Related Links

Feature

How to handle defective PV cells and wafers


Michael van Dooren

It is important from a cost standpoint to check the quality of incoming wafers and cells before they reach the production process. For example, reducing the occurrence of micro cracks, which cause wafer and cell breakage during production can save money. Michael van Dooren, of test equipment manufacturer RUV Systems, explains.

Ultrasonic Technologies Inc. (US) and RUV Systems BV (Netherlands) recently carried out research into PV manufacturing, in cooperation with cell and module manufacturers in Europe and Asia. As part of this research, Ultrasonic Resonance Vibrations (RUV) technology was used to test cracked wafers and cells, to determine their effects on breakage during the production process. The goal was to determine what part of breakage during production is caused by process equipment, and what part by the quality of incoming products.

The results were striking. It was confirmed that in general the majority of breakage in cell and wafer manufacturing was caused by cracks in incoming wafers or cells. By rejecting damaged wafers and cells the breakage rates in production decreased considerably (a decrease of 21% was noted, but up to a decrease in some cases of around 88%).

The research also found that the other reason for breakage was the process equipment itself.

Focus on incoming cell and wafer inspection

This means that money is being wasted in the solar PV manufacturing industry every day, as in some cases money is invested in wafers and cells that will not make it to the end of the production line. For example, cracks in cells and wafers at a cell manufacturing line with a capacity of, say, 1,800 cells per hour, could lead to losses of between 500,000–2 million Euros each year. The losses due to quality problems caused by cracks at a later stage, as well as claims on reduction of module capacity, are not included in this figure.

Although tuning of process equipment can and will improve yield in general, the focus for process improvement should be on the incoming and outgoing inspection of wafers and cells.

Ultrasonic analysis

A lot of effort was originally invested into finding ways to identify cracks in solar cells and silicon wafers through ultrasonics. The breakthrough was made by a Ukrainian researcher – Dr. Sergei Ostapenko – who had begun academic research in 2002 at the University of South Florida in Tampa, USA.

This resulted in the development of Resonance Ultrasonic Vibrations (RUV) technology, able to measure the resonance vibration modes in silicon wafers at high speed, as well as the crucial software needed to analyse the data automatically.

And in 2008 RUV Systems BV in The Netherlands was established, with the aim of introducing RUV technology in a production environment, in silicon cell and module manufacturing. Test results demonstrated clear progress in detecting inline cracks more reliably.

Earlier this year, a research paper was published, describing the tests and results of improvement research carried out with the aid of ultrasonic analysis, based on the frequency response curve of crystalline.

What did the research conclude?

The silicon wafer is a large contributor to the overall cost of the solar cell. In addition, until recently the cost of silicon had risen dramatically due to a worldwide shortage of polycrystalline silicon feedstock. Despite silicon prices falling back to a more realistic level recently, a key trend in the PV industry remains. To drive down the costs of the technology by reducing the amount of silicon material used.

Nowadays, solar silicon wafers are sliced thinner and thinner, with thicknesses down to 150 microns, and wafer areas increasing.

These trends of course make wafer handling more challenging, and reduce the yield and throughput of solar cell lines due to increased wafer and cell breakage.

The RUV technique was developed for in-line non-destructive crack detection. Its methodology relies on a deviation of the resonance frequency response curve measured on cracked and on identical non-cracked wafers. According to Sergei Ostapenko, “you can use a cracked wine glass as a metaphor. This glass will sound different from an undamaged glass. With wafer and cell inspection it's generally similar, however instead of audio range, it's an ultrasonic frequency range which the human ear cannot detect. At higher frequencies the acoustic wavelength is shorter and the sensitivity to detect small cracks is greatly improved”.

The RUV technique has a basic distinction from most other crack detection methods, In that these are based on visual or optical imaging. This makes them less ideal that RUV because they are either unreliable for detecting cracks, or not fast enough to meet the inline production cycle time specs.

RUV technology allows mechanically unstable silicon wafers to be rejected after ingot cutting before they are introduced into further cell processing; wafers and cells with mechanical defects (such as cracks) to be spotted during production, thus avoiding their in-line breakage; detection of cracked cells before they are laminated into modules (to avoid panel efficiency reduction and product return from the field); the evaluation of equipment on crack initiation; and the evaluation of wafer and cell vendors with regard to cracks.
Some Companies that co-operated with the live RUV tests were Helios Technology in Italy and Photovoltech in Belgium (both cell manufacturers); Ubbink Solar in The Netherlands (member of the Centrosolar Group) and Solar Fabrik in Germany, both module manufacturers. The initial goal was breakage rate reduction in wafer-to-cell, and cell-to-module manufacturing.

Tests were conducted with a minimum of 2,000 (and up to 5,000) wafers or cells from the same ingot or batch for optimum comparison, and divided into two equal batches; a test batch and a control batch as a reference. Obviously the test batch was tested for cracks with RUV, the control batch was not. Then both of the batches were put into production to monitor breakage.

To bring about an optimum sensitivity setting, tests were repeated to tune the RUV test unit in a way that the rejection rate of cracked wafers or cells was as close to the breakage reduction rate as possible. In both the cell and module production environments, these tests realised an actual breakage rate reduction of up to 88%, meaning the difference in breakage rate between the test batch and control batch. Apart from direct cost reduction for wafer, cell or module manufacturers (as well as equipment manufacturers) it shows what part of the breakage is caused by wafer or cell quality with respect to cracks, and what the equipment might contribute to this breakage factor.

For reducing breakage rates the RUV unit does not need to function at the top of its capabilities. According to Ostapenko, “depending on the frequency variation between cells or wafers, roughly 50% of its sensitivity can be sufficient”. This means RUV can do more than reduce breakage. It finds smaller cracks than those that will break during production. In general within the solar PV market there is still insufficient knowledge about the correlation between crack length and long term effects; like the effect on cell efficiency and in what way a small crack propagates during time and temperature changes in the field. In other words, what crack length is acceptable and what is not.

When the necessity for rejecting smaller cracks occurs, RUV will demonstrate added value. Earlier tests in Spain, the United States and China showed that RUV was able to detect small cracks in wafers with a reliability of 95%. Also some provisional test results, which still have to be confirmed, showed an increase of module efficiency when cells with smaller cracks were eliminated from production using RUV.

Responsibility

RUV Systems is a joint venture between Ultrasonic Technologies Inc. and Rimas Systems BV.

Rimas Systems' ceo Mark Verstraten added, “the reason we started this quest for a more reliable crack detection method is simple. Rimas supplies turnkey module manufacturing lines. As an integrator we see a growing desire to share responsibility for the end quality of the modules that our manufacturing lines produce. This means we need to control the quality of the incoming cells, which constantly seemed to change. Optical methods did not convince us, so we searched for reliable alternatives. We then ran into Sergei Ostapenko, who was looking for partners to industrialise ultrasonic technology”.

NB: A fully automated offline system is readily available as a standard configuration. For inline purposes RUV is designed to the customer's requirements and specifications.

Further information

The test research paper, as well as test results, recent papers concerning the technology basics and comparison with other crack detection technologies, are available at the download page of the RUV Systems website: www.ruvsystems.nl


 

Share this article

More services

 

This article is featured in:
Energy infrastructure  •  Photovoltaics (PV)