate the flow chart of zone refining
Fig. 3.1 The flow chart of unidirectional solidification
Fig. 3.2 The flow chart of zone refining. After unidirectional solidification and zone refining, it is available that the purity of Si ingot will reach to the level of multicrystal Si(6N) for solar energy using.
As mentioned above, a large scale experiments are done. The experimental results proved that the hydrometallurgical process we invented is completely feasible. After dealing with the hydrometallurgical process, most of SiC and other impurities in the slurry were removed without bringing in any other harmful ions. The purity of Si in powder can arrive to more than 85% as long as the saw slurry is drying in time、protected from oxygen and held properly. Then, in semi-industrial scale, Si ingot can be obtained by melting the purified
powder by our hydrometallurgical process. The ultrapure Si(6N) would be acquired as long as the later refining experiments could be fulfilled. Meanwhile, a quantities of slurry will be needed to support the subsequent experiments of casting ingot, unidirectional solidification and zone refining. Finally, it’s important to mention that the work speed will be greatly accelerated unless sufficient funds are provided.