Wafer Manufacturing
Qualitative electrical wafer quality determination
Problem:
A high resolution image of an as-cut wafer is required to qualitatively determine the electrical quality of a wafer. High resolution is required to see small local defects such as dislocations. Using methods such as ยต-PCD this is very slow, ranging from 30 minutes to several hours per wafer, depending on resolution.
Solution:
PL Imaging provides high resolution images of both local defects and large areas of poor electrical performance in as-cut wafers in seconds.
Data:
Wafer with a high density of Dislocations
Wafer with a high density of impurities
Wafer with Edge/corner impurities
Links:
Performance Limitations of mc-Si Solar Cells Caused by Defect Clusters
Bhushan Sopori et al, National Renewable Energy Laboratory
This paper describes characteristics of defect clusters, and shows, through theory and experiment, that defect clusters typically lower cell efficiency by 3 to 4 absolute percentage points. To recover this efficiency loss, it is necessary to getter precipitated impurities.
Comparing Luminescence Imaging with Illuminated lock-in thermography and carrier density imaging for inline inspection of silicon solar cells
Jonas Haunschild et al, Fraunhofer Institute for Solar Energy Systems
Effective lifetimes measured during wafer processing do not sale with the open circuit voltage of the cell. PL images taken on as-cut wafers already reveal defects from the crystallization which limit the cell efficiency. In particular, crystalline defects, e.g. dislocations have a significant influence on the open circuit voltage of a solar cell. Thus, by recognizing these features, an electric rating for the wafers can be done before production.
Incoming Wafer Electrical Quality Control
Problem:
Wafers incoming to the cell factory are not sorted by electrical quality today, and all wafers enter the cell line. Physical inspections are done, and wafers failing basic specs (e.g. gross cracks or chips) are rejected from the line. All other wafers are processed into cells. The electrical performance is determined only at the end of the cell process, in the IV tester. At this stage, sorting based on wafer quality is too late since the cost to process all the wafers is already at risk.
Solution
With the QS-W1, it is possible to inspect all wafers incoming to the cell line and sort based on electrical quality, or reject wafers below a user defined electrical quality spec. In this way, wafers can be rejected before the cell processing cost is spent. This sorting of wafers enables cell makers to increase the electrical output of the line. Poor quality wafers are rejected to wafer suppliers.
Outgoing Wafer Electrical Quality Control
Problem
Wafers outgoing from the wafer factory are not sorted by electrical quality today. Physical inspections are done, and wafers are graded into bins of physical quality. Wafers within a given physical quality bin are essentially sold to cell customers at the same price per wafer. If wafer makers were able to distinguish the electrical performance of their wafers, and demonstrate superior performance to their competitors, they would be able to charge a premium for their wafers.
Solution
Wafer manufacturers can use BTI tools to qualify the electrical quality of wafers after wafer manufacturing is complete. BTI tools can be a complementary addition to existing practices. In addition to sorting based on traditional metrics such as thickness, shape and roughness, wafers can now be sorted based on electrical entitlement.
Qualitative Imaging
Qualitative luminescence imaging is a powerful application that can be extremely useful in both production and in research. No particular analysis is performed on the images in this case, but rather features that are observed in the images are correlated, for example, with problems of specific processing equipment or processing conditions. BTi has used its technology in this manner to identify a large number of problems, such as handling issues, unsuitable hardware components, equipment contamination or poor processing recipes (see figure below).
Such qualitative imaging is particularly effective during the start up phase of new production lines, during the testing of new processing methods, to evaluate the quality of new wafer supply, or for the development of new cell concepts. The short data acquisition time in PL imaging is an enormous benefit in that context, as it allows a large number of samples to be measured in short time, even with manual sample loading.
Qualitative PL images taken at UNSW on silicon wafers, displaying a variety of processing problems: (a) sample/furnace contamination, (b) Crack, (c) dislocations resulting from non-ideal cleaving, (d) tweezers handling, (e) four point probe measurement marks, (f) circular marks from a micrometer and dome shaped defect region caused by a dent in the substrate holder of a PECVD machine.