非金屬材料專業(yè)畢業(yè)設(shè)計(jì)（論文）外文翻譯-太陽能級多晶硅長晶速率和雜質(zhì)分布（英文）

1、Materials Science and Engineering A 413–414 (2005) 545–549Growth rate and impurity distribution in multicrystalline silicon for solar cellsRannveig Kvande ?, Øyvind Mjøs, Birgit RyningenNorwegian University of

2、Science and Technology, Department of Materials Science and Engineering, Alfred Gets vei 2, 7491 Trondheim, NorwayReceived in revised form 1 July 2005AbstractThe solidification rate of multicrystalline silicon made by di

3、rectional solidification has been determined by in situ measurements of the solid/liquid interface position in a pilot-scale furnace. Two experiments were conducted where silicon was solidified vertically from the bottom

4、 with a nearly planar interface and cooled after solidification at two different rates. The average solidification rate was found to be 4 × 10?6 m/s, which fits well compared to values calculated from temperature me

5、asurements beneath the crucible. The solidified silicon was examined to determine the carbon and oxygen distribution and the electron lifetime vertically in the ingots. The carbon distribution was quite similar in both i

6、ngots with a concentration of about 4 ppma in the middle of the ingots. A higher oxygen concentration was found in the ingot with slow cooling. This was a result of poor coating which increased oxygen diffusion from the

7、crucible. The electron lifetime was found to be about 10 ?s in the material with fast cooling, whereas the material with slow cooling had an electron lifetime of 2 ?s. More diffusion of iron from the crucible may be the

8、reason for the low lifetime in the material with slow cooling. © 2005 Elsevier B.V. All rights reserved.Keywords: Directional solidification; Multicrystalline silicon; Solidification rate; Impurity distribution; Ele

9、ctron lifetime1. IntroductionMulticrystalline silicon is the most used material in solar cells with a share of more than 50% of the shipped PV modules world-wide [1]. Directional solidification is a com- mon method to ca

10、st multicrystalline silicon for use in solar cells. Here, the silicon is melted and solidified from below by extraction of heat through the crucible bottom, resulting in a nearly planar solid/liquid interface. Most impur

11、ities are segregated towards the top, and the final crystal structure is dominated by large columnar grains parallel to the solidification direction. The solar energy conversion efficiency of multicrystalline solar cells

12、 is typically in the range of 12–15%. The efficiency is mainly limited by minority carrier recombination at disloca- tions and intragranular defects such as impurities, small clusters of atoms, or precipitates [2]. Since

13、 recombination at pure dis- locations is known to be relatively weak, it has been suggested? Corresponding author. Tel.: +47 73 59 68 67; fax: +47 73 55 02 03. E-mail address: rannveig.kvande@material.ntnu.no (R. Kvande)

14、.that the decoration with metallic impurities or precipitates is responsible for the enhanced recombination in these regions [3–5]. It is well known that the solidification process strongly affects the efficiency. The so

15、lidification controls the structure of the material,whilepostsolidificationcoolingisbelievedtoaffectthe dislocation density. The curvature of the solidified interface may affect the crystal morphology and the location of

16、 impurities in the solidified material. Carbon and oxygen are, besides nitrogen, the main impurities in multicrystalline silicon, where carbon originates primarily from insulation and heating elements in the furnace, whi

17、le oxygen diffuses into the melt from the quartz crucible. The present paper describes experiments made in order to determine the growth rate of silicon during directional solidi- fication in a pilot-scale Bridgman furna

18、ce, and to establish the influence of cooling rate on the minority carrier lifetime which is affected by the dislocation density. The growth rate is also cal- culated from temperature measurements in the pedestal beneath

19、 the crucible. The material from both experiments was character- ized to determine the carbon and oxygen distribution vertically in the ingots.0921-5093/$ – see front matter © 2005 Elsevier B.V. All rights reserved.

20、 doi:10.1016/j.msea.2005.09.035R. Kvande et al. / Materials Science and Engineering A 413–414 (2005) 545–549 547Fig. 3. Measured and calculated solidification length as function of time.Combining Eqs. (1)–(3) and assumin

21、g QSi,s = Qcr = Qg = Q, gives the following expression:Q = A(Tm ? Tp)tg kg + tSi,skSi,s + tcrkcr(4)The thermal conductivity coefficients are assumed constant during solidification. The pedestal temperature Tp is measured

22、 by a thermocouple during solidification. Heat transfer through the solidified silicon can be expressed as:QSi,s = Qheat from reaction + Qheat from liquid= ?HSiρSiMSi vA + AkSi,lδ (TSH ? Tm) (5)?HSi is the latent heat fo

23、r silicon, ρSi the silicon density, MSi the molar mass for silicon, and v is the growth rate. By setting Eq. (4) = Eq. (5) the following equation was derived for the growth rate:v = MSiρSi?HSi?? Tm ? TptSi,s kSi,s + tcrk

24、cr + tgkg? kSi,liqδ (TSH ? Tm)?? (6)Eq. (6) is simplified by neglecting the second term under the assumption of no melt superheat, TSH = 0. The thermal conduc- tivity of the quartz crucible is uncertain and the constant

25、kcr in Eq. (6) includes also in reality the thermal conductivity of the coating, which is dependent of thickness and density of the coating. It is therefore difficult to quantify kcr and the growth rate is calculated at

26、two different conductivities. The calculated solidification length corresponds well with the measured val- ues when assuming the thermal conductivity of the crucible to be 1.2 W/mK, as shown in Fig. 3 The measured solidi

27、fication length as function of time was very similar in the two exper- iments, indicating stable and reproducible thermal conditions. An average growth rate of 4 × 10?6 m/s is calculated from the measured values. Th

28、e curvature of the interface during solidification was in both experiments slightly convex with an approximately height difference of 5 mm between centre and halfway out to the sides.Fig. 4. Carbon content as function of

29、 vertical position in ingot, measured and calculated from Scheil equation.3.2. Carbon and oxygen contentThe carbon content, shown in Fig. 4, increases in vertical direction as a result of segregation and exceeds the solu

30、bility limit at the end of solidification. This results in precipitation of SiC, and the last measurement for both ingots will therefore deviate from the real carbon content since FTIR can only mea- sure substitutional c

31、arbon. The carbon distribution in the two ingots was quite similar with a concentration of about 4 ppma in the middle of the ingots. The carbon content was also calcu- lated by Scheil equation using a distribution coeffi

32、cient of 0.058 [10]. The start concentration was set to 1.7 ppma. The measured carbon concentrations fit well with the concentration profile cal- culated from Scheil equation. The oxygen profile results from the competit

33、ion of three mechanisms: diffusion from the crucible, solidification rejec- tion, and evaporation from the melt surface. The oxygen con- centration is particularly high at the bottom for both ingots and decreases from bo

34、ttom to top as shown in Fig. 5. The high con- centration in the lower part is a result of oxygen diffusion into the melt from the quartz crucible. The contact area between melt and crucible reduces as solidification proc