A COMPREHENSIVE STUDY OF THE PERFORMANCE OF SILICON SCREEN-PRINTED SOLAR CELLS FABRICATED WITH BELT FURNACE EMITTERS A. Ebong 1 , V. Yelundur 1 , V. Upadhyaya 1 , B. Rounsaville 1 , A. Upadhyaya 1 , , K. Tate 1 , A. Rohatgi 1 and J. Kalejs 2 1 University Center of Excellence for Photovoltaics Research and Education, School of Electrical and Computer Engineering, Georgia Institute of Technology, 777 Atlantic Drive, Atlanta, GA 30332-0250 2 RWE Schott Solar, Inc., 4 Suburban Park Drive, Billerica, MA 01821 ABSTRACT: In this paper we report on the screen-printed solar cells fabricated on three types of silicon materials; float zone (FZ), HEM multicrystalline and EFG ribbon with POCl 3 and belt furnace diffused emitters. The belt furnace diffused emitters involved one- and two-side phosphorus spin-on to assess the contaminating effect of the IR belt. The solar cells with POCl 3 emitters and co-firing of screen-printed contacts produced efficiencies of 17.3% on FZ, 16.4% on HEM and 15.5% on EFG ribbon silicon. Solar cells with two-side phosphorus emitters diffused on the belt furnace, produced efficiencies of 17.2%, 16.0%, and 15.1%, respectively, on FZ, HEM and EFG ribbon silicon. However, appreciably lower efficiencies of 15.5%, 15.5%, and 14.1% were obtained, respectively, on FZ, HEM and EFG ribbon silicon for belt-diffused emitters with only one-side phosphorus spin-on with the other side on the belt. This difference in efficiency is reflected in V oc loss for the belt-diffused emitters compared to the POCl 3 emitter cells. The IQE measurements supported that solar cells with belt- diffused emitter with two-side phosphorus spin-on and POCl 3 emitter cells had comparable J sc . However, the cell with phosphorus spin-on on one-side gave much lower IQE because of poor bulk lifetime or the contamination due to direct contact with the belt. These results indicate that the belt emitters can account for appreciable loss in the performance of the many current commercial cells; however, this loss can be regained by applying phosphorus dopant to both side of the wafer. Keywords: spin-on, belt-emitter, screen-printed, solar cell, EFG, HEM 1.0 INTRODUCTION The silicon solar cell performance is controlled by the quality of the p-n junction and its impact on the bulk lifetime during the phosphorus deposition and drive-in. Phosphorus emitters for solar cells can be formed by spray, spin or print deposition of dopant followed by a belt furnace drive-in, or by a liquid source using POCl 3 in a conventional tube furnace. Although belt emitters are simple and cost-effective, they can lead to performance degradation due to contamination from the metal belt and junction leakage. In order to quantify the contamination due to direct contact of silicon with the belt during belt- diffusion on solar cell performance we fabricated solar cells using a) belt-diffused emitters with one-side phosphorus spin-on, with the back of the wafer in direct contact with the belt during drive-in b) belt-diffused emitters with two-side phosphorus spin-on, where the back of the wafer is separated from the belt by the phosphorus glass and c) the conventional tube furnace POCl 3 emitters where phosphorus is diffused on both sides of the wafer. Also, in each case end lifetime in the finished devices was assessed through photoconductance measurements. This study was performed on three different materials including float zone (FZ) silicon, multicrystalline silicon (mc-Si) grown by the heat exchanger method (HEM) and edge defined film fed growth (EFG) ribbon silicon. 2.0 Experimental Three types of silicon materials, including six FZ wafers, six HEM mc-Si and six EFG ribbon, were cleaned in 1:1:2 H 2 SO 4 :H 2 O 2 :H 2 O for 5 minutes, followed by a 3-min rinse in de-ionized (DI) water. This was followed by a clean in 1:1:2 HCl:H 2 O 2 :H 2 O for 5 minutes and a 3-min rinse in DI water. Next the wafers were etched in 15:5:2 HNO 3 :CH 3 COOH:HCl for 5 minutes followed by a 5-min rinse in DI water. A final dip in 10% HF for 2 minutes was performed, followed by a 30 second rinse in DI water. After the cleaning, the wafers were divided into three groups (A, B, C as depicted in Table 1) of six wafers including two FZ, two HEM and two EFG in each group. The Group A wafers were spun with phosphorus on the front side only and Group B samples had spin-on phosphorus on both sides followed by baking at 200°C for 20 min in an oven, and a 6-minute belt furnace diffusion at 925 o C. This resulted in a 40-45 Ω/square emitters with a junction depth of about 0.25 µm and peak concentration of 2.6x10 19 atom/cm 3 . The Group C emitters were formed in conventional tube furnace using POCl 3 at a set temperature of 877 o C, which resulted in a 40-45 Ω/square emitter. After the phosphorus glass removal and DI water rinse, a single layer low frequency PECVD SiN anti-reflection coating was deposited on the front at 400 o C. Next, the Al back contact was printed and dried at 200 o C. The wafers were separated into groups for the study of lifetime and solar cells. This was followed by front Ag grid printing (for the solar cell samples) and dried at 200 o C. Next, the two groups of samples were subjected to the same co-firing cycle in the IR belt furnace. The solar cells were isolated using a dicing saw to define an active area of 4 cm 2 and annealed in forming gas at 400°C for 18 minutes. After the co-firing step, the Al metal was removed from selected samples for lifetime measurement in an Al- etch solution followed by the silicon nitride etch in hydrofluoric acid. The samples were etched in buffered hydrofluoric acid to remove the n + region from the front and the p + region from the back. Each sample was placed in the iodine-methanol containing zip-lock bag to perform the lifetime measurements. Group Number Emitter type & process A Belt-Diffused and one-side phosphorus spin-on B Belt-Diffused and two-side phosphorus spin-on C Conventional tube furnace using POCl 3 Table 1: Summary of experimental design