CHARACTERIZATION OF CZOCHRALSKI SOLAR CELLS GROWN WITH A RECHARGE PROCESS

G. Mihalik and B. Fickett

Siemens Solar Industries, Vancouver, WA

J. Palm

Siemens Solar Industries, Munich, GDR

Abstract

Siemens Solar Industries (SSI), in cooperation with the Northwest Energy Efficiency Alliance (NEEA), has developed a multiple batch recharge system.  The multiple batch recharge Czochralski process dramatically reduces costs, energy consumption, material handling, and labor while increasing yields, throughput, and process capabilities.  Data will show the levels of metals, Carbon, and Oxygen in ten sequential ingot sections grown at SSI.  Solar cells and modules manufactured from recharge ingots show consistent performance with those grown from standard batch processes.  

Goals

1)Reduce power consumption by 40% per run (kwh/kg)

2)Reduce Argon consumption by 50% per run (kwh/kg)

3)Increase productivity by 15% (mm/day)

4)Improve or maintain the quality of the ingot produced.

5)Transform the market place, specifically semi-conductor crystal growers in the Northwest, into lower consumers of power and Argon.

Achievements

1)Reduced power consumption by 51% per run (kwh/kg)

2)Reduced Argon consumption by 85% per run (kwh/kg)

3)Increased productivity by 35% (mm/day)

4)5% higher valued solar cells-better axial Oi

5)Agreement with Wacker and in discussions with MSA.

• 30% lower ingot costs

• 15% higher yields

Introduction

In April of 1998, Siemens Solar Industries entered into a jointly funded project with the NEEA to develop a more energy efficient approach to crystal growing.  The established project objectives, shown right, were very ambitious.  Under the scope of the project, SSI’s hotzones were extensively modified with an energy efficient hotzone (EEH) to meet the technical goals shown above.  However, the addition of insulation resulted in a 20% reduction of the initial charge capacity.[i]  Therefore, SSI developed a system to add material to the crucible during meltdown.  The system was called a topping off system.  With the topping off system in place it was also possible to develop a recharge or semi-continuous process.  In addition, a market analysis of the Northwest semiconductor manufacturers, engineers, and managers showed much of the interest in the project focused on recharge technology. [ii]  With that in mind, SSI shifted its efforts from topping off the initial charge to refining multiple batch recharges.  The achievements, shown right, have been demonstrated for systems with an improved hotzone and recharge process. 

Results and Discussion

1.        GDMS Analysis

The GDMS is capable of detecting trace levels,  ( > 5 ppb ) and in some cases ( > 2 ppb), of metallic contaminants.  Samples were cut, rinsed, and labeled before shipping.  The laboratory etched the wafers and conducted the tests.  The GDMS analysis looked for detectable levels of Fe, Cr, Mo, Al, Ni, and Ba.  These metals were selected because of their segregation levels and detriment to the ingot’s electrical characteristics.  Segregation would suggest the highest level of impurities would be found in the bottom of the last ingot section analyzed.  The investigation of all recharge samples with GDMS showed the accumulation of impurities did not reach the detection limits for any of the metals. 

       2.        FTIR Analysis

Samples grown from randomly selected recharge sections were processed into p-line solar cells, shown left, using the SSI manufacturing facility in Camarillo, California.  A total of 11 recharge sections were processed into 315 micron wafers and subsequently solar cells.  The 11 sections included material from the first, middle, and ends of recharge runs.  A batch of approximately 7500 wafers (Test) were compared to cells prior (Pre) to the experiment and after the experiment (Post).  The cells processed before and after the experiment were from ingots grown in the standard batch process.  A chart of current at a rated voltage (Ivr) shows the consecutive cell results from Pre, Test, and Post solar cells. 

The cells were then assembled into SP75 modules and tested for light induced degradation (LID).  Degradation would be expected if excessive metallic impurities were accumulating in the subsequent recharge sections.  While the tests show a slightly higher level of degradation for power and Isc, the test modules performed consistently with one another.  Again, the affects of recharging on LID are comparable with standard ingots.