Today’s advanced wire saws are making it possible for manufacturers of crystalline silicon solar panels to respond to competition from lower-cost thin film modules by literally thinning their own costs. Progressive improvements in the performance and reliability of wire saws since their advent 25 years ago have resulted in a reduction of practical wire diameter to as little as 120-100?m from 180-160?m, with 120?m now the standard for slicing wafers for photovoltaic cells.
This has contributed to shrinking the thickness of PV production wafers to as little as 160μm, or roughly half of what was achievable just a few years ago. This shrinkage has been accomplished while reducing kerf loss − the silicon sawing residue generated by the slicing action — thereby maximizing wafer output per ingot. As silicon wafers represent over 40% of the cost of a c-Si PV module, and silicon, including all aspects of wafer-making, comprises most of that wafer cost, reducing silicon content lowers the total module cost.
 |
| Figure 1. Thickness of PV wafers and perspective [1]. |
Thinner Still
Wafers for the PV industry likely will get thinner still in the next few years, in a probable range of 140-100μm. Progress continues to be incremental (Fig. 1) [1].
The impediment isn’t the wire saw, but source material and manufacturing issues. We have cut 90μm wafers with other materials, but current PV grade silicon sawed that thin is too brittle. As it is, cell makers often struggle to achieve acceptable results with silicon wafers thinner than 180μm. Their capacity to maximize wafer yield from an ingot is only part of the total business case for advanced wire saws. Such saws have the largest load capacity yet, able to slice two ingots (up to 1600mm in total) with one wire web. At the same time, they offer improved energy efficiency, longer parts life, less maintenance down time, lower wafer spoilage rates while producing wafers with superior surface characteristics for making PV cells (Fig. 2) [2].
Speed Considerations
If there’s a divergence on how to optimize their performance, it involves wire speed. Maximum practical wire speeds vary from 10-14m/s. Sawing at very high speeds is more wearing on key machine parts components like wires and wire guides, pulleys and main bearings, and is more energy intensive than lower speed sawing. At high speeds, the mechanical forces exerted on the ingot are that much greater, elevating the risk of wire and wafer breakage, as well as more surface imperfections and a deeper sub-surface damage layer. A broken wire represents a lost production run, wasted ingot, as well as downtime to rewire the web. For some end users, that’s an acceptable tradeoff for greater throughput.
 |
| Figure 2. Yield of PV wafer production (wafers/kg silicon); mono ingots diameter is 208mm, mono PSQ wafer 156mm x 156mm [2]. |
Over the long haul, however, lower speed sawing can be more cost effective and provide a greater total benefit of ownership for many wafer makers. Sawing at 6-8m/s results in superior surface quality and thinner sub-surface damage layer, with no cracks or chips (Ra <0.6μm), which represents a critical advantage as post processing is not required for PV cell-making like it is for semiconductor production. In comparison, older wire saws can leave a uniform depth of damage to 15μm, while inner diameter (ID) saws produce a variable damage field of up to 20-30μm [3]. Sawing at lower speeds also means longer life on parts such as wire guides, bearings, pulleys and capstans, and lower energy consumption.
Longer Parts Life
Defining each machine’s optimal processing parameters is a virtual science for which manufacturers develop machine-specific table feed rates and slurry flow profiles. Using a 120μm wire at 7.5m/s, our AWSM Series 3800.6 will slice bricks up to 1,200mm into 3,500 156 x 156mm wafers in ~9 hours (or 1600mm bricks into 4,700 125 x 125 wafers in less than eight hours) and use more than 30-40% less electricity (an average draw of 65kW), less wire and less slurry than an older saw, while delivering much longer component life. Wire guide groove life is rated at one million wafers (about three months’ production) and main bearings at six million wafers.
Wire sawing has been around for centuries, but today’s high speed industrial wire saw for slicing silicon ingots is a relatively recent development. The general concept is the same for all wire saw makers whether they employ a single or dual wire setup: the wire, which can be hundreds of kilometers long, is wound over wire guides to form a web of parallel slicing wires that carry an abrasive slurry (oil-, water- or glycol-based) that performs the actual cutting at speeds of 300 – 450μm/min. The wire wraps around wire guides that have hundreds of grooves to create a triangular or square/rectangular web of parallel wires moving at high speed that initiate a series of precision cuts the length of the silicon brick.
Collateral Benefits
Despite their 14-15-tonne weight, advanced wire saws have a smaller footprint than older models and have many improvements, such as the latest in automated wire tensioning systems. Although advanced wire saws are rated for wire tension as high as 14 to 33N, a consistent lower wire tension extends the life of the principal mechanical components and enables the use of different wire types, facilitating the trend favoring the use of diamond wire as a slicing medium. Some systems have automatic detection of uncuttable bodies inside multi ingots, like hard ceramics and SiC inclusions that can cause wire breaks.
 |
| Figure 3. Viscosity of glycol-based slurry at various temperatures [4]. |
Advanced wire saws also offer more slurry options. A glycol (PEG) base is the current gold standard for slicing wafers for PV cells, but the energy consumption of an oil base, preferred in semiconductor production, is roughly half that of using a PEG slurry. There’s increasing end user interest in water-based slurry for cost and environmental reasons, including the potential for recovery and re-use. The technical requirements are different, but new wire saws can be adapted for it and there are water-based products available that can produce nice quality wafers. With further refinements, water-based slurry is likely to gain wider acceptance in PV wafering. There is only a small change of the viscosity of water-based slurry with increasing temperatures (Fig. 3) [4]. Other major advantages of water-based slurry are better and cheaper cleaning process and reduced energy consumption for cooling and slicing.
Advanced saws are largely or fully automatic, requiring fewer hands on operator interventions and less supervision than their predecessors. In any linear manufacturing process where a raw material (ingot) goes through multiple processing stages (cropping, squaring, gluing, slicing) to become a finished product (wafer), the process stages can be integrated to form an automated production line.
References
1. KUKA Systems internal research.
2. KUKA Systems internal research.
3. Kao, V. Prasad, Ji. Li., M. Bhgavat, “Wafer Slicing and Wire Saw Manufacturing Technology,” State U. of New York at Stony Brook, NY, http://dove.eng.sunysb.edu/~kao/Wiresaw/NSF-97.pdf
4. Julian Prölß, “High Heat Capacity Cutting Slurry Development for Sub-180μm Poly-Crystalline PV-Wafers,” Proc. of 24th European Photovoltaic Solar Energy Conf., 21- 25 Sept. 2009, Hamburg, Germany, p. 1259 – 1260
Karel Vojtechovsky received his PhD from the Technical U., Zlin, CZ and RNDr from Masaryt U., Burno, CZ and is R&D Manager at KUKA S-Base s.r.o. ,1.maje 2633, 75661 Roznov p.R., Czech Republic, Phone 011-420 602 730 149, email [email protected].
